Scientific Report 1999

Transcription

1

2 Scientific Report

3 4 Patron HM The Queen Beatrix

4 Scientific Report The Netherlands Cancer Institute CANCER RESEARCH LABORATORY AND CANCER HOSPITAL

5 6 Scientific Report 1999 Illustrations and unpublished data in these reports should not be used without permission of the author. Copyright ; The Netherlands Cancer Institute Antoni van Leeuwenhoek Huis Plesmanlaan CX Amsterdam The Netherlands Phone Fax ISSN

6 Contents 7 Board Members 8 Research Divisions 9 Introduction 12 Education in Oncology 19 I Division of Cell Biology 22 II Division of Molecular Carcinogenesis 32 III Division of Cellular Biochemistry 41 IV Division of Immunology 51 V Division of Molecular Biology 61 VI Division of Tumor Biology 69 VII Division of Molecular Genetics 76 VIII Division of Experimental Therapy 86 IX Division of Radiotherapy 98 X Division of Medical Oncology 111 XI Division of Surgical Oncology 126 XII Division of Psychosocial Research and Epidemiology 136 XIII Division of Diagnostic Oncology 145 Biometrics Department 155 Research Facilities 159 Cancer Hospital 164 Ongoing Trials 167 Invited Speakers 177 Projects 179 Personnel-Project Index 187

7 8 Board Members WF Duisenberg, President of Board of Governors International Scientific Advisory Board JR Bertino, American Cancer Society Professor of Medicine and Pharmacology, Yale University School of Medicine, New Haven, USA RA Flavell, Professor of Immunobiology, Yale University, New Haven, USA S Hellman, AN Pritzker Distinguished Service Professor, The University of Chicago, Chicago, USA WGJ Hol, Professor of Molecular Biology University of Washington, Biomolecular Structure Center, School of Medicine University of Washington, Seattle, USA J Mendelsohn, President, MD Anderson Cancer Center, The University of Texas, Houston, USA P Nurse, Professor of Microbiology, Director-General of Imperial Cancer Research Fund, London, UK R Nusse, Professor of Developmental Biology Stanford University; Investigator, Howard Hughes Medical Institute, Stanford, USA HL Ploegh, Edward Mallinckrodt, Jr. Professor of Immunopathology, Harvard Medical School, Boston, USA RA Weinberg, Professor of Biology Massachusetts Institute of Technology, Whitehead Institute, Cambridge, USA C Weissmann, Professor of Molecular Biology University of Zürich, Institut für Molekularbiologie Zürich, Switzerland National Scientific Advisory Board LA Aarden, Professor of Molecular Immunology, Amsterdam D Bootsma, Professor of Cell Biology and Genetics, Rotterdam SWJ Lamberts, Professor of Internal Medicine, Rotterdam B Löwenberg, Professor of Hematology, Rotterdam CJLM Meijer, Professor of Pathological Anatomy, Amsterdam CJM Melief, Professor of Immunohematology, Leiden HM Pinedo, Professor of Clinical Oncology, Amsterdam FH Schröder, Professor of Urology, Rotterdam GNJ Tytgat, Professor of Gastroenterology, Amsterdam AJ Van der Eb, Professor of Fundamental Tumor Virology, Leiden PC Van der Vliet, Professor of Physiological Chemistry, Utrecht Board of Directors P Borst, chairman and director of research (until ) AJM Berns, chairman and director of research (from ) L Neeleman, director organization and management S Rodenhuis, director clinical research and development Board of Governors WF Duisenberg, president ML Frohn-De Winter, vice-president (until ) HCJ Van der Wielen, vice-president (from ) JF Visser, treasurer R Hazelhoff JHM Temmink GNJ Tytgat PF Van der Heyden S Van der Kooij J Van der Meer MWM Vos-Van Gortel Scientific Advisory Council P Borst, president (until ) AJM Berns, secretary (until ), president (from ) AM Kruisbeek, secretary (from ) R Bernards S Rodenhuis A Sonnenberg FE Van Leeuwen

8 Research Divisions 9 I Cell Biology E Roos (head) J Calafat JG Collard K Jalink A Sonnenberg II Molecular Carcinogenesis R Bernards (head) E Kriek (honorary staff member) TK Sixma H Te Riele M Van Lohuizen III Cellular Biochemistry WH Moolenaar (head) J Borst N Divecha P Ten Dijke (from ) WJ Van Blitterswijk LN Vernie (until ) IV Immunology AM Kruisbeek (head) MJ Kersten TNM Schumacher H Spits CJM Vennegoor (stationed at the Free University) FA Vyth-Dreese K Weijer V Molecular Biology RHA Plasterk (head) P Borst APM Jongsma (stationed at the Free University) HGAM Van Luenen VI Tumor Biology J Neefjes (head) PC Hageman (honorary staff member) J Hilkens RJAM Michalides A Tulp (deceased ) AA Van der Gugten AA Verstraeten (stationed at the Free University) VII Molecular Genetics AJM Berns (head until ) P Demant (head from ) D Barlow R Beijersbergen (from ) P Krimpenfort M Snoek MA Van der Valk VIII Experimental Therapy AC Begg (head) KMG Haustermans JHM Schellens AH Schinkel FA Stewart MJ Van de Vijver LJ Van t Veer IX Radiotherapy H Bartelink (head) BMP Aleman JSA Belderbos J Borger IAD Bruinvis EMF Damen RW De Boer LGH Dewit RLM Haas AAM Hart KMG Haustermans RB Keus JV Lebesque EAH Masselink BJ Mijnheer LMF Moonen NS Russell JG Salverda BNFM Van Bunningen M Van Herk FW Wittkämper

9 10 RESEARCH DIVISIONS X Medical Oncology S Rodenhuis (head) JW Baars P Baas EM Bais JH Beijnen W Boogerd H Boot PF Bruning GC De Gast SP Israëls MJ Kersten H Neering JFM Pruijt (until ) JHM Schellens JH Schornagel BG Taal WW Ten Bokkel Huinink JJ Van der Sande M Van der Weide AMC Witte (from ) N Van Zandwijk APE Vielvoye-Kerkmeer Research Nurses D Batchelor AC Dubbelman MMJ Holtkamp GAM Linthorst ME Schot M Swart W Uyterlinde XI Surgical Oncology BBR Kroon (head) AJM Balm CL Blackburn KE Bos D Buitelaar JB De Boer (until ) FJM Hilgers S Horenblas MM Kaag W Meinhardt OE Nieweg RP Noordanus (until ) EJTh Rutgers PFE Schutte IB Tan M Van Beurden F Van Coevorden N Van der Vange ACM Van Lindert J Visscher LAE Woerdeman (from ) FAN Zoetmulder XII Psychosocial Research and Epidemiology NK Aaronson (head) MA Rookus FSAM Van Dam FE Van Leeuwen XIII Diagnostic Oncology MJ Van de Vijver (head) APE Besnard JMG Bonfrer D De Jong MPW Gallee K Gilhuis CA Hoefnagel FBL Hogervorst W Koops R Kröger SH Muller PM Nederlof WJ Nooijen FA Pameijer JL Peterse L Schultze Kool M Smid-Geirnaerd RA Valdes Olmos O Van Tellingen LJ Van t Veer MLF Van Velthuysen Biometrics Department OB Dalesio (head) H Van Tinteren

10 Heads of General Research Services Audiovisual Service JM Lomecky Biophysics GJF Blommestijn 11 RESEARCH DIVISIONS Central Cancer Library S Bakker Financial Administration B Simmelink General Facilities R Clement Laboratory Animal Department RGM Ten Berg Research Coordination AJM Berns, laboratory research coordinator (until ) AM Kruisbeek, laboratory research coordinator (from ) H Van Luenen, laboratory research manager EJ Vos, clinical research manager

11 12 Introduction Director of Research A. Berns The Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital (NKI/AvL) is an integrated cancer institute, combining a hospital and research laboratories under one roof in a single independent organization. The hospital comprises 180 beds, a large Radiotherapy Department and outpatient clinics. Facilities for patient research include a large patient database, clinical data management and active research groups in epidemiology and psychosocial oncology. The laboratory covers all major areas of cancer research, with special emphasis on mouse tumor models, mouse (reverse) genetics, cell biology, immunology and translational research requiring close collaboration between clinical and basic scientists. This Scientific Report deals mainly with the clinical and basic science in the NKI/AvL. Information on patient care can be found in our General Report will be primarily remembered as the year that Piet Borst stepped down as Director of Research after holding this position for almost 17 years. During his directorship the NKI/AvL metamorphosed from an organization that had lost contact with the fast developing area of molecular and cellular biology into a high profile research institute with cutting edge research in basic and clinical disciplines. The institute now hosts many strong research groups, often headed by young group leaders. They constitute a vibrant faculty with the ambition to further build on the solid foundation laid in the last decade. In recognition of Piet Borst s excellent leadership, a two-day symposium Killing Cancer Cells was held in Amsterdam. This well-attended meeting, with prestigious speakers, was a dignified and fitting tribute to Piet Borst. During the festivities Piet Borst received the Royal honor Commander in the Order of the Dutch Lion. The Mayor of the City of Amsterdam, Schelto Patijn presented the decoration. We are privileged that Piet Borst will remain at the Cancer Institute as a group leader for at least another 5 years, continuing his own research. At the turn of the Millennium At the turn of the millennium we look back on a fascinating period in biomedical research. The last fifty years have yielded fabulous insights in how nature works. We now know in some detail how cancer arises, and have identified many of the genetic lesions instrumental in causing cancer. However, there are still many unknowns. I want to mention some that are particularly relevant for cancer as they pose a challenge for us in the future. 1) We still have little understanding of most of the genetic traits that determine the level of susceptibility or resistance of individuals to the development of malignancies. 2) We do not know how these genetic traits interact with environmental factors, such as lifestyle and environmental exposure, giving us little background for preventive measures. 3) Although we have identified many of the lesions frequently occurring in specific tumors, this information is still insufficient for a reliable prognosis and prediction of response to treatment. 4) Our knowledge of the precise effects of the genetic lesions on tumor cell biochemistry is still sketchy. A more detailed insight is required for effective intervention strategies. There is, however, also ample reason for optimism. In a number of instances our understanding is beginning to reach the level that permits the design of new, highly specific drugs that will selectively attack cancer cells while leaving normal cells untouched. New treatments using these designer drugs are expected to differ profoundly from the conventional chemotherapies in that they are more specific and therefore cause fewer side effects. They will complement and may eventually replace the current chemotherapies, providing a much more acceptable supplement to surgery and radiotherapy, the most effective treatments for localized and operable tumors. Such a detailed insight has to come both from high quality basic research and the thorough analysis of clinical tumor samples. This will lead to markers for better diagnosis and to new targets for intervention. To achieve this, it is important to obtain a catalog of molecular features of human tumors, using array technologies for DNA, RNA and proteins. The NKI/AvL is well positioned to contribute to these areas, some of which are mentioned below. 1) Understanding the details of how a cell works, how it divides and how it migrates. We have invested heavily in cell biology research and this discipline will be of critical importance understanding how genetic instruc-

12 tions translate into specific cell characteristics. This information is also indispensable for designing and testing intervention strategies as they can define new targets for drug treatment or lead to strategies to effectively mobilize the immune system against tumor cells. 2) Understanding the heritable traits which control resistance and susceptibility to cancers. We have developed unique animal model systems, fully in the tradition of the NKI/AvL, that allow us to mimic and study tumor development as it occurs in man but under wellcontrolled laboratory conditions. These models will also be invaluable to test existing and new intervention strategies. 3) We have unique expertise with the formulation of drugs, and with pharmacokinetic, and pharmacodynamic measurements in patients and model systems. In addition, there are strong groups within the institute studying drug resistance mechanisms of cancer cells and how to circumvent these. All this expertise is pivotal for clinical research purposes and it is also invaluable for preclinical testing of new therapies. 4) Over the decades we have collected the records and stored tumor samples of thousands of cancer patients. Pairing the information from these records with thousands of molecular parameters of each patient sample using high throughput micro-array technologies should lead to the identification of molecular markers that can accurately classify tumors. This should lead to better diagnosis, more rational selection of therapies and, eventually, to the design of new therapies and early intervention strategies. These projects need the participation of both clinicians and basic researchers. They present unique opportunities for exciting research. While expressing this optimism for the future, we have to realize that current successful cancer treatments depend almost entirely on surgery, radiotherapy, and chemotherapy. These treatment modalities will remain the backbone for the treatment of cancer for quite some time and therefore further optimizing and careful pairing of these existing treatment modalities remain important objectives of our research. Figure 1 Quality of research The quality of research is monitored in several ways. Our scientific productivity as based on bibliometric parameters (citations and impact of scientific articles published by the NKI staff) has shown a steady increase since the beginning of the eighties. In the last years the number of citations and impact are leveling off (Tables I and 2) with yearly fluctuations suggesting that we have reached a steady state. Our competitiveness for obtaining grants is another measure of quality. We have been quite successful during the last five years in obtaining grant support, scoring on average 2-3 fold better than our competitors. This translates in funding of the vast majority of our grant applications. In 1999 our success rate was somewhat diminished but still almost twofold higher than the average. The first impression for 2000 indicates that the score is up again. The third measure of quality is based upon external site visits, in which international leaders in a particular field of research review the work of a division or research groups with a similar theme on a quinquennial basis. In February 1999 the work of two divisions was reviewed. The Division of Diagnostic Oncology was visited by D Sidransky (Director Head and Neck Cancer Research, Table 1 Short-term citations and impact of scientific articles published by the NKI research staff Publication year Citations* Impact** INTRODUCTION * In the two years after publication, excluding self citations. Starting with 1989, the citation analysis has been carried out on-line. This allows detection (and elimination) of all self citations. Before 1989 this pruning was limited to first authors. ** The impact factor is the average number of citations per year of an article in a given journal. The total impact is the sum of the impact of all articles published that year.

13 14 INTRODUCTION Table 2 Number of NKI publications in journals with an impact factor 10 and higher in the period 1989 to See Table 1 for definition of impact factor. IF97 Journal Total Nat Genet Cell Nat Med N Engl J Med Nature Science Immunity Genes Dev Curr Opin Cell Biol Immunol Today Lancet J Exp Med EMBO J Ann Intern Med J Cell Biol J Natl Cancer Inst Trends Cell Biol Mol Cell Biol Total Johns Hopkins, Baltimore, USA), P Selby (Director of Clinical Research for ICRF, Leeds, UK), and M Stratton (Section Chairman of Cancer Genetics, Institute of Cancer Research, Surrey, UK). The site report was positive with suggestions to further augment the research of this still young division. The Division of Psychosocial Research and Epidemiology was reviewed by P Ganz (University of California, Los Angeles, USA), P Patrick (University of Washington, Seattle, USA) and L Bernstein (USC-Norris Comprehensive Cancer center, University of South California, Los Angeles, USA). Again, the review was favorable, with part of the research identified as world-class. A number of constructive suggestions were made for improvement. This year we also received the results of a completely independent evaluation of biomedical sciences in the Netherlands conducted by the Dutch Academy of Art and Sciences. The mostly excellent rating of the NKI/AvL could now be put in the perspective of the evaluation of biomedical research elsewhere in the Netherlands. The evaluation showed that overall biomedical research in the Netherlands is of good to very good quality. Twelve percent of the programs score was excellent and 30% very good. Four of the five NKI programs scores were excellent and one very good. While it is pleasing that outside reviewers give us such a high rating, there is always room for improvement. That will be the challenge for the coming five years. Research Highlights It is always difficult for a research director to select research highlights to mention in the introduction. Rather than give an overview of highlights of all the divisions, I have decided to focus on a few discoveries by mostly young investigators of the institute. The availability of the complete sequence of a range of genomes will revolutionize the way we do research. Now we have to make sense of this information using reverse and forward genetics in combination with cell biological and biochemical analyses. Work in many of the research divisions relies on a combination of these approaches. An exciting example of this is represented by a discovery in the group of Ronald Plasterk (Division V). While all known natural isolates of the worm C. elegans contain multiple copies of the Tc1 transposon, which are active in the soma, Tc1 transposition is fully silenced in the germline of many strains. Plasterk and colleagues mutagenized one such silenced strain and isolated mutants in which Tc1 had been activated in the germline ( mutators ). Interestingly, many other transposons of unrelated sequence had also become active. Most of these mutants appeared resistant to RNA interference (RNAi) (Ketting et al. Cell 1999; 99: ). They found one of the mutated genes, mut-7, to encode a protein with homology to RNaseD. This provided support for the notion that RNAi works by dsrna-directed, enzymatic RNA degradation. This led them to propose a model in which MUT-7, guided by transposon-derived dsrna, represses transposition by degrading transposon-specific messengers, thus preventing transposase production and transposition. This represents an intriguing mechanism of gene regulation and gene expression inheritance. Whether similar mechanisms of gene regulation are employed in mammalian cells remains to be seen. Genetic screens can also be very rewarding in cells in culture, especially with the increasing availability of primary cells of knockout mice. The group of Maarten Van Lohuizen (Division II) showed that the oncogene BMI1 silences the tumor suppressor genes Ink4a and P19Arf in mouse embryo fibroblasts in vitro and in mice in vivo. The absence of BMI1 in Bmi1 knockout fibroblasts resulted in higher levels of INK4A and P19ARF leading to premature senescence of the cells (Jacobs et al. Nature 1999; 397: ). His group has now used the poor growth of these fibroblasts to perform a rescue screen with a retroviral cdna library. CDNAs were identified which could restore growth of these fibroblasts in the absence of BMI1. At least some of the cdna s appeared to specifically impair the level of p19arf but not INK4A, thereby adding new regulatory components to this cell cycle regulatory network. Furthermore, Van Lohuizen and colleagues have shown that downregulation of Ink4a-Arf by BMI1 underlies its ability to cooperate with MYC in tumorigenesis (Jacobs et al. Genes & Dev 1999; 13: ). Heterozygosity for Bmi1 inhibits lymphomagenesis in Eµ-Myc mice by

14 enhancing c-myc-induced apoptosis. They observe increased apoptosis in Bmi1(-/-) lymphoid organs, which can be rescued by deletion of Ink4a-Arf or overexpression of BCL2. Furthermore, BMI1 collaborates with MYC in enhancing proliferation and transformation of primary embryo fibroblasts (MEFs) in an INK4a-ARFdependent manner, by prohibiting MYC-mediated induction of p19arf and apoptosis. Strong collaboration was observed between the Eµ-Myc transgene and heterozygosity for Ink4a-Arf, with the concomitant loss of the wild-type ink4a-arf allele and formation of highly aggressive B-cell lymphomas. Together, these results reinforce the critical role of BMI1 as a dose-dependent regulator of INK4A-ARF. It is expected that unraveling of the pathway in which BMI1 acts will uncover additional genes that will be critical for tumorigenesis in man. Hein Te Riele and colleagues demonstrated that mice carrying a disruption in MutS homolog Msh5 show a meiotic defect, leading to male and female sterility (De Vries et al. Genes & Dev 1999; 13:523-31). Histological and cytological examination of prophase I stages in both sexes revealed an extended zygotene stage, characterized by impaired and aberrant chromosome synapsis, that was followed by apoptotic cell death. Thus, murine MSH5 promotes synapsis of homologous chromosomes in meiotic prophase I. They also found that mice carrying a partial defect in mismatch repair by ablation of the mismatch recognition protein MSH6, are highly cancer prone, but that intestinal cancer is suppressed by redundant functions of MSH3. This observation predicts that families segregating a defect in MSH6 show a high incidence of cancer but are often not recognized as HNPCC families due to suppression of intestinal tumorigenesis by residual MMR activity. Ed Roos and colleagues (Division I) showed that the chemokine SDF-1 is essential for T lymphoma metastasis to many tissues. This indicates that lymphoma dissemination is, at least in part, determined by chemokines present in those tissues (Soede et al. J Immunol 1999; 163: ). Stromal cell-derived Factor 1 (SDF-1) acts by inducing migration but, in addition, by activating integrin adhesion molecules through heterotrimeric G- protein G q. LFA-1 dependence is only seen at low SDF-1 concentrations. The LFA-1 signal can be propagated in a ZAP70 dependent fashion to other LFA-1 molecules on the same cell, which then bind to ICAM-2 on other cells. This causes cell aggregation that was also blocked by dominant-negative ZAP-70. Thus, an LFA-1 signal involving ZAP-70 activates other LFA-1 molecules, suggesting that the chemokine signal can be amplified by multiple cycles of LFA-1 activation. This represents an important step forward in our understanding of the molecular mechanisms and tissue specificities observed in metastatic growth. Kees Jalink (Division I) found, using a GFP-bound sensor, that the lipid PIP2, an important signaling molecule, is localized in patches near the plasma membrane. He generated a membrane-bound construct to detect PIP2 by intramolecular FRET (Fluorescent Resonance Energy Transfer). This should permit the highly sensitive detection of local and transient changes in PIP2 levels during dynamic processes such as cell migration. The possibility to measure protein-protein and protein-substrate interaction in real time and space offers unique opportunities for following molecular signaling events within a single cell upon exposure to external signals. The group of Ton Schumacher (Division IV) examined the role of cytotoxic T cell memory in the subsequent response against mutant epitopes in vivo (Haanen et al. J Exp Med 1999; 190: ). Applying soluble tetrameric peptide-mhc complexes, these experiments show that the CD8 T-cell mediated response against a large series of random mutants of a viral antigen is completely dominated by cross-reactive CD8 cells that recognize both the wild-type and the mutant epitopes. As these cross-reactive T cells can also differentiate into cytotoxic effector cells, these findings suggest that epitope mutations in TCR-exposed residues will not readily promote CTL escape. In future studies, the group will dissect the molecular requirements that lead to selective expansion of cross-reactive T cells. The groups of Jannie Borst and Wim Van Blitterswijk (Division III) studied the role of sphingolipid ceramide in apoptosis signaling. They examined ceramide formation induced by CD95, etoposide, or gamma-radiation (IR) in relation to caspase activation and mitochondrial changes in Jurkat T cells (Tepper et al. J Clin Invest 1999; 103: 971-8). They demonstrated that ceramide, contrary to common belief, does not play a triggering role in apoptosis. Instead, ceramide production appears to result from loss of plasma membrane symmetry during the effector phase of apoptosis; such lipid scrambling may facilitate membrane blebbing, vesicle shedding and apoptotic body formation. Alfred Schinkel and colleagues (Division VIII) have shown that P-glycoprotein (P-gp) inhibitors could potentially be used to increase the fetal penetration of P-gp substrate drugs (Smit et al. J Clin Invest 1999; 10: ). They used mice with a targeted disruption of the Mdr1a and Mdr1b genes. Mdr1a(+/-)/1b(+/-) females were mated with Mdr1a(+/-)/1b(+/-) males to obtain fetuses of 3 genotypes (Mdr1a(+/+)/1b(+/+), Mdr1a(+/-)/ 1b(+/-), and Mdr1a(-/-)/1b(-/-)) in a single mother. Intravenous administration of the P-gp substrate drug paclitaxel to pregnant dams revealed that 16-fold more drug entered the Mdr1a(-/-)/1b(-/-) fetuses than entered wild-type fetuses. Furthermore, placental P-gp activity could be completely inhibited by oral administration of the P-gp blockers PSC833 or GG918 to heterozygous mothers. This finding implies that the placental drugtransporting P-gp is of great importance in limiting the fetal penetration of various potentially harmful or therapeutic compounds and demonstrates that this P-gp function can be abolished by pharmacological means. The latter principle could be applied clinically to improve 15 INTRODUCTION

15 16 INTRODUCTION pharmacotherapy of the unborn child. Other studies by the Schinkel group showed that even low levels of P-gp and Mrp1 contribute considerably to tumor drug resistance, suggesting that multidrug transporter inhibitors may also increase the sensitivity of previously untreated (naïve) tumors to chemotherapy. The group of Marcel Van Herk (Division IX) developed a generic method for 3-D evaluation of target volume delineation in multiple imaging modalities to assist accurate high-dose radiotherapy (Remeijer et al. Med Phys 1999; 26: ). The evaluation includes geometrical and statistical methods to estimate observer differences and variability in defining the Gross Tumor Volume in relation to the diagnostic CT and MRI modalities. The statistical method distinguishes observer and modality related uncertainties, which are expressed in terms of various error components: random observer deviations, systematic observer differences and systematic modality differences. The method was successfully applied to a group of prostate cancer patients, where it was demonstrated that delineation variability is non-homogeneous, with the largest variations occurring near the seminal vesicles and the apex. This permits reliable treatment plans for a three- field technique giving a dose of 78 Gy to the target volume (prostate). This approach is expected to have a direct impact on both local cure and survival of the patients. Floor van Leeuwen and coworkers (Division XII) found that physically active women had a 30% lower risk of breast cancer than inactive women. This suggests that physical activity might be one of the first life-style risk factors for breast cancer amenable to preventive measures. The clinical pharmacology group (Division X) has invested in facilitating the oral availability of taxanes by inhibition of the P-glycoprotein multi-drug transporter that is responsible for the fact that orally administered taxanes have a very low bioavailability. The group has been successful in increasing the bioavailability of both paclitaxel and docetaxel by prior administration of cyclosporin A. These findings in phase I studies have led to a phase II study of oral Docetaxel in advanced breast cancer. Oral administration is not only of practical advantage, it also circumvents the adverse reactions of patients against Cremophor EL used for the formulation of paclitaxel for intravenous injection. One of the unique features of the institute is the natural interaction between disciplines and divisions. This has led to the integration of a unit for chemoradiation therapy in one of the hospital wards. The institute has played a prominent role in the relatively recent recognition that concurrent chemo- and radiotherapy has curative potential in a large number of epithelial tumors, whereas the sequential application of these treatment modalities shows no advantage. Examples include treatments of stage III NSLC with concurrent cisplatin and radiation and, more recently, intra-arterial cisplatin and radiotherapy in advanced head-and-neck cancers. The new clinical unit will make it possible to perform clinical studies in a range of tumor types. A boost for translational research in 2000 The expression profile of many thousands of genes in a single tumor by micro-array technologies is expected to contain important parameters for diagnosis and might predict response to treatment and clinical outcome. In order to establish correlations between expression profiles and biological characteristics of tumors, retrospective studies have to be conducted in which the profiles of thousands of tumors are compared with the clinical records. The NKI/AvL has collected tumor samples for many decades and has, in recent years, extracted RNA and DNA from several hundreds of samples. We are currently investigating several options to analyze a substantial number of these tumor samples. For this we will establish collaborations with expert centers elsewhere. In this way we hope to acquire the necessary expertise for data collection and bioinformatics in the course of Once reliable correlations between expression profiles and prognostic parameters have been established, array analyses can be introduced as a routine screening methodology. Patients will immediately profit, as this information will permit a more rational choice in the treatment options. Mouse models for cancer become more and more sophisticated. Switching genes on and off in a tissue specific and time controlled fashion permits the induction of specific tumors at high incidence within a narrow time frame in these mice. Since multiple mutations can be introduced simultaneously, these models permit more accurate determination of genotype-phenotype relationships. Moreover, since we can introduce in these models the same genetic lesions as found in human tumors, they probably will more closely mimic the human condition and therefore be better suited to the initial testing of intervention protocols. In that way they can help to bridge the gap between basic and clinical research. A number of clinicians have shown interest in these models and in the course of 2000 several clinicians are expected to start working with them. Honors Piet Borst received the Royal decoration of Commander in the Order of the Dutch Lion and the golden Van Leeuwenhoek Microscope of the Netherlands Cancer Institute during the festivities in honor of his retirement as Director of Research. Ronald Plasterk (Division V) received the prestigious Spinoza Award of the Netherlands Organization for Research, worth fl 3,000,000 Dutch guilders to be spent on research of his choice. In the spring of 2000, Plasterk will become Director of the Hubrecht Laboratory, a highly

16 regarded research institute of The Netherlands Academy of the Arts and Sciences, fully devoted to developmental biology. Maarten Van Lohuizen (Division II) received the Pioneer Stipendium of the Netherlands Organization for Research, a high profile research career grant of the Netherlands Organization for Research. Ada Kruisbeek became member of EMBO and Anton Berns was elected as member of the Academia Europeana. Jaqueline Jacobs, a graduate student in the group of Maarten Van Lohuizen, received the Antoni van Leeuwenhoek Prize 1999, for elucidating the role of the oncogene BMI1 in the cell cycle. She showed that BMI1 exerts its effect by suppressing the expression of the tumor suppressor genes Ink4a and p19arf, thereby preventing cell cycle arrest and senescence. New NKI professors The NKI/AvL cannot award university degrees. However, many of our staff members hold special parttime chairs in one of the Dutch Universities. This facilitates the supervision and awarding of degrees to graduate students receiving their training at the Netherlands Cancer Institute. A particularly large number of staff became professors in Jan Schellens was appointed as Professor of Drug Toxicology at the University of Utrecht. Piet Borst was re-appointed as Professor of Clinical Chemistry at the University of Amsterdam. Jacques Neefjes became Professor of The Biology of Antigen Processing and Presentation at the University of Leiden, Bin Kroon Professor of Surgical Oncology at the University of Amsterdam, Jannie Borst Professor of Experimental Oncology at the University of Amsterdam, and Peter Peters Professor of Cell Biology at the Free University of Amsterdam. Changes in NKI/AvL academic staff In January we were notified of the sudden death of Ab Tulp, a dedicated, if idiosyncratic biochemist who specialized in cell separation techniques. He was on the staff of the Netherlands Cancer Institute (Division of Tumor Biology) from Besides his originality as a scientist, he was renowned for his artistic and humorous drawings explaining biochemical processes or exposing political controversies within the research community at the NKI/AvL. We have lost in him a loyal and fine colleague. A number of staff members retired in In addition to Piet Borst as Director of Research, Lou Smets, Professor of Experimental Oncology, also retired. Lou joined the NKI/AvL as head of Experimental Cytology in 1970 and served the institute for almost 30 years. He was committed to leukemia research and studied drug induced apoptosis of tumor cells. Lou was a rigorous and original scientist. He was not easily convinced, and was known for his critical reviews. He pointed out the discrepancies between the outcome of short-term apoptosis assays and long-term clonogenic survival of cells defective in regulation of cell cycle control. Lou was always keen to see his scientific findings applied in the clinic. The concept of lowering the ph in tumors was explored under his supervision, for the treatment of carcinoid tumors by MIBG. He was also a devoted teacher and served as chair of the teaching committee of the institute for over 15 years, including coordinating the rotation of students. A postdoctoral course and symposium of the European School of Oncology and the International Society of Pediatric Oncology, of which Lou was an organizer for many years, was held in Amsterdam to honor Lou s contribution to cancer research. Leen Vernie (Division of Cellular Biochemistry) retired as staff member after having served the institute for more than 30 years. Leen investigated the anti-carcinogenic properties of selenium in the early seventies. His interest has remained with this early work although he was willing to take on many other tasks. In the eighties he became involved in the purification of G-proteins and subsequently in the synthesis of custom peptides and special derivatives. He has served the NKI/AvL as a dedicated, helpful and pleasant colleague. Ton Van der Gugten also said farewell to the NKI/AvL after being a staff member for more than 30 years. He worked in the area of tumor virology and hormone dependence of tumors. In the last 10 years he became involved in the mathematical modeling of biochemical processes. New staff joining the NKI/AvL includes Peter Ten Dijke who was recruited from the Ludwig Institute for Cancer Research in Uppsala, Sweden. He is well known for his work on TGF-β signaling and SMAD tumor suppressor proteins. He is a senior staff member of the Division of Cellular Biochemistry. Roderick Beijersbergen was recruited to the institute as an AvL fellow. He did his post-doctoral work in the laboratory of Bob Weinberg (Whitehead, Cambridge, USA) and has now joined the Division of Molecular Genetics where he will focus on the regulation of expression of human telomerase. Internal and external appointments With my appointment as Director of Research, my previous positions (head of the Division of Molecular Genetics and Laboratory Research Coordinator) became vacant. We are pleased that Ada Kruisbeek, head of the Division of Immunology, has accepted the position of Laboratory Research Coordinator. Peter Demant will succeed me as head of the Division of Molecular Genetics. I will continue to supervise my own research group within that division. Jannie Borst has taken over the chair of the teaching committee from Lou Smets. 17 INTRODUCTION

17 18 INTRODUCTION National and international activities Staff of the NKI/AvL fulfilled numerous functions in national and international organizations, on the scientific boards of scientific journals, as members of study sections, as organizer or co-organizer of scientific meetings, workshops and congresses. International organizations included: AACR, CIBA, EACR, EBCTCG, EMBO, EORTC, ESSO, ESTRO, ICRU, MRC, OECI, WFSOS, WHO. Staff members also served on boards of organizations such as EORTC cooperative groups, International Association for Breast Cancer Research, Mesotheliomenwerkgroep van de Nederlandse Vereniging van Artsen voor Longziekten en Tuberculose, Landstein Stichting voor Bloedtransfusie Research, Nederlandse Commissie voor Stralingsdosimetrie, Dutch Head and Neck Cancer Cooperation Group, Oog en Orbita Commissie, The Netherlands Working Party on Melanoma, International Psycho-oncology Society, European Community Committee on Palliative Cancer Care, Comprehensive Cancer Center Amsterdam, European Society for Therapeutic and Radiation Oncology, the WHO Quality of Life Group, The International Academy of Pathology, the Gezondheidsraad, The European Cancer Center, National Advisory Board on AIDS, Netherlands Health Council Committee on Home Care for Cancer Patients, Scientific Council on Social Oncology (WRSO) of the Dutch Cancer Society, KNAW committee Discipline adviesplan Medisch Onderzoek, KNAW selectiecommissie KNAW fellows, Walree Fonds, Heineken Prijs selection committee, the General Motors Award committee for Cancer research, KEMO, CCMO, Scientific Committee for the Cancer Research Campaign (UK), EMMA (European mouse repository), Education and Psychosocial Research Committee of the Cancer Research Campaign, the Dutch and European Societies of Surgical Oncolgy, the World Federation of Surgical Oncological Societies, Wetenschappelijke Kommissie van het Fonds voor Wetenschappelijk Onderzoek België, Deutsche Forschungsgemeinschaft. Staff members also served as editors or on editorial boards of scientific books or journals such as Genes & Development, BBA reviews in cancer, European Journal of Surgical Oncology, Genes to Cells, Experimental Cell Research, Mammalian Genome, Carcinogenesis, Immunogenetics, Quality of Life Research, European Journal Immunology, International Immunology, Journal of Biological Chemistry. Other activities included organizing national and international oncology meetings, workshops and congresses including: Killing Cancer Cells (Amsterdam, 17&18 November, 1999), Thymus Workshop Rolduc (Kerkrade, May, 1999), INSERM Meeting on the T Cell Receptor/ CD3 Complex (Aix-les-Bains, October, 1999), First International Congress on the Sentinel Node in Diagnosis and Treatment of Cancer (Amsterdam, 1999), Symposium on neck node metastasis of an unknown primary (1999) and STATs and SMADs (Seoul, October, 1999). Finally, staff members participated in teaching for the European School of Oncology and the ESTRO teaching course on Radiation Physics for Clinical Radiotherapy and Basic Clinical Radiobiology. Outlook and acknowledgements We are making steady progress in our plans for a new hospital building and renovation of the old premises. We received a positive review from several governmental agencies, including the Ministry of Health, regarding our plans for the new hospital building. We expect that we can now proceed expeditiously and lay the foundation for the new building in the course of the year This will at least temporarily solve the ever increasing need for additional space to accommodate our expanding basic and clinical research activities. For all our research we depend on the financial support of the government, the Dutch Cancer Society and of many individuals. Only with their help can we continue to develop new ideas that should lead to prevention, early detection, and more effective treatment. Ton Berns Director of Research

18 Education in Oncology 19 Education in fundamental and clinical research and training in patient care are regular activities of the institute. Several senior staff members have joint appointments as professors at Dutch universities and contribute to the regular curriculum at these universities. In addition, many staff members teach in courses for graduate students, either within the institute or at universities. The research departments attract many undergraduate students from universities throughout the country, who contribute practically to ongoing scientific projects and receive in house theoretical training. The NKI has a formal liaison with the Biology Faculty of the University of Amsterdam and is committed to present a course in Experimental Oncology to the third year students in Medical Biology of this university. The institute participates in the Oncology Graduate School Amsterdam, together with of the University of Amsterdam and the Free University. A large body of PhD students are assigned to this graduate school, all for a four-year period. These graduate students generally make significant contributions to the scientific program. They receive additional practical and theoretical training within the institute. toral teaching and training in clinical oncology and patient care is the major educational task of the cancer hospital. Collaboration with the local medical faculties and hospitals has increased continuously by shared activities of their graduate schools and joint projects in clinical and fundamental research. Table 1 Lecture course in Experimental Oncology, Fall 1999 Cancer, diagnostics, treatment Epidemiology Psychosocial research Pathology Surgery Chemotherapy Pharmacokinetics Drug resistance Radiotherapy DNA damage and repair Apoptosis Antigen presentation Immunology Intracellular transport Clinical immunology Genetic predisposition for cancer Genome project Cell cycle Transcription Developmental biology Cell senescence Cell adhesion Signal transduction Tumor progression Medical genetics S Rodenhuis F Van Leeuwen F Van Dam D De Jong O Nieweg J Schellens J Beijnen A Schinkel R Haas H Te Riele J Borst J Neefjes H Spits P Peters J Haanen P Demant R Plasterk R Michalides R Bernards M Van Lohuizen R Beijersbergen E Roos W Moolenaar A Berns L Van t Veer Undergraduate students The undergraduate program in Experimental Oncology attracts students of all national universities in partial fulfillment of the obligations of their curriculum. Students usually have a background in Medical Biology/Health Sciences, Biology, Chemistry or Medicine. The undergraduate program offers combined practical and theoretical training in various aspects of experimental oncology. Practical training includes participation in ongoing research projects for a minimum period of 4 months, after which the students deliver oral and written accounts of the results obtained. In 1999 about 35 undergraduate students did a 4-9 month placement at the basic research divisions of the NKI. The majority of these students came from either of the two universities in Amsterdam or from Utrecht University, in equal distribution. There were three students from Leiden University, one from Wageningen and one from Nijmegen. Most studied either Medical Biology or Biology. The core element of theoretical training is the lecture course in experimental oncology, given twice yearly by staff members of the institute. This course is also part of the regular curriculum of third year students in Medical Biology of the University of Amsterdam. It is much appreciated for giving an accurate account of recent developments in basic oncology (Table 1). This year, the course was supplemented by visits to the clinical departments of Radiotherapy, Radiobiology and Pathology, under supervision of department staff. All teaching activities are supervised by a Teaching Committee. Specialized tasks are assigned to individual

19 20 EDUCATION members of the scientific staff. These tasks include the organization of lecture courses, guidance of scientific visits and the distribution of information regarding undergraduate studies at the NKI. Information undergraduate school: Teaching Committee NKI; secretariat H3, phone Dean: Prof. J Borst Graduate students About 75 PhD students of the Netherlands Cancer Institute are members of our Graduate School: the OOA (Onderzoekschool Oncologie Amsterdam: Oncology Graduate School Amsterdam). These are scientists-intraining, who receive complementary theoretical and practical training covered by a 4-year contract. At the end of this period, the student can acquire the PhD degree from any national university. The tutorship for these graduate students is covered by the graduate school, headed by Prof. R Plasterk. The Oncology Graduate School Amsterdam is a joint activity with the medical schools of the Free University and the University of Amsterdam. All OOA graduate students participate in the yearly meeting of graduate students, a 3-days retreat during which they present the highlights of their ongoing investigations. This year s retreat was held in the monastery of Rolduc in Kerkrade. In 1999 several combined practical and theoretical courses were organized (Table 2). Scientists of international reputation in the respective areas were invited to teach at these courses. We have introduced a new element to the programme of the NKI part of OOA: a Special Seminar Program and Semester Courses. For the Special Seminar Program leading scientists were invited to the institute. We take full advantage of the visit by advanced preparation for the visit and by organizing a discussion session after the seminar. In the preparation sessions the graduate students discuss an article selected by the guest; the lecture itself is open to the Table 2 general public, it is advertized like all seminars, but the discussion sessions afterwards are only open to the NKI OOA students. In 1999, 50 special seminars have been organized. We have continued the experiment with a series of semester courses. These courses consist of 10 sessions in which the two course organizers and participants discuss in great depth key publications that have had a major impact on the field. The students prepare each session by writing a three page synopsis of the publication, and these synopses are corrected in detail (for style as well as content) by the course instructors before the discussion session. The following course were given: 1. Cell biology (instructors: R Bernards and M Van Lohuizen fall 1998); 2. Genetics (instructors: A Berns and R Plasterk, spring 1999); 3. Immunology (instructors: A Kruisbeek and T Schumacher, fall 1999). Information graduate school: secretariat H8, phone Dean: Prof. RHA Plasterk Postdoctoral training The NKI research department continually hosts a number of postdoc fellows and trainees from the Dutch Cancer Foundation, The European Cancer Center, Amsterdam or abroad. Staff members lecture on various postdoctoral training courses such as the Elementary Course in Oncology of the Dutch Oncology Society and various courses and seminars organized by the European School of Oncology. Moreover, the NKI is co-organizer of national, postdoctoral Oncology Days for clinical oncologists. Teaching for medical students and doctors in training At the medical schools of the University of Amsterdam and Free University, staff members lecture on fundamental and clinical oncology in the undergraduate study programs. Two groups of 15 undergraduate medical students of the University of Amsterdam participated in a Courses at the OOA Graduate School January Pathogenesis of malignant lymphomas C Meijer, S Pals January-Feb Determinants of tumor behavior and G Peters therapeutic effects April Translational research in pediatric oncology P Voûte, L Smets September Fluorescence microscopy: new developments in genetic analysis P Van Diest, G Meijer, M Hermsen October The anatomy and histology of the house W Lamers, C Van Noorden mouse October Detection, treatment and biology of minimal residual disease in human solid tumors R Brakenhoff, G Van Dongen October Annual graduate student retreat R Plasterk, I Deltrap Oct-Dec Immunology (semester course) T Schumacher, A Kruisbeek November Immunotherapy R Scheper, A Van den Eertwegh, B De Gast November Killing Cancer Cells (symposium) A Berns, S Rodenhuis

20 three week institutional training course with lectures given by staff physicians covering the whole scope of clinical oncology. As a spin off medical students apply for a 3 months senior internship being actively involved in daily clinical practice and retrospective analyses of patients series. Clinical oncology courses were given to the residents of the institute and trainees in radiotherapy (NKI/AvL), general surgery (Academic Medical Center), plastic & reconstructive surgery (Academic Medical Center) and urology (Free University). Medical specialists participated in the NKI/AvL postgraduate oncology fellowship programs of medical oncology, surgical oncology and head & neck oncology- and surgery. Cancer Nursing Education The educational department of the NKI/AvL organized a post-basic, core curriculum course in cancer nursing for internal and external candidates. This course promotes the development of cognitive, psycho-motorial, interactive and reactive skills to practice oncology nursing with confidence and professionalism. The theoretical program consists of 175 hours of classroom instruction, including the the following compulsory modules: 1) introduction of the educational system and instruments; 2) professional quality and quality of care; 3) continuity of care; 4) communication and collaborative skills; 5) introduction to oncology; 6) cancer nursing and medical treatment. In addition to these modules, the students make a choice out of two non-compulsory modules. The on-site practical training consists of 15 months internship which takes place on both medical and surgical wards. During this period the student will encounter various aspects of oncology nursing, including diagnosis, treatment, palliative and terminal care as well as participation in clinical research trials. Further developments within the educational department include regular nursing grand-rounds which involve the participation of all nursing departments. Themes directed to development in cancer nursing care, cancer nursing practice and new developments are discussed. The educational department also organized various post-graduate refresher courses in the treatment and care of cancer patients. Clinical post-graduate courses The Department of Radiotherapy has a full training program for radiotherapists and radiographers. Foreign radiotherapists are regular guests at the department, as well as residents of other hospitals. The Departments of Medical Oncology and Surgery offer 6-12 months elective residencies for internists and surgeons in training, in cooperation with various universities and training clinics. Both departments have 1-3 year training programs for qualified surgeons and internists to specialize in oncology. The Department of Plastic and Reconstructive Surgery participates in the training program of the Academic Medical Center and has a full time position for a resident. The Department of Pathology has a 3 month training program for residents in diagnostic cytopathology, including fine-needle aspiration cytology, in cooperation with the universities of Amsterdam, Utrecht and Rotterdam. In addition, postdoctoral courses in diagnostic cytology and histopathology are regularly organized by the department. The Department of Surgery organized the yearly Surgeon s Week. The program involves multidisciplinary teaching lectures, demonstrations of surgical techniques, demonstration of patients and discussion of new modalities in diagnosis and treatment. A cyclic program for continuing education in Medical Oncology has been developed by the Department of Medical Oncology. It consists of a 2-year course of 10 meetings concluded by a symposium and meets the requirements of the European Examination in Medical Oncology. 21 EDUCATION

21 22 I Division of Cell Biology Division head Ed Roos Introduction The major cause of cancer deaths is metastasis, i.e. the dissemination of cells throughout the body, leading to tumor growth in distant, often vital organs. Metastasis is due, at least in part, to defective regulation of cell adhesion and motility, the main research theme in this division. We study adhesion to the extracellular matrix, mainly mediated by integrins, and between cells, mediated by cadherins. Examples are the assembly of hemidesmosomes, which are structures required for stable adhesion of epithelial cells to the basement membrane, and the loss of intercellular adhesion induced by oncogenes and by integrin signals. In addition, we focus on the transient activation of adhesion molecules as required for movement of cells, and essential for invasion into tissues during metastasis formation. Adhesion and motility are regulated by signals triggered by chemokines and growth factors that bind to cell surface receptors. In addition, integrins themselves act as such receptors and elicit signals upon binding to their ligands. These signals not only lead to immediate responses such as the further activation of other integrins during invasion, but also have long-term effects on gene transcription, cell proliferation and cell morphology. Elucidation of the latter effects is crucial for the understanding of loss of anchorage-dependence for proliferation and loss of intercellular adhesion in cancer cells, and this goal is pursued in the groups of both A Sonnenberg and J Collard. Rho-like GTPases, which control the activities of the cytoskeleton, have an essential role in metastasis. The group of J Collard was amongst the first to demonstrate this. These GTPases also affect gene transcription, and feature in virtually all signal pathways that regulate adhesion or are triggered by adhesion molecules. These GTPases are therefore studied in all three main research groups. Relevant signals for invasion, such as triggered by chemokines acting on G-protein-coupled receptors, studied in the group of E Roos, are elicited locally and quite transiently and therefore hard to detect with standard biochemical assays. This problem should be overcome by real-time imaging of signaling events in living cells as they move. Development of the required methodology is a major aim of K Jalink. One of the crucial regulators of cytoskeletal activity is the lipid PIP2, but its role is poorly understood. Jalink devised a method for real-time imaging of PIP2 in the plasma membrane, which should be of great help in elucidating this role in the near future. Adhesion mechanisms in metastasis E Roos PhD RDM Soede PhD 1 PJM Stroeken MSc 2 IS Zeelenberg MSc 3 MK Kamp 1 EMF Ruuls-Van Stalle 3 EAM Van Rijthoven 2 YM Wijnands 1 S Ligthart N Van Poppel M Boter J Kamphorst Group leader Graduate student Graduate student Undergraduate student Undergraduate student Trainee technician Trainee technician Role of the chemokine SDF-1 in lymphoma metastasis We showed previously that the pertussis toxin catalytic subunit, expressed in a T-cell hybridoma, blocked metastasis to all tissues. We have now observed the same for a B-lymphoma and a myeloid leukemia except that, remarkably, colonization of the bone marrow by the myeloid cells was not affected. The toxin blocks G i protein-transduced signals from certain receptors, including those for chemokines. A prominent candidate chemokine for being involved was SDF-1 because it is constitutively expressed in many of the invaded tissues. To study this, we made cells that no longer displayed the SDF-1 receptor CXCR4 on the cell surface. This was achieved by transfecting SDF-1 with an attached KDEL sequence that binds to the KDEL-receptor, which retains endoplasmic reticulum (ER) proteins in the ER. Newly synthesized CXCR4 binds to SDF-1-KDEL, it is also retained in the ER and transport of CXCR4 to the surface is thus prevented. This is illustrated in Figure I.1. Chemotaxis towards SDF-1, but not towards the chemokine TARC, was abrogated. Invasion into fibroblast monolayers was blocked but was restored when fibroblasts were pretreated with TARC, demonstrating the specificity for SDF-1. Most importantly, metastasis to many tissues was prevented. The cells grew normally

22 upon i.p. injection showing that the lack of metastasis is not due to immune rejection. The dissemination of the B- lymphoma and myeloid leukemia cells probably involves other chemokines. This is being investigated. Thus, dissemination of hematopoietic malignancies may be determined in part by expression of appropriate chemokine receptors. However, the signaling pathways triggered by these receptors should also be intact, as demonstrated by BW5147 T-lymphoma cells which are not invasive and do not metastasize. These cells express CXCR4 but do not respond to SDF-1 in a chemotaxis assay. Figure I.1 Generation of cells without the CXCR4 receptor for the chemokine SDF-1. Transfected SDF-1 with an attached KDEL sequence binds to the KDEL-receptor, which retains endoplasmic reticulum (ER) proteins in the ER, and to CXCR4 and thus blocks the transport of CXCR4 to the surface. As a result, the cells no longer respond to SDF-1. G i and G q proteins and Rho family GTPases: role in chemokine-induced LFA-1 activation Chemotaxis towards low concentrations of SDF-1 (1 ng/ml) through a filter coated with ICAM-1, a ligand for the integrin adhesion molecule LFA-1, was dependent on LFA-1 function, as was invasion. We studied the involved signaling pathways using pharmacological inhibitors as well as retroviral transduction of toxins or dominantnegative mutants of signaling molecules. Whenever invasion was blocked, this low SDF-1 chemotaxis was also affected. In contrast, chemotaxis towards high levels (100 ng/ml) of SDF-1, which is not dependent on LFA-1, was often not affected. In both cases, i.e. high and low SDF-1 levels, G i proteins were required as well as the RHO-like GTPase CDC42. However, LFA-1-dependent chemotaxis (low SDF-1) as well as invasion required in addition G q proteins, the RHOA GTPase and myosin activity. Remarkably, this shows that high SDF-1-induced chemotaxis proceeds without contraction of actin filament networks by myosin. Metastasis of T-cell hybridoma cells expressing dominant-negative mutants of CDC42, RHOA and G q was much reduced. In vivo growth upon i.p. injection was normal, showing that this was not due to immune rejection or an inherent inability to grow in vivo. The retroviral vector contained an intra-ribosomal entry site and GFP, to monitor expression levels and select cells with high expression. Cells expressing similar GFP levels from an empty vector metastasized normally, excluding an effect of a possible immune response against GFP. We have also introduced a dominant-negative mutant of G q into the myeloid leukemia cells. This blocked metastasis to liver and spleen as well as the bone marrow. Thus, whereas G i protein signals are only required for liver and spleen metastasis of these cells (see above), G q appears to be required for invasion of the myeloid cells into all tissues, including the bone marrow. LFA-1 on lymphoid cells needs to be activated. Inhibition of T-cell hybridoma invasion by pertussis toxin can be overcome by direct LFA-1 activation, indicating that the chemokine signal activates LFA-1. This signal is too weak and transient to be assessed in standard adhesion assays, but it can be mimicked by AlF 4 which activates heterotrimeric G-proteins including G i and G q. Indeed, similarly to low SDF-1 chemotaxis and invasion, AlF 4-induced adhesion to ICAM-1-coated substrates was dependent on CDC42, RHOA and myosin activity. The existence of a G q-rhoa-myosin pathway was further supported by the effect of a constitutively active G q mutant which caused the constitutive activation of a large proportion of RHOA. The transfectants aggregated extensively, due to activation of LFA-1 that binds to ICAM-2 expressed by these cells, and this was blocked by myosin inhibitors. LFA-1-induced LFA-1 activation: the role of ZAP-70 We showed previously that invasiveness of T-cell hybridoma cells was blocked by the ZAP-70 inhibitor piceatannol and by a truncated dominant-negative ZAP- 70 mutant. ZAP-70 is probably not involved in the chemokine-triggered LFA-1 activation, described above, since neither the inhibitor nor the mutant affected AlF 4- induced adhesion. We propose that ZAP-70 acts downstream of LFA-1, based on its involvement in LFA-1 mab-induced aggregation. At subsaturating concentrations, certain LFA-1 blocking mabs induce aggregation, mediated by the unoccupied LFA-1 molecules that bind to ICAM-2. Cross-linking of LFA-1 by the mabs is required. This phenomenon can be explained if the triggering of LFA-1 elicits a signal that leads to the activation of other LFA-1 molecules. This LFA-1-to-LFA-1 signal appears to involve ZAP-70 since aggregation is blocked by piceatannol and dominant-negative ZAP-70. CDC42, RHOA and myosin contraction were not required for this aggregation, in contrast to the chemokineinduced LFA-1 activation. We therefore propose a model for invasion, involving two different modes of LFA-1 activation. ZAP-70 is essential in T-cell receptor (TCR) signaling. The ZAP-70 tandem SH2 domains bind to ITAM sequences in the TCR-CD3 complex, that have been phosphorylated on tyrosine by Src-like kinases. Using SH2 mutants, we found that both SH2 domains are required for ZAP-70 activity in invasion, suggesting that 23 CELL BIOLOGY

23 24 CELL BIOLOGY binding to a phosphorylated ITAM is required. However, the T-cell hybridoma we used does not express the TCR- CD3 complex. Furthermore, the lack of effects of specific inhibitors and dominant-negative mutants indicate that Src-like kinases are not involved. Identification of the protein that binds to the ZAP-70 SH2 domains, as well as the kinase that phosphorylates this protein, is now a main aim. Role of the ß1 integrin cytoplasmic domain in metastasis of ESb lymphoma cells ESb lymphoma cells metastasize to liver and spleen, as do T-cell hybridoma cells, but they use the α4β1 integrin instead of LFA-1 (αlβ2). We previously made β1- deficient double knock-out (DKO) mutants by targeted deletion of both β1 genes. DKO cells had greatly reduced metastatic capacity, which was regained upon restoration of β1 expression by retroviral transduction. ESb cells invade into monolayers of the BMS2 murine bone marrow stromal cell line, and this invasion was completely blocked by α4 and β1 mabs. In agreement with this, DKO cells did not invade. We introduced a series of β1 cytoplasmic domain mutants into DKO cells and found that invasiveness into BMS2 monolayers correlated with metastatic capacity. Deletion of the C- terminal five residues, as well as conservative replacements (T>A) of the threonines between the two NPXY domains, blocked invasiveness and metastasis. The two mutants have in common that they can not bind to the recently identified ICAP-1 protein that interacts specifically with the β1 cytoplasmic domain. For the threonine mutant we demonstrated this by yeast two-hybrid analysis. We are presently investigating the possible role of this protein. Remarkably, conservative replacement (Y>F) of the two NPXY tyrosines had no effect on invasion and metastasis but abrogated α4β1-mediated adhesion to fibronectin, induced by the integrin activator Mn 2+. The same was observed with a β1β2 chimera, containing a β2 cytoplasmic domain with NPXF domains instead of NPXY as in β1. This shows that the requirements for invasion and adhesion are not the same. Conceivably, α4β1 has mainly a signaling function in invasion which may not require interactions with all proteins that are essential for stable adhesion. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Adhesion mechanisms in metastasis Soede RDM, Driessens MHE, Ruuls-Van Stalle L, Van Hulten PEM, Brink A, Roos E. LFA-1 to LFA-1 signals involve the ζ-associated protein-70 (ZAP-70) tyrosine kinase: relevance for invasion and migration of a T-cell hybridoma. J Immunol 1999; 163: Stroeken PJM, Van Rijthoven EAM, De Boer E, Geerts D, Roos E. Cytoplasmic domain mutants of β1 integrin, expressed in β1- knockout lymphoma cells, have distinct effects on adhesion, invasion and metastasis. Oncogene (in press). JG Collard PhD LCJM OomenPhD 4 A Malliri PhD 3 L Price PhD 5,2 T Reid PhD 6 EE Sander PhD 7 S Van Delft PhD 2 G Zondag PhD 8 A Bathoorn 2 L Janssen 9 JP Ten Klooster 8 RA Van der Kammen MR Ahmadian PhD 1 C Olivio PhD Genetic control of invasion and metastasis Group leader Academic staff Guest Guest RHO family proteins control dynamic cytoskeletal changes Rho-like GTPases, which include CDC42, RAC, and RHOA, have been implicated in the control of a wide range of biological processes. In particular, RHO-like proteins act as key control molecules in signaling pathways that determine the reorganization of the actin cytoskeleton in response to receptor stimulation. Similar to RAS proteins, RHO-like proteins cycle between the active GTP-bound state and the inactive GDP-bound state. The dynamic cytoskeletal changes regulated by RHO-like GTPases determine the morphology, adhesion and motility of cells, processes required for the metastatic spread of tumor cells. The invasion-inducing Tiam1 gene encodes an activator of Rac To identify genes involved in invasion and metastasis, we employed retroviral insertional mutagenesis in combination with in vitro selection of invasive T-lymphoma variants. This led to the identification of the invasioninducing Tiam1 gene which encodes an activator (GEF) of the RHO-like GTPase RAC. TIAM1 is highly expressed in brain and testis and at moderate levels in most other organs. The gene is also expressed in almost all tumor

24 cell lines that we have analyzed, such as B- and T-cell lymphomas, melanomas, neuroblastomas and carcinomas. Overexpression of TIAM1 in NIH3T3 fibroblasts causes a flat, pancake-shaped morphology with extensive membrane ruffling and many pinocytotic vesicles, similar to V12Rac1. We have developed pull down assays using GST-PAK- and GST-Rhotekin fusion proteins, to determine the activity state of CDC42, RAC and RHO in cells, in response to receptor stimulation or ectopic expression of genes. Using these assays, we confirmed that TIAM1 predominantly regulates the activity of RAC and demonstrated an unexpected cross-talk between RHO family proteins, as discussed below. Moreover, we have searched for other GEF proteins capable of activating RAC. One of the known GEFs, hpem-2, showed a very high homology to TIAM1 within the catalytic DH domain. HPEM-2 is a 70 kda protein and features an SH3 domain followed by a DH-PH domain combination. The gene is highly expressed in the brain but turned out to encode a specific activator of CDC42 instead of RAC. RHO-like GTPases and invasion of T-lymphoma cells Similar to TIAM1, constitutively active V12RAC induces an invasive phenotype in T-lymphoma cells. Activated V12CDC42 also induces invasion of T-lymphoma cells which is not caused by CDC42-mediated activation of RAC, suggesting that one of the many common downstream signaling pathways of CDC42 and RAC is instrumental in the induction of invasion. Activated V14RHOA potentiates invasion but fails, by itself, to mimic RAC and CDC42. However, C3-transferase treatment inhibits invasion of Tiam1- and V12Rac1-transduced cells, indicating that RHOA function is required, but not sufficient. We found that TIAM1/RAC-mediated effects on the actin cytoskeleton and on cell adhesion play a major role in T-lymphoma invasion. Expression of TIAM1 and V12RAC1 promotes integrin-mediated adhesion to various substrates and correlates with invasive capacity, suggesting that enhancing cell-substrate adhesion of lymphoid tumor cells promotes their invasion. TIAM1-RAC signaling inhibits invasion of epithelial cells In epithelial carcinoma cells, invasion and metastasis is often associated with reduced E-cadherin-mediated cellcell adhesion. Ectopic expression of TIAM1 in epithelial cells inhibits HGF-induced cell scattering and cell migration by increasing these cell-cell adhesions. In addition, TIAM1-RAC signaling inhibits invasion and migration of fibroblastoid Ras-transformed MDCK(-f3) cells by restoring E-cadherin-mediated adhesion and the epithelial phenotype. Interestingly, these TIAM1/RACinduced cellular responses are dependent on integrinmediated cell substrate interactions. On fibronectin and laminin-1, TIAM1/RAC signaling inhibits migration of MDCK-f3 cells by restoring E-cadherin-mediated cell-cell adhesion, whereas on collagen, RAC activation promotes motile behavior. Invasion and migration of epithelial cells are thus determined by a balance between invasioninhibitory cell-cell interactions and invasion-promoting cell-substrate interactions, both mediated by TIAM1-RAC signaling. Antagonistic role of RAC and RHO Earlier findings suggested that TIAM1 is involved in regulating cytoskeletal reorganization required during neuronal cell migration and/or neurite extension. Indeed, overexpression of TIAM1 induces cell spreading and affects neurite outgrowth in N1E-115 neuroblastoma cells. Cells overexpressing TIAM1 no longer respond to LPA-induced neurite retraction and cell rounding, processes mediated by RHO. We found that RAC- and RHO-mediated pathways oppose each other during neurite formation, and that a balance between these pathways determines neuronal morphology. We show that RAC antagonizes RHO by regulating threonine phosphorylation of the myosin II heavy chain (MHC). This RAC-mediated phosphorylation of MHC is Ca 2+ - dependent and accompanied by cell spreading and loss of cortical myosin II. The antagonistic role of RAC and RHO in the dynamic changes of the actin-myosin cytoskeleton may thus be regulated by RAC-induced phosphorylation of the myosin II heavy chain and RHOinduced phosphorylation of the myosin II light chain. Recently we found that RAC can also antagonize RHO directly by downregulation of RHO activity at the level of the GTPase. Cross-talk between RHO family proteins A linear activation cascade has been proposed for CDC42, RAC, and RHO as deduced from cytoskeletal changes induced by these GTPases. By determining the actual activation state we found, however, that both transient PDGF-induced and sustained RAC activation by TIAM1 or V12RAC downregulate RHO activity. Activation of CDC42 can also lead to downregulation of RHO activity. Neither V14RHO nor N19RHO affects RAC activity, suggesting unidirectional signaling from RAC towards RHO. Downregulation of RHO activity occurs independently of RAC-induced cytoskeletal changes and cell spreading. Moreover, RAC effector mutants which are defective in inducing cytoskeletal changes or Jun-kinase activation both downregulate RHO activity, suggesting that neither of these RAC signaling pathways are involved in the regulation of RHO activity. A model on the crosstalk between RHO family proteins is given in Figure I.2. RHO GTPases and epithelial-mesenchymal transition In fibroblasts, activation of RAC by TIAM1 induces an epithelial-like morphology. TIAM1/RAC signaling promotes functional cadherin-based adhesions (P-cadherin and N- cadherin) and inhibits migration of fibroblasts, as demonstrated in epithelial cells. This epithelial-like phenotype is characterized by RAC-mediated downregulation of RHO activity. Restoration of RHO activity in TIAM1-expressing cells by expression of V14RHO results in reversion of the 25 CELL BIOLOGY

25 26 CELL BIOLOGY Publications: Genetic control of invasion Geijsen N, Van Delft S, Raaijmakers JA, Lammers JW, Collard JG, Koenderman L, Coffer PJ. Regulation of P21Rac activation in human neutrophils. Blood 1999; 94: Gimond C, Van der Flier A, Van Delft S, Brakebusch C, Kuikman I, Collard JG, Fässler R, Sonnenberg, A. Induction of cell scattering and activation of Rho-like GTPases by expression of β1 integrins in β1-deficient epithelial cells. J Cell Biol, (in press). Michiels F, Collard JG. Rho-like GTPases: their role in cell adhesion and invasion. In: Cell behaviour: control and mechanism of motility, Portland Press, Biochem Soc Symp 1999; 65: Figure I.2 Model showing the crosstalk of RHO-like GTPases regulating cell spreading and cell migration. For explanation see text. epithelioid phenotype towards a migratory, fibroblastoid morphology. The reciprocal balance between RAC and RHO activity determines cellular morphology and migratory behavior in NIH3T3 fibroblasts. The proto-oncogene Ras is frequently mutated in epithelial tumors resulting in uncontrolled growth and transition towards an invasive, mesenchymal phenotype. In epithelial cells, we found that sustained oncogenic RAS signaling permanently downregulates RAC and upregulates RHO activity which is accompanied by epithelial-mesenchymal transition. Oncogenic RAS provokes changes in RAC and RHO activity through sustained activation of the RAF/MAP-kinase pathway, which causes transcriptional downregulation of Tiam1. Reconstitution of RAC activity by exogenous expression of TIAM1 decreases RHO activity and restores the epithelial phenotype in mesenchymal V12RAS- or RAF- CAAX-transformed cells. These findings reveal a novel mechanism by which oncogenic RAS regulates RAC and RHO activity to achieve epithelial mesenchymal transition. Reid T, Bathoorn A, Ahmadian MR, Collard JG. Identification and characterization of hpem-2, a Guanine Nucleotide Exchange Factor specific for Cdc42. J Biol Chem 1999; 274: Sander EE, Collard JG. Rho-like GTPases: Their role in epithelial cell-cell adhesion and invasion. Eur J Cancer 1999; 35: Sander EE, Ten Klooster J-P, Van Delft S, Van der Kammen RA, Collard JG. Rac downregulates Rho activity: reciprocal balance between both GTPases determines cellular morphology and migratory behavior. J Cell Biol 1999; 147: Stam JC, Collard JG. The DH protein family, Exchange Factors for Rho-Like GTPases. In: Progress in Mol and Subcell Biol Cytoskeleton and Small G proteins. Springer Verlag, 1999; 22: Van Leeuwen FN, Van Delft S, Kain HE, Van der Kammen RA, Collard JG. Rac regulates phosphorylation of the Myosin-II heavy chain, actinomyosin disassembly and cell spreading. Nature Cell Biol 1999; 1: Notes 1 Funding: Dr Mildred Scheel Foundation for Cancer Research, Project Funding: Dutch Cancer Society, Project NKI Funding: EEC-Bio-CT Department of Biophysics. 5 Funding: Netherlands Organization for Scientific Research, Project Funding: EEC Project ERBFMBICT Funding: Deutsche Forschungs Gemeinschaft (DFG), SA 763/ Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI

26 27 A Sonnenberg PhD E Danen PhD 1 L Fontao PhD 2 HAM Geerts PhD 3 C Gimond PhD 4 M Nievers PhD 5,11 H De Jonge MSc 6 J Koster MSc 7 LMTh Sterk MSc 6 A Van der Flier MSc 4 S Van Leeuwen MSc 9 MR Van Leusden MSc 8 D Kramer 4 M Kreft 10 I Kuikman 5 P Sonneveld 1 Receptors for matrix adhesion Group leader Graduate student Graduate student Graduate student Graduate student Graduate student Graduate student proteins that interact with the β1d splice variant. One of the interacting clones, isolated from a human skeletal library, encodes FHL1/SLIM1, a four and a half LIM-only protein. The zinc finger-like LIM motif is present in many proteins, including homeodomain-containing transcription factors, cytoskeletal proteins (e.g. paxillin and zyxin) and kinases (e.g. LIM-kinase) and mediates protein-protein interactions. FHL1/SLIM1 is expressed at high levels in skeletal and cardiac muscle and other tissues. Immunohistochemical analysis revealed that the protein is localized mainly at the Z-lines in striated muscle fibers and in the nuclei of myoblasts and various other cell types. This nuclear staining probably represents a nuclear-targeted FHL1/SLIM1 splice variant, as recently reported by others. GFP-tagged FHL1/SLIM1 is localized in focal contacts of a subpopulation of transfected cells. Unfortunately, we could neither reproduce the binding of β1d to FHL1/SLIM1 nor demonstrate this interaction in biochemical assays. In collaboration with Dr V Wixler and M Aumailley (Institute for Biochemistry, Cologne, Germany), we studied the interaction of integrins with another member of the LIM-only protein family. We found that FHL2/DRAL binds to the cytoplasmic domains of different α and β subunits and is colocalized with integrins in focal contacts. CELL BIOLOGY Role of the ß1 integrin splice variants in mechanotransduction We generated a mouse strain in which the β1d integrin subunit splice variant in skeletal and cardiac muscle was replaced by β1a. This had no effect on the histology or ultrastructure of these tissues, but the mice displayed mild functional abnormalities of the heart. The levels of atrial natriuretic peptide (ANP) and the β myosin heavy chain (β-mhc) were significantly increased in the ventricles. Increased expression of these proteins is also induced by high blood pressure in normal mice and is part of the physiological response to mechanical stress. There is increasing evidence that integrins are mediators of this response. Our findings with the β1d knockout mice suggest that β1a is more efficient than β1d in transducing mechanical stress into biochemical signals. To investigate whether the presence of β1a (instead of β1d) in combination with a high blood pressure would lead to a pathological condition of the heart, we crossed our β1d -/- mice with transgenic mice carrying the rat angiotensin gene (collaboration with M Bader, Max-Delbrück Centrum for Molecular Medicine, Berlin, Germany). Several mice were obtained that expressed β1a instead of β1d in skeletal and cardiac muscle and were positive for the transgene. However, despite high blood pressure, none of these mice developed a pathological defect of the heart. Thus, replacement of β1d by β1a and high blood pressure do not seem to have a cumulative effect. Identification of proteins involved in integrin-mediated signal transduction We used the yeast two-hybrid system to search for Role of the β4 cytoplasmic domain in hemidesmosome formation Hemidesmosomes are multiprotein complexes that mediate adhesion of epithelial cells to the basement membrane. The transmembrane components include BP180 and the integrin α6β4, which connect, via HD1/plectin and BP230, to the keratin filaments. The α6β4 integrin, which is a receptor for laminin-5, has a central role in hemidesmosome assembly. In patients with Figure I.3 Schematic representation of a hemidesmosome, indicating the molecular interactions potentially involved in its assembly. The α6ß4 integrin interacts extracellularly with laminin-5 and intracellularly with plectin to form a core to which BP180 and BP230 attach. The hemidesmosome is then stabilized by multiple protein-protein interactions.

27 28 CELL BIOLOGY junctional epidermolysis bullosa associated with pyloric atresia, this integrin is absent due to mutations in the genes of either α6 or β4. The β4 cytoplasmic domain plays an essential role. It is over 1,000 amino acids long and contains two pairs of fibronectin (FNIII) repeats separated by a connecting segment. Previously, we showed that the recruitment of plectin by α6β4 is mediated by sequences in the first and second FNIII repeats and the connecting segment. Our suggestion that plectin redistribution involves a direct interaction with β4 has now been confirmed. Using yeast two-hybrid analysis and a dot blot overlay assay, we showed that this β4 region interacts directly with an N-terminal plectin fragment, which contains an actin-binding domain. Further studies indicated that the β4 and actin binding sites on plectin overlapped and that binding of β4 to plectin, prevents interaction with F-actin. This mutually exclusive binding of F-actin and β4 to plectin explains why F-actin does not associate with hemidesmosomes and provides a molecular mechanism for a switch in plectin localization, from actin filaments to hemidesmosomes, when β4 is expressed. Finally, we confirmed that plectin interacts with different cytoskeletal networks by mapping of the C-terminally located binding site for several different intermediate filament proteins on plectin using yeast two-hybrid assays and transfection of plectindeficient MD-EBS keratinocytes. A model summarizing the protein interactions in hemidesmosome assembly is presented in Figure I.3. Ligand-independent hemidesmosome formation requires binding of HD1/plectin to the ß4 integrin subunit We previously showed that α6β4 can induce hemidesmosome assembly independently of binding to its ligand laminin-5. Our results suggested a role for plectin. Using MD-EBS keratinocytes derived from a patient lacking HD1/plectin, we show that HD1/plectin is indeed required for ligand-independent but not for ligand-dependent hemidesmosome assembly. No clustering of α6β4 or of BP180 and BP230 was seen in MD-EBS keratinocytes on fibronectin or collagen. Moreover, a chimera of the interleukin-2 receptor (IL2R) and a β4 cytoplasmic domain with the R1281W mutation, which abrogates β4 binding to HD/plectin, was diffusely distributed when expressed in β4-deficient PA-JEB keratinocytes. In contrast, a β4 R1281W mutant that binds ligand, in association with α6, was directed to sites where laminin- 5 was deposited and induced clustering of hemidesmosomal components, including BP180 and BP230. Thus, ligand-independent hemidesmosome formation requires binding of HD1/plectin to β4. Intriguingly, for recruitment of the IL2R/β4 chimera to pre-existing hemidesmosomes of plectin-deficient keratinocytes, the same β4 region is required that binds to plectin (residues ). Because this β4 region lacks the BP180 binding site and is unable to interact with itself or with endogenous β4, we suggest that it becomes incorporated into hemidesmosomes via interaction with an as yet unidentified hemidesmosomal component. Control of cell cycle progression by integrin-mediated adhesion Interaction with the extracellular matrix (ECM), mediated by integrins, is a prerequisite for cell cycle progression in most normal cells whereas many malignant cells can proliferate in the absence of ECM adhesion. In collaboration with Ken Yamada (NIH, Bethesda) we are studying the relevant integrin signals in 3T3 fibroblasts. We observed that the cells respond to growth factors and proceed to S-phase when attached to fibronectin via the integrin α5β1 but not on other ECM components. MAP kinase activation and cyclin D induction are similar for all substrates whereas subsequent Rb phosphorylation and cyclin A induction occurs only on fibronectin. Differences in the activity of the Rho GTPase play a role in the different responses. Currently, we are further investigating the nature of the stimulus provided by fibronectin. We created a reporter construct in which GFP is under control of the cyclin A promoter which allows us to isolate, by FACS sorting, cells that enter S-phase. We will use this system for a genetic screen to identify proteins involved in the regulation of G1 cell cycle progression by the ECM. Induction of cell scattering by RHO-like GTPases upon expression of β1 integrins in β1-deficient epithelial cells Adhesion receptors play a crucial role in morphogenesis. We studied how intercellular adhesion molecules and β1 integrins influence each other using two β1-null cell lines, epithelial GE11 and fibroblast-like GD25 cells. Expression of either β1a or the cytoplasmic splice variant β1d induced the disruption of intercellular adherens junctions followed by cell scattering in both GE11 and GD25 cells. In GE11 cells, the morphological change correlated with the redistribution of ZO-1 from tight junctions to adherens junctions at high cell confluency. In addition, the expression of β1 integrins caused a dramatic reorganization of the actin cytoskeleton and of focal contacts. Interaction of β1 integrins with their respective ligands was required for a complete morphological transition towards the spindle-shaped fibroblastlike phenotype. The expression of an IL2R-β1A chimera and its incorporation into focal adhesions also induced disruption of cadherin-based adhesions and the reorganization of ECM-cell contacts but failed to promote cell migration on fibronectin, in contrast to full-length β1a. This indicates that disruption of cell-cell adhesion is not simply the consequence of stimulated cell migration. Expression of β1 integrins in GE11 cells resulted in decreased cadherin and α-catenin levels accompanied by their redistribution from the cytoskeleton to the detergent-soluble fraction. Regulation of α-catenin protein levels by β1 integrins probably plays a role in the morphological transition, since overexpression of α- catenin prior to β1 prevented disruption of intercellular

28 adhesion and cell scattering. In addition, we found that expression of β1a, β1d or IL2R-β1A in GE11 or GD25 cells triggers activation of both RHOA and RAC1, but not of CDC42. Moreover, dominant-negative RAC1 (N17RAC1), RHOA (N19RHOA) or CDC42 (N17CDC42) inhibited disruption of cell-cell adhesion and scattering when expressed prior to β1. All three GTPases may be involved since expression of either N19RHOA, N17RAC1 or N17CDC42 reversed cell scattering and partially restored cadherin-based adhesion in GE11-β1A cells. Our results indicate that β1 integrins regulate the polarity and motility of epithelial cells, by inducing RAC- and RHO-mediated reorganization of both intercellular and ECM adhesion. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Fondation pour la Recherche Médicale (FRM). 3 Funding: Dutch Cancer Society, Project NKI Funding: Netherlands Heart Foundation, Project Funding: Dutch Cancer Society, Project NKI Funding: Dutch Kidney Foundation, Project C Funding: Dystrophic Bullosa Research Association. 8 Funding: Netherlands Organization for Scientific Research (NWO), Project Funding: Netherlands Heart Foundation, Project Funding: Dutch Cancer Society, Project NKI Funding: EEC, Project CHRX-CT Publications: Receptors for matrix adhesion Borradori L, Sonnenberg A. Structure and function of hemidesmosomes: more than simple adhesion complexes. J Invest Dermatol 1999; 112: Cohn RD, Mayer U, Saher G, Herrmann R, Van der Flier A, Sonnenberg A, Sorokin L, Voit T. Secondary reduction of a7b integrin in laminin α2-deficient congenital muscular dystrophy supports an additional transmembrane link in skeletal muscle. J Neurological Sci 1999; 163: De Melker AA, Sonnenberg A. Integrins: alternative splicing as a mechanism to regulate ligand binding and integrin signaling events. BioEssays 1999; 21: Geerts D, Fontao L, Nievers MG, Schaapveld RQJ, Purkis T, Wheeler GN, Lane EB, Leigh IM, Sonnenberg A. Binding of integrin α6β4 to plectin prevents plectin association with F-actin but does not interfere with intermediate filament binding. J Cell Biol 1999; 147: Gimond C, Van der Flier A, Van Delft S, Brakebusch C, Kuikman I, Collard JG, Fässler R, Sonnenberg, A. Induction of cell scattering and activation of Rho-like GTPases by expression of β1 integrins in β1-deficient epithelial cells. J Cell Biol, (in press). Nievers M, Schaapveld RQJ, Sonnenberg A. Biology and function of hemidesmosomes. Matrix Biol 1999; 18:#5-17. Sonnenberg A, Nievers M, Schaapveld RQJ, Geerts D, Niessen C, Borradori L. Interaction of BP180 and α6β4. J Invest Dermatol 1999; 112: Van der Neut R, CachaHo AS, Thorsteinsd\ttir S, Gimond C, Janssen H, Prins D, Bulthuis, J Van der Valk M, Calafat J, Sonnenberg A. Partial rescue of epithelial phenotype in integrin β4 null mice by a keratin-5 driven human integrin β4 transgene. J Cell Science 1999; 112: Wixler V, Laplantine E, Geerts D, Sonnenberg A, Petersohn D, Eckes B, Paulsson M, Aumailley M. Identification of novel interaction partners for the conserved membrane proximal region of integrin a cytoplasmic domains. FEBS Lett 1999; 445: K Jalink PhD J Van der Wal R Habets Biophysics in cell signaling Group leader Undergraduate student Cellular signals that are either very local or short-lived can easily escape detection by conventional biochemistry. A detailed understanding of these signals requires techniques to study cellular processes with high spatiotemporal resolution in single living cells. Biophysical techniques (electrophysiology, microspectrometry, flash photolysis) offer this high resolution, but were, until recently, limited to parameters such as membrane potential and intracellular ion concentrations. The introduction of Green Fluorescent Protein (GFP) has rapidly increased the number of parameters that can be monitored, by protein tagging. In addition, spectral GFP mutants allow direct imaging of molecular interactions in the cell by Fluorescent Resonance Energy Transfer (FRET). FRET occurs when suitable donor and acceptor fluorophores come into close proximity, and serves as a cellular ruler with Ångström resolution. We have established a biophysical laboratory with the sensitive equipment to study these phenomena, and have characterized the physicochemical behavior of FRET between two wavelength-shifted GFP mutants (the Cyan-FP/Yellow-FP FRET pair). Detection of membrane phosphatidylinositol bisphosphate (PIP2) levels in vivo in single cells Inositolphospholipids in the plasma membrane play important roles in many signaling events. Upon activation, 29 CELL BIOLOGY

29 30 CELL BIOLOGY phospholipase C cleaves PIP2 to yield the second messengers diacylglycerol (DAG) and inositol-trisphosphate (IP3). PIP2 itself also has a signaling role, in particular in the regulation of the cytoskeleton. In vitro, it binds to a number of cytoskeletal proteins, including gelsolin, ezrin, cortactin and profilin. In the dynamic cytoskeleton, signals are short-lived and localized, and thus need to be investigated with high resolution techniques in living cells. We have adapted a strategy that uses a fusion construct of GFP with the PIP2-binding pleckstrin homology (PH) domain from PLC-γ1 (GFP-PH). Upon transfection with this construct, cells are imaged using confocal microscopy. At rest, the construct binds to PIP2, and cells show a brightly fluorescent plasma membrane. Following stimulation with agonists, PIP2 is hydrolyzed and GFP fluorescence rapidly translocates to the cytosol. Translocation is complete within 10 seconds. Subsequently, the majority of GFP-PH gradually returns to the membrane in 2-5 minutes. In line with a localized signaling role, GFP-PH fluorescence does not appear to be distributed uniformly over the membrane. We observed enhanced fluorescence in lamellipodia, filopodia and growth cones, as well as bright isolated patches throughout the membrane. PIP2 spots are dynamic, showing slow changes in size and location. When cells are stimulated with bradykinin, the spots disappear from the membrane. Remarkably, after desensitization of the signaling cascade, PIP2 reappears at the same hot spots. These spots thus reflect the existence of functionally specialized lipid patches in the membrane (opposing a view in which PIP2 is distributed uniformly in the membrane by passive diffusion). Origin and cellular function of these patches are currently under study. Improved PIP2 sensing using FRET Before PH-domain sensing can be routinely used to image PIP2 in vivo, a number of potential pitfalls have to be addressed. First, local enrichment in GFP-PH conceivably reflects local differences in PIP2 accessibility rather than concentration. Second, on and off rates depend, at least in part, on diffusion of GFP-PH, limiting the time resolution to seconds. Third, the specificity for PIP2 over other cellular molecules has to be rigorously demonstrated. We are currently addressing the first two issues by exploiting the FRET principle. We constructed a novel sensor that remains attached to the plasma membrane at all time and exhibits intramolecular FRET as a function of membrane PIP2 content. This sensor is distributed uniform over the PM and reports local PIP2 concentration as a shift in emission wavelength, thus demonstrating the existence of PIP2 patches. This construct will function as a template for the development of other lipid sensors by substituting the PH domain with protein motives with different specificity. FRET sensing also exhibits superior time resolution (Figure I.4). The potential of this approach to increase throughput of in vitro lipid biochemistry is also being investigated. Publication: Biophysics in cell signaling Miyawaki A, Llopis J, Mizuno H, Jalink K, Tsien RY. Cameleons as cytosolic and intra-organellar calcium probes. In: Calcium Signalling: A Practical Approach, Oxford University Press (in press). Immuno EM facility Figure I.4 FRET-based PIP2 sensing. A. A single cell expresses both CFP- PH and YFP-PH. Initially, both chimeras are bound to membrane PIP2 and thus located within FRET range. During PIP2 hydrolysis, the constructs dilute out into the cytosol and FRET ceases. B. Simultaneous emission recording at 475 nm and at 530 nm from a N1E-115 neuroblastoma cell, excited at 430 nm. The ratio between these two signals (lower trace) depicts PIP2 content. J Calafat PhD JWRM Janssen Academic staff Immuno-EM study on ARF6 and RAC1-mediated ruffling in macrophages 1 Activation of RAC1, a member of the RHO family of GTPases, is associated with multiple cellular responses,

30 including membrane ruffling and focal adhesion formation. The mechanisms by which RAC1 is coupled to these functional responses are not well understood. To determine whether ARF6 is required for RAC1-dependent cytoskeletal responses in macrophages, we studied macrophages that co-expressed wild-type (WT) or guanine nucleotide binding-deficient (T27N) ARF6 and activated RAC1 (Q61L), or that were stimulated with colony stimulating factor-1 (CSF-1). ARF6 T27N, but not ARF6 WT, inhibited ruffles induced by RAC1 Q61L or CSF-1. In contrast, ARF6 T27N did not inhibit RAC1 Q61L-induced focal adhesion formation and did not impair CDC42 Q61L-induced formation of filopodia. Cryoimmunogold electron microscopy demonstrated: 1) the presence of ARF6 in membrane ruffles induced by either CSF-1 or RAC1 Q61L; 2) addition of CSF-1 to macrophages led to the redistribution of ARF6 from the interior of the cell to the plasma membrane, suggesting that this growth factor triggers ARF6 activation; 3) direct targeting of RAC1 to the plasma membrane did not bypass the blockade in ruffling induced by ARF6 T27N, indicating that ARF6 regulates a pathway leading to membrane ruffling that occurs after the activation and membrane association of RAC1. Subcellular localization of the bactericidal/permeabilityincreasing protein (BPI) in neutrophils 2 BPI binds to lipopolysaccharides (LPS) and is bacteriostatic and bactericidal for many Gram-negative bacteria. We investigated the subcellular location of BPI in neutrophils. BPI co-localized with myeloperoxidase (MPO), a marker for azurophil granules, and had the same distribution pattern as CD63, a transmembrane protein. This suggests that BPI is membrane-associated in the azurophil granules in neutrophils. Induction of selective release of azurophilic granules by the Na + - ionophore monensin resulted in fusion of endosomes with azurophil granules, leading to the formation of large vacuoles containing MPO, CD63, and BPI. After phagocytosis of serum-treated zymosan (STZ), BPI was detected in phagosomes, both in association with membranes as well as in the lumen, suggesting the release of BPI into activated compartments. The results show that BPI is present in azurophil granules, is probably primarily membrane-associated and is relocated after activation, following the same route as MPO and CD63. Collaboration within the Institute We provided immuno-em analysis for other groups in the Institute. In collaboration with A Sonnenberg (this Division), we localized hemidesmosome components. With J Neefjes (Division of Tumor Biology), we studied the architecture of the Class II compartments (MIIC), the factors involved in retention and release of MIIC and the mechanism of fusion of MIIC with the plasma membrane. In collaboration with P Borst, we localized the variant transferrin receptor of T Brucei. Notes 1 In collaboration with S Greenberg of the Columbia University, New York. 2 In collaboration with A Egesten of the University Hospital MAS, Malmo, Sweden. Publications: Immuno EM facility Calafat J, Janssen J,. Knol EF, Egesten A. The bactericidal/- permeability-increasing protein (BPI) is membrane-associated in azurophil granules of human neutrophils, and relocation occurs upon cellular activation. APMIS (in press). Fernandez-Borja M, Wubbolts R, Calafat J, Janssen H, Divecha N, Dusseljee S, Neefjes J. Multivesicular body morphogenesis requires phosphatidyl-inositol 3-kinase activity. Curr Biol 1999;9: Gromme M, Uytdehaag FG, Janssen H, Calafat J, van Binnendijk RS, Kenter MJ, Tulp A, Verwoerd D, Neefjes J. Recycling MHC class I molecules and endosomal peptide loading. Proc Natl Acad Sci USA 1999; 96: Van den Oudenrijn S, De Haas M, Calafat J, Van der Schoot CE, Von dem Borne AE. A combination of megakaryocyte growth and development factor and interleukin-1 is sufficient to culture large numbers of megakaryocytic progenitors and megakaryocytes for transfusion purposes. Br J Haematol 1999; 106: Van der Neut R, CachaHo AS, Thorsteinsd\ttir S, Gimond C, Janssen H, Prins D, Bulthuis J, Van der Valk M, Calafat J, Sonnenberg A. Partial rescue of epithelial phenotype in integrin β4 null mice by a keratin-5 driven human integrin β4 transgene. J Cell Science 1999; 112: Zhang Q, Calafat J, Janssen H, Greenberg S. Arf6 is required for growth factor- and Rac-mediated membrane ruffling in macrophages at a stage distal to Rac membrane targeting. Mol Cell Biol 1999; 19: Secretary Division of Cell Biology NS Borsje-Maeyer Research staff positions (full time equivalents) Scientific permanent: 5.0 Scientific project: 22.0 Technical permanent: 10.0 Technical project: CELL BIOLOGY

31 32 II Division of Molecular Carcinogenesis Division head René Bernards Introduction The Division of Molecular Carcinogenesis consists of four independent research groups that study cell cycle regulation, DNA mismatch repair, the polycomb group of transcriptional repressors and three dimensional protein structure. Even though these subjects appear quite diverse at first glance, there has been a considerable convergence of research interests between the groups within the division. Most notably, the group of M Van Lohuizen identified a connection between the polycomb family protein BMI1 and the cell cycle machinery. The generation, by the group of H Te Riele, of fibroblast cell lines that lack all three members of the retinoblastoma family (prb, p107 and p130) by gene knockout technology has led to productive collaborations with the cell cycle group of Bernards. In addition to the overlap in research interests, the members of the division benefit from joint technology development. Both the groups of Van Lohuizen and Bernards have in the past year used functional genetic screens in mammalian cells to identify novel genes in cancer-relevant pathways. The further study of these genes promises to make the first year of the new millennium as exciting as the past year. Functional analysis of dominant and recessive oncogene products R Bernards PhD R Agami PhD 1 RM Kerkhoven PhD 2 DS Peeper PhD 3 A Shvarts PhD 4, 5 J Van Der Sman PhD 6 D Whyatt PhD 1 RML Zwijsen PhD 7 K Berns MSc 8 T Brummelkamp MSc 8 H Masselink MSc9, 10 Graduate J Pouwels MSc 6 BGPH Scheijen MSc 2 PM Voorhoeve MSc 5 R Oosterkamp MD 11 M Bronk 6 S Douma 3 EM Hijmans CJM Loomans 7 T Kwaks N Vastenhouw Group leader Graduate student Graduate student student Graduate student Graduate student Graduate student Clinician Undergraduate student Undergraduate student Cell cycle control proteins: E2F and cyclin D11, 2, 6, 11 E2F transcription factors regulate the expression of groups of genes that act in the G1 and S phases of the cell cycle and in the induction of apoptosis. E2F DNA binding activity consists of a heterodimeric complex that contains one of five related E2F proteins and one of two DP proteins. To study the activities of the five E2Fs in control of cell growth and differentiation in vivo, we have generated transgenic mice that express the five E2Fs under the control of a lymphoid-specific promoter. E2F-1 transgenic mice were small due to a specific defect in endochondral bone formation (Figure II. 1). However, no defects in lymphoid differentiation or apoptosis were seen. Because of the poor fertility of E2F1-transgenic animals, we generated a second series of animals in which the E2F1 transgene is inducible by the CRE recombinase. These animals are currently being crossed to several mouse strains that express the CRE recombinase in a tissue-specific fashion. This approach allows us to direct expression of the E2F1

32 transgene to selected tissues without the problems associated with the constitutive E2F1 transgene. Elevated expression of E2F4 and 5 is well tolerated in transgenic animals and does not lead to any detectable pathology. This may be due to the fact that the inactivating pocket protein partners of these E2Fs (p107 and p130) are present in molar excess to the E2Fs, leading to functional inactivation of the over-expressed E2F. To test this model directly, we crossed E2F4 transgenic mice to p107 -/- and p130 -/- mice to generate compound E2F4 transgenic/ pocket protein knockout mice. Interestingly, the presence of both the E2F4 transgene and p107 null was an early embryonic lethal combination due to massive apoptosis in the peripheral nervous system, whereas the E2F4 transgene and p130 null combination was tolerated during development. D-type cyclins are major downstream targets of extracellular signaling pathways which act to transduce mitogenic signals to the cell cycle machinery. Transcriptional induction of D-type cyclins occurs in response to a wide variety of mitogenic stimuli, including the RAS signaling cascade and the APC-β-catenin- TCF/LEF pathway. In addition, cyclin D1 protein turnover and subcellular localization is highly regulated during the cell cycle. Because of their critical role in linking cytoplasmic signals to nuclear responses, it is perhaps not surprising that D-type cyclins are frequent targets of mutagenesis in various types of cancer. We recently found that genotoxic stress also leads to rapid proteasome-mediated degradation of cyclin D1, but not of cyclin D2 or D3. Interestingly, this genotoxic stress induced degradation does not involve phosphorylation of threonine 286 of cyclin D1, which serves as the target of GSK-3β-mediated degradation of cyclin D1. Through mutational analysis, we identified a novel motif in cyclin D1 that mediates genotoxic stress-induced degradation. Insertion of this motif in cyclin D2 rendered cyclin D2 sensitive to genotoxic stress, which underscores the relevance of the identified motif. We propose that rapid degradation of cyclin D1 after exposure to genotoxic stress allows transfer of CDK inhibitors like p21 and p27 from cyclin D1/CDK complexes to cyclin E/CDK2 complexes, which causes rapid cell cycle arrest independent of the presence of functional p53. Our current experiments are aimed at testing this model. Cellular gene products that interact with viral oncoproteins9, 10 Several years ago, we identified a novel cellular protein, named BS69, based on its ability to interact with the adenovirus E1A oncoprotein. BS69 has several protein motifs that suggest a role in transcriptional regulation, including a PHD finger, a bromo domain and a MYND domain. We therefore tested the ability of BS69 to modulate synthetic reporter genes and found that BS69 has strong transcriptional repression activity, which required the presence of the MYND domain. Our data indicate that the MYND domain in BS69 mediates interaction with the co-repressor N-CoR. Interestingly, E1A interferes with BS69 repression, suggesting that E1A can interfere with BS69-containing chromatin remodeling complexes through interaction with BS69. Cell cycle regulation by c-myc 8 The c-myc gene encodes a transcription factor that acts to stimulate progression through the cell cycle. Several transcriptional targets of c-myc have been identified but it is not clear which of these targets are 33 MOLECULAR CARCINOGENESIS Figure II. 1 Defect in bone formation in E2F1 transgenic mice. Delayed endochondral ossification in the skull bones of E2F1 transgenic mice at E17.5. Skull bones were stained for cartilage (light gray) and bone (darker gray). Note that the interparietal bone (designated: I) and the supraoccipital bone (designated: S) show delayed ossification, whereas the frontal bone and parietal bone, which are formed by intramembranous ossification, ossify normally in E2F transgenic animals.

33 34 MOLECULAR CARCINOGENESIS crucial to the growth-promoting effects of c-myc. Rat1 fibroblasts in which both alleles of the c-myc gene have been inactivated by gene targeting manifest a stable slow growth phenotype. To identify genes which can complement the growth defect of c-myc null cells, we infected these cells with retroviral cdna expression libraries. Infected cells were plated at low density and large, fast-growing, variants were isolated. In total we identified some 30 fast-growing variants using three different cdna libraries. PCR amplification of proviral cdna inserts allowed identification of the cdnas with rescue activity. These data revealed that two thirds of the rescued clones harbored a full-length c-myc cdna and the remaining one third harbored a N-Myc cdna. These data suggest that only Myc family members can rescue the growth defect in c-myc null fibroblasts and that none of the identified cell cycle targets of c-myc alone is able to substitute for c-myc in cell proliferation. Functional screens to identify genes that rescue Rasinduced senescence 3, 4, 5 Expression of an activated RAS oncogene in primary cells leads to premature senescence, whereas expression of RAS in immortalized cells induces transformation. Several lines of evidence indicate that genetic inactivation of the p53 pathway allows transformation by RAS. In the past year, we have invested in the development of novel strategies to identify genes that rescue senescence induced by an activated RAS oncogene in cultures of primary cells. We infected primary mouse fibroblasts with retroviral cdna expression libraries and a retrovirus encoding an activated RAS oncogene and selected for foci of transformed cells that grew in a background of senescent cells. At this point we have established several cell lines that continue to grow in the presence of an activated RAS oncogene. Importantly, mobilization of the integrated provirus followed by second round selection in primary cells led to a dramatic increase in the number of transformed foci. This suggests that during the first round of selection the population of cdna library plasmids was significantly enriched for cdnas that allow rescue of RAS induced senescence. At present we have isolated several candidate genes that we are studying in more detail for their effects on RAS-induced senescence. Notes 1 Funding: EMBO-fellowship. 2 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Biomed 2-program. 5 Funding: Human Frontiers Science Program, Project RG M. 6 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Prolifix Ltd, UK. 10 Funding: Center for Biomedical Genetics. 11 Division X. Publications: Functional analysis of oncogene products Bernards R. CDK-independent activities of D type cyclins. Biochim Biophys Acta 1999; 1424: M Di Fiori B, Guarguaglini G, Palena A, Kerkhoven RM, Bernards R, Lavia P. Two E2F sites control growth-regulated and cell cycleregulated transcription of the Htf9-a/RanBP1 gene through functionally distinct mechanisms. J Biol Chem 1999; 274: Voorhoeve PM, Watson RJ, Farlie PG, Bernard R, Lam EW. Rapid dephosphorylation of p107 following UV irradiation. Oncogene 1999; 18: Voorhoeve PM, Hijmans EM, Bernards R. Functional interaction between a novel protein phosphatase 2A regulatory subunit, PR59, and the retinoblastoma-related p107 protein. Oncogene 1999; 18: Voorhoeve PM [Dissertation]. Dephosphorylation of the retinoblastoma family of proteins by protein phosphatase type 2A. Utrecht: University of Utrecht, H Te Riele PhD L Jansen PhD 1 N Claij MSc 2 J-H Dannenberg MSc 3 SS De Vries MSc 1 C Brouwer HMJ Dekker L Schuijff 3 Y Van Klink 2 K Van t Wout 3 N Saridjan Mouse models for hereditary cancer Group leader Graduate student Graduate student Graduate student Undergraduate student Research in our group is focused on 1) cancer predisposition by loss of DNA mismatch repair functions; 2) the role of the retinoblastoma gene family in cell cycle control and tumor suppression. The principle tools include gene inactivation in embryonic stem (ES) cells and analyses of the phenotypic consequences in homozygous, heterozygous and chimeric knockout mice and in cell lines derived thereof. DNA mismatch repair 1, 2 Cancer predisposition in HNPCC, the non-polyposis form of familial colon cancer, is caused by defects in

34 post-replicative DNA mismatch repair (MMR). This system recognizes and restores mismatched nucleotides that can arise by erroneous DNA replication, bypass replication of damaged DNA or incorporation of modified nucleotides. In addition, mismatches can arise in heteroduplex regions that are formed at initial stages of recombination between homologous but non-identical DNA molecules. Mismatch recognition activity in mammalian cells is attributed to two heterodimeric protein complexes composed of different MutS homologs (referring to the paradigmatic Escherichia coli muts,l postreplicative MMR system): MutSα, a dimer of MSH2 and MSH6, and MutSβ, a dimer of MSH2 and MSH3. These complexes have specific and redundant mismatch recognition capacity. Thus, while MSH2 deficiency ablates the activity of both dimers, leading to strong cancer predisposition in mice and men, loss of either MSH3 or MSH6 function only causes a partial MMR defect with possibly less severe consequences. This may explain the rarity of MSH6 and absence of MSH3 germline mutations in HNPCC families. To test this hypothesis, we have studied the phenotypic consequences of single and combined homozygous defects in mouse Msh3 and Msh6 genes. Msh6-deficient mice were highly cancer prone, the majority of animals developing lymphomas or epithelial tumors originating from the skin and uterus. Most lymphomas appeared within 30 weeks; in this period the tumor incidence in Msh6 -/- animals (9/12) was as high as that in Msh2 -/- Figure II. 2 Tumor susceptibility of mismatch repair deficient mice. Msh6- deficient mice are highly cancer prone; however, intestinal tumorigenesis in these animals requires simultaneous loss of Msh3. animals (31/19). In contrast to Msh2-deficient animals, Msh6-deficient mice only rarely developed intestinal tumors: two were found among 22 animals at a mean age of 52 weeks. Msh3 deficiency did not predispose to cancer, nor did it accelerate lymphomagenesis in Msh6- deficient animals. However, combined Msh3/Msh6- deficiency strongly predisposed to development of intestinal tumors, which were found in eight out of thirteen Msh3-/-Msh6-/- animals at a mean age of 30 weeks (Figure II. 2). In addition to mutation avoidance, we have previously demonstrated two additional functions of MMR in ES cells. 1) Msh2 deficiency caused tolerance to methylating agents; 2) In wild-type ES cells, but not in Msh2-deficient cells, homologous recombination was strongly suppressed by base sequence divergence as small as 0.6%. We now observed that Msh6 deficiency also caused loss of mismatch-dependent anti-recombination, whereas Msh3-deficient cells behaved identically to wildtype cells. Furthermore, we found that Msh6 deficiency, but not Msh3 deficiency, caused tolerance to N-methyl- N'-nitro-N-nitrosoguanidine. Thus, besides redundant functions, Msh6 also has specific functions that are not shared with Msh3. Previous studies by others have addressed the mismatch recognition capacity of Msh2/Msh3 and Msh2/Msh6 complexes: base/base mismatches and single unpaired nucleotides are preferentially targeted by MutSα, while MutSa and MutSβ are redundant in loop repair with possibly some specificity of MutSb for larger (four/five nucleotides) loops. The activity of MutSβ towards single nucleotide loops is controversial, with some assays contradicting and others supporting MutSbdependent repair capacity. Moreover, we observed in Msh6 -/- cells residual binding activity to DNA oligonucleotide sequences carrying a G/T mismatch or an unpaired dinucleotide, which completely disappeared on concomitant inactivation of Msh3. Thus, both MutSa and MutSβ can bind to G/T and extrahelical dinucleotide mismatches, whereas free Msh2 can not. Although MutSa activity exceeds that of MutSβ, on an equimolar basis the two complexes appear to have similar affinities for both types of mismatches. This may be indicative of at least some base/base mismatch repair capacity of MutSβ. Our data demonstrate two MMR functions to be specific for Msh6: mismatch-dependent anti-recombination and mediation of toxicity of methylating agents. Since lack of MutSα alone is sufficient to cause strong predisposition to development of lymphoid tumors and epithelial tumors of the uterus and skin, these tumors apparently ensue from genetic alterations that are specifically suppressed by MutSα function. These may include point mutations, or perhaps frameshifts in runs of mononucleotides, but also methylation-damage-induced point mutations and mitotic recombination events. On the other hand, the redundancy of MutSa and MutSβ functions was hypothesized to account for the rarity of MSH6 and lack of MSH3 germ line mutations in HNPCC 35 MOLECULAR CARCINOGENESIS

35 36 MOLECULAR CARCINOGENESIS families. This view is supported by our studies in mice: simultaneous loss of MutSα and MutSβ activity is required to induce intestinal tumors. Thus, specific mutations, presumably frameshifts that are suppressed by MutSβ, are rate limiting in intestinal carcinogenesis. Based on these observations, we speculate that families segregating a defect in MSH6 are cancer-prone but often not recognized as HNPCC families due to suppression of intestinal tumorigenesis by MutSβ activity. The retinoblastoma gene family 3 Loss of function of the retinoblastoma suppressor prb is a common event in the development of many tumor types in humans. prb belongs to the family of so-called pocket proteins which also includes p107 and p130. These proteins may functionally overlap in blocking progression through the cell cycle in response to exogenous factors. Indeed, we found evidence for functional overlap of tumor suppression functions of prb and p107 in mice: development of retinoblastoma required simultaneous ablation of both proteins. Since early embryonic lethality associated with prb deficiency in mice precluded a study of the consequences of prb deficiency at later developmental stages and adulthood, we choose an alternative approach. We generated embryonic stem (ES) cell lines carrying inactivating mutations in Rb alone and in both Rb and p107, and studied of their developmental and tumorigenic capacity in chimeric mice. While Rb -/- chimeras developed normally into adulthood, most Rb -/- ;p107 -/- chimeric embryos died in utero. Surviving chimeras, however, developed retinoblastoma. This result shows that in mice, p107 acts as a suppressor of oncogenic transformation of prb-deficient retinoblasts and is the first demonstration that p107 can act as a tumor suppressor. This result also demonstrates that the chimera approach provides a powerful tool to study the consequences of combinations of mutations, which in normal crossings would cause embryonic lethality. To further optimize this approach, genetically modified ES cells were labeled with a ubiquitously expressed LacZ marker gene, which allows their descendants to be precisely traced in chimeric mice. This refinement allowed us to demonstrate that the vast majority of Rb/p107-deficient retinoblasts disappeared from the developing retina through apoptosis. We speculate that retinoblastoma development requires at least one additional event that counteracts apoptotic cell death. We have now generated an isogenic panel of LacZlabeled ES cell lines with the following genotypes: wild type, Rb-/-; double knockouts Rb -/- ;p107 -/-, Rb -/- ;p130 -/- and p130 -/- ;p107 -/- and triple knockout Rb -/- ;p107 -/- ; p130 -/-. None of these genotypes affected the growth potential or characteristics of ES cells. Unlike wild-type ES cells, the triple knockout cells failed to arrest upon in vitro induction of differentiation, yielding a cell line with high growth potential. Also, unlike ES cells, these cells had lost the capacity to grow in nude mice, suggesting that some sort of acquired tumor suppressor function substituted for the absence of prb family members. We are currently studying the interplay between prb, p107 and p130 control pathways in development and oncogenic transformation by following the fate of mutant ES cells in chimeric embryos. The proliferative capacity of these cells will be assessed by combining LacZ staining with BrdU incorporation and apoptosis assays on histological sections. Specific antibodies will be used to identify specific cell types. In addition, we have used mutant ES cells to derive primary mouse embryonic fibroblasts (MEFs) from mid-gestation chimeric embryos, taking advantage of the presence selectable marker genes in mutant but not blastocyst-derived cells. MEFs could be obtained from all genotypes including the triple knockout. Thus far, we observed that all single and double knockouts underwent replicative senescence upon prolonged passaging which, like in wild-type cells, was associated with increased levels of p19. Interestingly, in late passage Rb -/- ;p107 -/- MEFs foci developed with enhanced growth capacity apparently as a result of an additional mutagenic event. This phenomenon was not observed in any of the other double knockout combinations. Triple knockout cells, however, appeared to be immortal from the beginning and grew to high density, despite abnormally high levels of p19. Thus, ablation of all three pocket proteins causes resistance to replicative senescence. We will use mutant MEFs to determine the effect of each genotype on the response of cells to a large number of growth inhibiting conditions and their sensitivity to oncogenic transformation. Detailed knowledge of the function of each of the pocket proteins in cellular growth control will provide insight into the question why abrogation of this pathway is a mandatory step in the development of human cancer. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Mouse models for hereditary cancer Claij N, Te Riele H. Microsatellite instability in human cancer: a prognostic marker for chemotherapy? Exp Cell Res 1999; 246: Claij N, Van Klink Y, Dekker M, Jansen L, Beijnen J, Te Riele H. Interactions of DNA mismatch repair and DNA damage: implications for the etiology and treatment of cancer. Mutat Res (in press). De Vries SS, Baart EB, Dekker M, Siezen A, De Rooij DG, De Boer P, Te Riele H. Mouse MutS-like protein Msh5 is required for proper chromosome synapsis in male and female meiosis. Genes & Dev. 1999; 13:

36 De Wind N, Dekker M, Claij N, Van Klink Y, Radman M, Riggins G, Van der Valk M, Van t Wout K, Te Riele H. HNPCC-like cancer predisposition in mice through simultaneous loss of Msh3 and Msh6 mismatch-repair protein functions. Nat Genet 1999: 23; Jansen L, Claij N, Dekker M, Van Klink Y, Van der Valk M, Van t Wout K, Te Riele H. Acceleration of lymphomagenesis in mismatch-repair deficient mice by exposure to genotoxic agents. Toxicology Letters (in press). Toft NJ, Winton DJ, Kelly J, Howard LA, Dekker M, Te Riele H, Arends MJ, Wyllie AH, Margison GP, Clarke AR. Msh2 status modulates both apoptosis and mutation frequency in the murine small intestine. Proc Natl Acad Sci USA 1999; 96: Bmi1, M33, Ring1b and Mll, respectively. We have previously shown that BMI1 interacts with M33, MPH1, RING1b and other Pc-G proteins in large complexes. More recently, we identified a separate Pc-G complex, containing ENX1, ENX2 and EED, that is required during early development when Pc-G repression is initiated. We recently investigated the subnuclear distribution and chromatin association of BMI1-containing Pc-G complexes in detail, using a combination of sensitive immunohistochemical methods and biochemistry. In primary human cells, BMI1 shows a fine-grain distribution over chromatin, whereas additional large subnuclear domains are observed in tumor cell lines. These were identified as associations with distinct heterochromatic regions of certain chromo- 37 MOLECULAR CARCINOGENESIS Role of mouse Polycomb-group genes in transcriptional repression and tumorigenesis M Van Lohuizen PhD P Keblusek PhD 1 A Lund PhD 2 M C Motta PhD 3 J W Voncken PhD 4 JJL Jacobs MSc 5 K Kieboom 5, 6 E Verhoeven M Verlaan 1 E Wientjens Group leader Graduate student We are studying the mechanism of stably inherited transcriptional repressed- or activated states of individual target genes by respectively the Polycomb-group (Pc-G) and Trithorax-group (Trx-G) protein complexes in the mouse, and the effects of deregulation of Pc-G and Trx-G genes on Homeobox (Hox) gene expression, development, cell cycle control and tumorigenesis. Functional and Biochemical characterization of Bmi1 and associated Pc-G protein complexes 4, 5 Bmi-1was found as an oncogene that synergizes with c-myc in mouse B and T cell lymphomagenesis. Bmi1 is the first example of the expanding mouse Polycomb-group transcriptional repressor family. Pc-G proteins and the counteracting Trithorax-group (Trx-G) of nucleosome remodeling factors are involved in maintenance of proper gene expression patterns during development at the level of chromatin structure. Important targets include the Hox gene clusters and critical cell cycle regulatory genes (see below and Figure II. 3). To study Pc-G and Trx-G function we focus on representative members of these groups: Figure II. 3 How Polycomb-group proteins connect to cell cycle regulation. Pc-G complexes recognize specific binding sites (PREs, Polycomb-responsive elements) and form, through an as yet poorly understood mechanism, a stable, repressive chromatin structure on specific target genes. Our recent results identified the tumor suppressors p16 and p19arf as critical cell cycle targets for Pc-G. Unbalanced mitogenic signaling mediated by the activation of oncogenes such as cmyc and H-RAS activate as a fail-safe mechanism these cell cycle repressors in normal cells, leading to senescence or apoptosis, rather than immortalization. In contrast, Pc-G complexes act to repress this tumor suppressor locus. Overexpression of the Pc-G protein Bmi1 leads to inadvertant downregulation of p16 and p19arf, leading to immortalization and powerful cooperation with cmyc and H- RAS in tumorigenesis, whereas inhibition of Pc-G function by deleting Bmi1 leads to overexpression of p16 and p19arf and induces premature senescence.

37 38 MOLECULAR CARCINOGENESIS somes. We recently discovered a marked cell cycle regulation of Pc-G-chromatin interaction: nuclear Pc-G staining dissipates in late S-phase and is rapidly reestablished in early G1 phase. Chromatin association of BMI1 and MPH1 inversely correlates with their phosphorylation status: at G1/S hypophosphorylated BMI1 and MPH1 is present in chromatin-associated nuclear protein fractions, whereas during G2/M, hyperphosphorylated BMI1 and MPH1 complexes dissociate from chromatin, although a small but significant fraction of BMI1 remains associated with specific sites on mitotic chromosomes. These results provide new insights in the dynamic regulation of Pc-G complex-chromatin association, and suggest that extra-cellular cues may influence this process through kinase/phosphatase pathways (collaboration with D Schweizer, University of Vienna). Generation and phenotypic analysis of new Polycomb-group mutant mice and genetic interactions of Pc-G and Trx-G4, 5, 6 We and others recently identified RING1b as a novel Pc-G member. RING1b interacts with a number of Pc-G proteins, including BMI1. For this interaction, the RING finger domains of both proteins are essential. To investigate the function of Ring1b null mutant mice have been generated (collaboration with P Krimpenfort and A Berns, Division VII). Unlike other single Pc-G gene knockout mice (which display various defects but are viable, see below), Ring1b null mutant mice die shortly after gastrulation (day 8-9). This may relate to the central positioning of the RING1b protein in the Pc-G complex, and in addition suggests that the closely related Ring1a gene cannot compensate for loss of Ring1b. This severe phenotype resembles that of the Eed null mutant mice (the only Pc-G homolog known so far to be represented by a single gene), and suggests that loss of Ring1b results in complete loss of Pc-G function. This phenotype is being studied in detail by in situ hybridization with early developmental markers (In collaboration with J Deschamps, Hubrecht laboratory, Utrecht). In addition, we are generating Ring1b-conditional knockout mice, which will be a valuable tool to study Pc-G function in relation to differentiation and development in a timed and cell type specific manner. To address the opposing roles of Pc-G and Trx-G in vivo, we collaborate with S Korsmeyer (Dana Farber, Boston). Recent results of this study, investigating the effects of combining Bmi1 and Mll (a Trx-G gene) mutations in mice, showed that Bmi1/Mll double knockout mice display a partial restoration of some of the Hox gene expression boundaries. This confirms the notion that also in mammals, Trx-G and Pc-G complexes act in opposition on specific (Hox) target genes. Pc-G genes and cell cycle regulation1, 5, 6 Bmi1 transgenic mice have expanded pro-b and pro-t cell compartments, culminating in a high incidence of B and T cell lymphomas. In contrast, Bmi1 knockout mice show severe proliferative defects of lymphoid and myeloid cells, resulting in severe lymphopenia. Proliferation defects are not restricted to lymphoid cells: primary embryo fibroblasts (MEFs) and neurons in the cerebellum of Bmi1 knockouts also show such defects, resulting in neurological disorders. We recently found that these defects are due to increased levels of the INK4aencoded tumor suppressors p16 and p19arf, that are critical regulators of the Cyc.D/CDK4,6;pRB and the p53 tumor suppressor pathways (see Figure II. 3). Conversely, overexpression of Bmi1 facilitates immortalization, and neoplastic transformation by repressing p16 and p19arf. The dramatic rescue of defects and proliferative capacity in INK4a/Bmi1 double knockout mice proved that p16 and p19arf are prime targets of BMI1. This connects Pc- G repression to cell cycle control and control of the replicative senescence checkpoint governed by these tumor suppressors. We have shown that loss of cellularity in Bmi1 knockouts is caused by INK4a/Arf-induced apoptosis. Moreover, we found that the basis for strong co-operativety between Bmi1 and cmyc in tumorigenesis is the capacity of Bmi1 to repress p19arf-mediated apoptosis. In addition, we found clear dosage effects of p16/p19arf in controlling proliferation and apoptosis, and showed that loss of p16/p19arf prevents apoptosis upon matrix-detachment of Myc+Ras transformed MEFs. To elucidate the molecular mechanism of Pc-G repression, we are studying the INK4a locus as a relevant in vivo target, for the presence of Pc-G responsive elements and Pc-G-associated alterations in chromatin structure. Genetic screens to identify new regulators of cellular senescence We have made use of the tight premature senescencearrest of Bmi1 knockout MEFs, to screen for mutations or overexpressed cdnas that can bypass the p16/p19arfmediated arrest. Application of MuLV proviral-insertional mutagenesis resulted in immortalized cell clones; these are currently being screened for common insertion sites. In a separate approach using tumor-derived retroviral cdna library transductions, we identified at least five cdnas other than Bmi1 that are capable of potently immortalizing Bmi1- null mutant MEFs (collaboration with G Daley, Whitehead Institute, Cambridge, USA). The first of these cdnas analyzed in detail, revealed a potent transcriptional repressor of p19arf. The role and mechanism of action of these putative oncogenes, and their interference with the cellular senescence/anti-immortalization fail-safe in mouse and human cells, is the topic of intensive study. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Danish Medical Research Council. 3 Funding: EC-Marie Curie training fellowship, Project MCFI Funding: Netherlands Organization for Scientific Research (NWO-ALW), Project Funding: Dutch Cancer Society, Project NKI Funding: Netherlands Organization for Scientific Research (NWO-MW) PIONIER, Project 0705/

38 Publications: Polycomb-group genes in transcriptional repression and tumorigenesis Hanson RD, Hess JL, Yu BD, Ernst P, Van Lohuizen M, Berns A, Van der Lugt NMT, Shashikant CS, Ruddle FH, Seto M, Korsmeyer J. Mammalian Trithorax and Polycomb-group homologs are antagonistic regulators of homeotic development. Proc Natl Acad Sci 1999; 96: Jacobs JLJ, Kieboom K, Marino S, DePinho RA, Van Lohuizen M. The oncogene and Polycomb-group gene Bmi1 regulates cell proliferation and senescence through the INK4a locus. Nature 1999; 397: Jacobs JLJ, Van Lohuizen M. Cellular memory of transcriptional states by Polycomb-group proteins. Cell & Dev Biol 1999; 10: Jacobs JLJ, Scheijen B, Voncken J-W, Kieboom K, Berns A, Van Lohuizen M. Bmi-1 collaborates with c-myc in tumorigenesis by inhibiting c-myc induced apoptosis via INK4a/ARF. Genes & Dev 1999; 13: Lessard J, Schumacher A, Thorsteinsdottir U, Van Lohuizen M, Magnuson T, Sauvageau G. Functional antagonism of the Polycomb-Group genes Eed and Bmi1 in hemopoietic cell proliferation. Genes & Dev 1999; 13: Van Lohuizen M. The trithorax-group and Polycomb-group chromatin modifiers: implications for disease. Curr Opin Genet Dev 1999; 9: Voncken JW, Schweizer D, Aagaard L, Sattler L, Jantsch MF, Van Lohuizen M. Chromatin-association of the Polycomb group protein BMI1 is cell cycle regulated and correlates with its phosphorylation status. J Cell Science 1999; 112: TK Sixma PhD K Brejc PhD 1 L Jovine PhD 2 P Knipscheer MSc 3 M Lamers MSc 4 J Enzlin 4 D Van der Riet 2 WJ Van Dijk H Winterwerp Protein Crystallography Group leader Graduate student Graduate student Development of cancer is generally due to errors in molecular pathways. Structural biology can help to understand these errors at the atomic level, by studying the proteins and the DNA involved. We use X-ray crystallography as a tool to provide three-dimensional structures. These will provide more insight in the molecular processes, but they can also be of great help in designing new drugs. In our group we focus mainly on proteins involved in DNA stability (transposition, mismatch repair) and in cell cycle control. Structure of proteins in cell cycle control 2, 3 Cyclin D1 is an important regulator of the cell cycle because of its effect on cyclin dependent kinases (CDKs). Recently, however, R Zwijsen in this institute showed that cyclin D1 can also positively regulate the estrogen receptor in a CDK-independent fashion and may thus have an additional role in regulating growth of breast epithelial cells. Up-regulation of estrogen receptor activity is a consequence of the direct physical interaction of cyclin D1 with the ligand-binding domain of the estrogen receptor. This binding and activation is seen both in the presence and absence of estrogens and anti-estrogens. Interestingly, cyclin D1 can also bind directly to the estrogen receptor coactivator SRC1. This suggests that cyclin D1 can act as a bridging factor between the estrogen receptor and its coactivators in the absence of hormone. This mechanism of estrogen receptor activation could contribute to growth of cyclin D1 amplified breast cancers, especially those that are refractory to the anti estrogen tamoxifen The interaction between cyclin D1 and the estrogen receptor and its coactivators are therefore attractive drug targets. We initiated the crystal structure determination of the cyclin D1/estrogen receptor ligand-binding domain in the presence of fragments of coactivator. Since cyclin D1 is a very labile protein it is difficult to produce it in sufficient quantities. Therefore we have resorted to coexpression of the three partners that we want to co-crystallize. Ubiquitin-mediated targeting is a mechanism for transmitting various signals in the cell, such as controlled breakdown of short-lived cellular proteins and targeting to the nucleus. Ubiquitin attachment is regulated through a ubiquitin-activating enzyme E1, a ubiquitin conjugating enzyme E2 (or Ubc) and a ubiquitin ligase E3. We have solved the structure of Ubc9, an E2 protein. Despite the fact that several E2 structures are known, the precise details of the regulation are not understood at the atomic level. The G1/S transition in the cell cycle is regulated by breakdown via the ubiquitin conjugating enzyme Ubc3 (Cdc34) and an E3 ligase, consisting of Skp1, a Cullin homolog (Cdc53) and an F-box protein (e.g., Cdc4, Grr1, Skp2). In this E3-ligase complex or SCF, the F-box protein is thought to mediate the target specificity. In humans cyclin A/CDK2 is part of this complex. We aim to determine the structures of the proteins in this ubiquitin-mediated pathway, in order to understand ubiquitin transfer in atomic detail. 39 MOLECULAR CARCINOGENESIS

39 40 MOLECULAR CARCINOGENESIS Structure determination of mismatch repair proteins 4 Mutations in DNA mismatch repair genes predispose for the most prevalent type of familial cancers, hereditary non-polyposis colon carcinoma (HNPCC). This type of DNA repair is specific for a single mismatch or a small stretch of unpaired bases in DNA. It involves a cascade of proteins that is highly conserved from bacteria to humans. The initial step of the repair is the recognition and binding of mismatched DNA by the MutS protein (E. coli) or its homologs (eukaryotes). This complex is then recognized by the MutL protein or its homologs. Although the proteins involved are known, it is unclear how even subtle mutations, as they are found in some HNPCC tumors, which involve the change of a single conserved amino acid in MSH2, can cause a deficiency in mismatch repair resulting in cancer. We study the structure/function relation of MutS, MutL and the eukaryotic homologs in detail by determining the X-ray structure of the proteins, alone and in complex with mismatched DNA. We have obtained crystals of MutS in complex to a mismatched DNA oligomer, diffracting to 2.7 Å. Initial problems in reproduction of the crystals have been overcome and an extensive derivative search has taken place, using both MIR and MAD derivatives. Nonisomorphism, pseudo-symmetry and twinning as well as the size of the complex (~230 kda or 460 kda in the asymmetric unit) are complicating the structure determination. Currently, systematic variation of crystallization conditions is used to improve the isomorphism of the crystals for solving the phase problem. Meanwhile we are studying the ATPase activity by mutations of the active site. Crystal structure of the DNA binding domain of Tc3 transposase in complex with DNA Transposable elements are small stretches of DNA that can move from one position in the genome to another. Excision and insertion of the transposon DNA into the genome is performed by the transposase protein, which is encoded by the transposon. The initial step in the process is recognition and binding of the transposase to the transposon DNA. In collaboration with the group of R Plasterk (Division V), we are extending our studies into the DNA binding domain of the Tc3 transposase (Figure II. 4). Crystals have been obtained of the bipartite DNA binding domain complexed to a 25-mer fragment of transposon DNA. Meanwhile we are studying the crystal structure of full-length Tc1. Notes 1 Funding: Technology Foundation STW VBI Funding: Dutch Cancer Society, Project NKI Funding: Netherlands Organization for Scientific Research (NWO-MW): Project Funding: Dutch Cancer Society, Project NKI Secretary Molecular Carcinogenesis M Sonne Research staff positions (full time equivalents) Scientific permanent 4.0 Scientific project 23.0 Technical permanent 4.8 Technical project 11.0 Figure II. 4 Dimer of DNA binding domains of transposase Tc3A bound to transposon DNA, as seen in the crystal structure. Further analsis is required for determining the biological relevance of this dimerization.

40 III Division of Cellular Biochemistry 41 Division head Wouter Moolenaar Introduction Major research topics in the Division of Cellular Biochemistry concern the mechanisms of signal transduction via cell surface receptors, particularly in the context of cell proliferation, differentiation and survival, and the role of lipid messengers, nuclear phosphoinositides and Smad transcription factors in determining cell behavior. The division comprises five research groups, each headed by a permanent staff member: J Borst, N Divecha, P Ten Dijke, W Moolenaar and W Van Blitterswijk. The group of Moolenaar investigates the multiple actions of the serum-borne mitogen lysophosphatidic acid (LPA). LPA acts through specific G protein-linked receptors belonging to the so-called Edg receptor family, which couple to activation of the Ras and Rho GTPases to stimulate cell proliferation and changes in cell morphology, respectively. Mitogenic signaling by LPA and other growth factors was shown to depend on dynamin, a key regulator of endocytosis. Dynamin function was found to be critical during the last step in the Ras-MAP kinase cascade. LPA signals through at least three distinct Edg receptors. Interestingly, the Edg-2 receptor was found to couple to activation of the Rac GTPase, thereby promoting cell spreading, whereas other LPA receptors preferentially couple to the Rho GTPase to stimulate cell contraction. Rho/Rac signaling is thought to be fundamental to tumor cell invasion. Newly identified Rho-binding proteins include a Rho-specific GDP/GTP exchange factor (p190rhogef) which binds to activated Gα q, and a scaffold protein (p116rip) that binds f-actin and probably participates in cytoskeletal responses to LPA. Furthermore, LPA receptors couple to activation of rapid conductance changes, notably an increase in chloride conductance that was found to be mediated by the G 13 protein. LPA also inhibits connexin-based gap junctional communication between adjacent cells, apparently via a Gαq signaling pathway involving both protein tyrosine kinase and phosphatase activity. Since impaired cell-cell communication is a hallmark of transformed cells, elucidation of how LPA regulates connexin function is of obvious importance. The group of J Borst studies the regulation of lymphocyte activation and survival. Central is the analysis of antigen receptor signaling and its interplay with signaling by members of the tumor necrosis factor (TNF) receptor family. With the discovery of c-cbl as an antigen receptor substrate, the group has entered the field of receptor ubiquitination and endocytosis. This line will be pursued, since it is becoming evident that regulated protein degradation is an important regulatory principle in signal transduction. Within the TNF receptor family, there are both death and survival promoting receptors such as CD95 and CD27, respectively. Variant T cells have been identified that fail to undergo CD95-mediated apoptosis. These cells are inhibited in the apoptosis signaling pathway at a newly defined control point, upstream from the mitochondria. It is the aim to identify both facilitating and inhibitory components in the trajectory leading to mitochondrial activation. In collaboration with the Van Blitterswijk group, it was established that ceramide does not play a second messenger role in this pathway. Instead, ceramide production was found to be secondary to redistribution of plasma membrane sphingolipids during the apoptotic effector phase. Finally, analysis of CD27-deficient mice has revealed that CD27 does not promote antigen receptorinduced cell cycle progression but mediates T cell survival. The molecular basis of this effect is currently under study. Van Blitterswijk and coworkers established that a newly cloned diacylglycerol kinase is a downstream target of RhoA and, furthermore, is found in both the cytosol and the nucleus. The latter finding raises the interesting possibility that diacylglycerol metabolism may play a role in nuclear signaling. Furthermore, in a collaborative project with M Verheij (Division IX), it was found that certain alkyllysophospholipids, used in the clinic as anti-cancer agents, trigger stress pathways and apoptosis in tumor cell lines, apparently by interfering with phosphoinositide metabolism at the plasma membrane. Other aspects of lipid signaling are explored by N Divecha and coworkers. This group studies phosphoinositide metabolism in the nucleus, particularly the role of inositol lipid kinases therein. While the exact function of inositol lipids in the nucleus is not yet clear, it seems likely they play an important role in transmitting signals from the cytosol to the nucleus. In September 1999, we welcomed our new staff member P Ten Dijke (formerly the Ludwig Institute for Cancer Research, Uppsala, Sweden). Research in the

41 42 CELLULAR BIOCHEMISTRY Ten Dijke group will focus on TGF-β serine/threonine kinase receptors and their intracellular effectors, termed Smad proteins. Now that so many new and exciting aspects of receptor-to-nucleus signaling pathways are being studied in the division, we trust that our science will flourish in the coming years. Signal transduction by G protein-coupled receptors WH Moolenaar PhD O Kranenburg PhD 1 H Lacerda PhD 2 FR Postma PhD 4 F Van Leeuwen PhD 3 BNG Giepmans MSc 4 J Mulder MSc 3 FPG Van Horck MSc 5 GM Hengeveld L Ran 6 I Verlaan I Van Etten Group leader Graduate student Graduate student Graduate student Lysophosphatidic acid (LPA) and sphingosine-1- phosphate (S1P) are serum-borne lysophospholipids that act on their cognate G protein-coupled receptors to trigger a host of cellular responses, ranging from rapid cytoskeletal changes and ionic conductance changes to stimulation of cell proliferation and survival. LPA and S1P signal through the so-called Edg family of receptors, comprising at least seven members (Figure III.1). Edg-2 receptor signaling The properties of the various Edg receptors are only just beginning to be characterized. We have obtained cdnas for three different LPA receptors (Edg-2, Edg-4 and Edg-7) and are now exploring their signaling properties in more detail, using selected cell lines which Figure III.1 The Edg family of G protein-coupled receptors. Edg2, Edg4 and Edg7 are specific LPA receptors, whereas Edg1, Edg3, Edg5 and Edg6 are receptors for S1P. are devoid of endogenous LPA receptors. Reorganization of the actin cytoskeleton is one of the earliest responses to Edg receptor activation and underlies changes in cell shape, adhesion and motility. LPA/S1P-induced cytoskeletal remodeling is mediated by the small GTPase RhoA. In neuronal cells, LPA induces RhoA-mediated cell rounding and neurite retraction. We have found that, in addition to RhoA, both the small GTPase Rac and calcium entry may play an important role in LPA-induced cytoskeletal remodeling. Ectopic expression of the Edg-2 receptor in B103 neuroblastoma cells restores calcium signaling and MAP kinase activation in response to LPA. LPA stimulation leads to prominent activation of Rac and subsequent cell spreading and increased motility. These effects appear to be G i/o-mediated. We conclude that LPA stimulation can either induce Rho-dependent cell rounding or Racmediated cell spreading, and that these differences reflect activation of different Edg family members. Rho signaling In neuronal cells, LPA and S1P induce rapid withdrawal of developing neurites and rounding of the cell body. In other cell types, LPA induces the formation of actin stress fibers. These effects are mediated by the RhoA GTPase which stimulates the so-called ROCK/ROK family kinases. These cause enhanced phosphorylation of the myosin light chain leading to contraction of the cortical actin cytoskeleton. Using transfection studies in both COS and neuroblastoma cells, we found that LPA signals RhoA activation and cytoskeletal contraction via the G 12/13 subfamily of heterotrimeric G proteins. RhoA, like other Ras-related small GTPases, is activated by guanine nucleotide exchange factors (GEF s) that stimulate the exchange of GDP for GTP. We have isolated a Rho-specific GEF (p190rhogef) that binds to and activates RhoA but not the related Rac or Cdc42 GTPases. In contrast to the isolated catalytic DH/PH domain, full length RhoGEF is inactive when tested in vitro. By deletion mutant analysis we are currently testing the nature of this intrinsic inhibitory activity. Interestingly, we observed that activated Gα q, but not Gα 12 or Gα 13, binds to the catalytic DH/PH domain of RhoGEF. We are currently testing the effect of this binding on RhoGEF activity (collaboration with MR Ahmadian, Dortmund). Activated Gα q does not promote Rho-like cytoskeletal rearrangements but, instead, exerts a destabilizing effect on actin stress fibers and promotes the neurite outgrowth. This raises the possibility that Gα q binding may serve to inhibit rather than stimulate RhoGEF activity. In neuronal cells RhoGEF localized to the cytoplasm, but also to discrete regions underneath the plasma membrane. In addition, we find a fraction of RhoGEF localized to microtubules. Of note, Rho activity is regulated by microtubule (de)polymerization. Thus, p190rhogef activity may be regulated by the polymerization state of microtubules. We will test the possibility that p190rhogef may serve a role in interconnecting the tubulin- and actin-based cytoskeletons.

42 Rho-regulated re-arrangement of the actin cytoskeleton involves not only the generation of force through the activation of myosin, it may also involve altered functioning of actin-binding proteins. We have identified a putative scaffold protein, termed p116rip, that directly interacts with both RhoA and actin filaments and, hence, is an attractive candidate to mediate changes in actin dynamics. The actin-binding domain of p116rip is subject to protein kinase C (PKC)-mediated phosphorylation with no apparent contribution of the ROCK/ROK kinases. PKC-activating phorbol ester has a dramatic effect on cytoskeletal architecture and receptor-mediated PKC activation has been implicated in the control of actin dynamics. A major goal in these studies is to assess the role of p116rip in the control of f-actin dynamics. We are currently testing the effects of purified full-length p116rip on microfilament behavior and the consequences of PKC-mediated phosphorylation of the actin-binding domain. Intriguingly, p116rip is not only localized to f-actin networks, it also resides in the nucleus. We are exploring a possible role for p116rip in nuclear signaling. Rac signaling: identification of myosin II heavy-chain kinase Myosin IIA is the major motor protein in non-muscle cells and associates with f-actin to regulate cell contraction. We have recently identified a Myosin II heavy chain kinase activity in rat PC12 cells and mouse NIE- 115 cells, which is transiently activated by agonists such as bradykinin and LPA. Activity of this kinase appears to be regulated by the small GTPase Rac and is dependent on calcium entry. Activation of this kinase correlates with enhanced cell spreading. Although the kinase involved has not been identified, current evidence suggests that Myosin II heavy chain phosphorylation serves to promote Myosin II filament disassembly and the breakdown of cortical actin/myosin, which leads to a loss of contractility and cell spreading. We are in the process of testing this hypothesis, with emphasis on investigating the regulatory role of calcium. Moreover, we assembled a partial cdna representing a putative human Myosin II heavy chain kinase. This cdna clone, which contains the complete catalytic domain of this kinase, was obtained from a human expressed sequence tag (EST) database. We are currently trying to acquire a full-length cdna and to express the kinase domain in E. coli. This should allow us to perform in vitro kinase assays using purified or recombinant Myosin IIA as a substrate. Ras-MAP kinase signaling: a role for dynamin-regulated endocytosis Mitogen-activated protein kinases (MAPKs) are key regulators of fundamental cellular processes like proliferation and differentiation. Ligands for tyrosine kinase receptors and G protein-coupled receptors, but also phorbol ester which bypasses cell-surface receptors, all stimulate MAPK activity. We have found that the signal transduction cascade from activated LPA receptors to MAPK activation involves the tyrosine phosphorylation of dynamin, a GTPase that regulates receptor endocytosis. At present it is unclear how tyrosine phosphorylation affects dynamin function. Yet, it appears that MAPK activation critically depends on the function of dynamin, as evidenced by the use of a dominant-negative dynamin-mutant (K44A) that interferes with clathrinmediated endocytosis. We find that dynamin is required for the activation of MAPK, but not for the activation of upstream elements including Ras, Raf and MEK. Indeed, we find that MEK is activated at the plasma membrane, even though the majority of MEK is cytosolic. It thus appears that internalization of activated MEK is a ratelimiting step in the MAPK activation cascade. We are currently examining the precise intracellular sites of activation of MEK and MAPK as well as the trafficking routes of these activated signaling intermediates. The G13 protein couples LPA and S1P receptors to chloride-mediated membrane depolarization LPA and S1P induce rapid neurite retraction and cell rounding in neuronal N1E-115 cells through G 12/13- mediated activation of the RhoA GTPase. We have discovered that LPA- and S1P-induced RhoA activation is accompanied by a prominent, long-lasting membrane depolarization (inward current) due to an increase in chloride conductance. As a result, action potential generation is suppressed for many minutes. The heterologously expressed Edg-2 receptor can mediate the effect of LPA. Ca 2+ -mobilizing agonists that do not activate RhoA, such as bradykinin and neurotransmitters, fail to activate the chloride current. Activation of the chloride current can be dissociated from RhoA activation and changes in cell shape, and requires neither inositol lipid hydrolysis nor Ca 2+ mobilization. Microinjected antibody against Gα 13 subunits prevents activation of the chloride current, whereas antibodies to Gα 12 and Gα i do not. Thus, the G 13 protein couples LPA and S1P receptors to a bifurcating signaling pathway leading to: 1) RhoA activation; 2) activation of chloride channels (Ca 2+ - and volume-insensitive) with consequent membrane depolarization. We propose that G 13-mediated membrane depolarization may serve to modulate neuronal activity and stimulate Ca 2+ entry during Rho-regulated neurite remodeling under both physiological an pathological conditions. Inhibition of gap junctional communication by LPA LPA rapidly disrupts connexin43 (Cx43)-based gap junctional communication in normal fibroblasts, through a pathway that involves the c-src tyrosine kinase but otherwise is poorly understood. Expression of an activated mutant of Gα q minics this effect of LPA. The downstream effector of Gα q is phospholipase C (PLC), which hydrolyses PtdIns(4,5)P2. We found that the Cx43 C-terminus can bind PtdIns(4,5)P2 in vitro. This suggests a model in which PtdIns(4,5)P2 levels modulate the accessibility of the Cx43 C-terminus to signalling proteins. Furthermore, we found that Cx43 co-precipitates with the receptor tyrosine phosphatase RPTPµ. Pervanadate 43 CELLULAR BIOCHEMISTRY

43 44 CELLULAR BIOCHEMISTRY treatment results in tyrosine phosphorylation of Cx43 and decreased cell-cell communication. How RPTPµ may affect connexin function remains to be elucidated. To search for additional binding partners of Cx43, cytosolic proteins were pulled down using a GST-fusion protein containing the Cx43 C-terminal tail from metabolically labeled cell lysates. A major Cx43-binding protein of around 55 kda was purified and, through microsequencing, identified as tubulin. How tubulin binding may affect Cx43 function in intact cells is currently under investigation. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: EMBO. 3 Funding: Dutch Cancer Society, Project NKI Funding: Dutch cancer Society, Project NKI Funding: Netherlands Organization for Scientific Research (NWO/SLW), Project Funding: Dutch Cancer Society, Project NKI J Borst PhD A De Melker PhD 4 L Smit PhD 4 J Hendriks MSc 2 AD Tepper MSc 5 A Werner MSc 1 I Bontjer 1 E De Vries G Van der Horst R Kikkert MSc 3 Regulation of lymphocyte activation Group leader Graduate student Graduate student Graduate student Guest Publications: Signal transduction by G protein-coupled receptors Feiken E, Van Etten I, Gebbink MFBG, Moolenaar WH, Zondag GCM. Intramolecular interactions between the juxtamembrane domain and the phosphatase domains of receptor protein tyrosine phosphatase RPTPµ: regulation of catalytic activity. J Biol Chem 1999 (in press). Kranenburg O, Verlaan I, Moolenaar WH. Gi-mediated tyrosine phosphorylation of Grb2-bound protein 2)-bound dynamin-ii by lysophosphatidic acid. Biochem J. 1999; 339: Kranenburg O, Poland M, Van Horck FPG, Drechsel D, Hall A, Moolenaar WH. Activation of RhoA by lysophosphatidic acid and Gα 12/13 subunits in neuronal cells: induction of neurite retraction. Mol Biol Cell 1999; 10: Kranenburg O, Verlaan I, Moolenaar WH. Dynamin is required for the activation of mitogen-activated protein (MAP) kinase by MAP kinase kinase. J Biol Chem 1999; 274: Moolenaar WH. Bioactive lysophospholipids and their G proteincoupled-receptors. Exp Cell Res 1999; 253: Moolenaar WH. Development of our current understanding of lysophospholipids. In: Annals NY Acad Sci, Goetzl EJ, Lynch KR, editors. Lysophospholipids and eicosanoids in cancer and in cardiovascular and neurodegenerative diseases. New York: NY Acad Sci 1999 (in press). Postma FR [Dissertation]. Cellular responses to lysophosphatidic acid and sphingosine-1-phosphate. Leiden: Leiden University, Zondag GCM, Reynolds AB, Moolenaar WH. Receptor tyrosine phosphatase RPTPmu binds to and dephosphorylates the catenin p120 ctn. J Biol Chem (in press). Lymphocyte activation is initiated by triggering of the T- or B- cell antigen receptors, which signal according to the protein tyrosine kinase (PTK) principle. We have established that the adaptor protein Grb2 links antigen receptors to the Ras-MAP kinase pathway, but also interacts with proteins that perform other functions. Among these, we have focused on c-cbl and the Wiskott-Aldrich syndrome protein (WASP). c-cbl is a major target for tyrosine phosphorylation as induced by antigen receptors and receptor PTKs. Cbl associates with these receptors and negatively regulates their function, in part by promoting receptor ubiquitination. Cbl s negative regulatory function impinges on an intact Ring finger domain; deletion mutations in this region turn c-cbl oncogenic. We set out to identify proteins interacting with the Ring finger. Three proteins were found, including an E2-type ubiquitin conjugating enzyme (UbcH7). This is consistent with c-cbl acting as an E3- type ubiquitin ligase, the component in the ubiquitination reaction that interacts with the substrate. Using the EGF receptor as a model, we are now monitoring receptor ubiquitination and endocytosis in transfected cells, its promotion by c-cbl and mutants, as well as by the newly identified Ring finger-interacting proteins. Patients with WASP mutations show defects in platelet and lymphocyte function. WASP regulates the actin cytoskeleton, but the receptor-linked pathways that control WASP function remain ill-defined. We have established that WASP associates with the EGF receptor via Grb2 in an EGF-dependent manner, and have identified a novel WW-domain protein that interacts with WASP. T-cell clones were generated from patients and reconstituted with WASP using retroviral gene transduction. In these clones, we will study the role of WASP and its binding partners in T cell receptor-induced cell polarization.

44 Regulation of lymphocyte survival While antigen receptors can prime lymphocytes for proliferation and differentiation, these same receptors can also induce apoptosis during repertoire selection and after antigenic challenge. We are examining the interaction between antigen receptor signaling and signals induced by members of the Tumor Necrosis Factor (TNF) receptor family. We have selected Jurkat cells for resistance to apoptosis induced by CD95, a deathinducing member of the TNF receptor family. These variant cells overexpress an, as yet, unknown molecule that inhibits CD95 signaling. By various approaches, including representational difference analysis of cdna (RDA), we are trying to clone this putative inhibitor. We are particularly interested in this regulation, since the CD95-resistant Jurkat cells are cross-resistant to DNA damaging anti-cancer regimens. Using isolated mitochondria, we have established that signaling to the mitochondria is blocked downstream from caspase-8 in these cells. This defines a novel control point in apoptosis signaling upstream from the mitochondria. We are in the process of identifying components of the death receptor- and DNA damage-induced pathways that evoke mitochondrial activation. Synthetic ceramide (Cer) can overcome apoptosis resistance in the variant Jurkat cells. Using cells overexpressing inhibitors of death receptor and mitochondrial function, we found that the endogenous Cer response, contrary to popular belief, has no causal role in the activation of inducer caspases or mitochondria but, rather, is a consequence of it. Cer was found to be generated from plasma membrane sphingomyelin by an a neutral sphingomyelinase. Since sphingomyelin is localized to the outer leaflet of the plasma membrane, the question arose of how it can serve as a substrate for a cytosolic enzyme. We found that Cer production correlates perfectly with the alterations in plasma membrane lipid distribution that occur in the effector phase of apoptosis as illustrated by phosphatidylserine exposure. With regards to the relevance of sphingomyelin to Cer conversion, we are examining the possibility that depletion of sphingomyelin alters the structural properties of the bilayer, which may have consequences for the ultimate generation of the apoptotic morphology. CD27 is a member of the TNF receptor family that lacks a death domain and associates with Traf-2. In vitro, CD27 costimulates TCR-induced responses of naive T cells. We are interested to define the exact nature of this costimulatory effect. In particular, we want to know whether CD27 affects cell survival and/or proliferation. We have generated a CD27-deficient mouse strain, in which lymphocyte development and thymic T-cell selection appear normal. Challenge of these mice with influenza virus has revealed a reduced responsiveness. In vitro, CD27-deficient T cells show decreased TCRinduced proliferative responses. Further analysis of cell cycle progression versus cell survival should reveal which pathways are regulated by CD27. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Netherlands Organization for Scientific Research (NWO) program grant. 3 Guest, graduate from University of Amsterdam. 4 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Regulation of lymphocyte activation Boesen-de Cock JGR, Tepper AD, De Vries E, Van Blitterswijk WJ, Borst J. Common regulation of apoptosis signaling induced CD95 and the by DNA damaging regimens etoposide and gamma radiation downstream from caspase-8. J Biol Chem 1999; 274: Borst J, Cope A. Turning the immune system on. Immunol Today 1999; 20: Medema,J-P, Borst J. T cell signaling: a decision of life and death. Human Immunol 1999; 60: Tepper AD, De Vries E, Van Blitterswijk WJ, Borst J. Ordering of ceramide formation, caspase activation, and mitochondrial changes during CD95- and DNA-damage-induced apoptosis. J Clin Invest 1999; 103: WJ Van Blitterswijk PhD PCJ Van der Hoeven PhD B Houssa MSc 1 GA Ruiter MSc 2 AD Tepper MSc 3 S Diks 4 A Los 4 JJM De Widt P Ruurs SF Zerp 2 Lipid metabolism in signal transduction Group leader Graduate student Graduate student Graduate student Undergraduate student Undergraduate student Function of diacylglycerol kinase 1 Diacylglycerol kinase (DGK) phosphorylates diacylglycerol (DAG) to yield phosphatidic acid (PA). Since DAG activates protein kinase C (PKC), DGK may regulate PKC activity. We have identified the DGKθ isotype which is highly expressed in brain. In collaboration with W Moolenaar, we found that DGKθ binds specifically to the effector loop of activated RhoA, in both COS and neuro- 45 CELLULAR BIOCHEMISTRY

45 46 CELLULAR BIOCHEMISTRY Figure III.2 Subcellular localization of DGKθ at different stages of cell division. A. MelJuso cells were microinjected with DGKθ cdna. Two microinjected cells show enhanced DGKθ staining in nuclear speckels, while surrounding non-injected cells show endogenous DGKθ. B. Cell in metaphase. DGKθ is located at the spindle pole bodies. Chromosomes at the metaphase plaque were stained with propidium iodine. C. Cells at late telophase. DGKθ clusters start to reform in the cytoplasm before reentering the nuclei of daughter cells. DGKq is also seen at the midbody. DGKθ was stained with a specific polyclonal antobody. blastoma cells. Strikingly, the binding of activated RhoA to DGKθ inhibits DGK catalytic activity. These results suggest that that DGKθ is a downstream effector of RhoA and that its activity is negatively regulated by RhoA. Through accumulation of newly produced DAG, RhoA-mediated inhibition of DGKθ may lead to enhanced PKC activity in response to external stimuli. DGKθ is present in the cytosol and the nucleus. Immunofluorescence analysis revealed that DGKθ in the nucleus clusters into so-called speckles (Figure III.2) where it colocalizes with splicing factors, suggesting a possible role for DGKθ in mrna processing. Phosphoinositide kinases and phospholipase C are also detected in such speckels, suggesting that these structures contain preassembled signaling complexes of a nuclear phosphoinositide cycle. DGKθ localization alters dramatically during cell division: in prophase and metaphase, DGKθ staining disappears and concentrates at the spindle pole bodies. At late anaphase and early telophase, DGKθ clusters reassemble in the cytosol and enter the nuclei of daughter cells. In addition, DGKθ acumulates at the midbody of cells undergoing cytokinesis, suggesting that DGKθ also acts in this process. PKC-ζ signaling In growth factor-stimulated Rat-1 cells, protein kinase C-ζ (PKC-ζ) phosphorylates and activates the Raf-1 kinase in a signaling complex. Formation of this complex is mediated by the scaffold protein Complex formation was much stronger with a kinase-inactive PKCζ mutant than with the wild-type enzyme, supporting the idea that kinase activation causes complex dissociation. Certain proteins could serve as substrates for PKC-ζ. Phosphorylation of by PKC-ζ negatively regulated their physical association. Our findings suggest that facilitates coupling of PKC-ζ to RAF-1 in a ternary complex, in an isotype-specific and phosphorylation-dependent manner. We propose that is a transient mediator of RAF-1 phosphorylation and activation by PKC-ζ. PKC-ζ can also associate with protein kinase B (PKB or AKT) in a mediated and phosphorylationdependent fashion (collaboration with P Van Bergen and Henegouwen, Utrecht University). Growth factor stimulation of COS cells decreased this association. Our results suggest that, contrary to its stimulatory activity on RAF-1 signaling, PKC-ζ may act as a negative regulator of the PKB/AKT survival pathway. Sphingomyelinase-mediated ceramide formation in apoptosis 3 (collaboration with J Borst) Apoptosis induction is generally accompanied with ceramide (Cer) production but cause-effect relationships remain unclear. As mentioned above, Cer formation in Jurkat cells by CD95 or DNA-damaging regimens appears to be associated with the effector phase of apoptosis. CD95-induced Cer production in Jurkat cells was temporally associated with sphingomyelin (SM) hydrolysis, phosphatidylserine (PS) exposure and the hydrolysis of the exogenously supplied fluorescent SM analog (NBD-SM), which intially inserts into the outer leaflet of the plasma membrane, to NBD-Cer. By contrast, NBD-phosphatidylcholine was not metabolized. Raji cells defective in calcium-induced phospholipid scrambling did not show CD95-induced PS externali-

46 zation nor Cer production, yet underwent apoptosis as assessed by loss of mitochondrial depolarization and nuclear fragmentation. These results support a causal link between two plasma membrane events, i.e. loss of phospholipid asymmetry and Cer formation. Moreover, it appears that Cer has no role in the induction of mitochondrial and nuclear changes. Direct induction of lipid scrambling by Ca 2+ ionophore resulted in concomitant PS exposure and NBD-SM hydrolysis in Jurkat cells but not in Raji cells, indicating that loss of phospholipid asymmetry is necessary and sufficient to generate a Cer response. We propose a novel mechanism of Cer formation, in which SM in the outer leaflet of the plasma membrane flips to the inner leaflet, where it becomes accessible to cytosolic neutral sphingomyelinase. SM has a high affinity for cholesterol, which is important for stability of the plasma membrane and the signaling function of specialized domains (rafts/caveolae) therein. Using methyl-cyclodextrin as a strong cholesterol-binding agent, we observed an enhanced cholesterol efflux from the plasma membrane upon exposure of Jurkat cells to bacterial sphingomyelinase, anti-cd95 mab or ionomycin. We postulate that lipid scrambling results in SM hydrolysis, loss of SM/cholesterol interaction and altered bilayer properties, which may facilitate membrane blebbing, vesicle shedding and apoptotic body formation. Accordingly, Raji cells defective in lipid scrambling did not show membrane blebbing, whereas exogenous NBD-SM prevented blebbing in CD95-stimulated Jurkat cells. This suggests that the breakdown of SM, rather than the production of Cer, is the determining factor for these morphological changes in apoptotic cells. Publications: Lipid metabolism in signal transduction Boesen-de Cock JGR, Tepper AD, De Vries E, Van Blitterswijk WJ, Borst J. Common regulation of apoptosis signaling induced by CD95 and the DNA-damaging stimuli etoposide and γ-radiation downstream from caspase-8 activation. J Biol Chem 1999; 274: Doornbos RP, Theelen M, Van der Hoeven PCJ, Van Blitterswijk WJ, Verkleij AJ, Van Bergen en Henegouwen PMP. Protein kinase C-ζ is a negative regulator of protein kinase B activity. J Biol Chem 1999; 174: Houssa B, De Widt J, Kranenburg O, Moolenaar WH, Van Blitterswijk WJ. Diacylglycerol kinase θ binds to and is negatively regulated by active RhoA. J Biol Chem 1999; 274: Ruiter GA, Zerp SF, Bartelink H, Van Blitterswijk WJ, Verheij M. Alkyl-lysophospholipids activate the SAPK/JNK pathway and enhance radiation-induced apoptosis. Cancer Res 1999; 59: Tepper AD, De Vries E, Van Blitterswijk WJ, Borst J. Ordering of ceramide formation, caspase activation, and mitochondrial changes during CD95- and DNA damage-induced apoptosis. J Clin Invest 1999; 103: Tepper AD, Van Blitterswijk WJ. Ceramide mass analysis by normal phase HPLC. Methods Enzymol 1999 (in press). Van Blitterswijk WJ, Houssa B. Diacylglycerol kinases in signal transduction. Chem Phys Lipids 1999; 98: CELLULAR BIOCHEMISTRY Apoptosis induced by alkyl-lysophospholipids (ALPs) 2 (collaboration with M Verheij, Division IX) Synthetic alkyl-lysophospholipids (ALPs) such as 1-Ooctadecyl-2-O-methylglycero-3-phosphocholine (ET-18- OCH 3) readily incorporate into the plasma membrane and induce apoptosis in various tumor cells by an unknown mechanism. ALPs act synergistically with ionizing radiation in inducing apoptosis in leukemic cells. ALPs rapidly activate the stress-activated protein kinase (SAPK/JNK) signaling pathway and inhibit two antiapoptotic pathways, i.e. the MAP kinase pathway and the PKB/AKT survival pathway. ALPs prevent ERK signaling in a RAS-independent manner, by interfering with phospholipase C and PKB signaling. Thus, ALPs may act by interfering with inositol lipid metabolism at the plasma membrane and possibly in intracellular organelles. Notes 1 Funding: Netherlands Organization for Scientific Research (NWO/SON ). 2 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI University of Utrecht. Van der Hoeven PCJ, Van der Wal J, Ruurs P, Van Dijk MCM, Van Blitterswijk WJ isotypes facilitate coupling of protein kinase C-ζ to Raf-1; negative regulation by phosphorylation. Biochem J 1999; 345: N Divecha PhD J Clarke PhD C D Santos PhD 1 J Halstead PhD 2 M Roefs Signaling through inositol phospholipids Group leader The inositol phospholipid pathway is an intracellular signaling cascade which, in response to growth factors and other stimuli, can generate a number of second

47 48 CELLULAR BIOCHEMISTRY messengers. Our research is focused on the regulation of the synthesis and cellular functions of phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2) and, in particular, the role of phosphatidyl-inositol-phosphate kinases (PIPkins) therein. Cloning has revealed that there are three sub-families of PIPkins able to phosphorylate different phosphatidylinositol phosphates at various positions. The type I PIPkin will phosphorylate PtdIns(4)P on the 5 position of the inositol ring, while the type II enzymes (or PIPkin C) have a specificity for PtdIns(5)P which they phosphorylate on the 4 position. These data demonstrate that eukaryotic cells have developed at least two distinct pathways for regulating the synthesis of PtdIns(4,5)P2. This highly regulated process probably reflects the cellular requirements for PtdIns(4,5)P2 in a number of diverse functions in discrete subcellular compartments. For example, it is the substrate for receptor-mediated generation of inositol(1,4,5)trisphosphate, which regulates cellular calcium, and diacylglycerol (DAG) which regulates the activity of protein kinase C. Genetic evidence suggests that PtdIns(4,5)P2 is important in membrane trafficking through the recruitment of coat proteins required for vesicle formation. PtdIns(4,5)P2 is also the substrate for a growth factor-stimulated PtdIns-3-kinase which leads to the formation of PtdIns(3,4,5)P3. Although PtdIns(3,4,5)P3 has pleiotropic effects, the demonstration that an enzyme (PTEN), which dephosphorylates this lipid is a tumour suppressor and is mutated in a number of tumors, suggests it is physiologically important in cell growth and survival. Finally, the phosphoinositide pathway is also present within the nucleus where the functions of PtdIns(4,5)P2 appear to be involved in the production of second messengers as outlined above, but also has a potential role in the regulation of other nuclear functions such as chromatin structure and nuclear transport. To better understand the regulation of these pools of PtdIns(4,5)P2 we are currently analyzing the subcellular localization and mode of activation of PIPkins as well as the regulation of the intranuclear pool of PtdIns(4,5)P2. Type I PIPkin There are three family members of the type I PIPkins identified to date and we have concentrated on the regulation of the type Iα. Our data suggest that it is regulated by interaction with at least two different proteins leading to its involvement in two important signaling pathways. The Rac GTPase regulates diverse cellular processes such as cytoskeletal and endocytic dynamics. We have demonstrated that RAC interacts with type I PIPkin. This interaction not only leads to the enhanced synthesis of PtdIns(4,5)P2, but also to an increase in the cellular PtdIns(3,4,5)P3 pool. Previous data have demonstrated that PtdIns(3,4,5)P3 is important in the regulation of protein kinase B (PKB), a cell survival factor. However, our data demonstrate that PtdIns(3,4,5)P3, derived from the interaction of RAC and Figure III.3 Porcine endothelial cells were microinjected with the type Iα PIPkin together with GFP-labeled PLD2. Three hours later, the cells were fixed and the type Iα was visualized after staining with a specific antibody. The figure shows the same cells viewed by confocal microscopy. the type Iα. PIPkin, is unable to activate PKB. This may reflect the intracellular location of this PtdIns(3,4,5)P3, which we are currently investigating in collaboration with K Jalink (Division I) using new fluorescent probes specific for this lipid. Phospholipase D (PLD) is important in the regulation of secretion and membrane trafficking. This enzyme has previously been shown to have an absolute requirement for PtdIns(4,5)P2. Using cloned PLDs (obtained in a collaboration with M Wakelam, Birmingham, UK), we have shown that the type Iα PIPkin interacts with, activates and co-localizes with PLD2. Interestingly, the coexpression of PLD2 appears to recruit the type Iα. PIPkin to vesicles (Figure III.3). The role of this interaction in terms of the regulation of membrane flow is currently under investigation (collaboration with J Neefjes, Division IV). Type II PIPkin In collaboration with R Irvine (University of Cambridge, UK) we have investigated the potential regulation of the type II PIPkins. This family is composed of at least three members. We have demonstrated that the α isoform is regulated by phosphorylation, by casein kinase 2, which leads to the translocation of the normally cytosolic protein to the plasma membrane. Whether this opens a cryptic Figure III.4 Hela cells were transfected with cdna encoding either GFP-type IIα or β. After 12 hours the cells were fixed and viewed by confocal microscopy.

48 membrane binding site or a protein interaction site is currently under investigation. We have previously demonstrated that a type II PIPkin is also present in the nucleus. Using GFP-tagged proteins we have shown that the β isoform is nuclear while the a isoform is cytosolic (Figure III.4). Chimeras between these proteins has allowed us to identify that the α helix-7 domain of the type IIβ. is important in this localization. This domain may be involved in the process of nuclear translocation or in binding to specific nuclear sites. The nuclear inositide pathway We have previously shown that the inositol phospholipid pathway also operates in the nucleus and that nuclear inositol lipids and the enzymes required to synthesize and metabolize them are regulated distinctly from those found at the plasma membrane. DAG is a key second messenger within the nucleus and its removal by overexpression of a DAG kinase leads to cell cycle arrest. By exploiting the differences in the fatty acid composition of PtdIns and PtdCho lipids, we have demonstrated that there exists within MEL (murine erythroleukemia) cell nuclei two discrete pools of DAG derived from the endogenous pools of PtdIns and phosphatidylcholine, respectively. By the use of specific inhibitors, we have shown that these pools are differentially regulated with respect to either cell differentiation or proliferation. We have shown changes in the levels of nuclear PtdInsP, PtdInsP 2 and phosphatidic acid which are consistent with these nuclear lipids having a function as cells move into the replicative phase of the cell cyle. To visualize this in real time, we have constructed cell lines which overexpress a fluorescent probe which is targeted to the nucleus and binds specifically to nuclear PtdIns(4,5)P2. Interestingly, we demonstrate that the mass levels of this lipid increases specifically at this time point and that the presence of the probe, which specifically binds to PtdIns(4,5)P2, appears to lead to a cell cycle arrest which does not occur if the probe is targeted to PtdIns(4,5)P2 present at the plasma membrane. These data suggest a important physiological function for nuclear PtdIns(4,5)P2 as cells progress through the cell cycle. We have characterized a pathway of nuclear PtdIns(4,5)P2 synthesis and now are able to pharmacologically inhibit this pathway resulting in a specific decrease in the nuclear PtdIns(4,5)P2 levels. This should shed light on the physiological functions of this pathway. Notes 1 Funding: EEC. 2 Funding: Dutch Cancer Society, Project NKI Publications: Signaling through inositol phospholipids Divecha N, Clarke JH, Roefs M, Halstead JR, D Santos C. Nuclear inositides; inconsistent consistencies. Cell Mol Life Sci 1999 (in press). D Santos CS, Clarke JH, Irvine RF, Divecha N. Nuclei contain two differentially regulated pools of diacylglycerol. Curr Biol 1999; 9: Fernandez-Borja M, Wubbolts R, Calafat J, Janssen H, Divecha N, Dusseljee S, Neefjes J Multivesicular body morphogenesis requires phosphatidyl-inositol 3-kinase activity. Curr Biol 1999; 9: Hinchliffe KA, Ciruela A, Letcher AJ, Divecha N, Irvine RF. Regulation of type IIalpha phosphatidylinositol phosphate kinase localization by the protein kinase CK2. Curr Biol 1999; 9: Smith GC, Divecha N, Lakin ND, Jackson SP. DNA-dependent protein kinase and related proteins. Biochem Soc Symp 1999; 64: P Ten Dijke PhD M-J Goumans PhD S Itoh PhD F Itoh MSc G Valdimarsdottir MSc 1 M Thorikay TGF-β/Smad signal transduction Group leader Graduate student Graduate student Transforming growth factor-β (TGF-β) is the prototype of a large family of multifunctional cytokines, which include TGF-βs, activins and bone morphogenetic proteins. Many of them have important functions during embryonic development and in adults they are involved in maintaining tissue homeostasis. Our research 2 is aimed at elucidating the signaling mechanisms by which TGF-β family members elicit their wide range of biological effects on many different cell types, including modulation of cell proliferation, differentiation, apoptosis and migration. An increased understanding in TGF-β signaling processes and in vivo function of the signaling components is essential for the treatment of various clinical disorders, including cancer, fibrosis, vascular and auto-immune diseases, that are caused by deregulated TGF-β activity. Our work focuses on the activation and regulation of the TGF-β serine/threonine kinase receptors and their 49 CELLULAR BIOCHEMISTRY

49 50 CELLULAR BIOCHEMISTRY intracellular effectors, termed Smad proteins. Upon binding to their specific receptors, each TGF-β family member activates a particular subset of Smad proteins, which form heteromeric complexes that translocate to the nucleus where they regulate gene transcription. We will study the role of Smads as transcriptional regulators of various TGF-β-induced responses by identifying transcription factors, co-activators and co-repressors with which Smads make functional interactions and identify and characterize Smad target genes. To elucidate how subversion of TGF-β signaling contributes to cancer, we will search for molecules that, when misexpressed or (inactivated) perturb TGF-β-induced growth arrest. To investigate the patho-physiological role of TGF-β in vasculogenesis, we will use mice lacking TGF-β receptors in endothelial cells and cells from patients with TGF-β receptor-associated vascular diseases, i.e. hereditary hemorrhagic telangiectasia 1. Notes 1 Funding: Netherlands Heart Foundation, Project Supported in part by EEC TMR grant (ERBFMRXCT980216) and Netherlands Organization for Scientific Research, Project ALW Publications: TGFβ/Smad signal transduction Heldin C-H, Ten Dijke P. SMAD destruction turns off signaling. Nature Cell Biol 1999; 1: E Ten Dijke P, Miyazono K, Heldin C-H. Signaling inputs converge on nuclear effectors in TGF-β signaling. Trends Biochem 1999 (in press). Secretary Cellular Biochemistry JM Overwater Research staff positions (full time equivalents) Scientific permanent: 5.0 Scientific project: 18.5 Technical permanent: 8.2 Technical project: 4.6

50 IV Division of Immunology 51 Division head Ada Kruisbeek Introduction This division spans a broad range of topics related to T-cell-mediated immune responses and T-cell development. Not all the research lines can be highlighted in this short introduction, but a few of the common interests should be mentioned. Since the group of T Schumacher introduced the technology for use of soluble tetrameric peptide-mhc complexes, essentially all projects related to tumor immunity and auto-immunity apply this technology to follow the fate of T cells in vivo and in vitro. As soluble tetrameric complexes bind to the variable portion of the antigen receptors on T cells, they can be applied to study the anatomy of the tumorspecific T-cell response, to study how T cells with receptors specific for self-proteins behave in vivo, and to study how variations in antigenic epitopes may affect escape from T-cell immunity. This year, T Schumacher and his co-workers discovered that the immune response to naturally arising variants of a dominant T-cell epitope in influenza A virus consists mainly of rare cross-reactive killer T cells, originating in the T-cell memory population. Another technology that is applied in virtually all projects in the division is retrovirus-mediated gene transfer, introduced here by the laboratory of H Spits some years ago. Lymphocytes are notoriously difficult to transfect through means that work well with most other cell types and, as lymphocytes are the mainstay of the division, it used to be difficult to perform signaling studies and gene complementation experiments ex vivo. Now, however, most of the T-cell development studies of both the Kruisbeek and the Spits laboratory rely heavily on ectopic expression of signaling molecules or transcription factors and their mutants for identification of crucial steps in the developmental pathway of T cells. H Spits reports on some exciting insights about the survival and proliferation versus differentiation signals in early T cells. In another recent discovery, the life span of human killer T cells could be extended tremendously through ectopic expression of the catalytically active component of telomerase. Ever since the finding that a self-antigen on human melanomas, tyrosinase, represents a tumor rejection antigen recognized by killer T cells, the ground rules of tumor immunology have changed. Rather than focusing on tumor-specific antigens, many laboratories now explore tissue-specific self-antigens as promising targets for T-cell mediated anti-tumor therapy. This has fostered an interest in defining conditions that can overcome the (naturally occurring) tolerance to self-antigens. Fortunately, tolerance is inherently imperfect, and a population of low avidity self-antigen specific T cells could easily be detected in the normal T-cell repertoire (Kruisbeek laboratory). The challenge for the future will be to define how we can make such cells work for us: it is clear that they have severe defects in their ability to produce IL-2 and proliferate in response to an antigenic challenge, and we are now in a good position to dissect the basis for these defects. Being able to correct them, however, may be another matter. In preparation of an alternative approach, the Schumacher laboratory has developed a very promising strategy for retroviral TCR display that allows selection of high affinity TCRs with a desirable fine-specificity. Application of this strategy may provide a way to circumvent the limitations of the low avidity self-specific T-cell repertoire. Tuning TCR signaling in T-cell development and tumor immunity AM Kruisbeek, PhD F Bischof MD 1 M Haks PhD 2 W Marissen PhD 3 E Pépin PhD 4 Q Valent PhD 5 Y Zevering PhD 6 T Cordaro MSc K De Visser MSc 7 I Klitsie 5 F Tirion 7 J Van den Brakel P Weder E Wijnands 5 A Bins S Smeele Group leader Graduate student Graduate student Undergraduate student Undergraduate student This laboratory studies how T-cell receptor (TCR) signaling determines T-cell fate, both early in development, and in the mature T-cell pool. Our work

51 52 IMMUNOLOGY centers around 2 major lines of investigation: 1) the role of the TCR and the pre-tcr in the development of T cells; and 2) the regulation of mature T-cell responsiveness in the context of tumorimmunity and autoimmunity. Recent results in these areas include the discovery that receptor triggering of CD8 T cells with low avidity TCRs results in an aborted immune response. We also found that pre-t cells with a dysfunctional pre-tcr fail to complete their differentiation program, even when their survival defect is restored. Pre-TCR and TCR controlled checkpoints of T-cell development Progression of T-cell development is monitored by antigen receptors at two checkpoints. The first is at the transition from the double negative (DN) to the double positive (DP) stage; this process is termed TCRβ selection and is regulated by the pre-tcr. The second checkpoint is at the DP to the SP stage and is driven by the mature TCR complex. Regulation of early T-cell development by the pre-tcr is one of the topics of our research, with special emphasis on the identification of molecular events modulated by signals from the pre-tcr. Our studies were the first to establish genetic evidence for the crucial role of the CD3γ chain in the pre-tcr (Haks et al. EMBO J 1998; 17: ). These findings were surprising, as they suggest that CD3γ is not functionally redundant in the pre-tcr complex. We also demonstrated a role for the p53 protein at the TCRβ selection checkpoint: introduction of p53 deficiency into CD3γ-deficient mice completely reverses the survival defect in the DN compartment and the DP compartment eventually reaches its normal size. Nevertheless, such rescued CD3γ-deficient thymocytes are unable to undergo certain subsequent developmental steps in αβ T-cell development. Proliferative and survival signals can thus be separated from the pre-tcr driven signaling events that regulate further differentiation at this stage of development. Given the previously established roles for the CD3δ and CD3ζ chains in the TCR complex at the positive selection checkpoint, our subsequent findings were even more surprising. When CD3γ-deficient mice are crossed to mice expressing a transgenic TCR, the thymus increases in size and forms a large DP compartment, but positive selection does not occur. Although DP cells lacking CD3γ do have the ability to initiate the first steps in positive selection, they are unable to progress through the entire differentiation program. The data further suggest that CD3γ-deficient TCRs are unable to deliver the prerequisite signals that DP thymocytes need to be rescued from programmed cell death. These findings underscore the complex relationship between TCR signaling and positive selection. Identification of crucial signaling events downstream of the TCR will be required to further dissect the multi-step program of positive selection. T-cell tolerance and antitumor immunity Questions such as how self-reactive T cells can survive, expand and differentiate into effector and memory populations are central to our understanding of both autoimmunity and antitumor immunity. Self-reactive T cells can cause both undesirable autoimmune pathology and desirable tumor rejection. We are therefore studying how such cells are generated and which conditions determine their in vivo response. In collaboration with the group of T Schumacher, we employ tetramer technology to determine the phenotypical and functional characteristics and in vivo fate of a self-reactive CD8 T-cell population specific for an ubiquitously expressed T-cell epitope. Contrary to expectations based on TCR transgenic studies, we find that a sizable population of self-reactive CD8 + T cells can be triggered to expand and differentiate into effector cells. These cells are not deleted, do not downregulate their antigen or CD8 receptors, are not inactivated, and can differentiate into memory cells. The main hallmark of the self-reactive CD8 + T-cell population that can be elicited from the polyclonal repertoire is its diminished avidity for the pertinent tetrameric MHC/peptide complex. The ability to identify and isolate these cells is then applied to compare the activation, survival, expansion and differentiation requirements of low versus high avidity CD8 + T cells. The results obtained so far show that, although low avidity T cells can be triggered to differentiate into effector CTL and IFNγ secreting cells, they differ from high affinity CD8 + cells in a number of ways. Low avidity cells display a dramatically reduced susceptibility to antigenic peptide, and cannot produce IL-2, regardless of the antigen dose used. In addition, their ability to be expanded after stimulation in exogenous IL-2 is severely limited. These studies illustrate that responsiveness is not dictated by signal strength alone. Future studies will be directed at acquiring a better molecular understanding of how binding characteristics of the TCR translate into different responses. Connected to this line of research are efforts to dissect how the CTLA-4 receptor regulates T-cell responsiveness, and how antigen processing pathways affect T-cell tolerance. It is clear that engagement of the CTLA-4 receptor on T cells is critically involved in terminating clonal expansion, but little is known about the underlying mechanisms. We have used biotinylated peptides of the inactive and active (i.e., phosphorylated) CTLA-4 cytoplasmic tail, in combination with chemical crosslinking reagents to search for CTLA-4 interacting proteins in T-cell lysates. This approach enabled us to confirm the documented CTLA-4 LCK interaction and to pick up two yet to be identified CTLA-4 interacting proteins. Apart from ongoing work on the identification of these proteins, a second major focus of this project is on the consequences of CTLA-4 engagement for TCR, CD28 and cytokine receptor signaling. In addition, we continue to study how self-antigen expression in different cellular and intracellular compartments affects T-cell repertoire development. In very

52 exciting recent experiments, we established that transgenic expression of a given (model) self-protein in the B- cell compartment, through use of an Ig-promotor, dramatically skews the repertoire towards generation of Th1-only responses. We shall investigate the molecular basis for this observation, plus its consequences for the susceptibility to the autoimmune pathology normally associated with responses to this protein. Notes 1 Funding: German Research Association, Germany. 2 Funding: Human Frontier Science Program Organization, Project RG0335/1998-M. 3 Funding: NWO (Netherlands Organization for Scientific Research), Program grant Funding: Association pour la Recherche sur le Cancer, France. 5 Funding: Dutch Cancer Society, Project NKI Funding: American Multiple Sclerosis Society, Project RG 2940-A-1. 7 Funding: Dutch Cancer Society, Project NKI Publications: TCR signaling Amsen D, Kruisbeek AM. Thymocyte selection: not by TCR alone. Immunol Rev 1999; 165: Amsen D, Osborne B, Kruisbeek AM. Costimulatory signals are required for induction of Nur77 during negative selection of CD4 + CD8 + thymocytes. Proc Natl Acad Sci USA 1998; 96: Haanen JBAG, Toebes M, Cordaro TA, Wolkers MC, Kruisbeek AM, Schumacher TNM. Systemic T-cell expansion during localized viral infection. Eur J Immunol 1999; 29: Haanen JBAG, Wolkers MC, Kruisbeek AM, Schumacher TNM. Selective expansion of cross-reactive CD8 + memory T cells by viral variants. J Exp Med 1999; 190: Haks MC, Oosterwegel MA, Blom B, Spits H, Kruisbeek AM. Cellfate decisions in early T-cell development: regulation by cytokine receptors and the pre-tcr. Semin Immunol 1999; 11: Haks MC, Krimpenfort P, Van Den Brakel JHN, Kruisbeek AM. Pre-TCR signaling and inactivation of p53 induces crucial cell survival pathways in pre-t cells. Immunity 1999; 11: Jacobs H, Krimpenfort P, Haks MC, Blom B, DÈmolliËre C, Kruisbeek AM, Spits H, Berns AM. PIM1 reconstitutes thymus cellularity in interleukin-7- and common γ chain-mutant mice and permits thymocyte maturation in Rag- but not in CD3γ-deficient mice. J Exp Med 1999; 190: Thesis Haks MC. Regulation of T-cell development by cytokine and antigen receptors. Amsterdam: Free University, H Spits PhD K Weijer DVM PhD D Clark PhD 1 E Hooijberg PhD 2 M Naspetti PhD 3 Y Yasuda PhD 4 W Wiegant MSc 5 AQ Bakker F Couwenberg 6 E Kueter 7 J Ruizendaal 2 A Voordouw B Visser D Overdijk Lymphoid cell development in humans Development and function of human T cells Group leader Academic staff Graduate student Undergraduate student Undergraduate student This research program focuses on two topics: 1) the role of cytokines, in particular of IL-7, in T-cell development; 2) investigation of the transcriptional control of human lymphoid development mediated by basic helix loop helix proteins. The role of the IL-7Rα chain in T-cell development The interleukin (IL)-7 receptor plays an essential role in T-cell development and IL-7 binding to the IL-7R activates PKB and STAT5 in precursor T cells. We introduced chimeric receptors into such precursors with a cytoplasmic domain of the IL-7R that is no longer able to activate PI-3K/PKB and STAT5 and tested the transduced cells in a fetal thymic organ culture. We also examined the T-cell precursor activity of progenitors expressing dominant negative forms of PI-3K (DNp85) or STAT5B. Ectopic expression of DNp85 in T-cell precursors strongly blocked proliferation and survival of these cells but did not affect their differentiation, as the few surviving DNp85 expressing cells developed normally into both TCRαβ and TCRγδ cells in fetal thymic organ cultures. These experiments revealed that PI-3K/PKB activation is essential for the survival and proliferation of T-cell precursors and suggest that STAT5 activated by IL- 7 mediates T-cell differentiation. Analysis of T-cell development in a normal thymic environment We have optimized methods to study the role of transcription factors and signaling molecules on development of human T cells in the natural thymic environment. Progenitors are transduced with dominant 53 IMMUNOLOGY

53 54 IMMUNOLOGY negative mutants of transcription factors or signaling components tagged with GFP and with a control tagged with a truncated form of the nerve growth factor receptor (DNGFR). These transduced cells are then mixed and injected into an established thymus graft in mice deficient for RAG1 and the common IL-2/IL7γ chain. By taking biopsies of the injected thymus and analyzing the NGFR and GFP expression, we can follow the kinetics of T-cell development. We applied this method to analyze the effect of dominant negative STAT5B on thymic T-cell development. This experiment revealed that DNSTAT5B strongly delayed T-cell development. In addition, in collaboration with N Taylor and N Noraz (University of Montpellier, France), we analyzed the effect of a dominant negative mutant of ZAP-70 and found that this mutant inhibits positive selection as assessed by evaluating expression of certain hallmarks of positive selection on TCR expressing thymocytes. A lymphoid origin of precursors of dendritic cell 2 (pdc2)/ natural interferon producing cells (NIPC) We have characterized (in collaboration with F Vyth- Dreese from the Immunotherapy group, and C Uittenbogaart, UCLA, USA) a DC precursor in the thymus which expresses CD4, CD45RA and high levels of IL-3Rα (CD123). The phenotype of this rare cell type is identical to the recently described pdc2 and NIPC, and is therefore referred to as pdc2/nipc. Importantly, these cells lack the surface markers characteristic for the myeloid lineage of DC, but express a number of markers typical for the lymphoid lineage. Strikingly, these cells also express high levels of pre T-cell receptor alpha (ptα) transcripts, suggesting a close relationship between these cells and cells of the T-cell lineage. A novel assay was developed in which we can determine the pdc2/nipc precursor activity of CD34 + populations. Using this assay we found that thymic precursors which mostly lack myeloid developing potential, have the ability to develop onto pdc2. These data indicate a lymphoid origin of pdc2/nipc. Control of human lymphoid cell development by bhlh transcription factors Basic helix loop helix (bhlh) transcription factors control cell fate decisions in a wide range of organisms, ranging from yeast to mammals. They act in concert with inhibitors of DNA binding (Id) proteins. These proteins possess the HLH motif and can form dimers with bhlh factors but, because they lack a basic DNA-binding domain, complexes with Id and bhlh proteins are transcriptionally inactive. We previously reported that ectopic expression of Id3 in hematopoietic progenitor cells inhibits the development of these cells into T and B cells at a very early stage. In contrast, development into NK cells was not affected. We now analyzed the effect of ID3 expression on development of hemopoietic progenitors into pdc2/nipc, and observed that development of these cells from both uncommitted stem cells and from early thymic precursors was completely inhibited by ID3. Ectopic expression of ID3 in progenitor cells thus inhibits not only development of T and B cells, but also of the lymphoid dendritic lineage. These data are consistent with a model in which Id proteins consolidate NK-cell development by limiting the developmental options of lymphoid precursors. Tumor immunology Isolation of antigen-specific T-cell clones from antigenactivated T-cell blasts infected with a self inactivating virus containing GFP under control of multiple NFAT binding sites Recent work in our tumor immunology group includes the development of a novel method for detection and isolation of antigen specific human T cells from a heterogeneous pool of T cells. We engineered into a self inactivating retroviral (SIN) vector DNA constructs carrying multiple NFAT binding sites, followed by the minimal IL-2 promoter and the reporter gene GFP. Jurkat cells, T-cell blasts and T-cell clones could be transduced with high efficiency with this vector (20-40%). Activation of transduced T-cell blasts with the superantigen SEB or of transduced antigen specific T-cell clones with cognate antigen, induced GFP expression which could be blocked by Cyclosporin-A and FK506. Following stimulation of a heterogeneous T-cell bulk culture with an HLA mismatched stimulator cell (JY), the GFP expressing cells were cloned. As expected, the cloning frequency of the antigen-specific GFP + cells was considerably higher than that of the total T-cell population. Most of the T-cell clones were either cytolytic, or proliferative towards JY stimulator cells. Interestingly, we also isolated T-cell clones which were non-cytolytic and non-proliferative towards JY cells, but specifically upregulated GFP after an overnight stimulation with JY. In future experiments, we will apply this technology towards the isolation and enrichment of tumor-specific T cells. Immortalization of antigen-specific T-cell clones Replicative senescence of T cells is correlated with erosion of telomere ends. As telomerase plays a key role in maintaining telomere length, it is thought that telomerase regulates the life span of T cells. To test this hypothesis we have (in collaboration with the group of J Walboomers, Free University, Amsterdam) ectopically expressed the catalytically active component of telomerase, telomerase end reverse transcriptase (TERT) tagged with GFP, into a Mart-1/HLA-A2 specific T-cell clone obtained from a melanoma patient. The transduction efficiency was low (5%), but the proportion of GFP-TERT + cells increased strongly upon further expansion of the cells by weekly stimulations in the presence of IL-2. After 5 weeks >95% of the cells were TERT +, indicating a strong growth advantage of the transduced cells. The telomere lengths of the TERT-transduced cells were longer than those of the untransduced cells, confirming the biological activity of the introduced TERT. In contrast to untransduced T cells, the TERT +

54 cells could be subcloned at least two times. After more than 120 population doublings, the subclones have so far maintained GFP expression, and they are still growing well, indicating the stability of the introduced gene. The phenotype of the TERT-transduced cells was virtually indistinguishable from that of the untransduced cells and, importantly, the specificity and the cytotoxic activity of the cells (determined by MART1/HLA-A2 tetramer staining and cytotoxicity assays) were not altered by introduction of TERT. In addition, the cells remained dependent on antigen-stimulation and cytokines for their expansion and survival, indicating that the cells retained their normal physiology and were not transformed. These observations prove that TERT directly regulates the replicative life span of human CD8+ T cells. Heemskerk MHM, Hooijberg E, Ruizendaal JJ, Van Der Weide MMC, Kueter E, Bakker AQ, Schumacher TNM, Spits H. Enrichment of an antigen-specific T-cell response by retrovirally transduced human dendritic cells. Cell Immunol 1999; 195: Jacobs H, Krimpenfort P, Haks MC, Blom B, Démoliere C, Kruisbeek AM, Spits H, Berns AM. PIM-1 reconstitutes thymus cellularity in interleukin-7 and common γ chain mutant mice and permits thymocyte maturation in Rag- but not in CD3γ-deficient mice. J Exp Med 1999; 190: Jaleco A-C, Stegmann APA, Heemskerk MHM, Couwenberg F, Bakker AQ, Weijer K, Spits H. Genetic modification of human B- cell development. B-cell development is inhibited by the dominant negative helix loop helix factor Id3. Blood 1999; 94: IMMUNOLOGY Notes 1 Funding: NWO (Netherlands Organization for Scientific Research), Project Funding: Dutch Cancer Society, Project NKI Funding: ARC, France. 4 Funding: Ciba-Geigy, Japan (till 11/99); Dutch Cancer Society, Project NKI (from 11/99). 5 Funding: NWO (Netherlands Organization for Scientific Research), Project Funding: Netherlands Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Development and function of human T cells Pallard C, Stegmann APA, Van Kleffens T, Smart F, Venkitaramen A, Spits H. Distinct roles of the phosphatidylinositol 3-kinase and STAT5 pathways in IL-7-mediated development of human thymocyte precursors. Immunity 1999; 10: Res PCM, Couwenberg F, Vyth-Dreese FA, Spits H. Expression of pre-tα mrna in a committed DC precursor in the human thymus. Blood 1999; 94: Res PCM, Spits H. Developmental stages in the human thymus Semin Immunol 1999; 11: In vitro and in vivo analysis of T-cell function Blom B, Heemskerk MHM, Verschuren MCM, Van Dongen JJM, Stegmann APA, Bakker AQ, Couwenberg F, Res PCM, Spits H. Overexpression of the helix-loop-helix factor Id3 in committed T cell progenitors blocks TCR αβ but not TCR γδ development. EMBO J 1999; 18: Blom B, Verschuren MCM, Heemskerk MHM, Bakker AQ, Van Gastel-Mol E, Wolvers-Tettero ILM, Van Dongen JJM, Spits H. TCR rearrangements and expression of the pre-t cell receptor complex during human T cell differentiation. Blood 1999; 93: Dardalhon V, Noraz N, Pollok K, Rebouissou C, Boyer M, Bakker AQ, Spits H, Taylor N. Green fluorescent protein as a selectable marker of fibronectin-facilitated retroviral gene transfer in primary human T lymphocytes Human Gene Ther 1999; 10: Galy A, Christopherson I, Ferlazzo G, Wesa A, Spits H, Georgopoulos K. Human lymphoid-related DC: A distinct population of antigen presenting cells requiring Ikaros for hematopoiesis. Blood (in press). Haks MC, Oosterwegel MA, Blom B, Spits H, Kruisbeek AM. Cellfate decisions in early T-cell development: regulation by cytokine receptors and the pre-tcr Semin Immunol 1999; 11: TNM Schumacher PhD JBAG Haanen MD PhD HH Kessels MSc 1 K Schepers MSc 2 G Stötter MSc 3 M Wolkers MSc M Toebes 2 M Van Den Boom 1 D Fontijn M Coccoris Group leader Academic staff Graduate student Graduate student Graduate student Graduate student Undergraduate student Undergraduate student The aim of our research is two-fold: 1) the development of novel strategies that can be used to examine protein interactions which should be of general value for biological and pharmaceutical research; 2) the use of such tools to unravel and manipulate the molecular processes underlying immune recognition by T lymphocytes. Amongst other things, in the past year this combination of technology development and biological research has resulted in a strategy for retroviral T-cell

55 56 IMMUNOLOGY Figure IV.1 A retroviral TCR display library, based on an influenza A-specific T-cell receptor (F5) was screened for variant T-cell receptors specific for the CTL epitopes from influenza strains A/NT/60/68 (NT-1) and A/PR8/34 (PR-1). Biological function of the selected T-cell receptors was assessed by monitoring ligand-induced T-cell activation, using a YFP reporter construct. Note that the in vitro selected TCRs display absolute ligand specificity for the T-cell epitope. receptor display that may be used to select TCRs with desirable specificities in vitro, and has led to an understanding of the role of memory T cells during the encounter of random antigenic variants in vivo. Molecular recognition Specificity and affinity in SH2-ligand interactions Protein-protein interactions are often mediated by the recognition of short continuous amino acid stretches by protein domains. Affinity-based selection strategies have successfully been used to define recognition motifs for a large series of protein domains, such as SH2, SH3 and PTB domains. However, it is unclear whether in vivo ligands do evolve towards an optimal affinity and, indeed, many physiological ligands only partially conform to the motifs that have been identified. We have developed a novel library screening technology that can be used to directly define ligands for protein domains based not only on affinity but also on specificity of binding. We have documented the value of this approach by the selection of peptide ligands that are either highly specific for the GRB2 SH2 domain, or that are cross-reactive between distantly related SH2 domains. Comparison of these selected ligands with previously identified physiological ligands for the GRB2 SH2 domain suggests that for many of these ligands, specificity of binding may be an important factor in ligand selection in vivo. Isolation of low affinity ligands The development of technologies that allow the assembly and screening of synthetic libraries of small molecules holds promise for the identification of novel ligands for target molecules of therapeutical interest. However, for most target molecules no high affinity ligands are represented in these molecular libraries. In addition, the available strategies for the identification of ligands from these libraries are ill-suited for the detection of low affinity ligands. The identification of such low affinity ligands would be of significant value, since the definition of these ligands would greatly facilitate the subsequent identification of high affinity ligands, by permitting a non-random walk through sequence space. We have developed a strategy that can be used to enhance the affinity of the interaction between library members and target molecules in an artificial manner using disulphide chemistry. This approach rests on simple thermodynamic principles and requires no prior knowledge about the binding site or ligand preference of the target molecule. We have demonstrated the value of this strategy for the identification of low affinity ligands for target molecules by the selection of peptide ligands for the N-SH3 domain of the adaptor protein c-crk. The low affinity ligands identified using this strategy should prove useful substrates for approaches for affinity optimization. Retroviral TCR display The diversity of the T-cell receptor (TCR) repertoire is limited due to the processes of positive and negative T- cell selection. In order to obtain T cells with specificities beyond the immune system s capacity, we have developed a strategy for retroviral TCR display. Using this approach, an in vitro T-cell library of variants of an influenza A-specific T-cell receptor was generated and screened for the occurrence of TCRs with a desired specificity. We have documented the value of TCR display by selection of variant T-cell receptors that either recognize the parental T-cell epitope, or that have acquired a novel specificity for a different influenza A strain. In collaboration with the group of H Spits we have shown that these manipulated TCRs initiate full T-cell activation upon ligand recognition in vitro (Figure IV.1). Retroviral TCR display opens the possibility for the generation of high affinity tumor specific TCRs, by circumventing T-cell self tolerance. This potentially important application of retroviral TCR display will be the subject of further studies.

56 Dissecting cytotoxic T-cell responses In situ detection of antigen specific T-cell immunity Over the past years, the development of novel strategies that can be used to visualize antigen specific T-cell responses by flow cytometry has radically changed our understanding of the dynamics and magnitude of antigen-specific T-cell immunity. Through the use of tetrameric MHC complexes, massive expansions of antigenspecific CD8 + T cells have been demonstrated and characterized for a variety of pathogens. However, the inability to use these reagents for the detection of antigen-specific T-cell responses in tissue sections has precluded a parallel anatomical dissection of T-cell immunity. In collaboration with F Vyth-Dreese from the Immunotherapy group, L Oomen from the Department of Biophysics, and the group of A Kruisbeek, we have recently developed a strategy that permits the use of MHC tetramers for the direct in situ visualization of antigen specific T-cell responses. The combined use of MHC tetramer technology and confocal laser scanning microscopy allows detection of both virus- and tumorspecific CD8 + T cells in lymphoid organs, peripheral tissues and tumor infiltrates. The value of this technique for the study of the anatomy of tumor-reactive T-cell immunity, in particular in melanoma patients, will be the subject of future studies. T-cell responses against antigenic variants The role of memory T cells during the immune response against random antigenic variants has not been resolved. In collaboration with the group of Ada Kruisbeek we have used a strategy employing the simultaneous staining of polyclonal T-cell populations with two tetrameric MHC-peptide molecules to address this issue. These experiments demonstrate that the polyclonal CD8 + T-cell response against a series of natural variants of the influenza A CTL epitope is completely dominated by infrequent cross-reactive T cells that expand from an original memory T-cell population. Based on both biochemical and functional criteria, these cross-reactive cytotoxic T cells productively recognize both the parental and the mutant epitopes in vitro and in vivo. These results provide direct evidence that polyclonal memory T- cell populations can provide protection against a range of antigenic variants, and suggest that random mutations of TCR-exposed residues of CTL epitopes are unlikely to result in efficient CTL escape. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: NWO (Netherlands Organization for Scientific Research), Project Publications: Analysis of T-cell function Haanen JBAG, Toebes M, Cordaro TA, Wolkers MC, Kruisbeek AM, Schumacher TNM. Systemic T-cell expansion during localized viral infection. Eur J Immunol 1999; 29: Schumacher TNM. Accessory to murder. Nature 1999; 398: Heemskerk M, Hooijberg E, Ruizendaal JJ, Van Der Weide MC, Kueter E, Bakker AQ, Schumacher TNM, Spits H. Enrichment of an antigen specific T-cell response by retrovirally transduced human dendritic cells. Cell Immunol 1999; 195: Diehl L, Den Boer ATh, Schoenberger SP, Van der Voort EIH, Schumacher TNM, Melief CJM, Offringa R, Toes R. CD40 activation in vivo overcomes peptide-induced peripheral cytotoxic T- lymphocyte tolerance and augments anti-tumor vaccine efficacy. Nature Med 1999; 5: Haanen JBAG, Wolkers MC, Kruisbeek AM, Schumacher TNM. Selective expansion of cross-reactive CD8 + memory T cells by viral variants. J Exp Med 1999; 190: Smyth Reynolds LA, Tyrrell R, Williams O, Norton T, Tybulewicz V, Schumacher TNM, Ley SC, Kioussis D. Altered peptide ligands and signalling in primary thymocytes. In: Cold Spring Harbor Symposia on Quantitative Biology, Vol LXIII. Signaling & Gene Expression in the Immune System, Cold Spring Harbor: CSH Press, 1999 (in press). GC De Gast 1 MD PhD D De Jong MD PhD 2 FA Vyth-Dreese PhD M Van Oijen PhD A Van Puijenbroek PhD W Vooys PhD N Verra MSc 2 TAM Dellemijn J Sein 3 WF Van de Kasteele CJC Van den Bogaard 3 S Elias Immunotherapy Group leader Academic staff Academic staff Graduate student Undergraduate student The role of T-cell co-stimulation in low grade B-cell non Hodgkin s lymphoma (B-NHL) In a previous study, immunohistochemical analysis of B-cell non Hodgkin s lymphomas (B-NHL) showed the 57 IMMUNOLOGY

57 58 IMMUNOLOGY presence of co-stimulatory receptors on the tumor B cells and their ligands on intermingled T lymphocytes. The preferential expression of these molecules, as well as cytokines, in low grade follicular lymphomas and extra nodal mucosa-associated lymphoid tissue (MALT) NHL suggested that these tumors are most optimally equipped for, and possibly dependent on, functional interactions with T cells (Vyth-Dreese et al, Immunology 1998; 94: 580-6). To test this hypothesis, patients with low grade, gastric MALT NHL, who were treated with antibiotics to eradicate Helicobacter pylori as anti-cancer therapy (in collaboration with H Boot, Department of Gastro-enterology, Division X), were studied. The H. pylori eradication treatment is known to result in remissions in about 70% of low grade MALT NHL patients, supporting a relative dependence of this B-cell lymphoma on H. pylori antigenic stimulation. A correlation was observed between the immunohistologically determined expression of co-stimulatory receptors CD80 and CD86, the histopathological grading of blast-like cells of type 2 and a clinical response to H. pylori eradication therapy. The implications of these observations are that in situ demonstration of co-stimulatory molecules in MALT NHL tumor tissues 1) indeed reflects a physiological process; 2) may predict whether patients will respond to subsequent H. pylori eradication treatment. Further analysis of tumor tissues from responding patients may reveal whether qualitative and/or quantitative alterations of intermingled T cells play a role in this response. Immunohistochemical analysis of renal cell carcinomas (RCC) from patients treated with immunotherapy In a dose-toxicity pilot-study (in collaboration with S Horenblas, Division XI), triple cytokine immunotherapy consisting of interleukin-2, interferon-α and granulocytemacrophage-colony-stimulating factor, has been given to metastatic RCC patients. From 8 out of 35 treated patients, tumor material was obtained after treatment. These tumor samples were compared to a control panel of 10 non-treated RCC for a large set of markers related to immunological reactions. In half of the treated patients, a more pronounced infiltrate was found compared to the control RCC. Furthermore, in 5 out of 8 patients, infiltrating T cells were partly located between the tumor cells, whereas in the control RCC, T cells were predominantly found in the stroma. Moreover, several markers were found to be expressed differentially in the treated tumors compared to the control RCC. An increase was found in treated RCC for dendritic cells and an adhesion marker. A decreased expression was found for resting macrophages/dendritic cells, an activation marker and several co-stimulation markers on T cells. These data suggest that triple cytokine immunotherapy induces a shift in the activation state of the tumor associated infiltrate, and increases the number of dendritic cells needed to initiate a specific immune response. Detection of dendritic cells of different subtypes and maturation stages in tumor tissues In collaboration with the group of H Spits, we described precursor lymphoid dendritic cells (DC) in the human thymus that have been suggested to have a function in the survival of activated T cells. We therefore searched for cells with the phenotype of precursor lymphoid DC in tumor tissues of patients with renal cell carcinoma (RCC) and head and neck squamous cell carcinoma (HNSCC), using immunohistochemistry and confocal laser scanning microscopy (CLSM) analysis. Cells expressing typical lymphoid DC markers were identified in 5/6 tumor samples from HNSCC patients, 4 of which also expressed the mature DC marker DC LAMP. Lymphoid DC were also identified in tissues from 5/8 RCC patients treated with combined immunotherapy with IL-2, GM-CSF and IFNα, but in 0/6 untreated patients. RCC patients did not show expression of mature myeloid DC markers regardless of whether they were treated or not. Further confirmation on the presence of precursor lymphoid DC in tumor tissues from treated RCC patients was obtained using 3-color immunofluorescence and CLSM analysis; myeloid origin could be excluded. This surprising presence of precursors of lymphoid DC in tumor infiltrates warrants further investigation of their functional role. Detection of antigen specific T lymphocytes in tumor tissues and blood In collaboration with the group of T Schumacher, a new strategy has been developed to visualize antigenspecific T cells in tissues through the use of tetrameric peptide/mhc complexes. This approach, when applied to patients, will provide a useful tool for further detailed studies on T-cell/tumor cell interactions in tumor infiltrates, and will also be applied in the monitoring of immunotherapy trials. Tetramer technology using HLA A2/melanosomal peptides is also used for the detection of CTLs in peripheral blood of melanoma patients before and after nonspecific immunotherapy with GM-CSF, low dose IL-2 and IFNα. In several patients, a ten-fold increase in the percentage of Mart-1 specific CTLs has been observed. This creates a good basis for following the fate of tumorspecific CD8 T cells in melanoma patients enrolled in different immunotherapy trials. Immunotoxins Reed-Sternberg (R-S) cells, the putative tumor cells of Hodgkin s disease, consistently express the antigens CD80 and CD86. The cells are very sensitive for immunotoxins directed to these antigens. Efficacy of immunotoxins is highly dependent on internalization of the complex by the target cell. Internalization experiments with iodinated anti-cd80 and anti-cd86 mabs indeed show that both antigens are internalized upon binding of the mab. Some R-S cell lines exhibit a strong induction of aggregation in the presence of anti-cd86 mab, but

58 not of anti-cd80. This aggregation does not match other previously described types of homotypic aggregation, as it is not inhibited by blocking mab to LFA-1 and ICAM-1, and is only partly dependent on Ca and Mg. A number of different anti-cd86 mab recognizing different epitopes, is capable of inducing aggregation and we speculate that this reflects CD86-mediated signaling. Indeed, Western blot analysis of cells treated with anti-cd86 mab revealed activation of the ERK1/2 pathway only in the cell lines that show aggregation. Since the CD86 molecule itself does not become phosphorylated on its serine residues in the intracytoplasmic tail, another molecule might be involved which is associated with the CD86 molecule. Such association has already been reported for CD80. Immunotoxins are also explored for treatment of multiple myeloma. The effect of an anti-cd138 plasma cell-specific immunotoxin (B-B4-SO6) in combination with the chemotherapeutic drug doxorubicin was evaluated in vitro on drug-sensitive and drug-resistant variants of the multiple-myeloma (MM)-derived cell line RPMI8226 and freshly isolated malignant myeloma cells. Drug-resistant RPMI8226 cells were still sensitive to the immunotoxin, although to a lesser extent than drug-sensitive cells. In the clonogenic assay, drug-sensitive RPMI8226 cells are several logs more susceptible to immunotoxin induced death than the drug-resistant RPMI8226 cells. When a sub-optimal dose of immunotoxin was combined with doxorubicin, an additive effect was found for drugsensitive RPMI8226 cells. The immunotoxin did not influence the sensitivity of resistant cells for doxorubicin. Both drug-sensitive and drug-resistant cells from MM patients were also sensitive to the immunotoxin. Doxorubicin was effective only on sensitive MM cells, and there was a tendency for an additive effect in the combination of immunotoxin and doxorubicin on these cells. These studies will form the basis for further dissecting the parameters dictating immunotoxin-sensitivity. Notes 1 Division X. 2 Division XIII. 3 Laboratory of Clinical Chemistry / Hematology- Immunotherapy Laboratory. Publications: Immunotherapy Post J, Vooijs WC, Bast BJEG, De Gast GC. Efficacy of an anti- CD138 immunotoxin and doxorubicin on drug resistant and drugsensitive myeloma cells. Int J Cancer 1999; 83: Res PCM, Couwenberg F, Vyth-Dreese FA, Spits H. Expression of ptα mrna in a Committed Dendritic Cell Precursor in the Human Thymus. Blood 1999; 94: MJ Kersten MD PhD M Schreurs PhD 1 P Van den Berk Immunotherapy of B cell non-hodgkin s lymphoma Group leader Advanced follicular non-hodgkin s lymphoma (FL) is at present an incurable disease, with a median survival of 8 years. FL forms an attractive target for immunotherapy because it represents a well-defined clinico-pathological entity with an indolent disease course, and because it possesses a tumor- and even patient-unique tumor antigen, the idiotype. Research of this recently formed group will focus on two projects (in collaboration with the group of H Spits, and with R van Oers, Department of Hematology, AMC, Amsterdam). Development of a novel lymphoma mouse model Current lymphoma mouse models using transplanted lymphoma cell lines or lymphomas that are virally or chemically induced, resemble acute leukemia or aggressive lymphoma more than FL. Major disadvantages of existing transgenic mouse models are that the introduced oncogene is present in the whole B-cell compartment, and that they often lead to aggressive lymphomas, which are more often of T-cell than of B-cell phenotype. In collaboration with P Krimpenfort (Division VII), we aim to develop a novel mouse model with an increased incidence of spontaneous lymphomas, with a known idiotype. For this purpose, anti-hel-ig transgenic mice will be crossed with bcl2 transgenic and/or p16 conditional knockout mice. Cre/Lox switching for the p16 tumor suppressor gene will take place either in vitro or in vivo. Other combinations of transgenes are currently being explored. Idiotype DNA vaccination Due to its flexibility, DNA vaccination offers an ideal model to study idiotype vaccination. Thusfar, idiotype vaccination has been shown to mediate tumor protection mostly through a humoral anti-idiotype response. In both a transplantation lymphoma mouse model, and in the transgenic mouse model, we aim at developing a DNA vaccine which will (through either targeting to antigen presenting cells, or targeting into the MHC class I pathway, or through combination with different cyto- or chemokines) elicit a protective CD8 + cytotoxic T-cell response directed against the lymphoma idiotype. Note 1 Funding: Academic Medical Center, Amsterdam. 59 IMMUNOLOGY

59 60 IMMUNOLOGY Publications: Immunotherapy of NHL Berger C, Van Baarle D, Kersten MJ, Klein MR, Samer Al-Homsi A, McQuain C, Van Oers MHJ, Knecht H. Carboxy terminal variants of Epstein-Barr virus encoded LMP1 during longterm HIV-infection: reliable markers for individual strain identification. J Infect Dis 1999; 179: Van Baarle D, Hovenkamp E, Kersten MJ, Klein MR, Miedema F, Van Oers MHJ. Direct Epstein-Barr virus typing on peripheral blood mononuclear cells: no association between EBV type 2 infection or superinfection and the development of acquired immunodeficiency syndrome-related non-hodgkinís lymphoma. Blood 1999; 93: Secretary Division of Immunology MA Van Halem Research staff positions (full time equivalents) Scientific permanent: 5.0 Scientific project: 18.0 Technical permanent: 6.1 Technical project: 7.0

60 V Division of Molecular Biology 61 Division head Ronald Plasterk Introduction Major research themes in the division are gene rearrangements, gene expression and signal transduction, functional genomics and drug resistance of tumor cells. The Borst group is working on two topics, antigenic variation in African trypanosomes, and multidrug resistance in tumor cells. The goal of the trypanosome project is to determine how trypanosomes manage to repeatedly change their surface molecules allowing them to escape immune destruction in the bloodstream of the host. Recent work focuses on the role of base J, β- glucosylhydroxymethylu, a new minor base discovered by the Borst group in trypanosomes. In 1999 the Borst group identified a protein that specifically binds to J in DNA. Disruption of the gene for this J-binding protein (JBP1) was found to have no gross effect on growth or gene expression, but reduced the level of J in the T. brucei genome 10-to 30-fold. It therefore appears that J- binding protein can set up a chromatin structure resulting in additional conversion of thymine into J. A second J- binding protein was identified and this JBP2 is being characterized. In the Borst multidrug resistance (MDR) group the emphasis is on MRPs (multidrug resistance proteins), a family of 6 related membrane proteins. Two of these, MRP1 and 2, have been shown to be organic anion pumps able to transport drugs conjugated to glutathione out of the cell. In 1999, the Borst group showed that cells overexpressing MRP3 can become resistant to etoposide and methotrexate, whereas overexpression of MRP5 results in resistance to purine base analogs, probably because MRP5 is an organic anion pump able to transport nucleotide analogs. This is an interesting new addition to the range of drug classes affected by drug pumps of the MRP family. The Plasterk group studies the nematode Caenorhabditis elegans as a model system, in particular the mechanism and regulation of DNA transposition and of signal transduction in the nervous system. This year the group identified a gene that is required for silencing of transposable elements in the germline of the animal; loss of function mutants of this gene (mut-7) show frequent mutations in progeny animals, as a result of jumping of Tc1, Tc3 and other transposons into genes. Surprisingly this gene, as well as many other similar mutator genes are also required for the effect of RNA interference (RNAi), the silencing of gene expression after experimental introduction of double stranded RNA. This suggests that the function of the RNA interference system may be to silence repetitive elements, such as transposons and viruses. Using the available complete sequence of the C. elegans genome, the group identified many SNPs (single nucleotide polymorphisms), and developed a method to use these SNPs for quick gene mapping (using so-called snip-snps, polymorphisms that can be scored because they change a restriction enzyme cleavage site). This method is now being used to identify genes involved in perception. A complete inventory was made of all G protein genes of the animal: for each of the 25 genes the expression pattern was studied, as well as the phenotype of knock out mutants; this showed that most G protein genes are specifically expressed in sensory neurons and are involved in perception. The division is undergoing major changes. P Borst gave up his directorship of the NKI/AvL, and will be able to focus his attention on his own research projects. Having reached the age of mandatory retirement by Dutch law, he will continue his own research group for at least five more years, as an honorary staff member and member of the division. R Plasterk will become the new director of the Hubrecht Laboratory in Utrecht and he and his research group will move to Utrecht early next year. In the coming year the division will therefore undergo major changes, and see new research groups move in.

61 MOLECULAR BIOLOGY 62 Caenorhabditis elegans RHA Plasterk PhD HGAM Van Luenen PhD 1 E Cuppen PhD 2 I Garcia-Saez PhD 3 Z Iszvak PhD 3 Z Ivics PhD G Jansen PhD 4 M Tijsterman PhD 5 C Van den Berg PhD 6 M Verheijen PhD 7 S Wicks PhD 8 E Berezikov SEJ Fischer 9 M Hilliard R Ketting 10 R Koch 11 C Moorman 8 F Simmer 12 S Van der Linden 2 E Hazendonk K Thijssen 2 M Van der Horst 8 T Haverkamp J Pothof E Wienholds Group leader Academic staff Graduate student Graduate student Graduate student Graduate student Graduate student Graduate student Graduate student Graduate student Undergraduate student Undergraduate student Undergraduate student Transposons in Caenorhabditis elegans and other organisms The Tc1/mariner superfamily of transposons is found in diverse species, from fungi to human, and seems to have spread by horizontal as well as vertical transfer. The elements are in essence single genes that encode the transposase protein, flanked by inverted repeats that the transposase protein recognizes to initiate the transposition process. The most active transposon in C. elegans is Tc1. Over the years the group has studied the mechanism and regulation of jumping of this element (plus the related Tc3 transposon). This year we finished a study to define the minimal cis-requirements for transposition; it turns out that approximately 100 base pairs of the termini of Tc1 and of Tc3 are sufficient for jumping. We compared the frequency of jumping of Tc1 and Tc3 in heterologous hosts (mouse NIH 3T3 cells and human HeLa cells) to that of the mariner element from Drosophila and the Sleeping Beauty transposon (reconstructed from several dead transposons in the fish genome). We found that Sleeping Beauty is 40 times more active. Most probably this indicates that Tc1 and Tc3, as naturally occurring elements, have been selected for sub-optimal activity (to keep their hosts alive), while the Sleeping Beauty element was a reconstructed element in its most aggressive state. The genome of the second isolate of C. elegans (Bristol N2) contains 33 copies of Tc1. We analyzed their sequences in detail, and found that virtually all have a distinctive set of SNPs (Single Nucleotide Polymorphisms). These can be used to identify and follow individual elements in the genome. Surprisingly there are three cases of elements with exactly the same SNP types; these are all within 2 map units of a sibling element, suggesting they once jumped to a neigbouring position in the genome. The Bristol N2 strain itself is not permissive for transposition: the pattern of Tc1 and other transposons remains exactly identical in consecutive generations. We previously described the isolation of an EMS-induced mutant derivative of Bristol N2 in which Tc1 and other transposons had become mobile (also see below). In this strain we see new Tc1 insertions appear, and we inspected their SNP patterns. One conclusion is that there is some preference for local jumping, so that new Tc1 insertions often have SNP types similar to that of a donor locus nearby in the genome. However, in some case we see new SNP types, that are combinations of SNPs of elements originally present in the N2 isolate. Apparently recombination has occurred between Tc1 types. Interestingly some of the resident Tc1 copies, still present at the position in the genome where they were in the starting strain, had also switched SNP type. This suggests that recombination may not occur to elements that are excising and reintegrating into new loci, but that it occurs during the DNA double strand break repair of the donor site after Tc1 excision. We refer to this phenomenon as transposon type switching, which seems mechanistically related to mating type switching in yeast. As mentioned above, we isolated multiple mutant Figure V.1 A model for MUT-7 action. A complex of proteins, indicated by X,Y, and MUT-7, binds a dsrna molecule. The dsrna specifically targets the complex to homologous mrna by an yet unknown mechanism, which needs at some level base pairing between the dsrna and the target mrna. This leads to degradation of the bound mrna by the Rnase function of MUT-7. Because of the homology to RNaseD, this is expected to proceed in a 3-5 direction. The dsrna could survive this process and target the nuclease complex to the next mrna molecule.

62 derivatives of N2 in which repression of transposition was lost. We now identified one of the genes, mut-7, at the molecular level. It encodes a putative nuclease, homolog of RNaseD, a 3 to 5 exonuclease. This may explain the intriguing observation that mut-7 mutants, as well 20 other mutator mutants we isolated, are resistant to the experimental silencing of gene expression by introduction of double stranded RNA, referred to as RNAi or RNA interference. We postulate that the MUT-7 protein, probably complexed to other proteins, and to a guide RNA which is double stranded, targets mrna molecules (based on homology to the dsrna), and degrades them enzymatically (see Figure V.1). This effect is lost in mut-7 mutants. The relation to transposon silencing may be that transposons, as soon as they occur in multiple copies in the genome, are bound to be transcribed by fortuitous read-through transcription from flanking promoters; the result would be RNA made from each of the transposon strands, allowing formation of dsrna. In this model the RNAi machinery is a surveillance system that recognizes repetitive sequences in the genome, sees them as potential threats, molecular parasites, and represses expression of their transposase. Signal transduction in the C. elegans nervous system We studied the entire family of G proteins, which are the connectors between membrane spanning receptors that recognize the outside world of cells and intra-cellular signal transduction cascades. The family consists of two genes encoding the beta subunits, two (recognized thus far) for the gamma subunits, and twenty for the alpha subunits. As reported before, we analyzed the expression patterns of all these genes, and their phenotypes, and we conclude that most G protein genes are (almost) exclusively expressed in sensory neurons of the amphid organs, and that their mutation (loss or gain of function) in several cases affect chemotactic responses. We performed mutant screens to search for genes specifically involved in chemo-perception of soluble tastants, and found several such mutants. These are currently being mapped, as well as a mutation previously identified by the lab of C Bargmann (University of California, San Francisco), adp-1, which affects adaptation to chemical tastants. Functional genomics of C. elegans We screened wild type isolates for Single Nucleotide Polymorphisms (SNPs) by shotgun sequencing of genomic DNA and comparison to the canonical sequence of the Bristol N2 isolate. We found an overrepresentation of SNPs in non-coding regions, and within coding regions in silent third-base positions. Less expected was that we found a strong clustering of SNPs in the arms of the chromosomes compared to the central regions. This fits with the observation that the centers contain a higher proportion of conserved genes (as scored by matches to yeast genes), and suggests that there are higher rates of evolutionary changes of genes in the chromosome arms. SNPs are ideal markers for gene mapping. We selected a set of what we call snip-snps, these are SNPs that change the restriction pattern of a DNA region, and can thus be recognized by PCR followed by a restriction enzyme snip. In collaboration with the group of R Waterston (Washington University, St. Louis) we selected a set of snip-snps that cover the entire genome, so that genes can now be mapped in a quick, easy and cheap fashion. These data are being made available to the research community via the world wide web, in an easily searchable fashion. Notes 1 Funding: Fellowship Human Frontier Science Program Organization (HFSPO). 2 Funding: Netherlands Organization for Scientific Research (NWO ). 3 Funding: EMBO fellowship. 4 Funding: Netherlands Organization for Scientific Research (NWO ). 5 Funding: SmitKline Beecham. 6 Funding: Netherlands Organization for Scientific Research (NWO ). 7 University of Utrecht, Utrecht. 8 Funding: Netherlands Organization for Scientific Research (NWO ). 9 Funding: Netherlands Organization for Scientific Research (NWO/Stichting Scheikundig Onderzoek: SON ). 10 Funding: Netherlands Organization for Scientific Research (NWO ). 11 Funding: Center for Biomedical Genetics. 12 Funding: Netherlands Organization for Scientific Research (NWO/Stichting Scheikundig Onderzoek: SON ). Publications: C. elegans Eijkelenboom APAM, Sprangers R, Hard K, Puras Lutzke RA, Plasterk RHA, Boelens R, Kaptein R. Refined solution structure of the C-terminal DNA-binding domain of HIV-1 integrase. Proteins 1999; 36: Fischer SEJ, Van Luenen HGAM, Plasterk RHA. Cis requirements for tramsposition of Tc1-like transposons in C. elegans. Mol Gen Genet 1999; 262: Jansen G, Thijssen KL, Werner P, Van der Horst M, Hazendonk E, Plasterk RHA. The complete family of genes encoding G proteins of Caenorhabditis elegans. Nature Genetics 1999; 21: Ketting RF, Haverkamp THA, Van Luenen HGAM, Plasterk RHA. mut-7 of C. elegans, required for transposon silencing and RNA interference, is a homolog of Werner syndrome helicase and RNaseD. Cell 1999; 99: Korswagen HC, Van der Linden AM, Plasterk RHA. G-protein 63 MOLECULAR BIOLOGY

63 64 MOLECULAR BIOLOGY hyperactivation of the Caenorhabditis elegans adenylyl cyclase SGA-1 induces neuronal degeneration. EMBO J. 1998; 17: Plasterk RHA. Hershey heaven and Caenorhabditis elegans. Nature Genetics 1999; 21: Plasterk RHA. The year of the worm. BioEssays 1999; 21: Plasterk RHA, Izsvak Z, Ivics Z. Resident aliens, the Tc1/mariner superfamily of transposable elements. Trends in Genetics 1999; 15: Plasterk RHA, Van Luenen HGAM, Vos C, Ivics Z, Isvak Z, Fischer S. Elements transposables: des outils permettant la mutagenese et la transgenese chez les vertebres. Pathologie Biologie 1998; 46: Van den Ent FMI, Vos A, Plasterk RHA. Dissecting the role of the N-terminal domain of human immunodeficiency virus integrase by trans-complementation analysis. J of Virology 1999; 73: P Borst PhD F Baas PhD 1 R Evers PhD 2 Z Hollo PhD 3 M Kool PhD 2 T Saeki PhD 4 A Van Helvoort PhD 5 P Wielinga PhD 2 J Wijnholds PhD 2 J Guo 6 N Zelcer 2 M De Haas 2 C Mol 2 L Van Deemter 2 Multidrug resistance Group leader Academic staff Graduate student Graduate student Multidrug resistance (MDR) of cancer cells is usually caused by drug pumps that lower the intra-cellular drug concentration by actively extruding drug from the cell. These pumps are large glycosylated plasma membrane proteins and of two types: P-glycoproteins (Pgps) and Multidrug Resistance (associated) Proteins (MRPs; see Figure V.2). Whereas Pgps extrude unmodified hydrophobic drugs, MRPs are also able to extrude drugs conjugated to glutathione (GSH), glucuronic acid, or sulphate. Carriers with this property are known as GS-X pumps. To study the physiological function of Pgps and MRPs we have disrupted the genes for these proteins in mice. In this way we have shown that these transporters have an important function in protecting mammals against hydrophobic drugs, but that they are not required for the well-being of mice in the protected environment of the NKI/AvL. The work on drug-transporting Pgps is being continued by the group of AH Schinkel (Division VIII). We are concentrating on the further characterization of the MRP gene family. The Multidrug Resistance (associated) Protein 1 To elucidate the physiological role of MRP1, we have generated mice lacking a functional Mrp1 gene, as well as mice with disrupted Mrp1, Mdr1a, and Mdr1b genes, in collaboration with P Krimpenfort and AJM Berns (Division VII). Mice homozygous for disrupted Mrp1, or Mrp1/Mdr1a/1b genes, are viable and fertile under laboratory conditions, but the Mrp1 -/- mice are hypersensitive to etoposide and the Mdr1a/1b -/-, Mrp1 -/- even more so. The hypersensitivity of Mrp1 -/- mice to etoposide is especially manifest in cells which contain substantial amounts of Mrp1 in wild-type mice. In collaboration with L De Lange and D Breimer (Pharmacology, University of Leiden) we have shown that the high Mrp1 concentration in the basolateral membrane of the epithelial cells of the choroid plexus protects the cerebrospinal fluid (CSF) against etoposide. In Mdr1a/1b -/-, Mrp1 -/- mice the etoposide level in the CSF rises 10-fold higher after i.v. administration than in the Mdr1a/1b -/- control mice. We have previously shown that P-glycoprotein in brain capillaries is important to protect the brain from amphipathic drugs. MRP1 appears to play a similar role for drugs that attempt to reach the brain via the choroid plexus and CSF. MRP3 In addition to MRP1, we identified 5 other members of the human MRP family, cmoat or MRP2, and MRP3-6. One of these, MRP2, has been characterized as an organic anion transporter responsible for secretion of bilirubin glucuronides and other organic anions from the liver into bile. For a more detailed analysis of MRP3, we isolated monoclonal antibodies specific for this protein (in collaboration with the group of R Scheper, Free University, Amsterdam) and constructed DNAs complementary to full-length mrna. These cdnas were stably introduced into various non-polarized and polarized cells by retroviral transduction. The only phenotype observed in transfected non-polarized cells is low-level resistance to etoposide and teniposide (2-to 4-fold) and high level resistance to brief (4 hour) incubations with high methotrexate (MTX) concentrations, but no resistance to low MTX during continuous exposure. In collaboration with G Jansen and coworkers (Medical Oncology, Free University, Amsterdam) we previously demonstrated similar results for cells overexpressing MRP1 and 2. It had long been known that mammalian cells contain several transporters able to export MTX. It now seems

64 Figure V.2: Membrane topology model of full-length MRP1 and MRP5. MRP1 consists of a MDR1-like core region containing two ATP binding sites (circles), two membrane-bound domains, and an N-terminal polypeptide region of about 280 amino acids, forming a membrane-bound domain and a cytoplasmic loop (TMD 0 and L 0, respectively). MRP5 lacks TMD 0. MRP1-3, but all our attempts to demonstrate MRP5- mediated resistance to MDR drugs have remained unsuccessful. However, we found low level resistance to sulphur-containing purine base analogs, 6-mercaptopurine (6-MP) and 6-thioguanine. When MRP5 transfectants are loaded with radioactive 6-MP, they export the radioactivity more rapidly than untransfected control cells. In the course of 1999 Schuetz and coworkers discovered that MRP4 is associated with resistance to nucleotide analogs, such as PMEA, 9-(2-phosphonylmethoxyethyl)adenine. We tested this drug on our MRP5-transfected cells and found low-level resistance. It is therefore possible that MRP5 is an organic anion pump with a preference for nucleotide analogs, resistance to 6- MP being due to the conversion of this base into its nucleotide analog, 6-thio-IMP. We have generated knock-out mice lacking MRP5. These mice are healthy and fertile showing that MRP5 has no essential function in mice in a protective environment. Whether these mice are hypersensitive to base/nucleoside/nucleotide analogs is under investigation. 65 MOLECULAR BIOLOGY likely that these transporters are the MRPs. Within the cell MTX is slowly polyglutamylated. We have found that the polyglutamylated forms are not transported by MRPs. The difference between short- and long-term MTX exposure is therefore probably due to competition between export by MRP and polyglutamylation by folylpolyglutatamate synthetase (FPGS), as illustrated in Figure V.3. During long-term MTX incubations enough MTX polyglutamates can accumulate to inhibit cell multiplication. In polarized cells overexpression of MRP3 gives rise to increased export of dinitrophenyl-glutathione at the basolateral side, but not of GSH itself (unlike MRP1 and 2). To study the substrate specificity of MRP3 in more detail we have expressed the protein in insect cells using a baculovirus MRP3 construct. Vesicles isolated from these cells transport estradiol-17-β-glucuronide in an ATP-dependent fashion. MRP3 levels are relatively high in small bile canaliculi and in the gut epithelium, and the low MRP3 level normally found in the liver is highly upregulated in liver disease. It is therefore possible that MRP3 plays a role in bile salt metabolism. We are generating mice with disrupted MRP3 genes to get more information about the physiological function of MRP3 and its potential for causing drug resistance. MRP5 MRP5 differs from MRP1, 2, 3 and 6 in that it lacks the TMD0 part of the protein (Figure V.2). We have cloned full-length MRP5 cdna and expressed this in transfected cells. In polarized cells MRP5 is routed to the basolateral membrane where it can mediate export of S- glutathionyl 2,4-dinitrobenzene and glutathione itself. MRP5 is also inhibited by benzbromarone, an inhibitor of other MRPs. MRP5 is therefore a GS-X pump, like Figure V.3 MRPs and folylpoly-γ-glutamate synthetase (FPGS) compete for MTX. MTX is taken up by the cells via the reduced folate carrier (RFC) and overexpression of MRP results in transport of unmodified MTX. Polyglutamylated MTX is formed in a relatively slow reaction by FPGS and this results in the efficient inhibition of dihydrofolate reductase (DHFR). Notes 1 Staff member Academic Medical Center, Amsterdam and honorary staff member NKI/AvL. 2 Funding: Dutch Cancer Society, Project NKI

65 66 MOLECULAR BIOLOGY 3 Funding: Netherlands Organization for Scientific Research, NWO-OTKA Funding: Japan Society for the Promotion of Science. 5 Funding: Netherlands Organization for Scientific Research, NWO Funding: Ministry of Education & Science of The Netherlands (O C & W). Publications: Multidrug resistance Allen JD, Brinkhuis RF, Wijnholds J, Schinkel AH. The mouse Bcrps/Mxr/Abcp gene: amplification and overexpression in cell lines selected for resistance to topotecan, mitoxantrone, or doxorubicin. Cancer Res 1999; 59: Arts HJ, Katsaros D, De Vries EG, Massobrio M, Genta F, Danese S, Arisio R, Scheper RJ, Kool M, Scheffer GL, Willemse PH, Van der Zee AG, Suurmeijer AJ. Drug resistance-associated markers P-glycoprotein, multidrug resistance-associated protein 1, multidrug resistance-associated protein 2, and lung resistance protein as prognostic factors in ovarian carcinoma. Clin Cancer Res 1999; 5: Borst P, Van Blitterswijk WJ, Borst J, Tepper AD, Schinkel AH. New physiological functions for drug-transporting P-glycoproteins? Drug Resistance Updates 1998; 1: Borst P. Multidrug resistance - A solvable problem? ECC: Biliopancreatic Malignancy: From Gene to Cure, Febr 4-6, Annals of Oncol. 1999; 10 S4: Borst P, Evers R, Kool M, Wijnholds J. The multidrug resistance protein family. Biochim Biophys Acta: special issue Structure and function of ABC transporters 1999; 1461: Borst P, Zelcer N, Van Helvoort A. ABC-Transporters in lipid transport. Biochim Biophys Acta; special issue on Intracellular Lipid Transport (in press). Van Helvoort A, De Brouwer A, Ottenhoff R, Brouwers JFHM, Wijnholds J, Beijnen JH, Rijneveld A, Van der Poll T, Van der Valk MA, Majoor D, Voorhout W, Wirtz KWA, Oude Elferink RPJ, Borst P. Mice without phosphatidylcholine transfer protein have no defects in the secretion of phosphatidylcholine into bile or into the lung airspaces. Proc Natl Acad Sci USA 1999; 96: Antigenic variation in Trypanosomes P Borst M Cross PhD 1 D Dooijes PhD 2 R Mussmann PhD 3 B Sabatini PhD 2 I Chaves 4 H Gerrits 2 S Ulbert 5 A Dirks 2 R Kieft 2 Group leader Graduate student Graduate student Graduate student Trypanosoma brucei, the causative agent of African Sleeping Sickness in humans, is a unicellular parasite transmitted to the mammalian host by the tsetse fly. In the bloodstream of the mammal, the trypanosome avoids immune destruction by repeatedly changing its Variant Surface Glycoprotein (VSG) coat. A given coat is encoded by a single gene from a repertoire of up to a thousand different VSG genes. To be expressed, these genes have to move to one of several expression sites, containing additional Expression Site Associated Genes (ESAGs) and located adjacent to a telomere (see Figure V.4). Antigenic variation occurs by replacing the VSG gene in the active expression site or by switching to another site. Hooijberg JH, Broxterman HJ, Kool M, Assaraf YG, Peters GJ, Noordhuis P, Scheper RJ, Borst P, Pinedo HM, Jansen G. Antifolate resistance mediated by the multidrug resistance proteins MRP1 and MRP2. Cancer Res 1999; 59: Kool M, Van der Linden M, De Haas M, Baas F, Borst P. Expression of human MRP6, a homologue of the multidrug resistance protein gene MRP1, in tissues and cancer cells. Cancer Res 1999; 59: Kool M, Van der Linden M, De Haas M, Scheffer GL, De Vree ML, Smith AJ, Jansen G, Peters GJ, Ponne N, Scheper RJ, Oude Elferink RPJ, Baas F, Borst P. MRP3, a new organic anion transporter able to transport anti-cancer drugs. Proc Natl Acad Sci USA 1999; 96: Paulusma CC, Van Geer MA, Evers R, Heijn M, Ottenhoff R, Borst P, Oude Elferink RPJ. The canalicular multispecific organic anion transporter (cmoat/mrp2) mediates low-affinity transport of reduced glutathione. Biochem J 1999; 338: Figure V.4 J in expression site. Schematic representation of an active and an inactive VSG gene expression site. The flag indicates the promoter, The modified DNA base J is present in the 50 bp repeat and telomeric repeat arrays flanking both active and inactive sites. In silent expression sites the VSG gene, 70 bp repeats and ESAGs are also modified. Control of VSG expression sites A central question in antigenic variation is how the bloodstream form trypanosome can silence all VSG expression sites but one, even though there are some 20 expression sites that look alike. Our earlier work suggested a model for expression site control involving three states: off, pre-active and active. Whereas the transition from off to pre-active or active is infrequent

66 (about 10-6), the transition from pre-active to active is sufficiently frequent to allow the selection of unstable double expressors. Such unstable double expressors might be intermediates in normal switching from one expression site to the next. To test this mechanism further we have marked a third expression site with a new resistance marker, blasticidin resistance. We have also succeeded in visualizing an expression site in living trypanosomes using the system developed by Belmont and coworkers (University of Illinois, Urbana, USA). This involves the insertion of a LAC operator gene array upstream of an expression site. When a construct encoding green fluorescent protein fused to a LAC repressor is introduced in the trypanosomes the operator gene array becomes a green fluorescent dot dancing in the microscope field in the living trypanosome. This system should allow us to keep track of expression sites when they are activated and inactivated. DNA modification and antigenic variation We have previously shown that bloodstream form T.brucei contains a novel DNA base b-d-glucosyl-hydroxymethyluracil, called J. To determine the location of J in the genome, we have generated J-specific antisera raised against JMP. By immunoprecipitation of J- containing DNA fragments, we showed that J is present in and around VSG genes in inactive expression sites, and never in transcribed VSG genes (see Figure V.3). Nevertheless, we have also found J in other repetitive sequences, which suggests that J might have a wider function and is not just involved in silencing of expression sites. This is supported by our identification of J in other kinetoplastids, including Leishmania, T.cruzi, and Crithidia, which do not undergo antigenic variation, and in Euglena. We did not find J in mammals, Drosophila, C.elegans, or yeast (Figure V.5). Figure V.5 Evolution tree. J is present in all kinetoplastida, in Diplonema and Euglena gracilis, but not elsewhere in nature. A J-binding protein The presence of J in DNA must somehow be translated into nuclear action (altered chromatin structure/transcription, etc.) and we expect this translation to be mediated by J-binding protein(s). We have found such a protein in nuclear extracts using a specific band shift assay of labeled oligonucleotides in acrylamide gels. The collaboration with J Van Boom s group in Leiden University, gave us access to a range of oligodeoxynucleotides containing either J, hydroxymethyluracil or thymine in defined sequences. We have purified the J- binding protein from Crithidia extracts and cloned the corresponding gene. J-binding protein is a novel protein with a predicted mass of 90.4 kda. By homology we also cloned the J-binding protein genes of T.brucei and Leishmania. To study the function of J-binding protein we have disrupted the two alleles of the gene in T.brucei. Remarkably, this has no effect on the multiplication of the trypanosomes, on their ability to keep silent VSG gene expression sites repressed, or on the rate of switching from one expression site to another. We find, however, that the J level of trypanosomes without J-binding protein is 10- to 30- fold lower than of wild-type trypanosomes. The difference is not due to turn-over of J. A possible explanation for this unusual result is that J-binding protein can change chromatin structure resulting in an additional conversion of Thy into J. As the effect of J- binding protein appears to be subtle, we are also disrupting the gene for J-binding protein in Leishmania major. This should allow us to test the effect of the absence of this protein in a complex competitive biological system, i.e. chronic infections in macrophages or mouse foot pads. By database search we identified a protein with significant homology to J-binding protein. We have cloned the gene for this large putative J-binding protein 2 and are testing its functions by gene disruption. The variant transferrin receptor of T. brucei Each VSG gene expression site contains a set of ESAG 6 and 7 genes (see Figure V.4) specifying a heterodimeric transferein (Tf) receptor (Tf-r), and the receptors encoded by different expression sites are similar but not identical. These small differences have dramatic effects on the ability to bind Tfs from different hosts. For instance, the Tf-R encoded by the 221 expression site does not detectably bind canine Tf, whereas the V02 site Tf-R does bind this Tf. Hence, we conclude that multiple VSG gene expression sites may primarily exist to allow the trypanosome to use a diversity of mammalian hosts. The ability to make multiple Tf-Rs binding the same host Tf may also help the trypanosome to evade the effects of accumulating antibodies in a chronic infection. To test this we have raised polyclonal antibodies against Tf-Rs encoded by different expression sites. We find that such antibodies inhibit Tf uptake by the homologous Tf-R more than by a heterologous Tf-R. This suggests that the 67 MOLECULAR BIOLOGY

67 68 MOLECULAR BIOLOGY ability to make multiple Tf-Rs may not only serve as an adaption to a diverse range of mammalian hosts, but also to allow antigenic variation of receptor during chronic infection in a single host. Notes 1 Funding: The Netherlands Cancer Institute. 2 Funding: Netherlands Organization for Scientific Research (NWO-Gebied Chemische Wetenschappen (CW ; CW ). 3 Funding: Fellowship Human Frontier Science Program Organization (HFSPO). 4 Funding: Gulbenkian PhD Program in Biology and Medicine. 5 Funding: Fellowship Boehringer Ingelheim. Publications: Antigenic variation Borst P, Chaves I. Mono-allelic expression of genes in simple eukaryotes. (Letter to) Trends in Genetics 1999; 15: Chaves I, Rudenko G, Dirks-Mulder A, Cross M, Borst P. Control of variant surface glycoprotein gene expression sites in Trypanosoma brucei. EMBO J 1999; 18: Cross M, Kieft R, Sabatini R, Wilm M, De Kort M, Van der Marel GA, Van Boom JH, Van Leeuwen F, Borst P. The modified base J is the target for a novel DNA-binding protein in kinetoplastid protozoans. EMBO J 1999; 18: McCulloch R, Barry D. A role for RAD51 and homologous recombination in T. brucei antigenic variation. Genes & Dev 1999; 13: Secretary Molecular Biology C Smit (until September) I Deltrap (from August) Secretary Director of Research I Van Lieverloo (till October ) B Van Houten Research staff positions (full time equivalents) Scientific permanent: 4.0 Scientific project: 31.0 Technical permanent: 4.5 Technical project: 2.8

68 68 MOLECULAR BIOLOGY ability to make multiple Tf-Rs may not only serve as an adaption to a diverse range of mammalian hosts, but also to allow antigenic variation of receptor during chronic infection in a single host. Notes 1 Funding: The Netherlands Cancer Institute. 2 Funding: Netherlands Organization for Scientific Research (NWO-Gebied Chemische Wetenschappen (CW ; CW ). 3 Funding: Fellowship Human Frontier Science Program Organization (HFSPO). 4 Funding: Gulbenkian PhD Program in Biology and Medicine. 5 Funding: Fellowship Boehringer Ingelheim. Publications: Antigenic variation Borst P, Chaves I. Mono-allelic expression of genes in simple eukaryotes. (Letter to) Trends in Genetics 1999; 15: Chaves I, Rudenko G, Dirks-Mulder A, Cross M, Borst P. Control of variant surface glycoprotein gene expression sites in Trypanosoma brucei. EMBO J 1999; 18: Cross M, Kieft R, Sabatini R, Wilm M, De Kort M, Van der Marel GA, Van Boom JH, Van Leeuwen F, Borst P. The modified base J is the target for a novel DNA-binding protein in kinetoplastid protozoans. EMBO J 1999; 18: McCulloch R, Barry D. A role for RAD51 and homologous recombination in T. brucei antigenic variation. Genes & Dev 1999; 13: Secretary Molecular Biology C Smit (until September) I Deltrap (from August) Secretary Director of Research I Van Lieverloo (till October ) B Van Houten Research staff positions (full time equivalents) Scientific permanent: 4.0 Scientific project: 31.0 Technical permanent: 4.5 Technical project: 2.8

69 VI Division of Tumor Biology 69 Division head Jacques Neefjes Introduction Within the Division of Tumor Biology, we use cell biology to understand intracellular transport processes, cell division, tumor progression and metastasis. Technologies to follow and manipulate these processes at virtually any level of resolution are operational in the division. These include molecular biological, biochemical and various morphological techniques such as confocal laser microscopy and electron microscopy. Active collaborations exist with various clinicians and scientists within and outside the NKI/AvL had a sad start because of the death of Ab Tulp, a loss that is still felt in the division. A Van der Gugten left because of early retirement. Three new projects were started this year. One on the regulation of transport of MHC class II molecules, one on the connection between the cytoskeleton and endo/phagosomes (both in the Neefjes group) and one on the routing of antigen uptake and presentation by the 4 different CD1 molecules that have recently been identified as antigen presenting molecules (Peters). Peters has now set up a lab for cryosectioning for immunoelectronmicroscopy. State-of-the-art technology is being introduced for this purpose. This already allows the production of top quality electronmicrographs which is recognized by many collaborating cell biologists. Real time confocal microscopy combined with micromanipulation of cells is a technique now operational in the lab of Neefjes. This allows the visualization of conformational changes and protein interactions in living cells. For example, Neefjes and collaborators have thus visualized conformational alterations in an ABC transporter, the peptide transporter TAP which is crucial for feeding MHC class I molecules with antigenic peptides. This single cell biochemistry approach is further used to study vesiclecytoskeleton interactions and the regulation of motor proteins that are essential in timing vesicle transport and DNA-segregation during mitosis (a process studied by Michalides). The combination of single cell biochemistry with electronmicroscopy will bring this type of research in a unique position. Antigen processing and presentation by MHC molecules J Neefjes PhD M Fernandez-Borja PhD 1 A Neisig PhD 2 SM Van Ham PhD 3 JC Vos PhD 4 RW Wubbolts PhD 5 I Jordens MSc 3 M Marsman MSc 6 EAJ Reits MSc 3 PJL Spee MSc 7 MJTM Van Lith MSc 3 S Dusseljee D Verwoerd Group leader Graduate student Graduate student Graduate student Graduate student Graduate student Antigen Presentation by MHC class I molecules and the action of the ABC transporter TAP MHC class I molecules present antigenic fragments that are mainly derived from degraded nuclear and cytosolic proteins. The fragments are subsequently translocated into the ER lumen for binding to MHC class I molecules. The ABC transporter TAP pumps the peptides into the ER. We had previously determined the peptide substrate specificity of TAP from human, mice and rat. We have now started a large series of biochemical experiments to determine the topology, subunit interactions and relative orientation of the two subunits of TAP. TAP has a head-head topology and three domains can be distinguished; a pore domain followed by a peptide-binding domain and, at the end, the two nucleotide-binding segments. ER-retention signals are found in the first trans-membrane segments of both TAP1 and 2 but other retention signals should be present as well in the pore segment. To follow the conformational changes in TAP during translocation, cells were made expressing GFP-labelled TAP. FRAP (Fluorescence Recovery After Photobleaching) experiments on living cells were performed to follow the activity of TAP (Figure VI.1). The lateral mobility of TAP was reduced by introduction of peptide and increased by inhibition of proteasomal activity as well as by ATP depletion. Experiments using the viral protein US6 or peptides with long side chains suggest that the alterations in mobility are the result of pore opening. These alterations were subsequently used to monitor the

70 70 TUMOR BIOLOGY Figure VI.1 Confocal analysis of GFP-labelled TAP molecules expressed in the endoplasmic reticulum of living cells. These cells were used to measure the lateral mobility of TAP in the presence or absence of peptides using bleaching protocol. peptide pool in living cells. We thus note a large increase in the intracellular peptide content upon an acute influenza infection. This is the result of additional degradation by the proteasome of influenza-encoded proteins. More notably, we then showed that a large content of peptides is derived from (endogenous and viral) proteins degraded very shortly after synthesis. The peptides that enter the ER can subsequently bind to a large set of chaperones, one of which is being used in vaccination trials. Using peptides with a photoaffinity label, we have identified other chaperones that are also tested in vaccination studies. The current dogma is that MHC class I molecules present peptides that are captured in the ER. In contrast to this we observed that a fraction of MHC class I molecules appears to enter endo/lysosomes and to recycle to the plasma membrane. Biochemical analysis showed that these were class I complexes that were able to exchange peptide at ph 5.0. Recycling class I molecules can thus exchange peptides and present novel peptides from endocytosed antigens. Antigen presentation by MHC class II molecules and the action of HLA-DO Unlike MHC class I molecules, MHC class II molecules present peptides derived from endocytosed antigens. Class II molecules are transported to early lysosomal structures termed MIIC where, with the help of a dedicated chaperone HLA-DM, they are loaded with peptide. HLA-DO binds to HLA-DM and is highly expressed in primary B cells. HLA-DO expression now appears to act as a ph sensor restricting the activity of HLA-DM to more acidic compartments and thus affecting the peptide repertoire associated to MHC class II Figure VI.2 Cytoplasmic tail of DOB contains internalizaition signal CD8 and CD8 DOB transfected into Meljuso stably transfected with DOαβGFP. molecules. HLA-DO is recycling along with HLA-DM between MIIC and the plasma membrane and rapid internalization from the plasma membrane is due to the tail of DMB as well as DOB (Figure VI.2). We have previously shown that retention and transport of MIIC is controlled by the counteracting activity of the motor proteins kinesin and dynein. We have now isolated a novel protein that induces the recruitment of the dynein-dynactin complex to MIIC and causes retention. We are currently studying the molecular basis for this. In order to present antigen, vesicles are internalized to meet the lysosome. These processes are critically dependent on the actin and microtubule network. We are studying the effect of a small GTPase called Rho B, which appears to inhibit an early step in endocytosis. Notes 1 Funding: EEC, Project TMR ERB FMRXCT Funding: The Alfred Benzon Foundation. 3 Funding: NWO, Project Funding: Dutch Cancer Society, Project NKI Funding: EEC, Project B Funding: NWO, Pioneer Project Funding: NWO, Project Publications: Antigen processing Beekman NJ, Van Veelen PA, Van Hall T, Neisig A, Sijts A, Camps M, Kloetzel P-M, Neefjes JJ, Melief CJ, Ossendorp F. Abrogation of CTL epitope processing by single amino acid substitution flanking the C-terminal proteasome cleavage site. J Immunol 1999 (in press).

71 Fernandez-Borja M, Wubbolts R, Calafat J, Janssen H, Divecha N, Dusseljee S, Neefjes J. Multivesicular body morphogenesis requires phosphatidylinositol 3-kinase activity. Curr Biol 1999; 9: Grommé M, Uytdehaag FG, Van Binnendijk RS, Kenter MJ, Tulp A, Verwoerd D, Neefjes J. Endosomal Antigen Presentation by Recycling MHC Class I Molecules. PNAS 1999; 96: Ultrastructural studies on membrane traffic in cytotoxic T lymphocytes and target cells 71 TUMOR BIOLOGY Haurum JS, Bjerring Hoier I, Arsequell G, Neisig A, Stevanovic S, Valencia G, Zeuthen J, Neefjes J, Rammensee H-G, Elliott T. Presentation of cytosolic glycosylated peptides by Human Class I Major Histocompatibility Complex Molecules in vivo. J Exp Med 1999; 190: Neefjes, J. CIIV, MIIC and other compartments for MHC class II loading. Eur J Immunol 1999; 29: PJ Peters 1 A Engering PhD 2 N Van der Wel PhD 3 E Tjin MSc 4 E Bos Y De Jong 3 D Fluitsma 5 E Van Donselaar Group leader Graduate student Neefjes J. Determining Protein Transport to the Plasma Membrane. Curr Prot Cell Biol 1999 (in press). Spee P, Subjeck J, Neefjes J. Identification of novel peptide binding proteins in the endoplasmic reticulum: Erp72, calnexin and GRp170. Biochem J 1999; 38: Tulp A, Verwoerd D, Neefjes J. Electromigration for separations of protein complexes. J Chromatogr B 1999; 722: Tulp A, Verwoerd D, Neefjes J. Lectin induced retardation of subcellular organelles during preparative density gradient electrophoresis: selective purification of plasma membranes. Electrophoresis 1999; 20: Vos JC, Reits EAJ, Wojcik-Jacobs E, J Neefjes. Subunit interactions visualized by ER-mobility and post-translational translocation define a head-head/tail-tail topology for the pore of the ABC transporter TAP. Current Biol (in press). Vos JC, Spee PJL, Momburg F, Neefjes J. The topology and dimerization of the two subunits of the Transporter associated with Antigen Processing reveal a three domain structure. J Immunol 1999; 163: Wubbolts R, Fernandez-Borja M, Jordens I, Reits E, Dusseljee S, Echeverri C, Vallee RB, Neefjes J. Opposing motor activities of dynein and kinesin determine retention and transport of MHC class II-containing compartments. J Cell Sci 1999; 112: Theses Grommé M. Aspects of conventional and alternative MHC Class I antigen processing and presentation, Leiden: State University, Wubbolts R. Dynamics in the intracellular transport of Major Histocompatibility Complex class II molecules, Leiden: State University, Cytotoxic T Lymphocytes (CTLs) are essential components of our immunological defense against tumor cells and virus infected cells. CTLs rapidly kill target cells via at least two distinct effector pathways: 1) the secretory pathway involving the polarized exocytosis of cytolytic granules; 2) the Fas pathway in which the Fas ligand on the CTL interacts with the Fas receptor on the target cell. Both these processes appear to work in concert and induce apoptotic changes in the target cell nuclei that culminates in the destruction of the target cell. Conceivably, regulated exocytosis will be pivotal to the cytolytic function of CTLs, however, the molecular machinery involved in this process is poorly characterized. We have previously shown, by use of the highresolution technique of cryo-immunogold electron microscopy, that CTL granules are secretory lysosomes that contain the lytic molecules perforin and granzymes and that their formation is coupled to the endocytic pathway. Mechanism of CTL degranulation 2 It is now well documented that members of the Ras superfamily of GTPases (Ras, ARF, Rho, Rab, Ran) control a wide range of biological processes including membrane trafficking and cytoskeletal remodeling. We are investigating the role of the GTPases of the ARF and Rho subfamilies in CTLs, since they are likely candidates to regulate various stages of the granule exocytosis pathway. Specifically, we are in collaboration with C D Souza Schorey, University of Notre Dame, IN, USA, V Hsu, Harvard Medical School, Boston, USA and H Spits (Division IV) investigating the role of ARF6 in CTL-TC adhesion and ARF6, Rac and Rho in granule exocytosis. SCAMP (secretory carrier membrane protein) compartment in CTL 4,5 In other systems such as the nerve termini of neurons, rapid advances have been made in understanding the process of synaptic vesicle exocytosis, vesicle-membrane fusion and in the identification of key components of the membrane fusion process. Preliminary investigations in

72 72 TUMOR BIOLOGY Figure VI.3 The prion disease is manifested when the prion protein, designated PrP, is converted from a benign cellular form to a disease causing form by a change in conformation. Thus, it has become increasingly important to establish the site(s) of conversion of PrP c (cellular prion protein) into PrP Sc (scrapie prion protein) and the subsequent site of accumulation. A better understanding of the subcellular trafficking routes of PrP c is a prerequisite for establishing the precise location where the conformational changes of PrP Sc takes place. To this end, we have analyzed the intracellular localization of PrP c in a Chinese Hamster ovary cell line that stably expresses PrP c, by using the technique of cryoimmunogold electron microscopy. This figure shows that PrP c (small dots) is enriched in caveolae labeled for caveolin (large dots). (A collaborative study with Stanley Prusiner, University of California, San Francisco, Medical School, San Francisco CA.) collaboration with D Castle, University of Virginia Health Sciences Center, Charlottesville, USA and with J Neefjes, have suggested that a comparable mechanism of membrane exocytosis may exist in CTLs. PCR cloning and Western Blot analysis using antibodies directed against Rab3, VAMP2, syntaxin 4, SCAMP (secretory carrier membrane protein) and SNAP-23 have indicated that these proteins or their T-cell homologs are present in activated CTLs. We have therefore begun to unravel the molecular components of the exocytic process in CTLs. An exciting finding is the observation that SCAMP redistributes from a non-granule membrane to the plasma membrane on encountering a target cell. We will investigate the role of SCAMP in CTL granule exocytosis and target cell lysis. The CD1 family of lipid antigen-presenting molecules in Target Cells 3 Four CD1 isoforms (CD1a,b,c and d) have been identified in humans as type 1 integral membrane proteins structurally related to MHC class I molecules. Functional studies revealed that CD1b presented mycobacterial antigens, either added exogenously or derived from endosome-resident live bacteria. Mycobacterial antigens presented in this way are known to enter the endocytic (phagocytic) pathways, previously appreciated to result in binding to MHC class II molecules. In collaboration with M Sugita, S Porcelli and M Brenner, Harvard Medical School, Boston, MA, USA we have demonstrated that CD1b was internalized from the cell surface via clathrin-coated pits/vesicles and transported to endosomes, eventually reaching late endosomal/lysosomal compartments where sampling of mycobacterial-derived lipid antigens occurred. The endosomal localization of CD1b, which is essential for its antigen presenting function, is directed by its short cytoplasmic domain containing a YXXZ sequence (Y, tyrosine; X, any amino acid; Z, a hydrophobic amino acid), as demonstrated by our studies showing that deletion of this cytoplasmic tail alters endosomal localization. This motif is found in other proteins localized to endosomes and is known to interact with adaptor protein (AP) complexes of clathrin-coated vesicles. Whereas CD1c and CD1d contain a similar motif and distribute to endosomes, CD1a is the only molecule of the human CD1 family that lacks this tyrosine-based endosomal targeting motif. Thus, CD1a is likely to follow an intracellular pathway distinct from that by the other CD1 isoforms. We hypothesize that the intracellular pathways followed by CD1a and CD1b are different. Such differences are likely to be of critical functional importance since they determine the compartments in which antigen loading can occur. By comparison, the differential intracellular trafficking patterns of MHC class I and class II molecules are central in determining their ability to bind endogenous or exogenous antigens, respectively. These differences also correlate with distinct effector functions of T cells activated by MHC class I (CD8+, cytotoxic) and class II (CD4+, Th1 and Th2). We are now trying to elucidate the pathway for antigen presentation mediated by the different CD1 molecules. Teaching at the Medical School, Free University This year PJ Peters started teaching at the Medical School of the Free University in Amsterdam. In addition to lecturing the Cell Biology class the task as semester chair for first year students is to co-ordinate the semester program From Molecule to Cell. Notes 1 PJ Peters has a joint appointment (20%) at the Medical School of the Free University Amsterdam. 2 Funding: Dutch Cancer Society, Project: UU Funding: Netherlands Leprosy Relief. 4 Funding: NWO, Project: Free University, Medical School, Amsterdam.

73 Publications: Ultrastructural studies Aoe T, Lee AJ, Van Donselaar E, Peters PJ, Hsu VW. Modulation of intracellular transport by transported proteins: insight from regulation of COPI-mediated transport. Proc Natl Acad Sci U S A 1998; 95: Cell cycle control genes and tumor progression 73 TUMOR BIOLOGY D Souza-Schorey C, Van Donselaar E, Hsu VW, Yang C, Stahl PD, Peters PJ. ARF6 targets recycling vesicles to the plasma membrane: insights from an ultrastructural investigation. J Cell Biol 1998; 140: Franco M, Peters PJ, Boretto J, Van Donselaar E, Neri A, D Souza-Schorey C and Chavrier P. EFA6, a sec7 domaincontaining exchange factor for ARF6, coordinates membrane recycling and actin cytoskeleton organization. EMBO J 1999; 18: Hochstenbach F, Klis FM, Van den Ende H, Van Donselaar E, Peters PJ, Klausner RD. Identification of a putative alpha-glucan synthase essential for cell wall construction and morphogenesis in fission yeast. Proc Natl Acad Sci U S A 1998; 95: Hsu VW, Peters PJ. Current views in intracellular transport: insights from studies in immunology. Adv Immunol 1998; 70: Hunziker W, Peters PJ. Rab17 localizes to recycling endosomes and regulates receptor-mediated transcytosis in epithelial cells. J Biol Chem 1998; 273: Jackman RM, Stenger S, Lee A, Moody DB, Rogers RA, Niazi KR, Sugita M, Modlin RL, Peters PJ, Porcelli SA. The tyrosine-containing cytoplasmic tail of CD1b is essential for its efficient presentation of bacterial lipid antigens. Immunity 1998; 8: Korinek V, Barker N, Moerer P, Van Donselaar E, Huls G, Peters PJ, Clevers H, Depletion of epithelial stem-cell compartments in the small intestine of mice lacking Tcf-4. Nat Genet 1998; 19: Peters PJ. Cryo-Immunogold Electron Microscopy. In: Bonifacino JS, Dasso M, Harford JB, Lippincott-Schwartz J and Yamada KM, editors. Current Protocols in Cell Biology. New York, John Wiley & Sons, 1999: Spada FM, Grant EP, Peters PJ, Sugita M, Melián A, Leslie DS, Van Donselaar E, Lee HK, Hanson DA, Krensky AM, Majdic O, Porcelli SO, Morita CT, Brenner MB. Self Recognition of CD1 by gamma delta T cells: Implications for Innate Immunity. J of Exp Med 1999 (in press). Sugita M, Grant EP, Van Donselaar E, Hsu VW, Rogers RA, Peters PJ, Brenner MB. Separate pathways for antigen presentation by CD1 molecules. Immunity 1999 (in press). Sugita M, Moody DB, Jackman RM, Grant EP, Rosat JP, Behar SM, Peters PJ, Porcelli SA, Brenner MB. CD1 a new paradigm for antigen presentation and T cell activation. Clin Immunol Immunopathol 1998; 87: R Michalides PhD Group leader A Balkenende T Verschoor 1 1 W Silvis Undergraduate student Tumor progression The proliferative effect of overexpression of cyclin D1 in differentiated cells of squamous cell carcinomas is counterbalanced by the induction of cyclin kinase inhibitor p21. In squamous cell carcinomas, proliferation marker Ki-67 was present in the basal layer of the tumor and absent in the more differentiated layers where cyclin D1 was co-expressed with p21 (a collaborative study with P Van Diest, Free University Amsterdam). Induced overexpression of cyclin D1 resulted in upregulation of p21 in MCF7 breast cancer cells. The balance between cyclin D1 expression and p21 may therefore, be associated with the differentiation status of tumor cells. In a collaborative study with M Barbareschi (Trento, Italy) we found that the prognostic value of cyclin kinase inhibitor p27 as described in the literature, is strongly influenced by a selective partitioning of breast cancer cases. p27 should therefore be used with caution as a marker for prognosis. Indicators for therapy Disturbed regulation of the G1 phase of the cell cycle in the Rb/cyclin D1/p16/cdk4 pathway alters chemo- and radiosensitivity of tumor cells (Figure VI.4). In MCF7 cells (wild type for Rb, cyclin D1, p53, p19arf, but with a mutated p16), we observed that induced overexpression of cyclin D1 increased sensitivity to γ-radiation by an enhanced transit through G2/M before a subsequent entry into apoptosis. Sensitivity towards paclitaxel in MCF7 cells is also enhanced by cyclin D1. Cyclin D1 also affected sensitivity of MCF7 cells towards simvastatin, a farnesyltransferase-inhibitor preventing activation of Ras and Rho (a collaborative study with A Weisz, Napoli, Italy). Overexpression of cyclin D1 resulted in a transient release of the simvastatin induced cell cycle block. This was also observed when MCF7 cells were treated with anti-estrogens ICI or tamoxifen. These findings indicate that overexpression of cyclin D1 affects the therapeutic window of treatment with antiestrogens, farnesyltransferase-inhibitors, γ-radiation and paclitaxel, which may well be of clinical relevance.

74 74 TUMOR BIOLOGY Schuuring E, Van Damme H, Schuuring-Scholtes E, Verhoeven E, Michalides R, Geelen E, De Boer C, Brok H, Van Buuren V, Kluin Ph. Characterization of the EMS1 gene and its product, human cortactin. Cell Adhesion and Communication 1998; 6: Genes involved in breast cancer metastasis Figure VI.4 Effects of overexpression of cyclin D1 on some therapeutic modalities. Adhesion-dependent proliferation Adhesion of cells onto extracellular matrix components activates the Rho-like GTPases Rac and CDC42. Constitutive activation of the Rho-like GTPases in porcine aortic endothelial PAE cells resulted in mitotic slippage, where nuclear replication becomes dissociated from cytokinesis. This results in polyploid, multinucleated cells, which ultimately succumb to apoptosis. The mechanism behind the prevention of cytokinesis by constitutive active Rac or CDC42 is under study. Note 1 Funding: Dutch Cancer Society, Project Publications: Cell cycle control Barbareshi M, Peterse H, Mauri FA, Veronese S, Van Tinteren H, Maisonneuve P, Scaioli M, Caffo O, Galligioni E, Doglioni C, Dalla Palma P, Michalides R. P27 KIP-1 expression in breast carcinomas: an immunohistochemical study on 512 patients with long follow-up. Int J Cancer 1999 (in press). Coco-Martin J, Verschoor T, Lallemand F, Michalides R. Overexpression of cyclin D1 increases sensitivity to g-irradiation in MCF7 cells. Cancer Research 1999; 59: De Jong JS, Van Diest P, Michalides R, Baak J. Concerted overexpression of p21 and cyclin D1 is associated with growth inhibition and differentiation in various carcinomas. J Clin Pathology: Mol Pathol 1999; 52: Germano D, Pacillio C, Cancemi M, Cicatiello L, Altucci L, Belsito V, Salzano S, Michalides R, Taya Y, Bresciani F, Weisz A. Cyclin D1 overexpression does not interfere with inhibition of human breast cancer growth by blockade of the mevalonate-protein prenylation pathway. Br J Cancer 1999 (in press). Michalides R. Cell cycle regulators: mechanisms and their role in etiology, prognosis and treatment of cancer. J Clinical Pathology (review) 1999; 52: Michalides R. Prognosis for cell cycle regulators: useful to predict course of disease and for assessment of therapy in cancer. J Path 1999; 188: J Hilkens PhD I Gaemers PhD 1 T Thingstad MSc 2 M Van der Velden MSc M Boer H Volders 1 J Hendriksen P Schneider Group leader Graduate student Graduate student Undergraduate student Undergraduate student Overexpression of membrane-associated mucins in breast cancer Episialin (also designated EMA, CA 15-3 antigen) encoded by the MUC1 gene, and epiglycanin belong to the class of membrane-associated mucins which are characterized by a heavily O-glycosylated mucin domain. Both are frequently highly overexpressed on carcinoma cells compared with normal epithelial cells. Biochemical and electronmicroscopy studies revealed that the mucin domains of both molecules project more than 200 nm above the plasma membrane. When these elongated and relatively rigid molecules are overexpressed they mask other much shorter cell surface molecules, thus impeding cell-cell and cell-matrix interactions and contributing to the invasive and metastatic potential of tumor cells. Epiglycanin has thus far only been identified in mouse mammary tumor cells. An antiserum against the nonglycosylated precursor of mouse epiglycanin crossreacted with human epiglycanin precursor, showing that epiglycanin is also present in carcinoma cell lines. The expression of epiglycanin in carcinoma cells appears to be sensitive to a component present in mouse and human cancer ascites fluids. The normal function and the mechanism of overexpression of both membraneassociated mucins are presently being studied. Episialin overexpression and metastasis We have previously shown that the formation of experimental metastases in the lung is strongly affected by overexpression of episialin. The effect of episialin overexpression on tumor progression has been further investigated in episialin transgenic FVB mice. N-ethyl-Nnitroso-urea (ENU) induced lung tumors in these mice express high levels of episialin. Histological examination

75 of the lungs of the ENU treated mice revealed more and larger tumors in the transgenic mice than in the FVB control animals, indicating that episialin indeed affects the progression of lung tumors in this animal model. Regulation of episialin gene expression Examination of the episialin promoter revealed several putative regulatory elements, including a STAT-responsive element (STAT: Signal Transducer and Activator of Transcription) and several glucocorticoid binding sites. Luciferase reporter gene assays and electrophoretic mobility shift assays (EMSA) revealed that STAT1, STAT3 and STAT5 are binding to the STAT element. Interestingly, in some breast carcinoma cell lines STAT3 is constitutively activated, which may contribute to the overexpression of episialin in these carcinomas. Hydrocortisone also stimulated the episialin promoter. Mutation analysis of two putative glucocorticoid binding elements (GRE) just upstream of the STAT binding element revealed that both sites are involved in hydrocortisone mediated regulation of the episialin promoter. Addition of hydrocortisone and IL-6 together synergistically stimulated episialin expression. We have elucidated the mechanism responsible for this synergism: supershift EMSA and co-immunoprecipitation using antibodies against the glucocorticoid receptor (GR) revealed a direct interaction between STAT3 and the GR. Moreover, mutations in the STAT binding element abolished the synergism and negatively affected the stimulation by hydrocortisone and visa versa, suggesting that DNA binding of both transcription factors stabilized the interactions. Identification of novel genes involved in mammary tumor metastasis We have developed a model system that can be used to identify novel metastasis genes in a BALB/c substrain that acquired mouse mammary tumor virus (MMTV) by foster-nursing (denoted BALB/c+ strain). Subcutaneous (s.c.) transplantation of lungs from tumor bearing animals provided a source of metastatic mammary tumor cells for genetic analysis which was not previously available. A series of independent sets of primary tumors and s.c. lung metastases has been collected and analyzed for MMTV proviral integrations. In about 1/3 of the subcutaneously expanded metastatic tumors extra MMTV copies were present that were not present in the primary tumor. These MMTV copies might be involved in activating metastasis genes by promoter or enhancer insertion. We have adapted the splinkerette PCR to specifically amplify the MMTV flanking sequences. Several of these flanks are currently being analyzed to detect metastasis related common integration sites. Publication: Genes in breast cancer metastasis Hilkens J, Bonfrer H Diagnosis and monitoring of bone metastases using serum tumor markers. In: Body JJ editor. Tumor Bone Disease and Osteoporosis in Cancer Patients, Pathophysiology, Diagnosis and Therapy. New York, Marcel Dekker, 1999: Secretary Tumor Biology MK Van der Velde Research staff positions Scientific permanent: 4.1 Scientific project: 16.0 Technical permanent: 6.0 Technical project: TUMOR BIOLOGY Notes 1 Funding: NIH grant 1R01-CA Funding: Biotag.

76 76 VII Division of Molecular Genetics Division head Anton Berns Introduction The Division of Molecular genetics currently comprises four research groups, headed by D Barlow, R Beijersbergen, A Berns and P Demant, respectively. In addition, our animal pathologist, M Van der Valk, is formally part of the division. Focus of the group of Barlow is on the mechanisms of imprinting using the Igf2 receptor gene locus as a paradigm. In addition, her group concentrates on the identification of genes involved in invasive growth properties of placental tissues and their possible role in tumorigenesis. R Beijersbergen joined the division in 1999 as an AvL fellow after having spent a post-doctoral period in the lab of R Weinberg at MIT, Boston, USA. His focus is on the regulation of telomerase activation during immortalization and tumorigenic transformation of human cells. In the group of P Demant the main emphasis is on the identification of Tumor Susceptibility Genes for intestinal and lung tumors. To this end, the group has developed the Recombinant Congenic Strains in which 12.5% of the genome of a susceptible donor strain is introduced onto a resistant background strain. The group is closing in on one of the susceptibility genes that was first identified (Scc1). Within this group M Snoek focuses on the identification of genes within the MHC that predispose to lung cancer. The group of A Berns concentrates on the generation and study of conditional tumor suppressor (TS) gene knockout mice with the aim to identify the contribution of distinct genetic lesions to specific tumor characteristics. Finally, there is an ongoing effort to improve current mouse model systems. Activities in this area include exploring strategies to mark cells after Cre/Lox switching, generation of better inducible Cre/Flp transgenes, the use of somatic gene transfer to achieve switching, development of reliable cell markers, and the generation of better insertional mutagens to disrupt and mark genes of interest. In the coming years emphasis in the division will be on the identification of new genes directly involved in or predisposed to tumorigenesis, unraveling of the regulation of telomerase, and on the development of a new generation of mouse models for human cancer. Molecular genetics A Berns PhD P Krimpenfort PhD* M Van der Valk MSc** J Jonkers PhD 1 S Lyons PhD 4 S Marino PhD 2 P Merel PhD 4 R Meuwissen PhD 1 K Quon PhD 3 M Vooijs PhD 5 A Loonstra MSc 1 C Martins MSc 6 H Mikkers MSc 8 R Van Amerongen MSc 7 D Hoogervorst 5 F Matthesius8 H Van der Gulden J Vink 1 F Bleeker P Knipscheer F Micelli Group leader Academic staff Academic staff Graduate student Graduate student Graduate student Graduate student Undergraduate student Undergraduate student Undergraduate student We have embarked on developing better mouse models for cancer using techniques that permit the conditional activation of oncogenes and inactivation of tumor suppressor genes in a tissue-specific fashion. Four different sets of mice have been generated with: 1) Conditional tumor suppressor gene alleles through the introduction of LoxP sites in the locus by gene targeting techniques (Rb, p53, Brca1, Brca2, Dcc, Nf2, Apc, Pten, E-cadherin, Ink4a, Atr, VHL). 2) Conditional and inducible (Tet-on/Tet-off) oncogenes permitting regulatable oncogene activation in a tissue-specific fashion. 3) Cre transgenic lines, some with hormone inducible Cre, either produced by conventional transgenesis or knockin strategies (POMC-Cre. K14-Cre, Actin-Cre, GFAP-Cre, IRBP-Cre, P0-Cre, Rosa26-Cre-MER, WAP-Cre). 4) Reporter lines permitting the monitoring of Cre activity by a LacZ reporter gene. These systems are used to produce compound mutant strains in which multiple alleles can be switched in a tissue-specific fashion. This will allow us to somatically induce, at will, a complex set of oncogenic mutations. In this way we expect to

77 reproduce more faithfully the mutations found in tumors in man. We can then study the correlation between tumor genotype and phenotype in a highly controlled experimental setting. These systems will be used to study tumor progression events, to test the effects of therapeutic intervention protocols, and to serve as a source of cells to study the effects of these mutations on cell cycle and differentiation in vitro. Many of these analyses are ongoing, utilizing various combinations of mutant alleles. Conditional tumor suppressor gene knockout mice Retinoblastoma A number of years ago, in collaboration with M Hooper in Edinburgh, we embarked on the targeted disruption of the retinoblastoma gene. Rb deficiency resulted in early embryonic lethality. Surprisingly, chimeric mice generated from injection of Rb -/- ES cells into blastocysts were relatively normal with significant contribution of Rb -/- cells to most tissues and no propensity to develop retinoblastoma. However, we did observe extensive apoptosis in the retinal layer explaining the low contribution of the Rb -/- ES cells to the adult retina. In collaboration with Te Riele (Division II), a series of compound chimeric mice was generated carrying various combinations of mutant alleles for Rb, p107. Loss of Rb and p107, but not p53 was required for retinoblastoma formation in the mouse. Tumors originated from the inner nuclear layer rather than the outer nuclear layer and exhibited markers of amacrine cells. Subsequently, we have generated conditional Rb mutants to better explore the role of Rb loss in tumors in mice. The Rb gene was fitted with either frt or loxp sites flanking exon 19. Crossing of Rbfrt/frt mice with POMC- Flp transgenic mice, which express the Flp recombinase in the intermediate lobe of the pituitary gland, resulted in inactivation of the Rb gene specifically in these cells leading to pituitary tumors at high incidence. Next, we generated compound mutant mice with the floxed Rb alleles on a p107 and/or p53 null background. A photoreceptor cell-specific Cre transgene was crossed into these mice. No retinoblastomas developed indicating that inactivation of Rb, p107 and p53 in photoreceptor cells is not sufficient to give rise to retinal tumors in mice. However, pineal gland tumors, tumors of the anterior lobe of the pituitary, and choroid plexus tumors were frequently found. Interestingly, some of these tumor prone tissues arise from the same neuroblastic precursor as the photoreceptors suggesting that this lineage is not indifferent to Rb loss. p53 loss was found to contribute only to a subset of these tumors (e.g. to anterior lobe tumors but not to intermediate lobe tumors). Epithelial switching of Rb in the presence or absence of a functional p107 allele resulted in hyperplasia of the skin. Basal layer cells showed increased proliferation and retained proliferative capacity during differentiation into the more differentiated cell types of the skin. Absence of p107 aggravated this phenotype considerably, leading to early death of the K14-Cre;Rbflox/flox;P107 -/- mice. Currently, the effects of additional mutations on a Rb conditional background are assessed in vivo and in vitro using cultured keratinocytes of these mice. Schwann cell tumors We have generated conditional Nf2 knockout mice in collaboration with the group of G Thomas, Institut Curie, in Paris. In view of the involvement of Nf2 in the induction of schwannomas, we have produced P0 transgenic mice that express the Cre recombinase in Schwann cells. P0Cre;Nf2flox/flox mice develop schwannomas after a long latency period. Tumorigenesis is greatly accelerated by the introduction of a mutant p53 allele. p53 and Nf2 are on the same chromosome. When a single conditional Nf2 allele is combined with a single p53 allele no tumors are found when both mutant alleles are in trans, while a low incidence of tumors do form when they are in cis. The much earlier tumor onset in Nf2flox/flox;p53 +/- mice suggests that Cre-mediated loss of Nf2 is critical for the development of schwannomas and accelerates loss of the wt p53 allele. Mesotheliomas Screening of human mesotheliomas for tumor suppressor gene mutations has shown the relatively frequent mutation/deletion of NF2 and INK4a. Since INK4a acts in the RB pathway, we have generated compound Nf2flox/flox;Rb/flox/flox mice and inactivated the Nf2 and Rb alleles in the mesothelial lining of the thoracic and peritoneal cavity by a single injection with AdenoCre virus. Tumors with the characteristics of mesotheliomas arose after a latency of 8-12 months. Most of the tumors showed a spindle cell morphology and the tumors are currently further characterized using a panel of antibodies. We will test a series of additional combinations of conditional tumor suppressor genes in order to determine how the different lesions affect the phenotypic characteristics of the tumors. Furthermore, we will derive cell lines from the mesothelial lining of the various mutant strains and study their growth and differentiation in vitro, with and without switching. In addition, we will accelerate tumorigenesis by carcinogen treatment (both ENU and asbestos fibers) and search for chromosomal abnormalities using RDA and LOH analyses. Epithelial tumors Emphasis is on the role of Rb, p107, Brca1, Brca2, p53, Ras, and E-cadherin in tumorigenesis in skin and mammary epithelium. Conditional alleles of Brca2 were combined with a conditional allele of p53 yielding Brca2flox/flox;P53flox/flox mice. In this background a K14-Cre transgene was introduced. The double conditionals carrying the K14-Cre transgene developed skin abnormalities in a subset of the mice. The penetrance of this phenotype appears to correlate with the level of switching of Brca2 and p53. None of the mice of the 77 MOLECULAR GENETICS

78 78 MOLECULAR GENETICS Brain tumors We have generated a mouse model for medulloblastoma by the Cre-mediated inactivation of both p53 and Rb in the cerebellar external granular layer (EGL). Using the promoter of the gene encoding glial fibrillary acidic protein (GFAP), Cre expression was directed to astrocytes and immature precursor cells of the EGL in the developing cerebellum. GFAP-Cre;Rbflox/flox;P53flox/flox mice developed aggressive embryonal tumors with typical features of the medulloblastoma, a malignant human childhood tumor. Neoplastic lesions were identified at the age of seven weeks on the outer surface of the molecular layer, corresponding to the location of the EGL during development. Interestingly, complete P53 deficiency was required for the development of this tumor. In GFAP-Cre;Rbflox/flox;p53 +/- animals these tumors were not found, whereas in GFAP- Cre;Rbflox/+;P53 -/- mice the tumors developed after a prolonged latency period with the concomitant loss of the wt Rb allele. Figure VII.1 4 -OHT regulated Cre recombination in vivo. I.p. administration of 4 OHT Tamoxifen to Rosa-Rep/Rosa- CreMer double transgenic mice results in Cre-mediated recombination in vivo. Whole mount beta-galactosidase staining of peritoneum of a single rosa-rep (A) and double rosa-rep/rosacremer without tamoxifen (B) and rosa-rep/rosacremer with tamoxifen (C & D).Rosa-rep and Rosa-CreMer mice were generated through a knockin of floxed lacz reporter construct in the rosa-26 locus (Rosa-reporter, P. Soriano) and a knockin of a Cre gene fused to the mutated estrogen receptor (Mer) binding domain (CreMer) into the rosa-26 locus (Marc Vooijs) K14-Cre;Brca2flox/flox or K14-Cre;p53flox/flox genotype developed this condition (Figure VII.1). Subsequently, papillomas and mammary tumors developed in the compound mutants. These are currently being analyzed in more detail. Similar compound conditionals are being produced with conditional alleles of Brca1 and p53. As Brca1 and Brca2 are implicated in DNA repair, the effects of DNA damage inflicted by carcinogens and radiation will be tested in these mice. When the K14-Cre transgene was combined with conditional Rb alleles widespread hyperplasia of the skin was observed. This phenotype was further aggravated in a p107 deficient background. Proliferating cells, expressing both basal and suprabasal markers were found in the suprabasal layer suggesting that Rb plays a role in cell cycle exit rather than in differentiation. This system will be used to further define the role of Rb and p107 in differentiation and proliferation of basal cells. We are defining conditions to switch conditional alleles in a number of cells types in vitro. Special attention has been given to mitigate the inhibitory effects that Cre expression appears to exert on cell growth. Identification of tumor suppressor genes We have obtained a highly metastatic human cell line C8161 (Welsch et al. Oncogene 1994; 9: ) whose metastatic capacity is suppressed by introduction of a single copy of a Neo-tagged human chromosome 6. We have shown that the spontaneous reversion of this cell line to the highly metastatic phenotype is sufficiently low to permit inactivation and subsequent identification of the responsible tumor suppressor gene by insertional mutagenesis with a replication-defective gene trap vector. We are currently in the stage of identifying the integration sites of proviruses in cells that regained enhanced growth properties upon transplantation into syngeneic hosts. Identification and characterization of collaborating oncogenes Transgenic mice expressing (putative) oncogenes can provide insight into the oncogenic potential of genes, disclose their tissue-specific activity and reveal their roles in transformation. Genetic crosses between transgenic mice carrying different oncogenes allow assessment of their cooperative activity and, when used in combination with proviral tagging, permit identification of genes whose gain- or loss-of-function synergizes with the transgene(s). Furthermore, by transplanting primary tumors induced in this fashion, genes contributing to tumor progression can be identified. Finally, by applying proviral tagging in compound mutant mice, we can mark genes acting in specific signal transduction pathways. Emphasis in this program is on the identification of oncogenes, on the unraveling of the biochemical function of the encoded proteins, their physiological role, and their contribution to the tumorigenic process. We have set up a high throughput screen for proviral insertion sites with the aim of achieving saturation tagging of genes contributing to tumorigenesis. In the coming years we expect to accumulate flanking sequences of thousands of insertion sites which will give access to the genes deregulated by these insertions. In this way we expect to identify the signaling networks critical for transformation of lymphoid cells in mice. This will facilitate the search for alterations in the corresponding pathways/genes in tumors in man.

79 Pim oncogenes Myc and Pim oncogenes strongly collaborate in tumorigenesis. However, it remains unclear what the mechanism is underlying this synergy. In search of the functions of PIM proteins, we have generated Pim deficient mice (Pim1, Pim2, and Pim3). Single Pim1 or Pim2 knockout mice show a rather subtle phenotype. The same holds for the Pim1;Pim2 double KO mice. The other pairs of double mutants and the triple compound mutant mice are currently being generated. There is good evidence that Pim1 and Pim2 act downstream of a number of interleukin receptors as these induce Pim1 and Pim2 transcription upon ligand binding, probably through the jak/stat pathway. Overexpression of Pim can partly counteract defective receptor signaling. Introduction of a Pim1 transgene into IL2Rgamma knockout mice, which have a very small thymus due to defective interleukin signaling, restores thymic cellularity. Interestingly, overexpression of Pim1 also leads to the formation of a CD4+8+ double positive compartment of normal size in a RAG1-deficient background whereas Pim1 is unable to rescue the strongly reduced cellularity of thymi in CD3gamma knockout mice, suggesting that PIM proteins might act as gain setters of signaling cascades. To uncover components in the PIM signal transduction pathway we have set up a genetic screen. Taking advantage of the strong collaboration between PIM and MYC we generated mice carrying the Eµ-myc transgene or a Pim1/Pim2-deficient background. In these mice lymphomas were induced by retroviral insertional mutagenesis. The concept explored here is that we expect to preferentially activate genes that can substitute for the function of PIM. Genes that have been found include Pim3, c-kit, tpl2, and cyclind2. None of these genes were found to be provirally activated in a Pim1/2 proficient background. We are currently assessing whether Pim proteins can suppress the phenotypic abnormalities seen in c-kit mutant mice. To determine the potential role of the human PIM genes in tumorigenesis we have analyzed a series of lymphoid tumors for aberrant expression of PIM genes. In a substantial fraction of these tumors, particularly large cell B lymphomas, high levels of either PIM1, PIM2 or both were found. However, so far we have no evidence for chromosomal alterations linked to one of the PIM genes. Frat1 Frat1 has been cloned as a tumor progression gene using retroviral tagging. Frat1 is a small intronless gene. The encoded protein does not exhibit obvious sequence motifs that suggest a function or points to specific interactions. Transgenic mice overexpressing Frat1 are not tumor prone but show an accelerated tumorigenesis rate when infected with MuLV or when crossed with Pim1 transgenic mice. This is in accordance with the contribution of Frat1 to later steps in tumorigenesis as was Figure VII.2 Frat funtion in the ß-catenin signaling pathway further substantiated by retroviral transduction of Frat1 into cell lines derived from primary tumors arising spontaneously in Pim1 transgenic mice. Tumor cells expressing Frat1 expanded preferentially upon subcutaneous inoculation. Interestingly, at older age Frat1 transgenic mice exhibit epithelial hyperplasia in their mammary glands. This is in agreement with the recent finding by the group of Kimelman et al. Cell 1998; 93: that FRAT binds to GSK3beta and thereby modulates the betacatenin pathway. We failed to detect this interaction, in spite of exhaustive two hybrid screens in which we did identify a number of other binders. We have confirmed that the mouse FRAT1 can confer beta-catenin signaling activity in Xenopus (Figure VII.2). In addition, co-immune precipitations show its binding to GSK3beta and Axin. We are currently testing the functionality of these interactions. Frat1 knockout mice do not show phenotypical aberrations, possibly due to the widely overlapping expression pattern of the related Frat2, and Frat3. We are currently generating Frat2 and Frat3 knockout mice to test this. Proviral tagging in Cdk inhibitor knockout mice We have started proviral tagging experiments in p21, p27, p15, p16, p19arf single and compound knockout mice to determine which proto-oncogene activations synergize with the Cdk mutants. In this way we expect to identify the complementation groups in transformation to which the CDK inhibitors belong. This will permit us to establish functional relationships between these different groups of genes. Proviral tagging in a p21 and p15 KO background resulted in the activation of genes belonging to the standard complementation groups (Pim, Myc, Gfi1) without a significant difference in tumor latency. Interestingly, in p27 and p21;p27 deficient mice, MuLV infection resulted in accelerated tumor induction compared to control mice; p27 +/- showed an intermediate latency period. Remarkably, proviral tagging in the p27 -/- mice showed in a much higher frequency of activation of Myc, whereas integrations near Pim and cyclind 2 were strongly reduced. This reduction was most pronounced in p21 -/- ;p27 -/- tumors, suggesting that PIM might, directly 79 MOLECULAR GENETICS

80 80 MOLECULAR GENETICS or indirectly, act to relieve the inhibitory function of p21 and p27 proteins. We are currently sequencing the flanks of the proviruses in this tumor panel. Further refinement of insertional mutagenesis Tissue-specific expression of L1 retrotransposons to achieve insertional mutagenesis was investigated. For this purpose, we have generated transgenic mice carrying this retrotransposon. None of these mice expressed the transposon. A knockin of the retrotransposon into the ROSA26 locus yielded a low level of expression. Mice carrying this insertion were crossed with tumor prone Pim1 transgenic mice to determine whether the transposon would accelerate tumorigenesis but this did not appear to be the case. We are also exploring the utility of the Sleeping Beauty (SB) DNA transposon for mutagenesis of somatic cells. This project is in collaboration with the Plasterk group (Division V). Transposition in tissue culture cells is found at an appreciable level. However, insertion of the transposon into an adenoviral vector and infection of cells with this vector did not give rise to a high frequency of transposition. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: EEC training grant ERBFMBICT Funding: EMBO fellowship. 4 Funding: EEC BMH4-CT Funding: Dutch Cancer Society, Project NKI Funding: Portugese PhD fellowship. 7 Funding: Centre of Biomedical Genetics. 8 STW royalties. * Head of transgenic/knockout service facility. ** Animal Pathology Laboratory. Publications: Molecular genetics Berns A. Mouse models for cancer at center stage.trends Genet : 177. Berns A. Turning on tumors to study cancer progression. Nat Med 1999; 5: Hanson RD, Hess JH, Yu BD, Ernst P, Van Lohuizen M, Berns A, Van der Lugt N, Shashikant CS, Ruddle F, Seto M, Korsmeyer S J. Mammalian Trithorax and Polycomb-group homologues are antagonistic regulators of homeotic development. Proc Natl Acad Sci USA 1999; 96: Jacobs H, Krimpenfort P, Haks M, Allen J, Blom B, Démollière C, Kruisbeek A, Spits H, Berns A. PIM1 reconstitutes thymus cellularity in interleukin 7- and common gamma chain-mutant mice and permits thymocyte maturation in rag- but not CD3gammadeficient mice. J Exp Med 1999; 190: Jacobs JJ, Kieboom K, Marino S, Depinho RA, Van Lohuizen M. The oncogene and Polycomb-group gene Bmi1 regulates cell proliferation and senescence through the Ink4a locus. Nature 1999;397: Jacobs J, Scheijen B, Voncken J-W, Kieboom K, Berns A, Van Lohuizen M Bmi-1 collaborates with c-myc in tumorigenesis by inhibiting c-myc induced apoptosis via INK4a/ARF. Genes & Development 1999;13: Jonkers J[Dissertation]. Identification and characterization of the Frat1 proto-oncogene. University of Amsterdam Jonkers J, Van Amerongen R, Van der Valk M, Robanus-Maandag E, Molenaar M, Destrée O, Berns A. Mechan of Devel 1999;88: Jonkers J, Weening JJ, Van der Valk M, Bobeldijk RC, Berns A.. Overexpression of Frat1 in transgenic mice leads to glomeruloscleroses and nephrotic syndrome, and provides direct evidence for involvement of Frat1 in lymphoma progresssion. Oncogene 1999;18: Konietzko U, Kauselmann G, Scafidi J, Staubli U, Mikkers H, Berns A, Schweizer M, Waltereit R, Kuhl D. Pim kinase expression is induced by LTP stimulation and required for the consolidation of enduring LTP. EMBO J 1999;8: Li L, Yuan H, Weaver CD, Mao J, Farr GH, Sussman DJ, Jonkers J, Kimelman D, Wu D. Axin and Frat1 interact with dvl and GSK, bridging Dvl to GSK in Wnt- mediated regulation of LEF-1. EMBO J 1999;18: Berns A, Mikkers H, Krimpenfort P, Scheijen B, Jonkers J. Identification and characterization of collaborating oncogenes in compound mutant mice. Cancer Res 1999;59: 1773s-1777s. Giovannini M, Robanus-Maandag E, Niwa-Kawakita M, Van der Valk M, Woodruff J, Berns A, Thomas G. Schwann cell hyperplasia and tumors in transgenic mice expressing a naturally occurring mutant NF2 protein. Genes Dev 1999; 13: Quon K, Berns A. Biochim Biophys Acta 1999; 1423(2): Robanus Maandag E [Dissertation]. Mouse models for the human hereditary cancer syndromes retinoblastoma and neurofibromatosis type 2. University of Amsterdam, Vooijs M, Berns A. Developmental defects and tumor predisposition in Rb mutant mice. Oncogene 1999; 18: Haks MC, Krimpenfort P, Van den Brakel JH, Kruisbeek AM. Pre- TCR signaling and inactivation of p53 induces crucial cell survival pathways in pre-t cells. Immunity 1999;11:

81 D Barlow PhD R Lyle PhD 1 R Zwart PhD 2 F Sleutels MSc 1 S Verhaagh MSc 2 I Van der Borch M Cuadrado Lopez Mammalian development and cancer genetics Group leader Graduate student Graduate student Undergraduate student The mechanism regulating imprinted gene expression 1 Imprinted genes are epigenetically silenced on one of the two parental chromosomes. Since both the active and inactive parental allele are retained in the nucleus, imprinting offers a particularly useful model to identify the epigenetic modifications that lead to gene silencing. Loss of imprinting (known as LOI) and parental-specific LOH (loss of heterozygosity) that indicates the presence of imprinted loci, are features of tumor progression in diverse tumors types, suggesting that imprinted genes can function as growth regulators in adult tissues. Our laboratory has studied gene imprinting for several years with the aim of understanding more about its biological function and epigenetic mechanism. We have identified and analyzed the imprinted maternal-specific expression of the Igf2r gene (Insulin-like growth factor type 2 receptor, also known as the cation independent mannose 6-phosphate receptor). This work has shown that the primary function of Igf2r is to reduce extracellular levels of the growth hormone IGF2 in the embryo and, thus, to decrease embryonic size. Other laboratories have shown that in humans, IGF2R is a tumor suppressor gene in liver and breast tumors. In an attempt to dissect the epigenetic mechanism imprinting this gene we have initially followed a transgenic approach in mice. Our previous work has demonstrated that the paternal allele of Igf2r is repressed by a 3.7 kb DNA element contained in intron 2 (Wutz et al. Nature 1997; 389: 745-9). This intron 2 element contains the promoter of an unusual antisense RNA that has the reciprocal imprinted expression compared to Igf2r. These results have allowed us to propose that expression-competition between neighboring promoters may be a common feature of the imprinting mechanism (Barlow, EMBO J 1997; 16: ). Work in progress on Igf2r imprinting Transgenic studies have demonstrated that imprinting of the Igf2r gene could be maintained in ectopic chromosomal positions on a 300 kb genomic fragment. However, this work also demonstrated that short fragments (1-14 kb) could not maintain imprinting of the antisense RNA promoter. We have continued to investigate the minimum elements required for Igf2r imprinting using transgenes. This work is still in progress but results indicate that multiple elements may be required for full imprinted expression. This year we have mainly focused on defining the extent of the antisense RNA. This has been done by generating a 130 kb sequence contig spanning from Igf2r intron 7 to the upstream flanking gene. This sequence was used to identify EST cdna clones from the database and this and other analyses shows that the antisense RNA is an exceptionally long and stable RNA that appears to lack an open reading frame. Whatever the final answer, the existing analysis of gene imprinting has already demonstrated the involvement of several key epigenetic players including DNA methylation and cis-repressor elements. The exact function of each of these is being actively investigated. Embryonic implantation as model of tumor metastases 2 Establishing a vascular connection to host tissue is essential for further growth and development of mammalian embryos. It is also a requirement of tumors. The mammalian embryo makes a vascular connection by initiating a tightly controlled invasion of the maternal uterus using a specialized invasive cell type (the trophoblast), and subsequently forms a direct connection with the maternal circulation through a specialized organ (the placenta). Several genes have already been identified by targeted inactivation as essential for implantation, however, many are cell lethals or not directly informative about the implantation process itself. In this project we are using a naturally-occurring mouse mutant named tw 73, to isolate a gene that appears to play a direct role in trophoblast differentiation. Embryos homozygous for the tw 73 gene fail to form a competent maternal-fetal connection and die in the peri-implantation period. The tw 73 gene will be used as a starting point to analyze implantation and to test our hypothesis that genes specifically regulating implantation are candidate oncogenes or tumor suppressor genes. Work in progress in isolating the tw 73 gene A positional cloning strategy is being used to isolate the tw 73 gene that we have previously mapped to a 500 kb critical region on mouse chromosome 17. Last year we described the identification of a novel cluster of three organic cation transporter genes (named Slc22 a1, a2, a3) spanning 150 kb of the 500 kb critical region. This year, in an attempt to further reduce the minimal critical region we have applied genetic tests to this gene cluster to determine if any were equivalent to tw 73. Slc22 a2 and a3 fulfill several of the criteria demanded for a candidate tw 73 gene. Both are expressed in implanting embryos and, in addition, Slc22a3 contains a tw 73 specific DNA polymorphism. Genetic rescue using a 300 kb YAC transgene (for Slc22 a1 & a2) and genetic complementation using a targeted-inactive locus (for Slc22a3) are 81 MOLECULAR GENETICS

82 82 MOLECULAR GENETICS now being performed. The Slc22a3 gene is exclusively expressed in placental tissue during embryonic development and is down regulated prior to birth. Other laboratories have suggested that Slc22a3 is a nonneuronal catecholamine transporter that limits catecholamine availability during embryonic development. Other laboratories have also shown that the catecholamines adrenaline and nor-adrenaline are essential for embryonic development and mice deficient for these neurotransmitters die shortly after mid-gestation. We are using our Slc22 a3 knockout mouse to examine the function of this gene in the embryonic placenta and in particular to test its role in placental catecholamine clearance(figure VII.3). Our final goal with this project is to use the isolated tw 73 gene as a molecular entry point to study the invasive phase of embryonic implantation. In addition, the human homolog of the tw 73 gene is expected to reside in the syntenic region on human chromosome 6q This region shows frequent LOH in liver, breast and ovarian tumors. The human tw 73 gene will provide a useful marker to define more accurately the physical maps of the minimum LOH region. It is also anticipated that identification of the human tw 73 gene may provide a molecular entry point into genes involved in cancer metastases. Notes 1 Funding: Dutch Cancer Society, Project Funding: Dutch Cancer Society, Project Publications: Mammalian development Sleutels F, Barlow DP. One more imprinting review? Book Review. Nat Genetics 1999; 23: 23. Verhaagh S, Schweifer N, Barlow DP, Zwart R. Cloning of the mouse and human solute carrier 22a3 (Slc22a3/SLC22A3) identifies a conserved cluster of three organic cation transporters on mouse chromosome 17 and human 6q Genomics 1999; 55: Control of telomerase activation R Beijersbergen PhD L Carlee MSc W Nijkamp S Cillessen Group leader Graduate student Undergraduate student Figure VII.3 Expression of the Slc22a3 gene is restricted to the labyrinth (L) layer of the embryonic placenta. Top panel: general organization of the day 13.5 placenta. Middle, RNA in situ analysis of Slc22a3 (10x magnification), bottom panel 20X magnification. S, spongiotrophoblast layer. Research in our group is primarily focused on the activation of telomerase in human cells. Normal human somatic cells, which generally lack telomerase activity, exhibit progressively shortened telomeres with repeated cell divisions and enter senescence after a limited number of cell divisions. When the life span of normal cells is extended beyond senescence, life span terminates at crisis, a point at which telomere loss results in

83 chromosome instability and cell death. However, telomere shortening is stopped by the expression of htert, the catalytic subunit of the telomerase enzyme, in those cells that have become immortal. htert mrna cannot be detected in most normal cell lineages but is present in a variety of tumor cell lines and is detectable in 90% of human tumors. The ectopic expression of htert is sufficient to restore telomerase activity and maintain or extend telomeres. The introduction of htert in several primary human cells prevented senescence, allowed them to bypass crisis and become immortal. The introduction of htert together with collaborating oncogenes, such as Large T and Ras, is sufficient for tumorigenic transformation of primary human cells. In contrast, inactivation of telomerase in immortal cells results in telomere shortening and cell death. Together these results indicate that htert expression represents the rate-limiting determinant that regulates the levels of telomerase enzyme activity in tumor cells and that the process of cell immortalization is closely or completely linked to the expression of htert. We are studying the molecular mechanisms responsible for derepression of the htert gene that occurs upon immortalization. We have isolated, sequenced, and analyzed genomic fragments containing the promoter and potentially upstream regulatory elements of the human htert gene. Our detailed analysis has revealed several candidate transcription factor-binding sites. We have identified the promoter region required for activation in telomerase positive cells. This region is currently being used to identify transcriptional activators and/or repressors of the htert gene. Through comparison of htert promoter regulation in telomerase positive and negative cells by deletion analysis, genomic footprinting, and electromobility bandshift analysis, we will identify elements involved in the regulation of htert expression. These elements will then be used to identify interacting proteins that can elucidate the regulation of the htert gene. Furthermore we have set up a htert-reporter system that allows for the identification of proteins or pathways that can either repress or activate htert expression through cdna expression library screens. The result of these studies may allow us to elucidate the molecular mechanism that enables cancer cells to overcome two barriers, senescence and crisis, before they become immortal. The identification of the proteins or pathways that are involved in regulation of telomerase expression may have implications for interference with the process of immortalization and provide us with a molecular event, specific in the generation of tumor cells. Publications: Control of telomerase activation Hahn W C, Counter C M, Lundberg A S, Beijersbergen R L, Brooks M W, Weinberg R A. Creation of human tumour cells with defined genetic elements. Nature 1999; 400: Hahn W C, Stewart S A, Brooks M W, York S G, Eaton E, Kurachi A, Beijersbergen R L, Knoll J H, Meyerson M, Weinberg R A. Inhibition of telomerase limits the growth of human cancer cells. Nat Med 1999; 10: P Demant PhD M Snoek PhD T Csikos PhD 5 N Tripodis PhD 1 M Van Kooy PhD 4 C Ruivenkamp MSc 2 T Van Wezel MSc 3 K De Groot 4 E Delzenne-Goette J De Moes E Dijsselbloem 5 A Klous 5 M Treur-Mulder H Van Vugt 2 M Gajewska H Havelkova I Schultz Cancer genetics Group leader Academic staff Graduate student Graduate student Guest Guest Undergraduate student The outcome of the processes initiated by the exposure to carcinogenic agents (chemical, physical, viral) is strongly modified by numerous genes of the host. This often results in large differences in cancer susceptibility between individuals. The identification of the modifier genes and the elucidation of their functions is the aim of this group. As the number of these genes is large, their detection in humans is virtually impossible and they have to be identified first in animal models. This effort starts with the establishment of linkage of the modifier genes to specific chromosomes. For this purpose we developed a specific genetic tool, the recombinant congenic strains (RCS) of mice which exhibit reduced genetic variability and therefore facilitate detection of modifier genes, even those with relatively small effects. In the next steps the modifier genes are mapped to a very short chromosomal segment (less than 0.5 centimorgan long), encompassing less than 1 Mb DNA. Subsequently, the candidate genes in the region are identified and the correlation of their polymorphism with their cancer-modifying effects has to be established and their function demonstrated in functional assays or by genetic modification. In addition to identifying the modifier genes we are also studying the nature of their effects in the segregating populations, as well as their impact on tumor phenotype. 83 MOLECULAR GENETICS

84 84 MOLECULAR GENETICS Our main interest resides in the study of colon and lung cancer, the two major causes of cancer morbidity and mortality. In collaborative projects genes modifying susceptibility to radiation-induced leukemias and genes involved in control of various stages of immune responses have also been mapped. MHC class III region and tumorigenesis We have mapped a tumor susceptibility gene for chemically induced alveolar lung cancer to an interval of 27 kb within the Major Histocompatibility Complex of the mouse. Detailed analysis of this segment has revealed the presence of three genes G7e, G7a and G7c, for which we have determined the complete genomic organization. The G7e encodes a viral envelop protein, G7a encodes valy-trna synthetase, while the G7c gene has no homologs in the database that could predict its function. We are currently studying the expression patterns and expression levels of the genes in the critical region. We generated genomic DNA lambda phage libraries and built contigs spanning the interval derived from DNA from the resistant B10.A(1R) and the susceptible B10.A(2R). Using shot gun subcloning of lambda inserts and in conjunction with automatic sequencing, we are determining the sequence of the complete interval of both strains. Comparison of the critical region between the susceptible and the resistant strain will lead to the finding of disease associated differences. Analysis of the genetic differences and study of the susceptibility gene involved might lead to the understanding of the possible mechanism underlying the phenomenon of predisposition to tumorigenesis. Since the class III region of the MHC is highly conserved between species, the assignment of a mouse tumor susceptibility gene might be of importance for the evaluation of human predisposition to lung cancer. Susceptibility to lung tumors The strain O20/A is susceptible and the strain B10.O20/Dem is relatively resistant to carcinogen induced tumors. Moreover, they differ in the histological type and rate of progression of the lung tumors. Previously, using crosses of several OcB recombinant congenic strains (containing 88% of genes from the strain O20 and the remainder from the strain B10.O20) we mapped 14 novel Sluc (Susceptibility to lung cancer) loci and demonstrated that they frequently engage in mutual inter-locus interaction. We proceeded with the fine mapping of these loci, as a step towards their positional cloning. Mice with recombinations at various sites of the regions bearing the Sluc loci 1,3,4,6, and 8 were produced for testing their susceptibility phenotype. Analysis of the genetic differences in the histological type of the lung tumors is ongoing. Morphometric systems and methods of statistical analysis were developed which allow recognition of strain specific patterns of various aspects of the tumor phenotype, as well as the study of their inheritance in segregating populations. This study led to mapping of several novel loci controlling the qualitative aspects of lung tumorigenesis. Susceptibility to colon tumors The analysis of genetics of susceptibility to colon tumors in the recombinant congenic strains led to the detection of 9 novel Scc (Susceptibility to colon cancer) loci (Scc1 - Scc9) located on chromosomes 1, 2, 3, 8, 10, 11, 17, and 18. Two pairs of loci were found to interact with each other: Scc4 - Scc5 and Scc7 - Scc8 (Van Wezel et al. Nature Genetics 1996; 14: and Van Wezel et al. Cancer Res 1999; 59: ). A positional cloning approach is being used to identify the Scc1 gene on chromosome 2. Previously, Scc1 has been mapped to an interval of less than 200 kb using recombinant haplotypes. The DNA from the BACs covering this interval was subcloned and shot-gun sequenced. The available sequence covers almost the whole region and its analysis revealed several genes and ESTs. The exon/intron organization of the genes has been partly unraveled and several polymorphisms between susceptible and resistant allele were defined. This resulted in the mapping of the Scc1 locus to less than 70 kb. An analogous strategy is being applied to the positional cloning of the interacting pair of the loci Scc4 (chromosome 17) and Scc5 (chromosome 18). A novel mapping cross between the strains CcS19 and BALB/c has been started, to confirm and further improve the original mapping result and a number of recombinant haplotypes in the chromosomal regions containing the two loci has been produced. The susceptibility testing of these recombinants will be performed in the coming period and may localize each of the loci to a 2-3 centimorgan interval, which will provide a starting point for fine mapping and positional cloning. Notes 1 Funding: NWO Project NKI Funding: Dutch Cancer Society, Project NKI Funding: European Commission, Contract CT Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Cancer genetics Havelkova H, Kosarova M, Krulova M, Demant P, Lipoldova M. T cell proliferative response is controlled by loci Tria4 and Tria5 on mouse chromosomes 7 and 9. Mammalian Genome 1999; 10: Kosarova M, Havelkova H, Krulova M, Demant P, Lipoldova M. Production of two Th2 cytokines, IL-4 and IL-10 is independently controlled by a locus Cypr1 and by loci Cypr2 and Cypr3, respectively. Immunogenetics 1999; 49: Mori N, Okumoto M, Yamate J, Stassen APM, Tsubura A, Akamatus T, Sakuma S, Demant P. Enhancement of glucocorti-

85 85 coid-induced apoptosis of thymocytes linked to the p53-deletion and interaction between the chromosomal segments containing Rapop1 and p53. Oncogene 1999; 18: Snoek M, Van Kooij M, De Groot K, Van Vugt H. The mouse p52 subunit of the transcription/dna repair factor TFIIH is not located in the MHC class III region of the H2 complex, but resides next to a G7a/Bat6 homologue in the telomeric part of the MHC. Immunogenetics 1999 (in press). MOLECULAR GENETICS Snoek M, Van Vugt H. The sequence and organization of the mouse valyl-trna synthetase gene G7a/Bat6 located in the MHC class III region. Immunogenetics 1999; 49: Szymanska H, Sitarz M, Krysiak E, Piskorowska J, Czarnomska A, Skurzak H, Hart AAM, De Jong D, Demant P. Genetics of susceptibility to radiation-induced lymphomas, leukemias and lung tumors studied in recombinant congenic strains. Int J Cancer 1999; 83: Van Wezel T, Ruivenkamp C A L, Stassen APM, Moen CJA, Demant P. Four new colon cancer susceptibility loci, Sscc6 to Scc9 in the mouse. Cancer Research 1999; 59: Secretaries G De Jong-Meijerink S Heussen K Van den Berg Research staff positions (full time equivalents) Scientific permanent: 5.1 Scientific project: 20.0 Technical permanent: 2.5 Technical project: 7.5

86 86 VIII Division of Experimental Therapy Division head Adrian Begg Introduction The goals of the Division of Experimental Therapy are to optimize and refine existing cancer therapies, develop and test new ones, and improve diagnosis and response prediction by cellular and molecular characterization of tumors. The division consists of three subdivisions: Experimental Chemotherapy, Experimental Radiotherapy and Molecular Pathology. Clinical links are emphasized in all projects, with active collaboration of scientists and clinicians. One new project was started this year; a collaboration between the Schellens and Schinkel groups on mechanisms of reduced cellular accumulation of topoisomerase I inhibitors. Lou Smets retired this year, bringing to an end an illustrious career in experimental chemotherapy. His scientific and social input will be missed. Highlights and future plans In studies on hypoxia (Begg group), an image analysis method was developed and validated in which IdUrd labeling around individual vessels was used to monitor vascular perfusion. Analysis of vasculature alone was found to be insufficient to accurately assess tumor hypoxia (comparison with the bioreductive marker pimonidazole). Projects to develop assays for predicting radiotherapy outcome will continue to concentrate on ways to measure chronic and acute hypoxia in human tumors, focusing on the combined use of bioreductive marker drugs, vascular derived parameters and vascular perfusion using IUdRbased and other techniques. Clinical studies will continue in head and neck cancer patients, which will include these hypoxia measurement methods. A recently started project has shown that transduction with a retroviral vector expressing a dominant negative of DNA polymerase beta together with EGFP (green fluorescent protein) markedly sensitized human tumor cells to radiation. These studies will continue, concentrating on DNA polymerase beta, for which significant effects have already been observed, and on ku80, with the eventual aim of testing the potential for therapeutic intervention. The Stewart group has shown a significant increase in glomerular vwf and leukocytes in the renal cortex after irradiation, although the increased glomerular vwf did not quantitatively predict renal dysfunction. Platelet aggregation could be inhibited with acetylsalicylic acid or Clopidogrel, but with only small effects on kidney function. Studies will continue on the role of endothelial cell and vascular damage in kidney pathogenesis after radiation, with emphasis on mechanisms and ways to ameliorate damage. Emphasis will be on anti-thrombotic agents which block the GPIIb/IIIa and ADP platelet receptors. In the photodynamic therapy (PDT) program, carbogen breathing ± nicotinamide improved po2 leading to significantly improved PDT response. In addition, preclinical and clinical studies have demonstrated greater PDT responses for illuminations at 1 to 2 days after sensitization than at 3 to 5 days. Future research will focus on the relationship between photosensitizer pharmacokinetics in tissues and plasma and PDT efficacy in animal and clinical studies. In a new preclinical study with Division XII, attempts to improve peritoneal lavage regimes will be made to increase cytotoxicity and reduce tumor re-seeding. The Schinkel Experimental Chemotherapy group has shown that placental P-gp limits the passage of various toxins and drugs into developing fetuses. It was also shown that P-gp inhibitors could potentially be used to increase the fetal penetration of P-gp substrate drugs. Other studies showed that even low levels of P-gp and Mrp1 contribute considerably to tumor drug resistance, suggesting that multidrug transporter inhibitors may also increase the sensitivity of previously untreated (naïve) tumors to chemotherapy. Further investigations will be carried out into the functional role of Bcrp1 in drug resistance, e.g. by trying to elucidate why Bcrp1 overexpression sometimes correlates with efficient resistance against anthracyclines, whereas othertimes only very low resistance is observed. Conditional and constitutive Bcrp1 knockout mice will be generated in order to study the physiological and possible pharmacological roles of Bcrp1. In addition, knockout mice for several other pharmacologically relevant proteins, including the Orct1 and Orct2 organic cation transporters and various drugmetabolizing enzymes, will be generated to better understand and predict the overall pharmacological behavior of drugs in an intact organism. The behavior and interactions of HIV protease inhibitor drugs will be further analyzed in Mdr1a/1b knockout mice, and possibly in additional knockout mouse strains.

87 The Schellens Experimental Chemotherapy group found that drug resistant human ovarian tumor cells expressed the Breast Cancer Resistance Protein (BCRP), which could efficiently transport Topotecan and mitoxantrone. One of several potential inhibitors for BCRP was capable of completely blocking outward transport of topotecan, completely resensitizing these cells to drug. Studies will continue on mechanisms of resistance to topoisomerase I inhibitors and mitoxantrone, in particular investigating the nature of the implicated novel BCRP efflux system present at the plasma membrane of resistant cells. In addition, studies on the interaction between platinum drugs and topoisomerase I inhibitors will continue, focusing on yeast as a model systems in which a variety of mutants will be employed, having defects in various aspects of DNA repair and enzyme topoisomerase levels. These studies have already indicated a role for rad52 but not rad4 in synergy. Finally, platinum-dna adducts, and topoisomerase I levels in WBC and tumor tissue, will be measured as pharmacodynamic endpoints in ongoing clinical studies with cisplatin, carboplatin and topoisomerase I inhibitors. Research in the Molecular Pathology group (Van t Veer, Van de Vijver) will remain focused on breast cancer pathogenesis and treatment. In studies on multidrug resistance proteins in breast cancer, MDR1/P-Gp involvement was shown to be unlikely. Preliminary experiments involving comparative genomic hybridization (CGH) showed that the number of genetic alterations involved in breast tumor progression (in situ to invasive) is low. These studies will continue in the coming year. Completing the case-control study on the involvement of radiotherapy in breast cancer susceptibility will enable us to assess the contribution of ATM heterozygosity to the risk of radiation-induced breast cancer. Marker genes for breast cancer cells, determined by the method of Serial Analysis of Gene Expression (SAGE), will be tested for their validity in minimal residual disease detection. SAGE profiles of a breast cell line transfected with either a normal or a dominant-active estrogen receptor will elucidate downstream effector genes of normal and variant receptors. Publications Smets group (now disbanded) Kuin A, Aalders M, Lamfers M, Van Zuidam DJ, Essers M, Beijnen JH, Smets LA. Potentiation of anticancer drugs at low intratumoral ph induced by the mitochondrial inhibitor m-iodobenzylguanidine (MIBG) and its analogue benzylguanidine (BG). Br J Cancer 1999, 79: Kuin A, Van der Valk M, Rutgers M, Beijnen JH, Smets LA. Bioavailability and toxicity of oral administration of m-iodobenzylguanidine (MIBG). Br J Cancer 1999, 79: Taal BG, Hoefnagel C, Rutgers M. Carcinoid tumors. N Engl J Med 1999, 341: Salomons GS, Smets LA, Verwijs-Janssen M, Hart AAM, Haarman EG, Kaspers GJL, Van Wering ER, Van der Does-Van den Berg A, Kamps WA. Bcl-2 family members in childhood acute lymphoblastic leukemia: relationships with features at presentation, in vitro and in vivo drug response and long-term clinical outcome. Leukemia 1999; 13: Prediction and manipulation of treatment outcome AC Begg PhD H Bartelink MD PhD 1 KMG Haustermans MD PhD NS Russell MD 1 JM Coco Martin PhD 2 C Vens PhD E Dahmen A Grummels I Hofland D Sprong I Van de Pavert 3 M Verwijs W Blyweert B Van der Schueren Group leader Academic staff Academic staff Academic staff Undergraduate student Undergraduate student Prediction of tumor radiosensitivity Attempts to measure chromosome damage in cells extracted from human tumor biopsies have continued. Separating fibroblasts from tumor cells is a major concern, and we previously reported on the use of an anti-human-fibroblast antibody and FACS sorting. We are now using an improved antibody together with magnetic bead separation (MACS). MACS was shown to have little or no negative effect on plating efficiencies and to produce purity of over 95% tumor cells and yields of around 70%. It appeared to be less toxic than FACS sorting and is now our method of choice. Premature chromosome condensation (PCC) using the phosphatase inhibitor calyculin A (CA) has proved feasible in cells from biopsies, although PCCs indices are lower than in cell lines. We showed that the majority of scorable PCCs were from G2 cells and few from G1, although this could be improved slightly by addition of activated cdc2. We compared 5h colcemid, which should collect all G2 cells in mitosis, with 1h CA in cells from biopsies. Approximately 4 times more scorable PCCs were found than metaphases, showing the superiority of chemical PCCs over conventional metaphase collection. The optimum culture time of biopsy cells for maximum PCC yield was found to be 1 week. Good growth of cells from a large fraction of tumor cells remains the most difficult and limiting factor. Prediction of tumor hypoxia There is strong evidence that tumor hypoxia limits the success of cancer therapy, including radiotherapy. Indirect evidence suggests acute hypoxia, caused by 87 EXPERIMENTAL THERAPY

88 88 EXPERIMENTAL THERAPY blood perfusion changes, can be equally important as chronic hypoxia, caused by diffusion limitations. Our immediate goal is to define the relative importance of acute and chronic hypoxia in human tumors in limiting success of radiotherapy. We developed an image analysis protocol to measure the Diffusion Limited Fraction (DLF), being the fraction of tumor area in a histological section greater than 120µ from the nearest blood vessel; a measure of diffusion limited hypoxia. Comparisons were made with pimonidazole staining (a bioreductive hypoxic marker drug) in 10 cervix tumors. The area stained by pimonidazole was significantly smaller than DLF for all tumors. When images from all tumors (n=123) were analyzed together, the correlation was highly significant (r=0.36, p<0.0001). The large scatter, however, implies that analysis of vasculature alone is insufficient to accurately estimate tumor hypoxia. To estimate vascular perfusion by a method applicable to histological sections, we have pursued the hypothesis that variations in labeling with the cell kinetic marker IdUrd reflect mainly perfusion variations between vessels. IdUrd flow cytometry of RIF1 mouse tumors in which blood flow was manipulated supported this hypothesis. In addition, IdUrd labeling around individual vessels was scored (imunohistochemistry) and the fraction of vessels with no surrounding label was found to correlate well with blood flow changes. The method is now being further validated, and image analysis routines developed. A clinical trial has been started (collaboration between NKI and Gasthuisberg, Leuven) in which pimonidazole and IdUrd are given to head and neck tumor patients 2 h before tumor resection. Sections are being analyzed for DLF, pimonidazole staining, IdUrd labeling around vessels, and for expression of the endogenous hypoxic marker HIF-1α (hypoxia inducible factor; collaboration with G Semenza, Baltimore). Six patients have been included to date. Other predictors Cisplatin: In collaboration with sections IX and XI, we are following up on our previous result showing that cisplatin-dna adducts in buccal cells significantly predicted outcome in NSCLC. Buccal cell adducts are being measured in an EORTC trial of concomitant versus sequential cisplatin and radiotherapy. Adducts are also being measured in head and neck cancer patients given the RADPLAT protocol (high dose intra-arterial cisplatin plus radiotherapy). Normal tissue prediction: In collaboration with K Borgmann and E Dikomey (Hamburg), we are following up on preliminary results showing that chromosome damage in lymphocytes after in vitro irradiation correlates with normal tissue morbidity after radiotherapy. Hamburg and NKI patients will be included, and damage measured by FISH and conventional cytogenetic assays. Mechanisms and modulation of radiosensitivity Lethal lesions after ionizing radiation are thought to be Figure VIII.1 Radiosensitization by a dominant negative polymerase beta protein (pol-dn) in human SQD9 tumor cells. Radiation survival curves are shown for LZ-G5 clone transduced with empty vector versus two clones with middle (pol8-g9) and high (pol8-d9) expression levels of the pol-dn protein. mainly unrepaired or misrepaired DNA double-strand breaks, ultimately leading to lethal chromosome aberrations. However, studies with radioprotectors and repair inhibitors also indicate that single strand breaks, damaged nucleotides or abasic sites can influence kill. The relationship between initial DNA damage, DNA repair and ultimate cell killing are not fully understood and this project aims to investigate these relationships in human tumor cells. We are focusing at present on DNA polymerase beta and Ku80 in order to determine the relative contributions of single and double strand break repair pathways after irradiation. Strategies include analysis of knockout cell lines and construction of dominant negatives influencing the activity of repair proteins by preventing DNA binding or heterodimerization. Human tumor cell lines were transduced with a retroviral vector expressing the dominant negative protein together with EGFP (enhanced green fluorescent protein) Single cell clones with varying EGFP and dominant negative protein expression, were sorted and analyzed for radiosensitivity changes by a clonogenic assay. We were able to sensitize SQD9 cells (human squamous carcinoma cells) by a dose enhancement factor of 1.8 using a protein comprising the DNA binding domain of polymerase beta (pol-dn)(figure VIII.1). Sensitization was dependent on expression level of the protein. Control experiments excluded artifacts such as multiplicity and altered cell cycle distributions. Clonogenic assays of a mixed population of pol-dn transduced cells exhibited similar sensitization to individual clones. In contrast, A549 human lung carcinoma cells showed a slight increase in resistance to irradiation in a pol- DN/EGFP expression dependent manner. Pol-DN levels were lower than in SQD9 cells, and we hypothesize that EGFP has a modest radioprotecting function under these conditions (being verified).

89 Notes 1 Division IX. 2 Funding: Dutch Cancer Society, Project NKI Biophysics. Publications: Prediction and manipulation of treatment outcome Bartelink H, Begg A, Coco J Martin, Van Dijk M, Van t Veer L, Van de Vaart P, Verheij M. Towards prediction and modulation of treatment response. Radiother Oncol 1999; 50: Begg AC, Haustermans K, Hart AA, Dische S, Saunders M, Zackrisson B, Gustaffson H, Coucke P, Paschoud N, Hoyer M, Overgaard J, Antognoni P, Richetti A, Bourhis J, Bartelink H, Horiot JC, Corvo R, Giaretti W, Awwad H, Shouman T, Jouffroy T, Maciorowski Z, Dobrowsky W, Struikmans H, Wilson GD. The value of pretreatment cell kinetic parameters as predictors for radiotherapy outcome in head and neck cancer: a multicenter analysis. Radiother Oncol 1999; 50: Begg AC, Hofland I, Van de Pavert I, Van der Schueren B, Haustermans K. The use of thymidine analogs to indicate vascular perfusion in tumours. Br J Cancer (in press). Bergstrom C, Begg A, Palmqvist R, Waites A, Denekamp J. Labelling indices in human tumours: to apply corrections or not - that is the question. Br J Cancer 1999; 80: Coco Martin JM, Mooren E, Ottenheim C, Burrill W, Nunez MI, Sprong D, Bartelink H and Begg AC. Potential of radiation-induced chromosome aberrations to predict radiosensitivity in human tumor cells. Int J Radiat Biol 1999; 75: Dittmann KH, Gueven N, Mayer C, Ohneseit P, Zell R, Begg AC, Rodemann HP. The presence of wild-type TP53 is necessary for the radioprotective effect of the Bowman-Birk proteinase inhibitor in normal fibroblasts. Radiat Res 1998; 150: Haustermans K, Fowler J, Geboes K, Lerut A, Van Cutsem E, Begg A. Cell kinetic measurements: principles, guidelines for treatment? Hepatogastroenterology 1999; 46: Van de Vaart PJM, Belderbos J, De Jong D, Sneeuw KCA, Major D, Bartelink H, Begg AC. DNA-adduct levels as a predictor of outcome for NSCLC patients receiving daily cisplatin and radiotherapy. Int J Cancer (in press). Van de Vaart PJM [Dissertation]. Interaction of platinum coordination compounds with radiation or hyperthermia: from laboratory to clinic. Amsterdam: Free University, FA Stewart PhD P Baas MD PhD 1 LGH Dewit MD PhD 2 N Van der Vange MD PhD 3 S Bapat PhD 4 A Kuin PhD EM Van Kleef PhD 5 BGJ Floot H Oppelaar 4 MC Ruevekamp JAM Te Poele A Van der Wal 5 RLP Van Veen 4, 6 JH Schouwink MD 7 Vascular mediated therapy and tissue damage Group leader Academic staff Academic staff Academic staff Guest Radiation induced renal injury 5 Irradiation has been shown to induce changes in endothelial cell (EC) function which initiates thrombotic and inflammatory cascades. We are investigating the role of EC and vascular damage in the development of late renal radiation injury with the aim of identifying possible mechanisms for pharmacological intervention. Markers of EC damage under investigation include von Willebrand Factor (vwf) and ADPase (mediators of platelet adhesion), and leukocyte accumulation, which occurs at sites of platelet activation. Quantitative immunohistochemistry of mouse kidneys demonstrated a significant increase in glomerular vwf and leukocytes in the renal cortex from 12 weeks after high single dose irradiation. Smaller increases were seen after more clinically relevant fractionated irradiation. The leukocyte accumulation was significantly related to both time and dose but increased glomerular vwf was independent of dose and did not quantitatively predict for the extent of renal dysfunction. Increased glomerular vwf could be due to upregulation of vwf synthesis in irradiated EC s, or to the presence of an increased number of glomerular EC s as a result of radiation induced compensatory proliferation. Increased glomerular proliferation was seen from 10 weeks after irradiation but the total cell number actually decreased (cell loss exceeded proliferation). This suggests that the observed increase in glomerular vwf was due to increased production of the protein in surviving EC s rather than proliferation. We have investigated the potential of anti-platelet therapy for ameliorating late radiation nephropathy in mice. Daily oral administration of the cyclooxygenase inhibitor acetylsalicylic acid (ASA) inhibited production of thromboxane and effectively inhibited platelet aggre- 89 EXPERIMENTAL THERAPY

90 90 EXPERIMENTAL THERAPY gation. ASA given continuously from the time of irradiation lead to a small (non-significant) protective effect and delayed the onset of renal functional damage by 2 to 6 weeks. The ADP receptor blocker Clopidogrel also provided effective inhibition of platelet aggregation but studies to date have not demonstrated any protection against radiation nephropathy. Proteinurea is an early sign of progressive renal damage. The increased vascular permeability may be due to inflammatory responses or increased expression of VEGF. We are currently studying these phenomena to see how they relate to the dramatic changes seen in renal vascular structure after irradiation. Major vessels in irradiated kidneys become enlarge and tortuous, whereas the fine vascular network is lost leading to underperfusion of large areas of the tissue. Oxygen radical scavengers are being evaluated as possible modifiers of proteinuria resulting from the inflammatory response. Short term administration of catalase and superoxide dismutase could not reduce existing proteinurea. Long term administration of n-actyl-cysteine is currently under investigation. of photosensitized rats did demonstrate increased cardiac muscle degeneration and myocardial hemorrhage for illumination at 1 day. Short illumination intervals, although effective, are probably only suitable for small volume PDT. In the clinical application of PDT for treatment of MM, integral illumination of the entire thoracic cavity is required. Real time fluence rate measurements made at various positions within the cavity have demonstrated that sinuses which form near the diaphragm are sometimes under-illuminated. To overcome this problem, a wedge shaped light delivery/detector applicator has been developed to give additional illumination in the sinuses and provide a more homogeneous dose distribution over the thorax (Figure VIII.2). One possibility for improving the selectivity of PDT is to target photosensitizer specifically to the tumor. In collaboration with M Vrouenraets and G van Dongen (Free University, Amsterdam), we have tested photoimmunoconjugates of Foscan coupled to mab s specific for Photodynamic Therapy 4 PDT is being evaluated in our clinic for the treatment of malignant mesothelioma (MM), head and neck cancer and basal cell skin cancer (BCC). The goal of this project is to identify optimal parameters for Foscan -mediated PDT. Oxygen is a necessary substrate for the photochemical reaction and deficits can arise during illumination, rendering the reaction inefficient. This is more likely to occur for illumination at high fluence rates and in poorly vascularized tumors. In situ measurements of po 2 and immunohistochemical staining of hypoxic areas, demonstrated hypoxia developing during PDT in some murine RIF1 tumors but not in human MM xenografts. Carbogen breathing ± nicotinamide (a vasodilator) improved po 2 in all tumors and lead to significantly improved PDT response (tumor growth delay and cure). Reducing the fluence rate also improved response for RIF1 tumors but not for MM xenografts. Our preclinical experiments have consistently demonstrated greater PDT responses for tumor illumination at 1 day after sensitization than at 3 to 5 days, despite the fact that tumor drug levels remained high throughout this period. PDT of patients with multiple BCC was also more effective for illumination at 1 to 2 days (70 to 85% CR) than at 3 to 4 days (<50% CR). We are currently exploring the relationship between PDT efficacy and sensitizer profiles in plasma, tumor and tissues to test the hypothesis that vascular mediated PDT damage is more important than direct tumor cell toxicity and that drug exposure of endothelial cells in vessels feeding the tumor determines PDT response. One problem with short drug/light intervals is the increased risk of normal tissue injury. This becomes a serious drawback for large area PDT. Whole thorax PDT Figure VIII.2 Applicator for light delivery and dosimetry in sinus during PDT of MM patients. A linear diffuser (L) is used for light delivery and is placed centrally at the bottom of the wedge through a catheter. The applicator contains two channels for dosimetry.

91 head and neck carcinoma (HNSCC). Biodistribution studies in mice bearing HNSCC xenografts demonstrated improved tumor selectivity (by >5x) of the conjugates, despite their rapid elimination from the blood. Preliminary results on in vitro efficacy of PDT showed that Foscan coupled to an internalizing Ab exhibited more phototoxicity than when coupled to a non-internalizing Ab. Intra-abdominal perfusion for treatment of minimal residual disease (MRD) Since 1995 an experimental treatment modality (OV)HIPEC is being investigated in our clinic for intraabdominally metastasized colorectal and ovarian cancer (see Division XII). In this new preclinical study we are using rat models to determine whether lavage regimes as used in the (OV)HIPEC procedure could be improved to 1) increase cytotoxicity against MRD; 2) reduce the risk of tumor re-seeding and growth at sites of surgically induced peritoneal defects. Peritoneal perfusion regimes to be investigated include hyperthermic cisplatinum, with the addition of compounds that inhibit cell adhesion (e.g. RGD-peptides) or cell infiltration (e.g. metalloproteinase inhibitors). Notes 1 Division X. 2 Division IX. 3 Division XI. 4 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Dr. Daniel den Hoed Cancer Clinic, Rotterdam. 7 Medisch Spectrum Twente, Enschede. Publications: Vascular mediated therapy Aalders MCG, Van der Vange N, Stewart FA, Klein MG, Van de Vijver MJ, Sterenborg HJCM. White light toxicity resulting from systemically administered 5-ALA under normal operating conditions. J Photochem Photobiol B 1999; 50: Stewart FA. Retreatment after full course radiotherapy: is it a viable option? Acta Oncol 1999; 38: Tan IB, Oppelaar H, Ruevekamp MC, Veenhuizen RB, Timmers A, Stewart FA. The importance of in situ light dosimetry for photodynamic therapy of oral cavity tumors. Head Neck 1999; 21: Van Kleef EM, Zurcher C, Oussoren YG, Te Poele JAM, Van der Valk MA, Niemer Tucker MMB, Van der Hage MH, Broerse JJ, Robbins MEC, Johnston DA, Stewart FA. Long-term effects of total body irradiation on the kidney of Rhesus monkeys. Int J Radiat Biol (in press). Vrouenraets MB, Visser GWM, Stewart FA, Oppelaar H, Stigter M, Postmus PE, Snow GB, Van Dongen GAMS. Development of mthpc-monoclonal antibody conjugates for photoimmunotherapy. Cancer Res 1999; 59: AH Schinkel PhD JD Allen PhD 1 JW Smit PhD 2 MT Huisman MSc 3 JW Jonker MSc 2 RF Brinkhuis 1 E Wagenaar Genes and proteins involved in anticancer drug resistance and pharmacokinetics Group leader Graduate student Graduate student Our research focuses on genes and proteins that can: 1) cause (multi-)drug resistance in tumors; 2) determine the pharmacological behavior of anticancer drugs. Insight into these systems may help improve current chemotherapy strategies used to treat cancer, as well as pharmacotherapy in a broader sense. To study the physiological and pharmacological roles of the relevant genes we generate knockout mice that lack one or more of the encoded proteins, and analyze the mice for resulting alterations. In addition, we are using cell lines obtained from such knockout mice as tools to search for new drug resistance genes. Mice lacking Mdr1-type P-glycoproteins Pharmacological experiments using the Mdr1a/1b knockout mouse model have taught us that the drugtransporting P-glycoproteins (P-gps) have an important role in protecting an organism against exogenous toxins. We previously found that P-gp in the blood-brain barrier and blood-testis barrier prevents accumulation of transported drugs in these organs, while in liver, kidney and intestine P-gp contributes to the total body clearance of these drugs. We recently found that, analogous to the role of P-gp in the blood-brain barrier, placental P-gp limits the passage of various toxins and drugs into developing fetuses (Figure VIII.3). We have further shown that the P-gp inhibitors PSC833 or GG918 can enhance the trans-placental movement of substrate drugs into the fetus by complete inhibition of placental P-gp. These findings imply that P-gp inhibitors should be used cautiously in pregnant women but that they may potentially be used to increase the fetal penetration of P-gp substrate drugs when this is therapeutically desirable. Since the fetal penetration of the HIV protease inhibitors saquinavir and indinavir was clearly diminished by placental P-gp, this principle could potentially be used to diminish the chance of vertical transmission of HIV. Characterization of new anticancer drug resistance genes Characterization of mouse fibroblast cell lines deficient for the multidrug transporters P-gp and Mrp1 demon- 91 EXPERIMENTAL THERAPY

92 92 EXPERIMENTAL THERAPY Amsterdam). The goal is to identify an analog that is both less toxic and more easily synthesized than the native compound. Bcrp1-mediated resistance in the mouse cell lines could also be effectively reversed by the known P- gp inhibitor GG918. The existence of this effective BCRP/Bcrp1 inhibitor, which also has a low toxicity, may allow us to quickly test possible in vivo roles of Bcrp1 in mice. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: AIDS fonds, Netherlands, Project Publications: Drug resistance and pharmacokinetics Figure VIII.3 Effect of placental P-glycoprotein (Mdr1a/1b) genotype on the distribution of the anticancer drug paclitaxel (inset) to fetuses. Heterozygous females were crossed to heterozygous males to obtain embryos of all three genotypes in one mother. The ratio of total fetal paclitaxel concentration and maternal plasma concentration of paclitaxel, 1 hour after an i.v. administration of paclitaxel, is plotted. P-glycoprotein-deficient fetuses accumulate 16-fold more paclitaxel than wild-type fetuses. strated that normal expression of these proteins contributes substantially to basal drug resistance. This suggests that even low levels of P-gp and Mrp1 could contribute to tumor drug resistance and that inhibitors of multidrug transporters, currently tested in clinical trials as drug resistance reversal agents, may also increase the sensitivity of previously untreated (naïve) tumors to chemotherapy. Selection of the P-gp/Mrp1 deficient cell lines for resistance to the drugs mitoxantrone, topotecan or doxorubicin resulted in amplification and consequent overexpression of the murine homolog (Bcrp1) of the recently identified human multidrug transporter BCRP (Breast Cancer Resistance Protein). Cloning and characterization of the mouse Bcrp1 cdna indicated that Bcrp1 closely resembles its human counterpart, and that mouse models will therefore be useful for investigating the normal physiological and possible pharmacological roles of BCRP. Inhibitors of drug transporters are valuable as reagents for analyzing drug resistance mechanisms in the laboratory, and potentially for reversal of clinical drug resistance in tumors. We are currently evaluating analogs of fumitremorgin C, a known inhibitor of BCRP, as potential inhibitors of Bcrp1 and BCRP. Compounds are produced by combinatorial synthesis in a collaboration with A Van Loevezijn and GJ Koomen (University of Allen JD, Brinkhuis RF, Wijnholds J, Schinkel AH. The mouse Bcrp1/Mxr/Abcp gene: amplification and overexpression in cell lines selected for resistance to topotecan, mitoxantrone or doxorubicin. Cancer Res 1999; 59: Huisman MT, Smit JW, Schinkel AH. Significance of P-glycoprotein for the pharmacology and clinical use of HIV protease inhibitors. AIDS (in press). Jonker JW, Wagenaar E, Van Deemter L, Gottschlich R, Bender HM, Dasenbrock J, Schinkel AH. Role of blood-brain barrier P- glycoprotein in limiting brain accumulation and sedative sideeffects of asimadoline, a peripherally acting analgaesic drug. Br J Pharmacol 1999; 127: Schuetz EG, Umbenhauer DR, Yasuda K, Brimer C, Relling MV, Schuetz JD, Schinkel AH. Altered expression of hepatic cytochromes P450 in mice deficient in one or more mdr1 genes. Mol Pharmacol 2000; 57: Smit JW, Huisman MT, Van Tellingen O, Wiltshire HR, Schinkel AH. Absence or pharmacological blocking of placental P-glycoprotein profoundly increases fetal drug exposure. J Clin Invest 1999; 104: Van Asperen J, Van Tellingen O, Schinkel AH, Beijnen JH. Comparative pharmacokinetics of vinblastine after a 96-hour continuous infusion in wild-type mice and mice lacking mdr1a P-glycoprotein. J Pharmacol Exp Ther 1999; 289: Van Asperen J, Van Tellingen O, Tijssen F, Schinkel AH, Beijnen JH. Increased accumulation of doxorubicin and doxorubicinol in cardiac tissue of mice lacking mdr1a P-glycoprotein. Br J Cancer 1999; 79:

93 Pharmacodynamics of anticancer agents JHM Schellens MD PhD Group leader JH Beijnen PhD 1 Academic staff AH Schinkel PhD Academic staff M Maliepaard PhD 2 RCAM Van Waardenburg PhD 3 LA De Jong 3 D Pluim 5 MA Van Gastelen 2 M Bohlander Undergraduate student Resistance to topoisomerase I inhibitors and mitoxantrone Investigations into the mechanism of resistance in the human ovarian tumor cell lines IGROV1, T8 and MX3 have revealed the presence of the Breast Cancer Resistance Protein (BCRP, also known as MXR and ABCP). Topotecan and mitoxantrone were shown to be transported out of the cells very efficiently by this transporter. The level of resistance in a series of partially Figure VIII.4 Northern blot of BCRP mrna in parent IGROV1 and in resistant T8 and MX3 cell lines, including T8 revertant cell lines (upper panel). Lower panel shows levels of resistance to SN-38 ( ), TPT ( ) and MX ( ) in T8 lines passaged without drug (left), and in revertants re-exposed to drug (right). mrna was isolated from passages indicated with an arrow. reverted T8 cells correlated with the expression level of BCRP (Figure VIII.4), strongly suggesting that BCRP has a pivotal role in the development of resistance to topotecan and mitoxantrone in these tumor cells. In attempts to circumvent this resistance, several potential inhibitors for BCRP were tested. One of them, GF120918, a compound originally developed as an inhibitor of P-glycoprotein, was capable of completely blocking the BCRP-mediated outward transport of topotecan in the T8 and MX3 cells. By doing this, cells were almost completely resensitized to topotecan and mitoxantrone. The resistance factor for topotecan in the T8 cell line decreased from 110 to 4.2 in the presence of GF and in the MX3 cells from 30 to 1.5. GF appeared to be a very potent inhibitor of BCRP, with maximum inhibition of topotecan cytotoxicity and transport present at concentrations as low as 100 nm. This means that GF is equally efficient in inhibiting P-gp and BCRP. Furthermore, we have assessed the substrate characteristics of BCRP in detail. Several other registered topoisomerase I inhibitors appear to be transported by BCRP. However, a couple of experimental topoisomerase I drugs are hardly transported by BCRP, and could therefore be potentially useful in circumventing BCRP-related resistance. In collaboration with Professor Scheper (Free University, Amsterdam), a monoclonal antibody was raised against BCRP. Further studies focus on cellular distribution of BCRP and expression in normal and tumor tissues. Interactions between platinum drugs and topoisomerase I inhibitors 4 Our laboratory previously reported a schedule dependent synergistic cytotoxicity of cisplatin (cddp) followed by topoisomerase I (topi) inhibitors in various human cancer cell lines (Ma et al., Cancer Chem. Pharmacol. 1998; 41: ). The most active schedule was addition of cddp followed by the topi inhibitor. This schedule dependent synergistic cytotoxicity was also found for other platinum derivatives such as transplatin, carboplatin, oxaliplatin, lobaplatin and JM-216 in combination with various topi inhibitors (camptothecin (CPT), topotecan (TPT) and SN-28). In contrast, BCNU or mitomycin C (also DNA-adduct forming agents) in combination with CPT and TPT showed no synergy. The synergistic effect appears to be platinum dependent. The kinetics of the major intra-strand (AG- and GG-) adducts were only slightly affected by the topi inhibitor and could not explain the observed synergistic cytotoxicity. Using the ICT (in vivo complex of topoisomerase I)-assay we observed that the amount of topi cleavable complexes increases after incubation with cddp, which could mean that Pt-DNA damage facilitates topi interaction with DNA. The significance of this observation in relation to the synergistic effect is currently investigated. To further investigate synergy between cddp and CPT, we have used the yeast S. cerevisiae as a model. Wild type, W303, (wt, repair proficient) cells and isogenic 93 EXPERIMENTAL THERAPY

94 94 EXPERIMENTAL THERAPY mutants defective in nucleotide excision repair (rad4), mismatch correction (msh2) and double strand break repair (rad52) were used. All strains contain a plasmid for overexpression of the topi gene to increase sensitivity for CPT, apart from the rad52 strain which was already 10 times more sensitive for CPT than the topi overexpressing strains. The cddp survival curve showed the expected pattern, i.e. the msh2 mutant was less sensitive for cddp compared to the wt, while the rad52 and the rad4 mutants were highly cddp sensitive. Combination experiments with cddp followed by CPT showed a synergistic effect at high doses in rad4 and wt strains, indicating no involvement of the rad4 pathway in synergy. In contrast, the rad52 mutant showed a highly antagonistic to additive effect, which was confirmed in the rad4,rad52 double mutant. We conclude that the rad52 pathway is required for synergy, while the nucleotide excision and mismatch repair pathways are not. This indicates that homologous recombination plays an important role, and its involvement is currently being investigated. Pharmacodynamic endpoints in clinical studies We initiated a randomized phase I study in patients with advanced non-small cell lung cancer to explore the toxicities and maximum tolerated dose (MTD) of cisplatin plus gemcitabine given weekly or every two weeks. Gemcitabine is administered on day one and cisplatin on day two. The study has not yet reached its defined endpoints. The total number of patients entered is 35. The bioanalytical, pharmacokinetic and pharmacodynamic results reveal that there is a significant interaction between gemcitabine and cisplatin. The clearance of unbound platinum is significantly reduced with increasing doses of gemcitabine. In addition, there appears to be a reduction of formation of platinum DNA-adducts in leukocytes by gemcitabine. The mechanism of the interaction and clinical implications are the subject of current investigations. The DNA-adducts are being measured by a sensitive 32Ppostlabeling assay that we have developed and validated. Another phase I and pharmacologic study is ongoing to explore the maximum tolerated dose, dose-limiting toxicities, pharmacokinetics and pharmacodynamics of the combination of carboplatin and topotecan. In this study tumor biopsies are also taken during course one to quantitate the formation of platinum DNA adducts and topo I activity and expression, by a catalytic activity assay and by Western blotting respectively. Levels will be correlated with levels in leukocytes and pharmacokinetic parameters of both drugs as well as with toxicity endpoints. In another phase I and pharmacologic study a novel ruthenium organic compound NAMI-A is being investigated for its toxicity and MTD. In this study the postlabeling assay is being used to quantitate ruthenium-dna complexes in leukocytes. Preliminary results indicate that ruthenium forms DNA adducts, but to a much lower extent than cisplatin and carboplatin. An HPLC assay with online radioisotope detection is being employed to support a recently initiated mass balance study with the novel disulphonamide anticancer agent E7070. Notes 1 Also University Utrecht, Faculty of Pharmacy. 2 Funding Dutch Cancer Society, Project NKI Funding Dutch Cancer Society, Project NKI Collaboration with J Brouwer, Department of Molecular Genetics, University Leiden. 5 Funding Dutch Cancer Society, Project NKI Publications: Pharmacodynamics De Graaff M, Maliepaard M, Pluim D, Floot BGJ, Slaper- Cortenbach IC, Schellens JHM. In vitro antagonistic cytotoxic interactions between platinum drugs and taxanes on bone marrow progenitor cell CFU-GM. Anti-Cancer Drugs 1999; 10: Maliepaard M, Van Gastelen MA, De Jong LA, Pluim D, Van Waardenburg RCAM, Ruevekamp-Helmers MC, Floot BGJ, Schellens JHM. Overexpression of the BCRP/MXR/ABCP gene in a topotecan-selected ovarian tumor cell line. Cancer Res 1999; 59: Pluim D, Maliepaard M, Van Waardenburg RCAM, Beijnen JH, Schellens JHM. 32P-postlabeling assay for the quantification of the major platinum-dna adducts. Anal Biochem 1999; 275: LJ van t Veer PhD H Bartelink MD PhD 2, 5 BR Pieters MD S Rodenhuis MD PhD 3, 6 NS Russell MD FE Van Leeuwen PhD 1, 4 A Broeks PhD 4 AC Lambrechts PhD 6 MAJ Van Dijk MSc 5 AJ Bosma 6 AN Floore JN Ridderbos 5 JHM Urbanus 4 E Dahler 1, 4 Molecular Pathology: Molecular markers of breast cancer group leader Academic staff Academic staff Academic staff Academic staff Academic staff Graduate student Research assistent

95 ATM heterozygous germline mutations contribute to breast cancer susceptibility 4 Approximately 0.5-1% of the general population has been estimated to be heterozygous for a germline mutation in the ataxia-telangiectasia mutated gene (ATM) and epidemiological studies have indicated that female heterozygous carriers have an excess risk of breast cancer (RR 3.9). The purpose of our study is to determine the contribution of ATM heterozygosity to the risk of (radiation-induced) breast cancer. The frequency of ATM germline mutations is examined in high-risk breast cancer families, and in three case-control studies of women who developed breast cancer subsequent to exposure to various levels of ionizing radiation: 1) breast cancer (n=250); 2) Hodgkin s disease (n=200); 3) mammography (n=300). Using the protein truncation test (PTT) as the mutation detection method, followed by genomic sequencing, we expect to have a mutation detection sensitivity of approximately 75%. At present, material of more than half of our selected study participants has been collected and analyzed. The results so far in our first case-control study indicate that ATM heterozygotes have an increased risk of developing a type of breast cancer characterized by frequent bilateral occurrence, early age at onset and long-term survival (approximate RR 9). In addition, all patients had received radiation for their first breast tumor and a trend was noted towards a higher incidence of ATM mutations for women with bilateral breast cancer (5/58 bilateral versus 4/175 unilateral breast cancer patients). This suggests that radiation might indeed be an induction trigger and warrants further investigation into the role of ATM heterozygosity in the pathogenesis of radiogenic cancers. Long-term survivors of Hodgkin s disease who received mantle-field irradiation at a young age have a strong increased risk of developing breast cancer. In the second case-control study we concluded that disease causing ATM germline mutations, however, are not a major component underlying this increased risk (32 cases tested). We did detect several alternative splicing events which might influence protein expression levels. Functional activity of the human estrogen receptor in primary breast cancer and normal breast tissue 5 The positive immunohistochemical staining of the estrogen receptor alpha (her) in fifty to sixty percent of all breast cancers correlates with a better prognosis when patients are treated with anti-estrogens (e.g. tamoxifen). However, a substantial proportion (40%) of these patients does not respond to tamoxifen therapy. Variant estrogen receptors (non-functional, dominantactive or dominant-negative receptors) cannot be discriminated with the routinely used immunohistochemical methods and might be responsible for this phenomenon. The aim of this project is to evaluate the functional activity of her in breast carcinogenesis and to correlate the functional activity of her with clinical response and recurrence-free/overall survival of patients treated with tamoxifen. We developed a functional assay in yeast (her-fasay) that can quantitatively detect the presence of her variants with an aberrant functional activity among normal, wild-type her in a tissue specimen. We found that the presence of wild-type her, relative to the total amount of her present, differs markedly (P<0.0001) between normal breast tissue (median 85% wild-type her) and breast tumors (median 74% wild-type her). Secondly, the her variants with altered function that are present in normal breast tissue are mainly one-exon deleted splicing variants (median 100%), whereas in breast tumors only half of all variants lacks just one single exon (median 50%) (P<0.0001). Our results suggest that her dependent estrogen responsiveness of breast tissue will change during tumor outgrowth, and indicate that specific her variants may play a role in breast cancer development or progression. At present, the resistance to tamoxifen treatment is being correlated with the functional activity of her in breast tumor specimens. Both short term clinical response in advanced disease and recurrent disease after adjuvant tamoxifen treatment are evaluated. Recently, a second estrogen receptor has been identified, estrogen receptor beta. Expression patterns and hormone-binding affinities are different from the known estrogen receptor alpha. The contribution of the newly identified estrogen receptor beta to tamoxifen resistance is also being assessed in our series of breast tumor specimens. To determine the consequences of variant and normal ER expression mechanistically, we have set up a SAGE (serial analysis of gene expression) analysis in the last year in which we compare the spectrum of activated target genes of the wild-type ER and a dominant active ER variant. Stable transfectants of the human cell line, MCF10A (ER-negative breast epithelium cell line), yielded tag-libraries of expressed sequences (on average 9000 tags per cell line), and computer analysis using the SAGE software program is now underway. Differently regulated target genes, as identified by SAGE, will be verified by semiquantative RT-PCR and Northern blotting experiments. Identification of new marker genes for breast cancer using serial analysis of gene expression6 Evaluation of the presence of breast cancer cells in blood and bone marrow has become increasingly important for the staging of breast cancer at diagnosis and in predicting relapse. Disseminated disease has been monitored using epithelial genes as markers in mrnabased methods but yielded conflicting results. We undertook a systematic approach to identify additional marker genes that are breast cancer specific and that are potentially useful for the detection of minimal residual disease (MRD) with molecular assays. Using serial analysis of gene expression (SAGE), we generated three different mrna expression profiles: from a pool of breast 95 EXPERIMENTAL THERAPY

96 96 EXPERIMENTAL THERAPY cancer tissues, from blood of healthy volunteers and from bone marrow of control individuals. Comparison of these three expression profiles resulted in the identification of a range of genes that were expressed in breast cancer, but were absent in the profiles of blood or bone marrow cells. The relative expression levels of 30 of these genes were studied by semi-quantitative RT-PCR. For twelve of these potential marker genes higher mrna expression levels were found in peripheral blood mononuclear cells (PBMC) of patients with metastatic breast cancer than in PBMCs of healthy volunteers. As an example, one of these genes (ps2) was subsequently tested by real-time quantitative PCR on 50 PBMCs of patients with metastatic breast cancer and on 50 PBMCs of healthy volunteers. The ps2 expression was significantly higher in PBMCs of patients with breast cancer than in PBMCs of healthy volunteers (p=0.038). The results indicate that SAGE analysis has the ability to identify suitable genes for detection of minimal residual disease in breast cancer patients. ps2 is the first example resulting from this selection process. Notes 1 Division XII. 2 Division IX. 3 Division X. 4 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Publications: Molecular markers of breast cancer Bartelink H, Begg AC, Coco Martin J, Van Dijk M, Van t Veer L, Van de Vaart P, Verheij M. Towards prediction and modulation of treatment response. Radiotherapy and Oncology 1999; 50, Broeks A, Russell NS, Urbanus JHM, Dahler EC, Van t Veer MB, Hagenbeek A, Noordijk EM, Crommelin MA, Van Leeuwen FE. Van t Veer LJ. Increased risk of breast cancer following irradiation of Hodgkin s disease is not a result of ATM germline mutations. Int J Rad Biol 1999 (in press). Broeks A, Urbanus JHM, Floore AN, Dahler EC, Klijn JGM, Rutgers EJTh, Devilee P, Russell NS, Van Leeuwen FE, Van t Veer LJ. ATM heterozygous germline mutations contribute to breast cancer susceptibility. Am J Hum Genet 1999 (in press). Laake K, Launonen V, Niederacher D, Gudlaugsdottir S, Seitz S, Rio P, Champème MH, Bièche I, Birnbaum D, White G, Sztan M, Sever N, Plummer S, Osorio A, Broeks A, Huusko P, Spurr N, Borg A, Cleton-Jansen AM, Van t Veer L, Benitez J, Casey G, Peterlin B, Olah E, Varley J, Bignon YJ, Scherneck S, Sigurdardottir V, Lidereau R, Eyfjord J, Beckmann MW, Winqvist R, Skovlund E, Børresen-Dale AL and the Breast Cancer Somatic Genetics Consortium. Loss of heterozygosity at 11q23.1 and survival in breast cancer; Results of a large European study. Gen Chr Cancer 1999; 25: Lambrechts AC, Bosma AJ, Klaver SG, Top B, Perebolte L, Van t Veer LJ, Rodenhuis S. Comparison of immunocytochemistry, reverse transcriptase polymerase chain reaction and nucleic acid sequence-based amplification for the detection of circulating breast cancer cells. Breast Cancer Res Tr 1999; 56: Van Dijk MAJ, Hart AAM, Van t Veer LJ. Differences in estrogen receptor alpha variant messenger RNAs between normal human breast tissue and primary breast carcinomas. Cancer Res (in press). MJ Van de Vijver MD PhD H Bartelink MD PhD 1 JL Peterse MD 2 E Robanus Maandag PhD 3 I Faneyte MD C Bosch P Kristel Molecular analysis of breast cancer Group leader Academic staff Academic staff Graduate student Multidrug resistance related proteins in breast cancer Many breast cancer patients die of metastasized disease that is unresponsive to cytotoxic treatment. The mechanisms causing drug resistance are unknown, and no good parameters are available to predict success of treatment. In vitro studies have shown the involvement of ATP binding cassette proteins in multidrug resistance. ABC proteins are over-expressed in MDR cells and their encoding genes, transfected into drug sensitive cells, confer the same MDR phenotype. Although it is clear that ABC proteins play a role in the physiologic defense against toxic compounds, it is as yet unresolved whether they are also involved in clinical drug resistance of breast cancer. To answer this question we have first validated a method to measure the ABC expression. With several RNA based techniques (RNase protection, semi-quantitative RT-PCR, Northern blot) we detected varying levels of expression of Multidrug Resisitance (associated) Protein (MRP)1, MRP2 and MRP3 in breast cancer cell lines and tumor samples. Immunohistochemical detection of these proteins was possible in the cell lines and samples that over-expressed the encoding RNA at least two-fold, relative to that of a housekeeping gene. The P-glycoprotein gene MDR1 was not detected in breast cancer cell lines, whereas the expression in tumor samples could be attributed to lymphocytes and macrophages. We conclude that it is unlikely that MDR1/P-Gp is involved in drug resistance in breast cancer. The possible role of the other ABC proteins will be investigated by comparing the expression of these proteins with response to chemotherapy.

97 97 Genetic alterations in ductal carcinoma in situ and invasive carcinoma of the breast In a previous project (see Division XIII) we have analyzed the genetic alterations in ductal and lobular carcinoma in situ of the breast. This study revealed that the spectrum of genetic alterations in the in situ tumors is comparable to that of the invasive carcinomas. However, the frequencies of the individual genetic alterations differed significantly between the two tumor categories. As most invasive carcinomas also contain an in situ component, we want to compare the genetic alterations in both components of the same tumor and, in this way, identify the genetic alterations that are involved in the progression from the in situ to the invasive stage. Of two invasive breast carcinomas with a relatively large in situ component, we have microdissected the invasive and adjacent in situ component, isolated DNA and performed comparative genomic hybridization. These pilot experiments showed only minor differences between the two components, suggesting that the number of genetic alterations involved in breast tumor progression is low. A larger group of invasive breast carcinomas with an in situ component is now being analyzed. EXPERIMENTAL THERAPY Notes 1 Division IX. 2 Division XIII. 3 Funding: Dutch Cancer Society, Project NKI Secretary Division of Experimental Therapy TEA Eggenhuizen Research staff positions (full time equivalents) Scientific permanent: 4.0 Scientific project: 15.2 Technical permanent: 10.5 Technical project: 13.0

98 98 IX Division of Radiotherapy Division head H Bartelink Head H Bartelink MD PhD Permanent academic staff BMP Aleman MD, JSA Belderbos MD, JH Borger MD PhD, IAD Bruinvis PhD, EMF Damen PhD, RW De Boer PhD, LGH Dewit MD PhD, RLM Haas MD, AAM Hart MSc, RB Keus MD, JV Lebesque MD PhD, EAH Masselink MD, BJ Mijnheer PhD, AWH Minken PhD, LMF Moonen MD PhD, NS Russell MD, JG Salverda MD, BNFM Van Bunningen MD, M Van Herk PhD, FW Wittkämper PhD Other academic staff LJ Boersma MD PhD, L Bos PhD, K De Jaeger MD MSc, N Dekker MSc, M Engelsman MSc, W Heemsbergen MSc, FJP Hoebers MD, M Hoogeman PhD, CW Hurkmans MSc, EPM Jansen MD, R Louwe PhD, GJ Meijer MSc, F Ong MD PhD, BR Pieters MD, LAJ Ploeger MSc, CRN Rasch MD, P Remeijer PhD, A Saarnak PhD, Y Seppenwoolde MSc, CJA Tissing MD, JF Ubbels MD, M Verheij MD PhD, C Vrieling MD Permanent technical staff TJ Minderhoud, PJH Van de Ven Other technical staff J De Goede, P Filius, C Goedbloed, E Kalkman, P Muller, MHP Smitsmans, B Smulders, H Uiterwaal, L Zijp Guests A Olszewska O Törzsök Secretary D Brandwijk E Hofmann Research funded positions (full time equivalents) Clinical permanent: 3.5 Clinical project: 14.3 Technical permanent: 4.4 Technical project: 6.3 Introduction During 1999 progress has been made in the implementation of research findings in the clinic. At the same time, fundamental research inspired by clinical observations of staff members is carried out in close cooperation with other research divisions. A typical example of application of research activities in clinical studies is a phase III randomized clinical trial in prostate cancer patients. In this trial sophisticated conformal radiation techniques are used to escalate the radiation dose to the prostate in order to improve both local cure and survival in these patients. This approach became possible through the input of the different research groups and by applying sophisticated image registration techniques. The treatment plan for individual patients is designed by applying treatment optimization procedures. This plan is executed by meticulous irradiation procedures, with a multileaf collimator on the linear accelerator, and accompanied by intensive quality assurance with dosimetry and megavoltage imaging. The results of this work have also made the application of this approach possible in patients with head and neck or lung cancer. Other studies in patients with breast cancer have led to a better insight into possibilities of reducing the side effects of irradiation and have led to the development of improved treatment approaches. This year was also characterized by an increased integration of the research activities into multidisciplinary projects. Epidemiological studies are ongoing to assess the risk of second cancers after radiotherapy and the possible contribution of ATM germ line mutations in the induction of breast cancer (see Division XII). Preliminary results of the investigation of genetic alterations involved in local recurrence and progression of invasive breast cancer and in situ ductal cancer after breast conserving therapy are reported by Division XIII. The observed improved local cure from adjuvant hormonotherapy with radiotherapy for patients with locally advanced breast cancer has lead to a joint investigation with Division VIII into the link between estrogen receptor b and tumor progression under tamoxifen treatment. Previous results have demonstrated improved local cure and survival in lung cancer patients treated with concomitant cisplatin and irradiation (Schaake Koning et

99 al. NEJM 1992; 327: 432-3). Our results have now been confirmed in various USA trials for patients with cervical cancer. More optimal use of this combined chemo-radiotherapy approach seems to be possible since our results show that cisplatin-sensitive patients can be identified using specific markers. Another successful approach appears to be increasing the tumor concentration of cisplatin by giving this drug intra-arterially (i.a.). An impressive local cure rate has been observed for patients with advanced tumors of the head and neck region treated with i.a. cisplatin plus radiotherapy. Studies are on going in cooperation with Division VIII on the role of endothelial cell and vascular damage in kidney pathogenesis after radiation, with emphasis on mechanisms and possible ways to ameliorate damage. Further collaboration with Division VIII can be seen in projects on the use of fluorescence in situ hybridization (FISH) for detecting radiation-induced chromosome damage, both for estimating intrinsic radiosensitivity of tumor cells (using human tumor biopsies) and in predicting normal tissue morbidity (using peripheral blood lymphocytes). Interesting results emerged from our work on influencing apoptosis by inhibiting the MAPK pathway while initiating pro-apoptotic signals through the SAPK/JNK cascade. The experiments show that alkyl-lysophospholipids increased apoptosis in a wide variety of tumor cells while normal tissue cells remain unaffected. In patients with bladder cancer, it was found that low base-line apoptotic index and over-expression of cyclin D1 was correlated with a poor local cure after irradiation. Three co-workers of the division successfully defended their theses this year and the ESTRO Varian Award for 1999 was presented to Conny Vrieling for her investigation into the dose-effect relationships in patients with early breast cancer with regard to the cosmetic result. Joos Lebesque was presented with the Glyn Evans Medal at the Annual Scientific Meeting of the Royal College of Radiologists, London for his lecture on conformal radiotherapy of prostate cancer in clinical practice. Harry Bartelink received the Marion Burgers Memorial Award for 1999 from the Egyptian Cancer Society. Future directions The clinical research is directed towards escalation of the radiation dose in patients with prostate or lung cancer. This will involve newly designed radiotherapy techniques with radiation dose-intensity modulation. Optimal use will be made of the increasing knowledge of normal tissue tolerance, for example lung tolerance, movement of organs during radiotherapy and improved tumor delineation. Reduction of the impact of organ movement which can enlarge the irradiated volume will be attempted by the so called ABC machine, which controls the breathing of the patient during radiotherapy. Some of these treatments will be carried out in combination with systemic treatment in order to increase the local tumor cell killing effect of radiotherapy. In close collaboration with the Radiology Department, new methods of 3-dimensional delineation of the tumor will be tested in breast cancer patients who underwent MRI examination and will undergo tumorectomy followed by breast conserving therapy. The physics research is directed at providing new methods of target volume delineation, new approaches in delivering intensity-modulated beams to patients and improving accurate dose delivery. Both static and dynamic methods of modifying beam intensity will be explored using various methods of verifying both patient position and beam intensity. Better knowledge of the target volume position during treatment and the incorporation of geometrical uncertainties into the treatment margin will further enhance the potential advantages of these intensity-modulated beams in the clinic. Dose calculations will be improved and verified to assure adequate target coverage with these high dose-high precision treatment techniques. Image acquisition and processing R De Boer, N Dekker, P Filius, S Muller 1, BJ Mijnheer, AR Peters, LS Ploeger, C Rasch 2, P Remeijer 2, MHP Smitsmans, AJ Van Dalen, PJH Van de Ven, RE Vijlbrief, L Zijp, H Bartelink, KGA Gilhuijs 1, JV Lebesque, M Van Herk EPID application for intensity modulated radiotherapy (IMRT) A first version of a field shape verification system for dynamic therapy was completed. In the matrix ionization chamber electronic portal imaging device (EPID), each scan line is recorded at a slightly different time due to the scanning readout. This means that the position of, for instance, a moving diaphragm, will appear in different positions for each line of the image. To verify the position of irradiated objects, the percentage integrated monitor unit (IMU) signal available from the accelerator is digitized once for each scan line and stored with the image. A filter was developed to compensate for the lag and the slow update rate of the IMU signal. Rendering software was developed to obtain the expected (reference) shape of the field outline on the EPID images. The dynamic MLC prescriptions are accepted both in Elekta and DICOM formats. Comparing the reference and measured field outlines demonstrated both random leaf position errors (about 1 mm standard deviation at 6 mm/s at the isocenter) and a significant lag in the position of the leaves. The lag corresponds with a delay of about 0.6 s, and it is partly explained by the lag of the IMU signal which is used to drive the collimator. In cooperation with the Royal Marsden Hospital, London, the feasibility of a number of commercial EPID systems was compared for verification of intensity modulated beam delivery. The preliminary results showed that the accuracy was similar for all systems. 99 RADIOTHERAPY

100 100 RADIOTHERAPY Patient setup errors: 3-D setup analysis for parotid gland tumors Twenty-eight patients were treated for parotid gland tumors with two oblique beams. Orthogonal localization images were used to determine setup deviations. The images were analyzed with a conventional 2-D registration technique and a true 3-D technique which uses CT data to determine patient displacement. For both techniques, standard deviations of the systematic and random translations along the three principal axes (leftright, cranial-caudal and anterior-posterior) were determined. The 3-D technique also quantified the rotations around the three axes. To determine the accuracy of the 2-D registration technique for various rotation angles, the differences between translations determined using 2-D and 3-D registration techniques were analyzed as a function of 3-D rotations. The standard deviations of the systematic and random translations along the three principal axes varied between 1.5 and 2.3 mm. The overall mean translations deviated significantly from zero in the left-right direction (0.8 mm) and the anteriorposterior direction (-1.2 mm). The standard deviations of the systematic and random rotations were about 1 degree. The overall rotations did not deviate significantly from zero. Surprisingly, the rotations are larger than those found in an earlier study for prostate patients, even though we do not apply any fixation of the prostate patients whereas this is the case for parotid gland patients. No correlation was found between the 2-D - 3- D translational differences and the rotation magnitude, which indicates that rotations do not systematically influence the measurement of the translations in 2-D. However, the variability of the 2-D - 3-D translation differences increased with larger rotation magnitudes. This finding demonstrates that the registration errors in 2-D become larger for large rotations. directions), which is not taken into account in the TRE equation. A simulation of marker matching with anisotropy in the errors confirmed this finding. With four re-applied markers, the TRE is 6 mm in most regions of the head. Image registration methods and application: clinical implementation An improved delineation tool has been developed for incorporation of two or more registered scans into treatment planning. The tool is based on concurrent processing, which means, for example, that each manual delineation of multiple image modalities on a CT display occurs simultaneously with the MRI display, and viceversa. A prototype of this system is shown in Fig IX.1. The tool can handle up to 10 scans simultaneously. There are two main challenges in this software. The first one is to keep the 10 images synchronized, for which special automated data management software was developed. The second challenge was to maintain the transformation matrices involved with the different coordinate systems. For this purpose a software component was developed that automatically maintains a matrix of NxN coordinate systems and all associated transformation parameters, thus simplifying the extension of the number of scans. This system is in experimental use for clinical problems such as measurement of fast tumor growth (Fig IX.1) and the assessment of recur- Image registration methods and application: marker matching We investigated the accuracy of 3-D matching using markers that are repeatedly applied to external anatomical landmarks on the head. The purpose of this study is to establish a lower limit of the errors that would occur in, for instance, MRI-SPECT matching, which can, sometimes, only be achieved using external landmarks. Marker matching was compared with (single modality) volume matching for 20 MRI scans using several marker combinations. The results were compared with a published expression for the Target Registration Error (TRE), which gives the 3D distribution of the mismatch between both scans. The main error source was identified as reapplication of the markers on the anatomical landmarks. The theoretical expressions describe the relative distribution of the TRE in space, but tend to underestimate the actual registration error. This deviation is due to anisotropy of the marker position errors (the error of the marker position in the direction perpendicular to the skin surface is much smaller than the error in other Figure IX.1 A prototype delineation tool, developed at the NKI, for multiple image modalities. Scan 1 shows a zoomed part of a planning CT scan of a brain tumor patient. Scans 2 and 3 show matched and resampled MRI scans. In this particular patient, fast tumor progression was observed on a diagnostic CT scan (4). This tool allows instantaneous delineation on any of these scans. Actions performed on any display are reflected in real time in all other displays: such as slicing, zooming and panning. In this particular case, the planning CT delineated edema region (black) and tumor margins (white) are smaller than the corresponding region on the diagnostic CT scan. On Scan 5 the high dose region of the planned dose distribution is the bright area. Since the edge of the edema was closely approaching the edge of the high dose region, the treatment fields were enlarged. Scan 6 contains the paint tool used to delineate the tumor.

101 rences, and also in an evaluation study in which helical CT scans made in different phases of the breathing cycle are compared with each other and with PET data which allows separation of active tumor tissue from other tissues (see also Division XIII). Notes 1 Division XIII. 2 Funding Dutch Cancer Society, Project NKI Dose calculation J Venselaar 1, BJ Mijnheer New formalisms, for monitor unit calculation for radiation treatments with photon beams from accelerators and 60 Counits, have been developed by ESTRO and a working group of the Netherlands Commission in Radiation Dosimetry (NCS). Both formalisms apply a coherent system for the use of scatter correction factors for dose calculations on the central axis of arbitrarily-shaped photon beams. The system is suitable for application in both the fixed source-surface distance (SSD) and in the isocentric treatment setup. In order to derive the relations in the formalism, we introduced a separation of the phenomena related to the energy fluence in air and to the phantom scatter contributions to the dose. It was shown that dose calculations can be performed with only one set of basic beam data, obtained at the reference depth of 10 cm. Problems related to measurements performed at the depth of maximum absorbed dose, due to the electron contamination of the beam, are avoided in this way. Dose calculations in a fixed SSD treatment set-up are straightforward. Application in the isocentric treatment setup needs simple conversion steps, while the inverse approach, from isocentric to fixed SSD, is also described. The new formalism requires collimator and phantom scatter correction factors, S c and S p, defined at the reference depth of 10 cm. These data can be obtained from measurements at that depth in a mini-phantom and in a full scatter phantom. Equations have been derived, giving the relation between these quantities and corresponding quantities obtained from measurements at the depth of dose maximum. It was shown that conversion of S c and S p determined at 10 cm depth to quantities defined at dose maximum such as (normalized) peak scatter factor, (N)PSF, (normalized) tissue-air ratio, (N)TAR, and vice versa is not possible without quantitative knowledge of the electron contamination. The difference in S c at d max resulting from this electron contamination compared with S c values obtained at a depth of 10 cm in a mini-phantom has been determined as a multiplication factor, S cel, for a number of photon beams of different accelerator types. Figure IX.2 shows that S cel may vary up to 5%. In the new formalisms, output factors are defined at a reference depth of 10 cm, therefore they do not require S cel data. The use of S c and S p values, defined at 10 cm depth, in combination with relative depth-dose data or tissue-phantom ratios is therefore recommended. For a transition period, the use of the new equations and S cel data might be required if, for instance, treatment planning systems apply S c data normalized at d max. Note 1 Dr Bernard Verbeeten Instituut, Tilburg. 101 RADIOTHERAPY Experimental dosimetry L Bos 1, M Engelsman 2, JH Lanson 3, G Laureijs 1, R Louwe 4, N Raaijmakers 5, H Uiterwaal, K Van Ingen 2, FW Wittkämper, BJ Mijnheer, M Van Herk Figure IX.2 Experimentally determined values of S cel(v c,d max) for four different photon beam qualities ranging from 60 Co to 25 MV. For measurements at d ref, S cel(v c,d ref) by definition equals unity. In vivo dosimetry Conformal radiotherapy requires accurate knowledge of the actual dose delivered to a patient. The impact of routine in vivo dosimetry has been analyzed for three conformal treatment techniques to evaluate its usefulness in daily clinical practice. Based on pilot studies, routine in vivo dosimetry quality control (QC) protocols were implemented in the clinic. Entrance and exit diode dose measurements have been performed during two treatment sessions for 378 patients having prostate, bladder and parotid gland tumors. Dose calculations were performed with a CT-based three-dimensional treatment planning system. In our QC-protocol we applied action levels of 2.5% for the prostate and bladder

102 102 RADIOTHERAPY tumor group and 4.0% for the parotid gland patients. When the difference between the measured dose at the dose specification point and the prescribed dose exceeded the action level, the deviation was investigated and the number of monitor units (MUs) adjusted. For 34 (9%) patients the difference between the measured and calculated dose was larger than the action level. Systematic errors in the use of a new software release of the monitor unit calculation program, limitations of the dose calculation algorithms, errors in the planning procedure and instability in the performance of the accelerator have been detected. It can be concluded that accurate in vivo dosimetry, using a diode measurement system, is a powerful tool to trace dosimetric errors during conformal radiotherapy in the range of %, provided that the system is carefully calibrated. The implementation of an intensive in vivo dosimetry program requires additional staff for measurements and evaluation. The patient measurements add only a few minutes to the total treatment time per patient and guarantee an accurate dose delivery, which is a prerequisite for conformal radiotherapy. Transmission dosimetry Several groups have presented procedures to use electronic portal imaging devices for in vivo dosimetry. We developed methods to determine transmission, exit and mid-plane dose distributions from portal images acquired with a liquid-filled electronic portal imaging device (EPID). Recently, a PC-based software system was developed for bringing portal in vivo dosimetry into full-scale clinical use. The implementation consists of computation modules written in C and a user interface written in Delphi. The following components have been incorporated: database and treatment planning system (TPS) integration, acquisition software, dose reconstruction algorithms and evaluation tools. The program was tested for a variety of combinations of input parameters and calibration data which resulted in a number of modifications. The intercomparison of film dosimetry and EPID dosimetry for phantoms has also been made possible. One complication is that the films and the EPID measurements often refer to different planes in space. For this reason, the dose planes are first matched with the patient s CT data in those planes and then compared in a three-dimensional grid. Phantom measurements are in progress and will be used to compare EPID dose distributions with ionization chamber and film measurements. Recently, a start has been made with the clinical use of the system for the measurement of midplane dose distributions during radiotherapy of the breast. The purpose of this study is to evaluate the occurrence and magnitude of dose inhomogeneities. Dosimetry of intensity-modulated beams Conformal radiotherapy of lung tumors can be performed in various ways, including the use of intensitymodulated beams (IMRT). In order to assess the limitations in the design of new techniques for IMRT of lung tumors, it is necessary to quantify the errors made by dose calculation algorithms currently available in commercial treatment planning systems (TPS). Both the absolute dose at the ICRU dose specification point and relative dose distributions were therefore assessed for a simple beam setup. Film and ionization chamber measurements were performed in an inhomogeneous polystyrene and cork layered phantom, with a spherically shaped polystyrene insertion simulating a tumor inside a lung. Single non-wedged conformal beams of different energies, shaped by a multi-leaf collimator, were used and measurements were compared with calculations using various treatment planning systems with different inhomogeneity correction algorithms. For this specific configuration, the dose calculation at the dose specification point, i.e. the monitor unit calculation, has a systematic uncertainty smaller than a few percent depending on field size, beam energy and dose calculations algorithm. Underestimation of the penumbra width may occur, depending on the algorithm used in a particular TPS, leading to an underdosage of the Planning Target Volume, PTV, up to about 20%. It can therefore be concluded that algorithms used for monitor unit calculations in lung cancer treatments have acceptable accuracy if the non-wedged fields are not too small. However, these algorithms do not yet take into account the lateral electron transport in low density material like lung tissue. This limitation will lead to unaccepable errors when choosing field sizes for conformal treatments particularly for high photon energies. An important tool for the verification of newly designed IMRT techniques is film dosimetry using a stack of films in specific phantoms. In order to determine the accuracy of dose determinations using radiographic film, we evaluated the reproducibility of two types of dosimetric films, TVS (CEA) and X-Omat V (Kodak). For the former type of film, we found a large intra-institutional variability of the sensitometric curve, in contrast to other studies evaluating the same type of film. This large variability is probably related to ageing of the film, in combination with a rather long period of analysis. The inter-institutional variability for the TVS (CEA) films was smaller than the intra-institutional variability, but markedly higher than the inter-institutional variability of the X-Omat V (Kodak) films. We concluded that the sensitometric curves of CEA films must be determined regularly in order to perform accurate film dosimetry. The sensitometric curves of Kodak films show less variability, which reduces the need for regular verification. Boron neutron capture therapy Consistent dosimetry is a requirement for the meaningful comparison of the results of treatment by boron neutron capture therapy (BNCT). International recommendations currently available for conventional photon and electron beam dosimetry and for fast neutron therapy are not applicable to BNCT. An EC project was

103 therefore started to draft a Code of Practice for the Dosimetry of BNCT. An important problem with the standardization of dosimetric procedures is which phantom material is most suitable for the dosimetry under reference conditions of neutron beams for BNCT. For this purpose, phantoms of various dimensions, composed of water, tissue-equivalent (TE) liquid, polyethylene (PE), polymethyl-metacrylate (PMMA) and water containing 10 B were irradiated using the Petten BNCT beam. Activation foils and a diode detector were used for the determination of the thermal neutron fluence rate. The gamma-ray dose rate and the fast neutron dose rate were determined using paired ionization chambers. In water, PMMA and TE-liquid, the absolute dose and fluence values agreed within 3% at the reference depth, with the exception of the gamma-ray dose rate in PMMA which was 12% lower than in water. Due to a higher hydrogen concentration in PE compared with water, the dose and fluence values in PE differed more than 30% from those in water. Only minor differences were observed between the percentage depth dose curves for the various dose components in water, PMMA and TE-liquid. The addition of clinically relevant amounts of 10 B to water resulted in a considerable decrease in the absolute thermal neutron fluence and a decreased penetration of thermal neutrons. For reference dosimetry of an epithermal neutron beam for BNCT both water and TE-liquid are suitable phantom materials. For practical reasons, water is therefore proposed as reference phantom material. For measurements requiring a solid phantom, PMMA is recommended. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Medisch Centrum Alkmaar. 4 Funding: Foundation for Technical Sciences, Project BGN University Medical Center, Utrecht. Treatment planning JSA Belderbos, JH Borger, K De Jaeger, BA Fraas 1, CW Hurkmans, JV Lebesque, DL McShan 1, TJ Minderhoud, BR Pieters, AE Saarnak, BS Smulders, MJ Steggerda, IAD Bruinvis, EMF Damen, BJ Mijnheer Dose and monitor unit calculation The use of multi-segment techniques for irradiation of advanced head and neck cancer and prostate cancer initiated the development of a more accurate beam description in the treatment planning system and of new methods to calculate monitor units for highly asymmetric fields. For the 6 and 8 MV beams of our accelerators, the parameters of the photon beam model in the treatment planning system have been refitted with emphasis on asymmetric fields. This procedure resulted in excellent agreement between measured and calculated dose profiles for wedged and open fields for both symmetric and asymmetric fields (including fields with offsets larger than half the field size). To accurately calculate monitor units for asymmetric fields, correction factors depending on field size and offset were measured and incorporated into the monitor unit calculation program. As a result of these modifications, the relative and absolute dose calculation is accurate to within 2% for all open and wedged fields, including highly asymmetric fields. Lung cancer Treatment planning for inoperable NSCLC patients was considerably refined by allowing up to six beams and a non-coplanar beam set-up. In addition, small beam segments are allowed to compensate the higher beam attenuation in the mediastinum. In the dose escalation study the patients are classified into five groups according to the calculated mean lung dose. The physical 3-D dose distribution of the complete treatment (i.e. including the dose contribution of fields used for electronic portal imaging) is converted to a biologically equivalent dose distribution given in fractions of 2 Gy using the Normalized Total Dose (NTD) concept with an α/β-ratio of 3 Gy. The mean (biologically equivalent) lung dose, NTD mean, is subsequently calculated for the lungs as one organ. The final treatment plan should fulfill strict criteria with respect to coverage of the Planning Target Volume (PTV), and total biologically equivalent dose to spinal cord, heart, and esophagus. Breast cancer In order to minimize the lung dose and to reduce the dose to the heart, a new irradiation technique for internal mammary node (IM) irradiation of breast cancer patients was developed. In the improved technique, an oblique electron and an oblique asymmetric photon field are combined to irradiate the IM lymph nodes. Three-dimensional treatment planning was performed for 8 patients for the standard and the new technique. Dose volume histograms and normal tissue complication probabilities were compared. Better IM planning target coverage was seen for the improved technique. The heart generally received less dose with the improved technique and a small, but acceptable, increase in lung dose was occasionally found. In order to determine the most accurate and efficient localization procedure of the IM nodes, lymphoscintigraphy, ultrasound, and CT-scan are being evaluated in a study group of 40 patients. Note 1 University of Michigan, Ann Arbor, USA. 103 RADIOTHERAPY

104 104 RADIOTHERAPY Studies of radiation response in normal tissue P Baas 1, K De Jaeger 2, C Goedbloed 2, E Kalkman 3, SH Muller 4, Y Seppenwoolde 2, JCM Theuws 3, J Belderbos, LJ Boersma, JV Lebesque Local dose-effect relations were determined by using Single Photon Emission Computed Tomography (SPECT) lung perfusion and ventilation scans matched with the accompanying CT scans for patients with breast cancer, malignant lymphoma and non-small cell lung cancer (NSCLC). Normalized dose-effect relations for perfusion and air-filled fraction were calculated (at 3 months followup) for different patient subgroups and different regions in the lung. The changes in perfusion as a function of dose were less for all NSCLC-patients than the changes for patients with healthy lungs, especially above 50 Gy. However, for the air-filled fraction, no deviation from the previously derived dose-effect relation for healthy lung tissue could be seen. When the group was divided into patients with and without reduced perfusion adjacent to the tumor, the group with homogeneous perfusion before radiotherapy showed greater changes in lung perfusion than the group with reduced perfusion adjacent to the tumor. Average dose-effect relations for well and poorly perfused regions of the lung (DE wp and DE pp) are shown in Figure IX.3. For the well-perfused lung regions there was a steeper dose-effect relation than for the poorly perfused areas; these last regions show a more chaotic dose response while in the lower dose bins some functional recovery could be observed (re-perfusion). For the reference group (lymphoma and breast cancer patients), there was no difference in response for the well and poorly perfused lung regions. For the air-filled fraction no significant differences in dose response could be observed between the patient subgroups and the different regions of the lungs. To obtain a dose-effect relation, valid up to 80 Gy, we extended the data of the reference group (lymphoma and breast cancer patients) with the data of the well-perfused parts of the lungs of the patients with NSCLC. This dose-effect relation could be described by a logistic fit with a D 50 of 63 ± 3 Gy and a steepness parameter k of 1.7 ± 0.2. For patients with breast cancer and malignant lymphoma we saw a recovery in local function (perfusion and ventilation) and air-filled fraction when we compared dose-effect relations measured at 3 and 18 months after the start of radiotherapy, with no further recovery from 18 to 48 months. The recovery from radiation induced changes in ventilation and air-filled fraction was similar for breast cancer and lymphoma patients, but lymphoma patients showed more recovery in perfusion. Regional differences in radiosensitivity were not present, except for the dorsal versus ventral region, which was attributed to a gravity-related effect in the measuring procedure. Notes 1 Division X. 2 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Division XIII. Conformal radiotherapy JSA Belderbos, LJ Boersma, JJP De Goede, K De Jaeger 1, WD Heemsbergen, M Hoogeman 2, P Koper 2, EA Lamers, GJ Meijer, BJ Mijnheer, P Muller 2, C Rasch, Y Seppenwoolde 3, BS Smulders, M Van Herk, HP Vijlbrief, D Yan 4, IAD Bruinvis, JV Lebesque Conformal radiation therapy (CRT) aims at increasing the dose to the target volume and/or decreasing the irradiated volumes of normal tissue. CRT is being used in Phase II and III studies to investigate the impact of delivering higher tumor doses on local control and survival. In 1999, 90 prostate, 12 head and neck, 20 parotid gland, 12 bladder and 38 lung CRT treatments were done. Figure IX.3 The average dose-effect relation for well perfused lung regions, DEwp (triangles) and the average dose-effect relation for poorly perfused lung regions, DEpp (open circles). The solid line (extrapolation is dashed) represents the doseeffect relation (logistic fit) for lymphoma and breast cancer patients (DEref.pat). Prostate cancer The accrual of patients in the phase III dose escalation trial was continued. Up to October 1999, 78 NKI and 195 Daniel den Hoed Cancer Center patients had been randomized. To study the results of this trial, with respect to tumor control and normal tissue complications in relation to the 3D dose distribution, a 3D image database was developed to store all acquired patient data. The database offers direct access to CT images, dose distributions, and delineated organs of all entered patients. Furthermore, the database is able to process the various uncertainties in the delivery of CRT and apply

105 these uncertainties to the analysis of dose-effect relationships for all patients in a uniform way. When all treatment data of a patient is available, the data is transferred to the database, where it is stored in a CD-ROM jukebox. A sophisticated user interface provides access to the different planning system and data exchange formats. This interface is used to check the consistency of the treatment planning data and, if necessary, fix the data. Altered and newly generated data are stored in an additional, format-independent database, whereas the user interface combines both databases. The trial database has been designed to incorporate uncertainties in dose delivery for the analysis of the relationships between the actual delivered dose distribution and local tumor control and complications of normal tissue. A recently implemented model is a mechanical representation of the rectum, which has a different shape for each delivered fraction. The model describes the shapes that a rectum can obtain during the course of a treatment. In this model the rectum is divided into slices. For each slice we determine the area, the difference of the area with respect to its two neighboring slices, and its position. Different rectum configurations are generated using a Monte Carlo technique, in which the position and the area of all rectum slices is randomly varied. The random variations are weighted by probability distributions of the computed parameters. The probability distributions are derived from a generic group of patients, which are provided by D. Yan 4. Each patient of this group received 1 planning CT scan and 17 repeat CT scans during the treatment. All repeat CT scans are matched on the planning CT scan using the bony anatomy. For each treatment fraction, a new rectum configuration will be computed using the Monte Carlo technique. The actual delivered total dose will be estimated on the basis of the obtained rectum configurations. The location of the prostate is correlated with the position and the shape of the rectum. The target volume motion will therefore also be included in the simulation. Lung cancer The phase I/II dose escalation study for inoperable NSCLC patients was continued. Up to October 1999, 23 patients had been entered in the study. In the three lowest risk groups for radiation pneumonitis, the next higher dose level of 81 Gy was attained. A collaboration with the AZVU was started to investigate the impact of metabolic imaging on target volume determination. In a number of patients with inoperable non-small cell lung cancer referred for radical radiotherapy, a thoracic transmission and emission 18 FDG-PET scan was performed. CT scans for radiotherapy planning were matched with the PET-transmission scans, using an adjusted root mean square algorithm. 18 FDG activity in lymph nodes and primary tumor was anatomically localized by overlaying the simultaneously registered emission PET scan on the matched CT scan. Preliminary analysis indicates that additional information from PET scan changes the CT-based target volume in approximately 30% of the patients. Correcting dose distributions distorted by set-up verification fields Electronic portal imaging is used to verify the set-up accuracy. Since the treatment field images are generally not suitable for matching purposes, additional orthogonal fields are used with a number of monitor units between 16 and 34. Both the absolute dose and the dose homogeneity over the planning target volume (PTV), are distorted by these extra fields. A method was developed to adjust the beam weights and wedge angles of the treatment plan in such a way that the dose homogeneity over the PTV is maintained. The method applies to three dimensions, which also allows application for noncoplanar beam setups. The dose gradient of a single beam can be described by a vector, whose magnitude is proportional to the beam weight of the beam. This problem can thus be solved using vector analysis methods using several coordinate transformations. The formalism resulted in clinically acceptable solutions for all coplanar and non-coplanar cases. Notes 1 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Dr. Daniel den Hoed Cancer Center, Erasmus University, Rotterdam. 4 William Beaumont Hospital, Royal Oak, MI, USA. Breast Cancer E Damen, AAM Hart, FJP Hoebers, CW Hurkmans, JV Lebesque, BJ Mijnheer, SH Muller1, JL Peterse 1, BR Pieters, AE Saarnak, L Van t Veer 1, C Vrieling, H Bartelink, JH Borger In EORTC trial 22881/10882 a total of 5,569 breast cancer patients was randomized for boost versus no boost irradiation following whole-breast irradiation. The cosmetic result was assessed for 731 patients after 3 years of follow-up. A panel scored the qualitative appearance of the breast using photographs. Quantitative digitizer measurements of nipple displacement were performed using these photographs for breast retraction assessment (BRA). For the panel evaluation the intraobserver agreement for the global cosmetic score was moderate, the interobserver agreement was fair. All specific cosmetic items scored by the panel were significantly related to the global cosmetic score; breast size and shape being the most important. For the digitizer measurements, the standard deviation from the average value of 30.0 mm was 2.3 mm (7.7%) for the intraobserver variability and 2.6 mm (8.7%) for the interob- 105 RADIOTHERAPY

106 106 RADIOTHERAPY server variability. The two methods moderately correlated. Both methods showed that the boost had a limited but significant adverse effect on the cosmetic result. The digitizer evaluation of cosmesis appears to be easier and more accurate than the panel evaluation, however, there are some treatment sequellae, such as disturbing scars and skin changes, that can not be evaluated by BRA measurements. Three-dimensional treatment planning was performed for 30 patients with left-sided breast cancer to assess the dependence of cardiac and lung complications on treatment technique and individual patient anatomy. Two locoregional techniques (A and B) and a tangential field technique (C), were planned and evaluated for each patient. Dose-volume histograms (DVHs) and normal tissue complication probabilities (NTCPs) for the heart and lung were compared for the three techniques. In the beam s eye view of the medial tangential fields the maximum distance of the heart contour to the posterior field border (MHD) was measured. NTCP values for excess cardiac mortality due to acute myocardial ischaemia varied considerably between patients, with minimum and maximum values of 0.2% and 8.9% (A), 0.3% and 7.2% (B) and 0.0% and 6.4% (C). NTCP values for radiation pneumonitis were very low for all techniques. The NTCP values for technique C could very well be described using the MHD concept. Therefore, the MHD is a simple and generally applicable method to estimate the risk for cardiac mortality, making it easy to estimate the cardiac risk using simulator films and thereby to modify the treatment technique. To evaluate the effectiveness and morbidity of primary axillary radiotherapy in breast conserving therapy in postmenopausal clinically node negative patients with early stage breast cancer, 105 patients were treated between 1983 and 1997 by wide local excision, followed by radiotherapy to breast and axilla. Fifty-five patients with no evidence of disease were evaluated for late side effects. Three patients had a local recurrence (3%). No isolated axillary recurrences were seen. In 2 patients axillary recurrence was accompanied by distant metastases (2%). Arm edema was reported subjectively by 4% of patients, objectively measured in 11%. An impaired shoulder function was subjectively noticed in 35%, objectively measured in 17%. No brachial plexus neuropathy was seen. It can be concluded that primary axillary radiotherapy can be considered as a safe alternative for axillary lymph node dissection in selected patients. Note 1 Division XIII. Brachytherapy M Boersma, JH Borger, AE Saarnak, MJ Steggerda, BNFM Van Bunningen, R Wolterink, BR Pieters A method to reduce treated volumes of interstitial breast implants using geometrically optimized stepping source techniques and alternative dose normalization The standard breast implant of our clinic, consisting of an uneven number of needles loaded with linear sources of equal lengths, was compared with an optimized implant applying an even number of needles. In the optimized situation linear sources were replaced by a stepping source with geometric volume optimization to the shape of the target volume. A cylindrical target volume of 48 cm 3 was chosen. The V(100), V(125) and V(150) (= volumes encompassed by the reference isodose (CRI) of 100%, 125% and 150%) were reduced by approximately 35% for the optimized implant. Despite substantial reduction of treated volumes, the target volume was completely covered by the RI in the optimized situation. The effect of dose normalization on the dose distribution, i.e. 85% of the mean central dose (MCD) vs 91% MCD, was subsequently studied for the optimized implant. As an example four different target volumes (varying from 15 cm 3 to 40 cm 3 ) were chosen. The average reduction of V(125) and V(150) when the dose is normalized to 91% instead of 85% of the MCD was 21.2% and 12.7 %, respectively. The V(100) showed little variation. This study shows that by altering the geometry of an implant, performing geometric volume optimization with a stepping source and improving the dose distribution by normalizing to 91% of the MCD, a substantial reduction can be achieved for treated volumes, whilst maintaining full coverage of the target volume. This optimized technique was recently implemented in the clinic using a high dose-rate stepping source. Inter-observer variation in bladder and rectum delineation on CT-images for brachytherapy of cervical cancer Intracavitary treatment of cervical cancer is planned using post-implant CT images. Bladder and rectum, which are the organs at risk, are delineated in order to estimate the absorbed dose using dose-volume histograms (DVHs). In order to determine the accuracy of the dose assessment in bladder and rectum, different aspects of the planning procedure are investigated. One of the studies concerned the inter-observer variation in delineation of bladder and rectum. Two dosimetrists and one radiation oncologist delineated bladder and rectum on CT scans of 10 patients. The dose to bladder and rectum volumes in the high-dose region was determined from DVHs of each delineated organ. In addition, the dose at points placed on the delineated bladder and rectum walls was determined. Variation in the dose due to delineation differences was about 10% both in the high-dose region of the DVHs and in the dose points. Variation in delineation occurred particularly in areas where organ borders are difficult to distinguish

107 on a CT image, such as between the bladder and cervix and the rectum and vagina. Since the filling of these organs influences the DVHs, we are now investigating whether dose-wall histograms or dose-surface histograms would give a better representation of the dose to these organs. Radiation-induced apoptosis L Moonen, GA Ruiter 1, 2, SF Zerp 1, 2, H Bartelink, WJ Van Blitterswijk 1, M Verheij Figure IX.4 T4 processus alveolaris carcinoma before treatment, tumor blush and 6 weeks after treatment. Tumor is in complete remission. Apoptosis is now recognized as a significant mode of cell death by ionizing radiation. Split-dose fractionation experiments in endothelial cells showed that the contribution of the second fraction to cell death increased with longer time intervals up to 32 hours. These data are in line with the concept of re-emergence of an apoptosissensitive sub-population between fractions. In order to further increase radiation-induced apoptosis, we use synthetic membrane-permeable alkyl-lysophospholipids (ALPs; Et-18-OCH 3, HePC and D-21266). These compounds have potent anti-tumor properties and some are currently undergoing clinical evaluation. ALPs primarily act on cell membranes where they interfere with mitogenic signaling pathways. We found that ALPs prevent MAPK/ERK signaling in a Ras-independent manner by interfering with the PLC/Ca 2+ /PKC pathway. Besides the inhibitory effect on MAPK/ERK, ALPs also suppress the anti-apoptotic Akt/PKB pathway and initiate pro-apoptotic signals through the SAPK/JNK cascade. Whereas ALPs were found to induce apoptosis in a wide variety of tumor cell lines, normal fibroblasts and quiescent vascular endothelial cells remained unaffected. In combination with radiation, ALPs act synergistically in inducing an apoptotic response. Both radiation- and ALP-induced apoptosis proceed in a caspase-dependent fashion. Our current investigations focus on identifying the upstream, membrane-associated, target(s) of ALP-induced apoptosis. In collaboration with M Gallee (Division XIII) we investigated whether pretreatment levels of apoptosis (AI), Ki-67 index and expression of p53, cyclin D1 and prb correlate with treatment outcome after radiotherapy in muscle-invasive bladder carcinoma (n=88). In this retrospective analysis, over-expression of cyclin D1 and, to a lesser extent, a low baseline AI were found to correlate with poor local control after external beam irradiation. Notes 1 Division III. 2 Funding: Dutch Cancer Society, Project NKI Combination of radiotherapy and chemotherapy AC Begg 1, JSA Belderbos, R Keus, C Rasch, L Schultze Kool 2, B Tan 3, PJM Van de Vaart 4, H Bartelink Assessment of cisplatin DNA adducts as prognostic factor In a study of 27 patients with non-small cell lung cancer (NSCLC) treated with concomitant daily cisplatin and radiotherapy, the tumor specimens were analyzed for p53 and bcl-2 expression. Staining intensity with antibodies against Ki-67 was also measured and the apoptotic index was estimated with the tunnel assay. Finally, in buccal cells cisplatin-induced DNA adducts were assessed immunohistochemically. The median follow-up was 41 months, and 21 patients (78%) have died. In a univariate analysis, age, tumor stage and cisplatin-dna adduct staining were the only factors significantly associated with survival. P53, bcl-2, Ki-67 and apoptosis showed no relationship with outcome. Multivariate analysis revealed that cisplatin-dna adduct staining remained an independent prognostic factor with shorter survival times for patients with a low adduct staining. The determination of the extent of platinum-induced DNA modification in buccal cells could provide a simple way to predict treatment outcome in patients with NSCLC treated with daily concomitant cisplatin and radiation. RADPLAT For inoperable advanced head and neck cancer the phase II study (RADPLAT) protocol of concomitant high 107 RADIOTHERAPY

108 108 RADIOTHERAPY dose intra-arterial chemotherapy with systemic rescue with sodiom thio-sulfate and conventional radiotherapy was continued. A total of 76 patients entered the study (67 T4, 9 T3 tumors, 3-10 cm). The majority of the patients were node positive (51 patients). The median follow-up was 11 months (20 30), the median age was 50 years (40-69). A complete remission (CR) of the primary tumor was achieved in 84%, the lymph node CR-rate was 88%, the remainder had partial remission (PR) at the primary site as well. Figure IX.4 shows an example. No stable or progressive disease occurred. Grade IV hematological toxicity (Common Toxicity Criteria) occurred in 17% but did not result in treatment interruptions. Grade III (RTOG) pharyngeal mucositis, upper gastrointestinal and skin toxicity and hearing loss occurred in 43%, 60%, 24% and 10%, respectively. The mean hearing loss was 5 db. At 20 months local and regional control was 73% and 80% respectively. Three patients experienced osteoradionecrosis, all in previous tumor invaded areas. Notes 1 Division VIII. 2 Division XIII. 3 Division XI. 4 Daniel den Hoed Cancer Center. Borger JH, Kroon BBR. Low risk of locoregional recurrence of primary breast carcinoma after treatment with a modification of the Halsted radical mastectomy and selective use of radiotherapy. Cancer 1999; 85: Brugmans M, Van der Horst A, Lebesque JV, Mijnheer BJ. Beam intensity modulation to reduce the field sizes for conformal irradiation of lung tumors: a dosimetric study. Int J Radiat Oncol Biol Phys 1999; 43: Coco Martin J, Mooren E, Ottenheim C, Burril W, Nunez MI, Sprong D, Bartelink H, Begg AC. Potential of radiation-induced chromosome aberrations to predict radiosensitivity in human tumour cells. Int J Radiat Biol 1999; 75: De Jong D, Aleman BMP, Taal BG, Boot H. Controversies and consensus in the diagnosis, work-up and treatment of gastric lymphoma: an international study. Ann Oncol 1999; 10: Dewit L, Verheij M, Valdés Olmos R. Regarding Maor, North, Cabanillas et al. in Int J Radiat Oncol Biol Phys 40, , 1998 [letter]. Int J Radiat Oncol Biol Phys 1999; 43: Essers M, Mijnheer BJ. In vivo dosimetry during external photon beam radiotherapy. Int J Radiat Oncol Biol Phys 1999; 43: Publications: Division IX Full papers Lanson JH, Essers M, Meijer G, Minken AWH, Uiterwaal H, Mijnheer BJ. In vivo dosimetry during conformal radiotherapy: requirements for and findings of a routine procedure. Radiother Oncol 1999; 52: Bartelink H, Begg AC, Coco Martin J, Van Dijk M, Van t Veer L, Van de Vaart PJ, Verheij M. Towards prediction and modulation of treatment response. Radiother Oncol 1999; 50: Bartelink H. Is neoadjuvant chemotherapy the answer for bladder cancer? [Commentary]. Lancet 1999; 354: Bartelink H, Roelofsen F, Eschwege F, Rougier P, Bosset JF, Gonzalez Gonzalez D, Peiffert D, Van Glabbeke M, Pierart M. Concomitant radiotherapy and chemotherapy is superior to radiotherapy alone in the treatment of locally advanced anal cancer: results of a phase III randomized trial of the European Organization for Research and Treatment of Cancer, Radiotherapy and Gastrointestinal Cooperative Groups. Classic Papers and Current Comments 1999; 3: Mageras GS, Fuks Z, Leibel SA, Ling CC, Zelefsky MJ, Kooy HM, Van Herk M, Kutcher GJ. Computerized design of target margins for treatment uncertainties in conformal radiotherapy. Int J Radiat Oncol Biol Phys 1999; 43: Meijer G, Van den Brink M, Hoogeman MS, Meinders J, Lebesque JV. Dose wall histograms and normalized dose surface histograms for the rectum. Int J Radiat Oncol Biol Phys 1999; 45: Moonen L, Van der Voet H, De Nijs R, Hart G, Horenblas S, Bartelink H. Muscle-invasive bladder cancer treated with external beam radiotherapy: pretreatment prognostic factors and the predictive value of cystoscopic re-evaluation during treatment. Radiother Oncol 1998; 49: Begg AC, Haustermans K, Hart G, Dische S, Saunders M, Zackrisson B, Gustaffson H, Coucke P, Paschoud N, Houer M, Overgaard J, Antognoni P, Richetti A, Bourhis J, Bartelink H, Horiot JC, Corvo R, Giaretti W, Awwad H, Shouman T, Jouffroy T, Maciorowski Z, Dobrowsky W, Struikmans H, Rutgers D, Wilson GD. The value of pretreatment cell kinetic parameters as predictors for radiotherapy outcome in head and neck cancer: a multicenter analysis. Radiother Oncol 1999; 50: Bijker N, Rutgers EJ, Peterse JL, Van Dongen JA, Hart AMM, Ptaszynski A, Van den Bogaert W, Horiot JC, Poortmans P, Fourquet A, Struikmans H, Bartelink H, Pierart M, Collette L, Van der Schueren E. Radiation dose homogeneity in an EORTC multicentre trial on breast irradiation. Acta Oncol 1999; 38 (suppl 13): Rasch CRN, Remeijer P, Koper PCM, Meijer G, Stroom JC, Van Herk M, Lebesque JV. Comparison of prostate cancer treatment in two institutions, a quality control study. Int J Radiat Oncol Biol Phys 1999; 45:

109 Rasch CRN, Barillot I, Remeijer P, Touw A, Van Herk M, Lebesque JV. Definition of the prostate in CT and MRI: a multi-observer study. Int J Radiat Oncol Biol Phys 1999; 43: Remeijer P, Rasch CRN, Lebesque JV, Van Herk M. A general methodology for three-dimensional analysis of variation in target volume delineation. Med Phys 1999; 26: Roelofsen F, Bartelink H. Combined modality treatment of anal carcinoma. Oncologist 1998; 3: Ruiter GA, Zerp SF, Bartelink H, Van Blitterswijk WJ, Verheij M. Alkyl-lysophospholipids activate the SAPK-JNK pathway and enhance radiation-induced apoptosis. Cancer Res 1999; 59: Theuws JCM, Kwa SLS, Wagenaar A, Seppenwoolde Y, Boersma LJ, Damen EMF, Muller SH, Baas P, Lebesque JV. Prediction of overall pulmonary function loss in relation to the 3-D dose distribution, for patients with breast cancer and malignant lymphoma. Radiother Oncol 1998; 49: Theuws JCM, Muller SH, Seppenwoolde Y, Kwa SLS, Boersma LJ, Hart G, Baas P, Lebesque JV. The effect of radiotherapy and chemotherapy on pulmonary function following treatment for breast cancer and lymphoma. J Clin Oncol 1999; 17: Van Tienhoven G, Voogd AC, Peterse JL, Nielsen M, Andersen KW, Mignolet F, Sylvester R, Fentiman IS, Van der Schueren E, Van Zijl K, Blichert-Toft M, Bartelink H, Van Dongen JA, on behalf of the EORTC Breast Cancer Cooperative Group and the Danish Breast Cancer Cooperative Group. Prognosis after treatment for loco-regional recurrence after mastectomy or breast conserving therapy in two randomised trials (EORTC and DBCG- 82TM). Eur J Cancer 1999; 35: Multidimensional assessment of voice characteristics after radiotherapy for early glottic cancer. Laryngoscope 1999; 109: Verdonck-De Leeuw IM, Keus RB, Hilgers FJM, Koopmans-Van Beinum FJ, Greven AJ, De Jong JM, Vreeburg G, Bartelink H. Consequences of voice impairment in daily life for patients following radiotherapy for early glottic cancer: voice quality, vocal function and vocal performance. Int J Radiat Oncol Biol Phys 1999; 44: Verheij M, Van Blitterswijk WJ, Bartelink H. Radiation-induced apoptosis: the ceramide-sapk signaling pathway and clinical aspects. Acta Oncol 1998; 37: Vrieling C, Collette L, Fourquet A, Hoogenraad W, Horiot JC, Jager JJ, Pierart M, Poortmans P, Struikmans H, Van der Hulst M, Van der Schueren E, Bartelink H. The influence of the boost in breast conserving therapy on cosmetic outcome in the EORTC boost versus no bost trial. Int J Radiat Oncol Biol Phys 1999; 45: Vrieling C, Collette L, Bartelink E, Borger J, Brenninkmeyer SJ, Horiot JC, Pierart M, Poortmans P, Struikmans H, Van der Schueren E, Van Dongen J, Van Limbergen E, Bartelink H. Validation of the methods of cosmetic assessment after breast conserving therapy in the EORTC boost versus no boost trial. Int J Radiat Oncol Biol Phys 1999; 45: Zelefsky MJ, Crean D, Mageras GS, Lyass O, Happersett L, Ling CC, Leibel SA, Fuks Z, Bull S, Kooy HM, Van Herk M, Kutcher GJ. Quantification and predictors of prostate position variability in 50 patients evaluated with multiple CT scans during conformal radiotherapy. Radiother Oncol 1999; 50: Full papers in press 109 RADIOTHERAPY Vander Poorten VLM, Balm AJM, Hilgers FJM, Bing Tan I, Loftus- Coll BM, Keus RB, Van Leeuwen FE, Hart AMM. The development of a prognostic score for patients with parotid carcinoma. Cancer 1999; 85: Bartelink H, Begg AC, Coco Martin J, Van Dijk M, Moonen L, Van t Veer L, Van der Vaart PJ, Verheij M. Translational research offers individually tailored treatments for cancer patients. Cancer J Sci Am 1999 (in press). Vander Poorten VLM, Balm AJM, Hilgers FJM, Bing Tan I, Loftus- Coll BM, Keus RB, Hart AMM. Prognostic factors for long term results of the treatment of patients with malignant submandibular gland tumors. Cancer 1999; 85: Venselaar JLM, Heukelom S, Jager HN, Mijnheer BJ, Van der Laarse R, Van Gasteren JJM, Van Kleffens HJ, Westermann CF. Effect of electron contamination on scatter correction factors for photon beam dosimetry. Med Phys 1999; 26: Venselaar JLM, Van Gasteren JJM, Heukelom S, Jager HN, Mijnheer BJ, Van der Laarse R, Van Kleffens HJ, Westermann CF. A consistent formalism for the application of phantom and collimator scatter factors. Phys Med Biol 1999; 44: Verdonck-De Leeuw IM, Hilgers FJM, Keus RB, Koopmans-Van Beinum FJ, Greven AJ, De Jong JMA, Vreeburg G, Bartelink H. Bartelink H, Horiot JC. Factors in decision making in the treatment of breast cancer [commentary]. Radiother Oncol 1999 (in press). Broeks A, Russell NS, Van t Veer L. Increased risk of breast cancer following irradiation for Hodgkin s disease is not a result of classical ATM germline mutations. Int J Radiat Oncol Biol Phys 1999 (in press). Meijer G, Bruinvis IAD, Mijnheer BJ, Lebesque JV. A treatment planning method to correct dose distributions distroted by setup verification fields. Int J Radiat Oncol Biol Phys 1999 (in press). Moonen L, Gallee M, Verheij M, Ong C, Hart G, Bartelink H. Apoptosis, proliferation and p53, cyclin D1 and retinoblastoma gene expression in relation to radiation response in muscleinvasive bladder cancer. Int J Radiat Oncol Biol Phys 1999 (in press).

110 110 RADIOTHERAPY Remeijer P, Geerlof E, Ploeger L, Gilhuijs KGA, Van Herk M, Lebesque JV. 3-D portal image analysis in clinical practice: an evaluation of 2-D and 3-D analysis techniques as applied to 30 prostate cancer patients. Int J Radiat Oncol Biol Phys 1999 (in press). Russell NS, Bartelink H. Radiotherapy: the last 25 years. Cancer Treat Rev 1999 (in press). Russell NS, Lara PC, Grummels A, Hart AMM, Coco Martin J, Bartelink H, Begg AC. In vitro differentiation characteristics of human skin fibroblasts: correlations with radiotherapy-induced breast fibrosis in patients. Int J Radiat Biol 1999 (in press). Van de Vaart PJ, Belderbos J, De Jong D, Sneeuw KCA, Major D, Bartelink H, Begg AC. DNA-adduct levels as a predictor of outcome for NSCLC patients receiving daily cisplatin and radiotherapy. Int J Cancer 1999 (in press). Van Leeuwen FE, Klokman WJ, Van t Veer MB, Hagenbeek A, Krol ADG, Vetter UAO, Schaapveld M, Van Heerde P, Burgers JMV, Somers R, Aleman BMP. Second cancer risk in young survivors of Hodgkin s disease: a 25 year follow-up study. J Clin Oncol 1999 (in press). Verheij M, Bartelink H. Radiation-induced apoptosis. Cell Tissue Res 1999 (in press). Local papers and theses Jansen EPM, Valdés Olmos R, Dewit L, Hamburger HL, Bartelink H, Hoefnagel CA. Residual pituitary adenoma visualized by functional 99mTc-HMPAO SPECT. Tijdschr Nucl Geneeskd 1999; 21: Koning C, Belderbos J, Van Zandwijk N. Radiotherapie. In: Demedts M, Dijkman J H, Hilvering C, Postma D S, redactie. Longziekten; Vol 1. Assen/Leuven: Van Gorcum/Universitaire Pers, 1999: Moonen L. Radiotherapy in bladder cancer [dissertation]. Amsterdam: Free University, Struikmans H, Van Tienhoven G, Jobsen JJ, Jager JJ, Borger J, Scheijmans LJEE. Loco-regionale radiotherapie van het mammacarcinoom na mastectomie en chemotherapie, commentaar bij twee gerandomiseerde trials. Ned Tijdschr Geneeskd 1998; 143: Theuws JCM. Pulmonary effects of radiotherapy; a clinical analysis [dissertation]. Amsterdam: Free University, Van de Vaart PJM. Interaction of platinum coordination compounds with radiation or hyperthermia: from laboratory to clinic [dissertation]. Amsterdam: Free University, 1999.

111 X Division of Medical Oncology 111 Division head Sjoerd Rodenhuis Head S Rodenhuis MD PhD Permanent academic staff JW Baars MD PhD, P Baas MD PhD, EM Bais MD, JH Beijnen PhD, W Boogerd MD PhD, H Boot MD PhD, PF Bruning MD PhD, GC De Gast MD PhD, SP Israëls MD, MJ Kersten MD PhD, H Neering MD PhD, JFM Pruijt MD (until September), JHM Schellens MD PhD, JH Schornagel MD PhD, BG Taal MD PhD, WW Ten Bokkel Huinink MD PhD, JJ Van der Sande MD PhD, M Van der Weide MD PhD, AMC Witte MD (from July), N Van Zandwijk MD PhD, APE Vielvoye-Kerkmeer MD PhD Other academic staff JP Baars MD, GM Bronner, F Custers MD, G De Klerk MD, B De Valk MD, JBAG Haanen, ALT Imholz MD PhD, CMF Kruijtzer MD, RAA Mathôt PhD, Y Oosterhoff MD PhD, J Schrama MD, J Wanders MD, C Weenink MD Permanent technical staff MJX Hillebrand, SE Jansen, FJ Koopman-Kroon, L Nan-Offeringa Other technical staff CGL Cleypool, M Ouwehand, M Profijt, M Tibben Graduate students ADR Huitema, B Nuyen, M Bouma, Th Kerbusch, H Bardelmeijer, NE Schoemaker, MM Malingré, DHJG Van den Bongard, Ch Van Kesteren Undergraduate students P Quadvlieg Research nurses D Batchelor, AC Dubbelman, M Holtkamp, GAM Linthorst, H Mallo, ME Schot, M Swart, W Uijterlinde Guest JA Burgers Secretaries Y Arts-Blom, MEG De Kwant Research funded positions (full time equivalents) Clinical permanent: 7.4 Clinical project: 10.9 Technical permanent: 9.9 Technical project: 4.0 Introduction In 1999, the clinical research program of the Division of Medical Oncology continued to concentrate on the development of innovative treatments and further strengthened its ties with the fundamental science groups in the institute. To increase the level of interaction between the laboratory and the clinic, several young specialists in internal medicine work for two years in basic research under supervision of a non-clinical scientist and than switch to the clinic for their formal training in Medical Oncology. Good clinical science is demanding in many respects and the excellent infrastructure of the institute is essential for its expansion. As a national cancer institute, we believe that interactions with research groups in universities can be beneficial to all parties and we have established close links with university centers inside and outside Amsterdam. In January 1999, JHM Schellens, the coordinator of our Clinical Pharmacology program, was appointed part-time Professor of Pharmacy at the University of Utrecht. This appointment is a further natural link in the Cancer Pharmacology network that has the Netherlands Cancer Institute at its core. A novelty for the clinical research program was the appointment of a Clinical Quality Assurance Officer to help in the implementation and execution of Good Clinical Practice (GCP) guidelines. The conduct of clinical studies is increasinly done under GCP and the Clinical Pharmacology Program is entirely done under GCP conditions. Clinical Pharmacology is the largest research group within the division. One major research topic is the the oral administration of taxanes, facilitated by the inhibition of P-glycoprotein, a drug transporter that is responsible

112 112 MEDICAL ONCOLOGY for the fact that orally administered taxanes have a very low bioavailability. The Pharmacology Group has been successful in increasing the bioavailabiltiy of both Paclitaxel and Docetaxel by prior administration of Cyclosporin A. These findings from phase I studies have led to a phase II study of oral Docetaxel in advanced breast cancer. This study is now accruing patients. The autologous transplantation group has completed patient accrual in the national study of high-dose chemotherapy in breast cancer patients with 4 or more tumorpositive axillary lymph nodes. A total of 885 patients have been randomized in this study, in 10 Dutch centers, making this the largest study internationally. It may well be that this trial will prove pivotal since it lacks a number of major problems that other large studies have: it is essentially a population-based study rather than a trial in highly selected patients, and there is hardly any toxic mortality. The first meaningful (progression-free) survival analysis can be done in 2002 and any advantage seen for the high-dose therapy must be regarded as uncertain until that time. The Thoracic Oncology Group presented the first analysis of a large chemo-prevention trial initiated in This EUROSCAN study involved nearly 2,000 patients at risk for a second primary tumor of the lung or in the head-and-neck area. The analysis showed no preventive effect at all for interventions with either vitamin A, or N-acetylcysteine, or both. The disappointing outcome of this very large and well-designed study is in line with several American chemo-prevention projects in aerodigestive tract cancers. Unsurprisingly, even in the new millennium, there is no substitute for smoking cessation. One of the unique features of the institute is the rapid and natural interaction between disciplines and divisions. Among its consequences is the establishment of a clinical unit for chemo-radiation therapy integrated in one of the two medical wards. The institute has played a prominent part in the relatively recent recognition that concurrent radiotherapy has curative potential in a large number of epithelial tumors, where the sequential application of these treatment modalities holds no special advantage. Examples include the treatment of stage III non-small cell lung cancer with concurrent cisplatin and radiation and, more recently, intra-arterial cisplatin and radiotherapy in advanced head-and neck cancers. The concurrent approach is demanding in terms of logistics and an optimal clinical setup is obviously required for ambitious clinical studies in a large number of tumor types, including cervical, anal, esophageal and urothelial cancers. This will be provided by the novel clinical unit. The immediate future for the clinical research program of the division will not dramatically differ from the activities in The growing emphasis on clinical studies that exploit the basic science progress will be a common theme, greatly facilitated by the presence of strong fundamental biology in the institute. It is satisfying that at last the understanding of disease at a molecular level is beginning to benefit individual patients with cancer. Clinical Pharmacology P Baas, EM Bais, OB Dalesio, G De Klerk, AC Dubbelman, ADR Huitema 1, Th Kerbusch, CMF Kruijtzer, J Lieverst 2, JM Maaskant, M Mahn 2, M Maliepaard 3, 4, 5, MM Malingré, H Mallo, R Mathôt, I Mandjes 2, D Pluim 3, S Rodenhuis, NE Schoemaker, JH Schornagel, M Schot, A Schrijvers 6, M Swart, Ch Van Kesteren, RCAM Van Waardenburg 7, N Van Zandwijk, JH Beijnen, JHM Schellens, WW Ten Bokkel Huinink Phase I studies and drug scheduling In 1999, the number of active studies has increased even beyond that of At present, 20 phase I and pharmacologic studies are being conducted, of which 15 are open for accrual. The patient recruitment in these studies totals over 170 in One of the major translational research activities continued to be the development of oral treatment schedules for paclitaxel and docetaxel in combination with oral cyclosporin A (CsA). The bioavailability of docetaxel after oral administration plus one single oral dose of CsA was 88 ± 36% (Figure X.1 and X.2). Importantly, the interpatient variability in systemic exposure after oral and i.v. administration was almost identical. The safety of the oral route was excellent. To date, a phase II study of weekly oral docetaxel plus CsA in second line advanced breast cancer has recruited 7 patients. All three patients currently evaluable for response had achieved at least a partial remission. A P-glycoprotein blocker GF was combined with oral Taxol in 6 patients and the results reveal that the effect of this blocker on the systemic exposure to paclitaxel is at least as pronounced as that of CsA. Figure X.1 Plasma concentration-time curve of docetaxel administered orally at a dose of 75 mg/m 2 in combination with an oral dose of 15 mg/kg of the P-glycoprotein blocker cyclosporin A (CsA) and at a dose of 100 mg/m 2 intravenously to the same patient.

113 Figure X.2 Mean (SD) dose-corrected area under the curve (AUC) values of docetaxel when administered intravenously, orally in combination with CsA (N=10) and orally without CsA (N=3). Another phase I study was initiated with a novel farnesyltransferase inhibitor R The maximum tolerated dose (MTD) was 300 mg twice daily without interruption. The dose limiting toxicity (DLT) was neutropenia. Two out of 30 patients developed significant primarily sensory neuropathy after 2 months of treatment. A partial remission was documented in platinum resistant stage IV non-small cell lung cancer (NSCLC) and activity was also noted in colon and pancreatic cancer. The area under the curve (AUC) of R proved to be significantly correlated with myelosuppression. Based on the results of our previous phase I study with continuous i.v. CPT11, we are currently performing a phase I and pharmacologic study with oral CPT11 daily times 14. The regimen is well tolerated and the MTD has not been reached at 20 mg/m 2 /day. The oral miltefosine analog D21266 (perifosine) was investigated in a dose-escalating study to determine the MTD, which was reached at 200 mg on a 21 day q 28 day schedule. Gastro-intestinal toxicity was dose-limiting, in particular nausea and vomiting. The randomized phase I study with weekly or twoweekly gemcitabine and cisplatin in patients with NSCLC is ongoing. To date, 35 patients have been recruited. The current dose level in the weekly schedule is 55 mg/m 2 of cisplatin plus 900 mg/m 2 of gemcitabine. The doseintensity is already over 50% higher than in the standard schedule of this combination. Activity has been noted in several patients. Pharmacokinetic and dynamic analysis revealed that the total plasma clearance of unbound platinum as well as the DNA-adduct formation in leukocytes is reduced by gemcitabine. E7070 is a novel disulphonamide anticancer agent interfering with the G1-S cell cycle transition, which is investigated in phase I trials within the framework of the Early Clinical Studies Group (ECSG). The schedule is a one hour i.v. infusion q 3 weeks. The dose-limiting toxicity is neutropenia. The pharmacokinetics are profoundly nonlinear. The AUC of E7070 is significantly correlated with neutropenia. Currently, a mass balance study with 14C-labeled E7070 is in progress. The food-drug interaction study with the 5FU oral prodrug S1 revealed that food significantly affected the pharmacokinetics of S1 and reduced the AUC of 5FU by 16%. The combination of carboplatin and topotecan is being investigated in a phase I and pharmacologic study. In addition to bioanalysis of topotecan and carboplatin, platinum-dna adduct levels in leukocytes and tumor cells as well as catalytic activity and expression of topoisomerase I are determined in leukocytes and tumor biopsy material collected during the first course of treatment. Biopsies have been collected in all 5 patients. Another phase I study with a polymer MAG-camptothecin formulation is actively recruiting patients at the daily times 3, q 4 weeks schedule at a dose level of 68 mg/m 2 /day. A phase I study with a novel organic ruthenium compound NAMI-A was recently initiated. The schedule is daily times 5 q 3 weeks. In pharmacologic studies, concentrations of ruthenium in plasma and DNA-adducts in leukocytes are being monitored. Phase II and pharmacologic studies Within the framework of the ECSG, the activity of UFT (an oral prodrug of 5FU) plus leucovorin was determined as the first chemotherapy for advanced gastric cancer. The drug was well tolerated. One out of 8 patients remained stable for 2.5 years while on study. In another ECSG study, the activity of S1 was investigated in the first line treatment of advanced colorectal and gastric cancer. Clinical phase II and III studies within the framework of the European Biomed 2 program are ongoing. The aim is to determine pharmacokinetic/tumor biologic-response relationships and pharmacokinetic-toxicity relationships with the combination of doxorubicine-cyclophosphamide in the first line chemotherapy of advanced breast cancer, and with the carboplatin-taxol regimen in the first line chemotherapy of advanced ovarian cancer. To date, 21 patients with breast cancer and 12 patients with ovarian cancer were included in our institute. Population pharmacokinetics (PK) and dynamics (PD) Population PK and PD is a subject of increasing importance within the pharmacology group. The Nonlinear Mixed Effects Models software program NONMEM is extensively used. NONMEM is employed to support the outlined Biomed 2 program and NONMEM is also used to explore the population PK and PD of ET743. This is a novel anticancer agent of marine source under development in sarcoma. Notes 1 Funding: Dutch Cancer Society, Project NKI Biometrics. 3 Division VIII. 4 Funding: Dutch Cancer Society, Project NKI MEDICAL ONCOLOGY

114 114 MEDICAL ONCOLOGY 5 Funding: Dutch Cancer Society, Project NKI Funding: Biomed 2, Project PL Funding: Dutch Cancer Society, Project NKI Pharmacy H Bardelmeijer 1, I Bedeker 2, M Bouma 2, CGJ Cleypool 3, RMW Hoetelmans 2, MJX Hillebrand, ADR Huitema 4, SE Jansen 2, Th Kerbusch, FJ Koopman-Kroon 2, MM Malingré, RAA Mathôt, L Nan-Offeringa 2, B Nuyen 2, M Ouwehand 1, M Profijt 2, H Rosing 2, NE Schoemaker, RW Sparidans 2, WW Ten Bokkel Huinink, MM Tibben 4, DHJG Van den Bongard, R Van der Put 2, R Van Gijn 2, RPG Van Heeswijk 2, Ch Van Kesteren, O Van Tellingen 5, AI Veldkamp 2, JH Beijnen, JHM Schellens The research programs of the Department of Pharmacy and Pharmacology are focused on pharmaceutical formulation and drug level monitoring of (investigational) anticancer drugs to support preclinical (see Division VIII) and clinical pharmacologic research. The investigations are conducted in the setting of a foundation (NLADF), which is a collaboration between the Netherlands Cancer Institute and Slotervaart Hospital was a very important year for the department. The ministry of Health, Welfare and Sport has granted us the GMP (Good Manufacturing Practice) license which is the official qualification required to manufacture investigational cytotoxic drugs for clinical trials, both on site and elsewhere. On June , the laboratory received its official recognition in compliance with the OECD Principles of Good Laboratory Practice (GLP) in the field of Analytical Chemistry. Formulation Current formulation research is mainly focused on cytotoxic agents originating from marine organisms and heavy metals (platinum, ruthenium). The following compounds are currently investigated to find a suitable parenteral formulation to be used in clinical phase I research: thiocoraline, spisulosine, isohomohallichondrin- B, kahalalide F and P-65. The formulation of aplidine (freeze-dried) was successfully transferred to the pharmaceutical industry with no problems of upscaling. We have designed and manufactured a freeze-dried product of the ruthenium-containing antitumor agent encoded NAMI-A. The drug is currently tested in a phase I trial in our institute. A HPMA-polymer bound platinum compound is now in its pre-formulation stage. Batches of freeze-dried products (MEN10755, BBR2778) were manufactured for clinical phase I testing. In a collaborative network with the EORTC, NCI and CRC we perform pre-formulation and quality control studies of novel chemical entities with in vitro anti-tumor activity to be selected for further development. An oral tablet formulation for the tyrosine kinase inhibitor BIBX1382BS has been designed and batches have been manufactured to support ECSG phase I clinical trials. Further optimization of our validation programs of the Formulation Laboratory to maintain the GMP status is a continuous process which significantly impacts on the work load. Bioanalysis and clinical pharmacology Bioanalytic and clinical pharmacologic research comprise the following compounds: topoisomerase-i inhibitors, taxoids, platinum compounds, the drugs in the CTC high dose regimen (carboplatin-thiotepa-cyclophosphamide), ifosfamide, perifosine and novel cytotoxics (ET-743, aplidine, thiocoraline, KRN 7000) originating from marine sources. We have developed HPLC assays to measure camptothecin in plasma after administration of polymerbound drug (MAG-CPT) to support an ongoing phase I trial. A large program has been set up to track down metabolites of ecteinascidin-743 (ET-743) by using coupled liquid chromatography and mass spectrometry. HPLC assays have been validated and implemented to measure 4-hydroxy- and nitrogen mustard metabolites of cyclophosphamide and ifosfamide. Close pharmacokinetic monitoring in the CTC-program has revealed that thiotepa inhibits the activation route of cyclophosphamide. The clinical relevance of this observation, however, is not clear. A new alkylating metabolite of thiotepa, thiotepa-mercapturate has been identified in the urine of treated patients. Pharmacokinetic individualization appear feasible in the CTC-scheme. A pharmacokinetic, population-based model has been designed to describe the ifosfamide auto-induction phenomenon. The triple quadrupole, tandem mass spectrometer (LC/MS/MS; Sciex API 365) has greatly expanded our Figure X.3 Mass spectrum of spisulosine (ES-285), a novel anti-tumor drug found in a North Atlantic clam, with selectivity against prostate and renal cancer and hepatomas. The compound is now in the developmental, pre-formulation phase. Mass spectrum: m/z MH+, m/z MH+ (-H2O), m/z (2 x M)H+ (dimer).

115 analytical capabilities (Figure X.3). We are now setting up LC/MS/MS assays for E7070, ET-743 metabolites and kahalalide-f to support clinical pharmacologic research. The equipment is also being used for quality control purposes of peptides synthesized in the institute. Notes 1 Funding: Dutch Cancer Society, Project NKI NLADF/Slotervaart Hospital. 3 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Division XIII. Immunotherapy D Batchelor, JBAG Haanen, SP Israëls, MJ Kersten, H Mallo, WJ Nooijen 1, JH Schornagel, J Sein 2, W Van de Kasteele 2, N Verra 1, FA Vyth-Dreese 2, GC De Gast Metastatic melanoma A phase II single center study was performed of chemo-immunotherapy consisting of temozolomide (150 mg/m 2 for 5 days in 6 patients), escalated to 200 mg/m 2 in 18 patients, and subsequently to 250 mg/m 2 in 18 patients), followed by combined immunotherapy of GM- CSF, low dose IL-2 and IFNa. In one year, 42 patients were included. Preliminary results show response, stable disease and progressive disease, each in approximately one third of the patient population. Of interest is the absence of brain metastases in patients with response or stable disease and the absence of lymphocytopenia, a well known toxicity of temozolomide. Half of the patients with 250 mg/m 2 temozolomide showed thrombocytopenia as dose limiting toxicity and some also had granulocytopenia. Patients with residual metastases underwent surgical removal (two times lobectomy, two times adrenalectomy) with normalization of S100 afterwards. Monitoring with serum S100b levels showed a good correlation with the clinical response. In patients with progression after chemo-immunotherapy, a phase I trial of increasing doses of temozolomide with standard cisplatin (75 mg/m 2 ) followed by the same immunotherapy was performed. Stabilization of the disease was achieved in the majority of patients. Renal cell carcinoma (RCC) A phase II multicenter study of combined non-specific immunotherapy was started in which Maastricht, Utrecht and Nijmegen participate. At this time, 30 patients are evaluable with 5 x CR, 3 x PR (27%), 11 x SD, 11 x PD. Vaccination with a G250 peptide will start in HLA-A2+ patients with progression after combined non-specific immunotherapy. Perioperative immunomonitoring In HLA-A2+ patients undergoing a lymphadenectomy for melanoma, the number of cytotoxic T lymphocytes (CTLs) with a T-cell receptor for melanocyte differentiation antigens was determined by tetramer technology. In 2 patients Mart- 1 specific CTLs were found in the CD8 population ( %) from the lymph nodes, that were not present in the blood. In patients with a nephrectomy after 2 cycles of treatment with combined immunotherapy for RCC, there was an increase of dendritic cells of several types and maturation stages, as well as T cells in the tumor. Lymphocyte reinfusion after high dose chemotherapy in metastasized breast cancer Patients with breast cancer receiving multiple courses of high-dose chemotherapy (tctc) with peripheral blood progenitor cell transplantation, continue to be immunosuppressed for prolonged periods after treatment. This leads to an increased incidence of infections, such as Herpes zoster, and is an obstacle for the development of immunotherapy-based treatments to eradicate minimal residual disease. To deal with this problem, the feasibility and efficacy of lymphocyte-reinfusions after second and third high-dose courses was explored. Three patients who had not received a lymphocyte reinfusion showed a profound lymphocytopenia of CD4+ T cells (< 100/µL) and CD8+ T cells (< 200/µL) for more than 6 months. In 9 patients in whom lymphocytes ( 2 x ) were harvested after standard chemotherapy, lymphocytes were reinfused after the second tctc course, which was followed by G-CSF, and after the third tctc course, which was followed by both GM-CSF and IL-2. Lymphocyte recovery after re-infusion was significantly better than after the first high dose course or than in the first 3 patients, with restoration of CD4 and CD8 T- cell levels to pre-treatment levels after 1 month. After the third tctc course, CD4+ T cell recovery was better. There was, however, no apparent decrease in Herpes zoster reactivations. Notes 1 Division XIII. 2 Division IV. Autologous hematopoietic progenitor cell transplantation program JW Baars, JH Beijnen, GC De Gast, O Dalesio, M Holtkamp, ARD Huitema, MJ Kersten, G Linthorst, WJ Nooijen 1, JFM Pruijt, JH Schornagel, J Schrama, ICM Slaper-Cortenbach 2, AM Westermann, S Rodenhuis Multiple courses of tiny -CTC In the past few years, about sixty patients have been treated in a series of clinical studies aiming to develop a high dose alkylating chemotherapy regimen that can be administered safely within in a tight time frame. These studies have led to the tiny -CTC regimen, which incor- 115 MEDICAL ONCOLOGY

116 116 MEDICAL ONCOLOGY porates cyclophosphamide (4 g/m 2 ), thiotepa (320 mg/m 2 ) and carboplatin (target AUC 13.3 mg.ml -1.min). Three subsequent courses of tctc can be administered every four to five weeks provided that autologous peripheral blood progenitor cells can be reinfused following each course. In 1999, a first relapse-free survival analysis could be done of a phase II study 3 in which patients with disseminated hormone refractory breast cancer received intensive chemotherapy (Figure X.4). A total of forty-one patients were scheduled to receive two courses of fluorouracil, epirubicin and cyclophosphamide. Responding patients went on to three subsequent courses of tctc. Twenty-nine patients either showed evidence of tumor regression as a result of FEC, or were not evaluable. For this subgroup, the projected four year disease-free survival was 30% (confidence interval 14-49%). This finding confirms that long-term disease-free survival is achievable for a subgroup of chemotherapyresponsive patients with breast cancer. The large majority of patients had achieved a complete remission at the end of the treatment sequence. The fact that most of them relapsed within the first year illustrates that a treatment strategy is required to deal with minimal residual disease. For this purpose, a feasibility study of post-transplantation immunotherapy was completed in In this study, patients received an autologous in vivo activated lymphocyte re-infusion, in addition to the stem cell transplantation. Although this led to rapid restoration of T-cell subsets, most patients developed herpes zoster infections in the first months after transplantion. Thus, it is questionable whether the enhanced immunoreconstitution has clinical significance. Alternative methods to manage the minimal residual disease situation after high-dose chemotherapy are being considered, including maintenance low-dose chemotherapy or anti-angiogenesis agents. The repeated high-dose alkylating chemotherapy is also the subject of detailed pharmacodynamic studies 4. In 1999, analytic methods to determine the plasma levels of cyclophosphamide and its many metabolites were developed and validated, as were methods for thiotepa and tepa. These methods and the assays for carboplatin developed earlier were employed to adjust the dose of the four day tctc-courses based on pharmacokinetic parameters determined on the first day of the course. The efficacy of this strategy to decrease end-organ toxicity is now being studied. High-dose alkylating therapy is potentially curative in germ cell cancer, as well as breast cancer. Patients with chemotherapy refractory testicular cancer do not usually benefit from high-dose therapy and are therefore excluded from the national high-dose study. Two of these patients received three closely spaced courses of tctc as salvage therapy. Both achieved a complete remission. One of them has remained free of disease and is possibly cured. A phase II study of triple tiny-ctc in refractory germ cell cancer is being considered. Figure X.4 Progression-free survival of 29 patients who responded to two conventional doses of chemotherapy, and subsequently underwent multiple courses of high-dose chemotherapy. All had stage IV breast cancer and were hormone-unresponsive. Multicenter high-dose chemotherapy studies coordinated by NKI/AvL By August 1, 1999 the national study of adjuvant highdose chemotherapy in high-risk breast cancer was closed 5. This trial, designed in our institute and coordinated in collaboration with Groningen University, has randomized a total of 885 patients and is thus the largest study of its kind. The relapse- free survival advantage seen for the patients who were randomized to the highdose chemotherapy arm could be the result of a statistical fluctuation and should thus not be considered as a rationale for high-dose therapy outside clinical studies. A second national study devised and led by our institute is the high-dose chemotherapy study for patients with germ cell cancer relapsing from a complete remission. A preliminary analysis was performed and published in 1999, showing that about 50% of these patients can be cured by high-dose therapy. Notes 1 Division XIII. 2 Central Laboratory of the Netherlands Red Cross for the Blood Transfusion Service, Amsterdam. 3 Supported in part by the Schumacher-Kramer Foundation. 4 Funding: Dutch Cancer Society, Project NKI Funding: Health Insurance Council Fund for Investigative Medicine (Ziekenfondsraad) Ontwikkelingsgeneeskunde)

117 Gastroenterology BMP Aleman 1, D De Jong 2, CA Hoefnagel 2, P Qaedvlieg, RA Valdés Olmos 2, F Van Coevorden 3, L Van t Veer 2, MLF Veldhuysen 2, AMC Witte, FAN Zoetmulder 3, H Boot, BG Taal Gastric non-hodgkin s lymphoma Treatment decisions are mainly based on histological grading and clinical staging (including endoscopic ultrasound). Helibacter pylori eradication is generally accepted as a non-toxic and effective treatment in stage I low grade MALT lymphoma. In a series of 22 patients, a complete remission was achieved in 13 (59%) and partial response in 2 (9%). Further histological grading was related to response: complete remission was observed in 12 of 15 purely low-grade tumors (type A), and in only 1 of 7 lymphomas with some blasts (type B) (p < 0.01). The expression of co-stimulatory markers was also predictive: two or all of the three markers CD40, CD80 and CD86 were expressed in 12 of 13 type A lymphomas, compared to 1 of 6 type B lymphomas (p < 0.001). Carcinoid tumors Carcinoid tumors are rare and account for approximately 300 new cases per year in the Netherlands. The prognosis is relatively good: the 5 year survival for locoregional disease is 53-80% and for metastatic disease 18%. Whether adequate palliative treatment has resulted in a survival benefit, will be studied in cooperation with the Cancer Registry (IKA). A tracer dose of [131I]MIBG is frequently employed for imaging of neuroendocrine tumors, and results in a positive test in 70% of the carcinoid patients. Radiolabeled MIBG treatment was shown to be effective and resulted in (long-lasting) reduction of symptoms, but it was not associated with a significant decrease in urinary 5HIAA excretion (biochemical response). Improved tumor targeting by predosing with unlabeled MIBG could be achieved in 6 of 10 cases; the subsequently applied combined treatment dose induced symptomatic improvement in 5 patients and a biochemical response in 3 of them. These encouraging results have been confirmed in animal models and will be built on in further studies. Hereditary cancer Among the familial cancers, HNPCC (hereditary nonpolyposis coli cancer) has received a great deal of attention: DNA tests are available and guidelines for surveillance have been developed. Less information is available regarding a genetic predisposition for gastric cancer. Using the same Amsterdam criteria that were developed for HNPCC, the number of families with gastric cancer is increasing. Currently the clinical data of 8 kindreds has been collected. The histological subtype (diffuse or intestinal type) may be correlated with DNA tests. Gastric metastases In 51 breast cancer patients, an endoscopic diagnosis of gastric involvement was made. Histology showed a predominance (72%) of the relatively infrequent subtype of lobular carcinoma. Symptoms were non-specific; 12% of patients presented with hemorrhage. All patients also had other sites of metastatic disease, particularly bone metastases (43%). Although gastric involvement is a late manifestation of metastatic disease, its response to chemotherapy or endocrine treatment was fairly good with an objective response rate of 46%. The remaining life-span was 10 months (median) and the 2 year survival 23%. Esophageal cancer In advanced disease with a short life expectancy several palliative treatment options have been explored (argon beam coagulation, alcohol injection and laser) among which the insertion of a stent is most frequently applied. Since the introduction of the self expandable stents, various modified types have become available to individualize the choice of stent per patient. The overall success rate is 90%, leading to immediate improvement of dysphagia. In order to improve the treatment results in esophageal cancer, combined modality approaches have been developed. In locally advanced disease, chemo-radiation seems to be superior to radiation therapy alone: currently a phase I study is in progress that investigates increasing doses of radiotherapy and concurrent low-dose cisplatinum. In patients with a local tumor amenable for surgery, pre-operative chemo-radiation has been applied in a pilot study of 8 patients. Full dose cisplatinum and 5FU at the beginning and end of the radiation scheme was employed. This concurrent chemo-radiation scheme has induced several histologically confirmed complete remissions and is tolerable in fit patients; long-term results have to be awaited. Notes 1 Division IX. 2 Division XIII. 3 Division XI. Thoracic Oncology AJM Balm 1, JSA Belderbos 2, F Custers, OB Dalesio, Y Oosterhoff, JHM Schellens, FA Stewart 3, FJ Van Schooten 4, LJ Van t Veer 5, H Van Tinteren, C Weenink, P Baas, N Van Zandwijk Non-Small Cell Lung Cancer (NSCLC) Locally advanced NSCLC stage III disease involves a subset of patients without distant metastases. For many years, radiotherapy has been accepted as the standard of care. A revision of the therapeutic strategies in such patients came about in the early nineties when better 117 MEDICAL ONCOLOGY

118 118 MEDICAL ONCOLOGY results were obtained with cisplatin added to radiotherapy as a radiosensitizer. The recent favorable results (EORTC 08955) of the gemcitabine/cisplatin (G/C) combination as a pre-operative and pre-radiotherapy regimen suggest that further improvement is possible. With an overall response rate of 84%, leading to complete surgical resection in 71%, this regimen may well be considered as a standard in combined modality treatment. Although it is true that the population of patients who received G/C as induction treatment was small and selected, we conclude that surgery alone is no longer an adequate standard for stage IIIA patients. The question of whether induction therapy should be followed by surgery or radiotherapy is still under investigation in EORTC For medically inoperable patients, a comparison between concurrent and sequential chemoradiotherapy is being made in EORTC This trial was activated in the beginning of 1999 and has now been adopted by several centers in the Benelux. The use of incremental doses of cisplatin (day 1) and gemcitabine (day 2) on a weekly or a two-weekly basis in mainly stage IV NSCLC patients has revealed that this intensive dose therapy is reasonably well-tolerated. Several objective responses have been documented after only 6 weeks of intensive therapy and these results encourage the further exploration of intensified therapy in patients with early NSCLC. Chemoprevention and early detection Euroscan 6 is a large scale clinical trial to assess the chemopreventive effects of Vitamin A and N-acetylcysteine in patients treated with curative intent for lung and head and neck cancer. Disappointingly, the first full analysis of the results has not revealed any preventive effect of a 2-year intervention period with Vitamin A or N- acetylcysteine after a median follow-up of 49 months. Taking into account the latency period of lung cancer, assumed to be 10 years or more, additional follow-up is needed before definite conclusions may be drawn. In essence, Euroscan has yielded a similar outcome as several other large chemoprevention studies that were recently published. Another project focused on premalignant epithelial changes in smoking volunteers and revealed that p53 mutations detected by FASAY may already be present when white light bronchoscopy appears normal. To explore new methods for early detection, an endoscopy program studying autofluorescence and ALA (aminolevulinic acid) induced fluorescence in patients with elevated cancer risk (Euroscan population) was initiated. Malignant Mesothelioma 7 The phase II study of intraoperative photodynamic therapy and pleuropneumonectomy in patients with limited mesothelioma and favorable prognostic factors was completed. The following conclusions can be drawn from the study: 1) This combined treatment is frequently accompanied by a prolonged recovery period and requires intensive postoperative care to prevent complications which occur at a higher rate than seen after standard pneumonectomy. 2) Although efficient local control could be obtained, some patients were confronted with isolated distant metastases. It is possible that in the past metastatic disease has been obscured by progression of locoregional disease. Another novel technique combining pleurectomy with intraoperative hyperthermic chemotherapy has been introduced in the clinic, as described in the report of Division XI. This treatment approach was explored in 4 patients and was shown to be feasible. Additional experience with this technique will be gained in Phase II studies employing new agents or combinations of agents continued. Objective responses have been recorded with the gemcitabine/cisplatin combination. Notes 1 Division XI. 2 Division IX. 3 Division VIII. 4 University of Limburg, Maastricht. 5 Division VIII and XIII. 6 Funding: European Communities (European Commission DG V). 7 Funding: Dutch Cancer Society, Project NKI Publications: Division X Full papers Berger C, Van Baarle D, Kersten MJ, Klein MR, Samer Al-Homsi A, McQuain C, Van Oers MHJ, Knecht H. Carboxy terminal variants of Epstein-Barr virus encoded LMP1 during longterm HIV-infection: reliable markers for individual strain identification. J Infect Dis 1999; 179: Beijnen JH, Meenhorst PL. Follow-up of ectasy intoxication [letter]. J Toxicol & Clin Toxicol 1999; 37: 343. Boven E, Jansen WJM, Hulscher TM, Beijnen JH, Van Tellingen O. The influence of P170-glycoprotein modulators on the efficacy and the distribution of vincristine as well as MDR1 expression in BRO/mdr1.1 human melanoma xenografts. Eur J Cancer 1999; 35: Cohen Stuart JW, Schuurman R, Burger DM, Koopmans PP, Sprenger HG, Juttmann JR, Richter C, Meenhorst PL, Hoetelmans RMW, Kroon FP, Bravenboer B, Hamann D, Boucher CA, Borleffs JC. Randomized trial comparing saquinavir soft felatin capsules versus indinavir as part of triple therapy (CHEESE study). AIDS 1999; 13: F53-8. Crommentuyn KML, Schellens JHM, Van den Berg JD, Beijnen JH. In-vitro metabolism of anti-cancer drugs, methods and applications: paclitaxel, docetaxel, tamoxifen and ifosfamide. Cancer Treat Rev 1998; 24:

119 De Graaff M, Maliepaard M, Pluim D, Floot BJ, Slaper-Cortenbach IC, Schellens JHM. In vitro antagonistic cytotoxic interactions between platinum drugs and taxanes on bone marrow progenitor cell CFU-GM. Anticancer Drugs 1999; 10: De Vries N, Van Zandwijk N, Pastorino U. Chemoprevention of head and neck and lung (pre)cancer. Recent Results Cancer Res 1999; 151: De Wit R, Collette L, Sylvester R, De Mulder PH, Sleijfer DT, Ten Bokkel Huinink WW, Kaye SB, Van Oosterom AT, Boven E, Stoter G. Serum alpha-fetoprotein surge after the initiation of chemotherapy for non-seminomatous testicular cancer has an adverse prognostic significance. Br J Cancer 1998; 78: De Wit R, Van Dam FSAM, Vielvoye-Kerkmeer A, Mattern C, Huijer Abu-Saad H. The treatment of chronic cancer pain in a cancer hospital in the Netherlands. J Pain Symptom Manage 1999; 17: Enting RH, Hoetelmans RMW, Lange JMA, Burger DM, Beijnen JH, Portegies P. Antiretroviral drugs and the central nervous system. AIDS 1998; 12: Gerrits CJ, Schellens JHM, Burris H, Eckardt JR, Planting AS, Van der Burg ME, Rodriguez GI, Loos WJ, Van Beurden V, Hudson I, Von Hoff DD, Verweij J. A comparison of clinical pharmacodynamics of different administration schedules of oral topotecan. Clin Cancer Res 1999; 5: Giaccone GC, Baas P. Mesotheliomas. In: Raghaven D, Brecher ML, Johnson DH, Meropol NJ, Moots PL, Thigpen JT, editors. Textbook of uncommon cancer; second edition. West Sussex: John Wiley, 1999: Herben VMM, Schellens JHM, Swart M, Gruia G, Vernillet L, Beijnen JH, Ten Bokkel Huinink WW. Phase I and pharmacokinetic study of irinotecan administered as a low-dose, continuous intravenous infusion over 14 days in patients with malignant solid tumors. J Clin Oncol 1999; 17: Herben VMM, Van Gijn R, Schellens JHM, Schot M, Lieverst J, Hillebrand MJX, Schoemaker NE, Porro MG, Beijnen JH, Ten Bokkel Huinink WW. Phase I and pharmacokinetic study of a daily times 5 short intravenous infusion schedule of 9-aminocamptothecin in a colloidal dispersion formulation in patients with advanced solid tumors. J Clin Oncol 1999; 17: Hoetelmans RMW, Profijt M, Meenhorst PL, Mulder JW, Van Heeswijk RPG, Beijnen JH. Co-trimoxazole and stavudine interference in a high-performance liquid chromatographic analysis for didanosine in human plasma. Ther Drug Monitor 1998; 20: Hoetelmans RMW, Van Heeswijk RPG, Profijt M, Mulder JW, Meenhorst PL, Lange JMA, Reiss P, Beijnen JH. Comparison of the plasma pharmacokinetics and renal clearance of didanosine in once and twice daily dosing in HIV-1 infected individuals. AIDS 1998; 12: F Huitema ADR, Holtkamp M, Rodenhuis S, Beijnen JH. Sampling technique from central venous catheters proves critical for pharmacokinetic studies. Ther Drug Monitor 1999; 21: Jansen G, Mauritz R, Drori S, Sprecher H, Kathmann I, Bunni M, Priest DG, Noordhuis P, Schornagel JH, Pinedo HM, Peters GJ, Assaraf YG. A structurally altered human reduced folate carrier with increased folic acid transport mediates a novel mechanism of antifolate resistance. J Biol Chem 1998; 273: MEDICAL ONCOLOGY Godschalk RW, Maas LM, Van Zandwijk N, Van t Veer LJ, Breedijk A, Borm PJ, Verhaert J, Kleinjans JC, Van Schooten FJ. Differences in aromatic-dna adduct levels between alveolar macrophages and subpopulations of white blood cells from smokers. Carcinogenesis 1998;19: Kroep JR, Giaccone G, Voorn DA, Smit EF, Beijnen JH, Rosing H, Van Moorsel CJA, Van Groeningen CJ, Postmus PE, Pinedo HM, Peters GJ. Gemcitabine and paclitaxel: Pharmacokinetic and pharmacodynamic interactions in patients with non-small-cell lung cancer. J Clin Oncol 1999; 17: Groenewegen G, De Gast GC. GM-CSF can cause T-cell activation; Results of sequential chemo-immunotherapy. Eur J Cancer 1999; 35: Herben VMM, Nannan Panday VR, Richel DJ, Schellens JHM, Van der Vange N, Rosing H, Beusenberg FD, Hearn S, Doyle E, Beijnen JH, Ten Bokkel Huinink WW. Phase I and pharmacologic study of the combination of paclitaxel, cisplatin, and topotecan administered intravenously every 21 days as first-line therapy in patients with advanced ovarian cancer. J Clin Oncol 1999; 17: Herben VMM, Rosing H, Ten Bokkel Huinink WW, Van Zomeren DM, Batchelor D, Doyle E, Beusenberg FD, Beijnen JH, Schellens JHM. Oral topotecan: bioavailability and effect of food co-administration. Br J Cancer 1999; 80: Kuin A, Aalders M, Lamfers M, Van Zuidam DJ, Essers M, Beijnen JH, Smets LA. Potentiation of anti-cancer drugs at low intratumoral ph induced by the the mitochondrial inhibitor m-iodobenzylguanidine (MIBG) and its analog benzylguanidine (BG). Br J Cancer 1999; 79: Kuin A, Rutgers M, Van der Valk MA, Beijnen JH, Smets LA. Bioavailability and toxicity of oral administration of m-iodobenzylguanidine (MIBG). Br J Cancer 1999; 79: Lambrechts AC, Bosma AJ, Klaver SG, Top B, Perebolte L, Van t Veer LJ, Rodenhuis S. Comparison of immunocytochemistry, reverse transcriptase polymerase chain reaction and nucleic acid sequence-based amplification for the detection of circulating breast cancer cells. Breast Cancer Res Treat 1999; 56:

120 120 MEDICAL ONCOLOGY Maliepaard M, Van Gastelen MA, De Jong LA, Pluim D, Van Waardenburg RC, Ruevekamp-Helmers MC, Floot BG, Schellens JHM. Overexpression of the BCRP/MXR/ABCP gene in a topotecan-selected ovarian tumor cell line. Cancer Res 1999; 59: Meerum Terwogt JM, Mandjes IM, Sindermann H, Beijnen JH, Ten Bokkel Huinink WW. Phase II trial of topically applied miltefosine solution in patients with skin-metastasized breast cancer. Br J Cancer 1999; 79: Meerum Terwogt JM, Schellens JHM, Ten Bokkel Huinink WW, Beijnen JH. Clinical pharmacology of anticancer agents in relation to formulations and administration routes. Cancer Treat Rev 1999; 25: Nannan Panday VR, De Wit R, Schornagel JH, Schot M, Rosing H, Lieverst J, Ten Bokkel Huinink WW, Schellens JHM, Beijnen JH. Pharmacokinetics of paclitaxel administered in combination with cisplatin, etoposide and bleomycin in patients with advanced solid tumors. Cancer Chemother Pharmacol 1999; 44: Post J, Vooijs WC, Bast BJEG, De Gast GC. Efficacy of an anti- CD138 immunotoxin and doxorubicin on drug-resistant and drugsensitive myeloma cells. Int J Cancer 1999; 83: Ravaud A, Borner M, Schellens JHM, Geoffrois L, Schoffski P, Wanders J, Hanauske AR. UFT and oral calcium folinate as firstline chemotherapy for metastatic gastric cancer. Oncology (Huntington NY) 1999; 13: Reiser M, Salzberger B, Stiepel A, Ivette A, Hoetelmans RMW, Fätkenheuer G. Virological efficacy and plasma drug concentrations of nelfinavir plus saquinavir as salvage therapy in HIV-infected patients refractory to standard triple therapy. Eur J Med Res 1999; 4: Reubsaet JLE, Beijnen JH, Belshof EHM, Bouyakhrichan M, Bult A, Hop E, Kellekule Y, Van Maanen RJ, Teeuwsen J, Underberg WJM. Qualitative and quantitative aspects of the degradation of several oligopeptides derived from the antitumour peptide antagonist [Arg 6,D-Trp 7,9,MePhe 8 ] substance P{6-11}. J Pharm Biomed Anal 1999; 19: Nannan Panday VR, Hoetelmans RMW, Van Heeswijk RPG, Meenhorst PL, Inghels M, Mulder J-W, Beijnen JH. Paclitaxel in the treatment of human immunodeficiency virus 1-associated Kaposi s sarcoma: drug-drug interactions with protease inhibitors and nonnucleoside reverse transcriptase inhibitors: a case report study. Cancer Chemother Pharmacol 1999; 43: Nannan Panday VR, Huizing MT, Van Tellingen O, Hakvoort RA, Willemse PHB, De Graeff A, Vermorken JB, Beijnen JH. Pharmacologic study of Cremophor EL7 administered in cancer patients with impaired hepatic function receiving paclitaxel. J Oncol Pharm Pract 1999; 5: Nannan Panday VR, Rosing H, Huizing MT, Koopman FJ, Van Warmerdam LJC, Ten Bokkel Huinink WW, Beijnen JH. Urinary excretion of paclitaxel and metabolites in a large series study. J Oncol Pharm Pract 1998; 4: Nannan Panday VR, Ten Bokkel Huinink WW, Vermorken JB, Rosing H, Koopman FJ, Swart M, Schellens JHM, Beijnen JH. Pharmacokinetics of paclitaxel administered as a 3-hour or 96- hour infusion. Pharmacol Res 1999; 40: Nannan Panday VR, Van Warmerdam LJC, Huizing MT, Rodenhuis S, Schellens JHM, Beijnen JH. A single 24-hour plasma sample does not predict the carboplatin AUC from carboplatin-paclitaxel combinations or from a high dose carboplatin-thiotepa-cyclophosphamide regimen. Cancer Chemother Pharmacol 1999; 43: Reubsaet JLE, Beijnen JH, Bult A, Van Maanen RJ, Marchal JAD, Underberg WJM. Analytical techniques used to study the degradation of proteins and peptides. Chemical instability. J Pharm Biomed Anal 1998; 17: Reubsaet JLE, Beijnen JH, Bult A, Van Maanen RJ, Marchal JAD, Underberg WJM. Analytical techniques used to study the degradation of proteins and peptides. Physical instability. J Pharm Biomed Anal 1998; 17: Reubsaet JLE, Beijnen JH, Van Bennekom WP, Bult A, Hoekstra AJ, Hop E, Van Os PJHJ, Teeuwsen J, Underberg WJM. Reduction of Cys 36 -Cys 42 and Cys 64 -Cys 74 disulfide bonds in recombinant human granulocyte colony stimulating factor. J Pharm Biomed Anal 1999; 19: Rodenhuis S, De Wit R, De Mulder PHM, Keizer HJ, Sleijfer DTh, Lalisang RI, Bakker P, Mandjes I, Kooi M, De Vries EGE. A multicenter prospective phase II study of high-dose chemotherapy in germ cell cancer patients relapsing from complete remission. Ann Oncol 1999; 12: Rodenhuis S, Huitema ADR, Baars JW, Westermann A, Holtkamp MMJ, Schornagel JH, Beijnen JH. Multiple courses of cyclophosphamide, thiotepa and carboplatin: managing toxicity by dose reduction and pharmacokinetic monitoring. In: Dicke KA, Keating A, editors. Autologous blood and marrow transplantation: proceedings of the ninth international symposium; Arlington, Texas, 1998: Nannan Panday VR, Van Warmerdam LJC, Huizing MT, Ten Bokkel Huinink WW, Schellens JHM, Beijnen JH. A limited-sampling model for the pharmacokinetics of carboplatin administered in combination with paclitaxel. J Cancer Res Clin Oncol 1999; 125: Rosing H, Hillebrand MJX, Jimeno JM, Gomez A, Floriano P, Faircloth G, Henrar REC, Vermorken JB, Cvitkovic E, Bult A, Beijnen JH. Quantitative determination of ecteinascidin 743 in human plasma by miniaturized high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry. J Mass Spectrom 1998; 33:

121 Rosing H, Ten Bokkel Huinink WW, Van Gijn R, Rombouts RFM, Bult A, Beijnen JH. Comparative open, randomized, cross-over bioequivalence study of two intravenous dexrazoxane formulations (Cardioxane7 and ICRF-187) in patients with advanced breast cancer treated with 5-fluorouracil-doxorubicin-cyclophosphamide (FDC). Eur J Drug Metab Pharmacokinet 1999; 24: Straathof CS, Van den Bent MJ, Loos WJ, Vecht CJ, Schellens JHM. The accumulation of topotecan in 9L glioma and in brain parenchyma with and without dexamethasone administration. J Neuro-Oncol 1999; 42: Taal BG, Hoefnagel C, Rutgers M. Carcinoid tumors. N Engl J Med 1999; 341: MEDICAL ONCOLOGY Rosing H, Van Zomeren DM, Doyle E, Ten Bokkel Huinink WW, Schellens JHM, Bult A, Beijnen JH. Quantification of topotecan and its metabolite N-desmethyltopotecan in human plasma, urine and faeces by high-performance liquid chromatographic methods. J Chromatogr B 1999; 727: Schouwink JH, Baas P, Rutgers EJTh, Zoetmulder FAN. Malignant pleural mesothelioma. Eur Respir J 1999; 13: 706. Schouwink JH, Korse CM, Bonfrer JM, Hart AA, Baas P. Prognostic value of the serum tumour markers Cyfra 21-1 and tissue polypeptide antigen in malignant mesothelioma. Lung Cancer 1999; 25: Schrama JG, Rodenhuis S. Dose-intense chemotherapy for locally advanced breast cancer. Curr Oncol Rep 1999; 1: Sessa C, Ten Bokkel Huinink WW, Du Bois A. Oxaliplatin in ovarian cancer. Ann Oncol 1999; 10 (suppl 1): Sneeuw KC, Aaronson NK, Sprangers MA, Detmar SB, Wever LD, Schornagel JH. Evaluating the quality of life of cancer patients: assessments by patients, significant others, physicians and nurses. Br J Cancer 1999; 81: Sparidans RW, Den Hartigh J, Cremers S, Beijnen JH, Vermeij P. Semi-automated liquid chromatographic analysis of pamidronate in urine after derivatization with 1-naphthylisothiocyanate. J Chromatogr B 1999; 730: Sparidans RW, Henrar REC, Jimeno JM, Faircloth G, Floriano P, Beijnen JH. Bioanalysis of thiocoraline, a new marine antitumoral depsipeptide, in plasma by high-performance liquid chromatography with fluorescence detection. J Chromatogr B 1999; 726: Sparidans RW, Kettenes-van den Bosch JJ, VanTellingen O, Nuyen B, Henrar REC, Jimeno JM, Faircloth G, Floriano P, Rinehart KL, Beijnen JH. Bioanalysis of aplidine, a new marine antitumoral depsipeptide, in plasma by high-performance liquid chromatography after derivatization with trans-4 -hydrazino-2-stilbazole. J Chromatogr B 1999; 729: Sparidans RW, Taal BG, Beijnen JH. Bioanalysis of meta-iodobenzylguanidine in plasma by high-performance liquid chromatography after derivatization with benzoin. J Chromatogr B 1999; 730: Taal BG, Westerman H, Boot H, Rankin E. Clinical and endoscopic features of melanoma metastases in the upper GI tract. Gastrointest Endosc 1999; 50: Ten Bokkel Huinink WW, De Swart CA, Van Toorn DW, Morack G, Breed WP, Hillen HF, Van der Hoeven JJ, Reed NS, Fairlamb DJ, Chan SY, Godfrey KA, Kristensen GB, Van Tinteren H, Ehmer B. Controlled multicentre study of the influence of subcutaneous recombinant human erythropoietin on anaemia and transfusion dependency in patients with ovarian carcinoma treated with platinum-based chemotherapy. Med Oncol 1998;15: Theuws JC, Muller SH, Seppenwoolde Y, Kwa SL, Boersma LJ, Hart GA, Baas P, Lebesque JV. Effect of radiotherapy and chemotherapy on pulmonary function after treatment for breast cancer and lymphoma: a follow-up study. J Clin Oncol 1999; 17: Van Asperen J, Van Tellingen O, Schinkel AH, Beijnen JH. Comparative pharmacokinetics of vinblastine after a 96-hour continuous infusion in wild-type mice and mice lacking mdr1a P-glycoprotein. J Pharmacol Exp Ther 1999; 289: Van Asperen J, Van Tellingen O, Tijssen F, Schinkel AH, Beijnen JH. Increased accumulation of doxorubicin and doxorubicinol in cardiac tissue of mice lacking mdr1a P-glycoprotein. Br J Cancer 1999; 79: Van Baarle D, Hovenkamp E, Kersten MJ, Klein MR, Miedema F, Van Oers MHJ. Direct Epstein-Barr virus typing on peripheral blood mononuclear cells: no association between EBV type 2 infection or superinfection and the development of acquired immunodeficiency syndrome-related non-hodgkin s lymphoma. Blood 1999; 93: Van Cutsem E, Cunningham D, Ten Bokkel Huinink WW, Punt CJ, Alexopoulos CG, Dirix L, Symann M, Blijham, GH, Cholet P, Fillet G, Van Groeningen C, Vannetzel JM, Levi F, Panagos G, Unger C, Wils J, Cote C, Blanc C, Herait P, Bleiberg H. Clinical activity and benefit of irinotecan (CPT-11) in patients with colorectal cancer truly resistant to 5-fluorouracil (5-FU). Eur J Cancer 1999; 35: Van Gijn R, Ten Bokkel Huinink WW, Rodenhuis S, Vermorken JB, Van Tellingen O, Rosing H, Van Warmerdam LJC, Beijnen JH. Topoisomerase I/II inhibitor intoplicine administered as a 24-hour infusion: phase I and pharmacologic study. Anticancer Drugs 1999; 10:

122 122 MEDICAL ONCOLOGY Van Gijn R, Van Tellingen O, Haverkate E, Kettenes-van den Bosch JJ, Bult A, Beijnen JH. Pharmacokinetics and metabolism of the staurosporine analogue CGP in mice. Invest New Drugs 1999; 17: Van Heeswijk RPG, Hoetelmans RMW, Harms R, Meenhorst PL, Mulder JW, Lange JMA, Beijnen JH. Simultaneous quantitative determination of the HIV protease inhibitors amprenavir, indinavir, nelfinavir, ritonavir, and saquinavir in human plasma by ion-pair high-performance liquid chromatography with ultraviolet detection. J Chromatogr B 1998; 719: Van Heeswijk RPG, Veldkamp AI, Hoetelmans RMW, Mulder JW, Schreij G, Hsu A, Lange JMA, Beijnen JH, Meenhorst PL. The steady-state plasma pharmacokinetics of indinavir alone and in combination with low dose of ritonavir in twice daily dosing regimens in HIV-1 infected individuals. AIDS 1999; 13: F95-9. Van Helvoort A, De Brouwer A, Ottenhof R, Brouwers JFHM, Wijnholds J, Beijnen JH, Rijneveld A, Van der Poll T, Van der Valk MA, Majoor D, Voorhout W, Witz KWA, Oude Elferink RPJ, Borst P. Mice without phosphatidylcholine transfer protein have no defects in the secretion of phosphatidylcholine into bile or into lung airspaces. Proc Natl Acad Sci USA 1999; 96: Van Lent-Evers NA, Mathot RAA, Geus WP, Van Hout BA, Vinks AATMM. Impact of goal-oriented and model-based clinical pharmacokinetic dosing of aminoglycosides on clinical outcome: a cost-effectiveness analysis. Ther Drug Monitor 1999; 21: Van Maanen MJ, Beijnen JH. Liquid chromatographic mass spectrometric determination of the novel, recently identified thiotepa metabolite, thiotepa-mercapturate, in urine. J Chromatogr B 1999; 732: Van Maanen MJ, Brandt AC, Damen JMA, Beijnen JH. Degradation study of thiotepa in aqueous solutions. Int J Pharm 1999; 179: Van Maanen MJ, Thijhof IM, Damen JMA, Versluis C, Kettenesvan den Bosch JJ, Heck AJR, Rodenhuis S, Beijnen JH. A search for new metabolites of N',N'',N'''-triethylenethiophosphoramide. Cancer Res 1999; 59: Van Tellingen O, Beijnen JH, Verweij J, Scherrenberg EJ, Nooijen WJ, Sparreboom A. Rapid esterase-sensitive breakdown of polysorbate 80 and its impact on the plasma pharmacokinetics of docetaxel and metabolites in mice. Clin Cancer Res 1999; 5: Van Tellingen O, Huizing MT, Nannan Panday VR, Schellens JHM, Nooijen WJ, Beijnen JH. Cremophor EL causes (pseudo)-nonlinear pharmacokinetics of paclitaxel in patients. Br J Cancer 1999; 81: Van Tellingen O, Kemper M, Tijssen F, Van Asperen J, Nooijen WJ, Beijnen JH. High-performace liquid chromatographic bio-analysis of PSC 833 in human and murine plasma. J Chromatogr B 1998; 719: Van Zuylen L, Schellens JHM, Goey SH, Pronk LC, De Boer- Dennert MM, Loos WJ, Ma J, Stoter G, Verweij J. Phase I and pharmacologic study of the arotinoid Ro in combination with cisplatin and etoposide in patients with non-small cell lung cancer. Anticancer Drugs 1999; 10: Veldkamp AI, Van Heeswijk RPG, Hoetelmans RMW, Meenhorst PL, Mulder JW, Lange JMA, Beijnen JH. Rapid quantification of delavirdine, a novel non-nucleoside reverse transcriptase inhibitor, in human plasma using isocratic reversed-phase high-performance liquid chromatography with fluorescence detection. J Chromatogr B 1999; 727: Vermorken JB, Ten Bokkel Huinink WW, Kobierska A, Van der Burg ME, Forni M, Piccart MJ, Van der Putten E. Phase I study of high-dose epirubicin in platinum-pretreated patients with ovarian carcinoma. Oncology (Basel) 1999; 57: Verschraagen M, Koks CHW, Schellens JHM, Beijnen JH. P-glycoprotein system as a determinant of drug interactions. The case of digoxin-verapamil. Pharmacol Res 1999; 40: Westermann AM, Holtkamp MMJ, Linthorst GAM, Van Leeuwen L, Willemsen EJM, Van Dijk W, Nooijen WJ, Baars JW, Schornagel JH, Rodenhuis S. At home management of aplastic phase following high-dose chemotherapy with stem cell rescue for hematologic and non-hematologic malignancies. Ann Oncol 1999; 10: Van Maanen MJ, Van Ooijen RD, Zwikker JW, Huitema ADR, Rodenhuis S, Beijnen JH. Determination of N,N,N -triethylenthiophosphoramide and its active metabolite N,N,N -triethylenphosphoramide in plasma and urine using capillary gas chromatography. J Chromatogr B 1998; 719: Woll PJ, Judson I, Lee SM, Rodenhuis S, Nielsen OS, Buesa JM, Lorigan PC, Leyvraz S, Hermans C, Van Glabbeke M, Verweij J. Temozolomide in adult patients with advanced soft tissue sarcoma: a phase II study of the EORTC soft tissue and bone sarcoma group. Eur J Cancer 1999; 35: Van Meerbeeck JP, Baas P, Debruyne C, Groen HJ, Manegold C, Ardizzoni A, Gridelli C, Van Marck EA, Lentz M, Giaccone G. A Phase II study of gemcitabine in patients with malignant pleural mesothelioma. European Organization for Research and Treatment of Cancer Lung Cancer Cooperative Group. Cancer 1999; 85: Zuetenhorst H, Taal BG, Boot H, Valdes Olmos R, Hoefnagel C. Longterm palliation in metastatic carcinoid tumours with various applications of meta-iodobenzylguanidin (MIBG): pharmacological MIBG, 131I labeled MIBG and the combination. Eur J Gastroenterol Hepatol 1999; 11:

123 Full papers in press Ackerstaff AH, Hilgers FJM, Meeuwis C, Knegt P, Weenink C. Pulmonary function, pre and post total laryngectomy. Clin Otolaryngol 1999 (in press). Boogerd W, Beijnen JH. Methylphenidate for congenital cerebral palsy with choreo-athetosis. Ann Intern Med 1999 (in press). Boogerd W, Tjahja IS, Van de Sandt MM, Beijnen JH. Penetration of idarubicin into malignant brain tumor tissue. J Neuro-Oncol 1999 (in press). De Gast GC, Klümpen H-J, Vyth-Dreese FA, Kersten MJ, Verra NCV, Sein J, Batchelor D, Nooijen WJ, Schornagel JH. Phase I/II trial of combined immunotherapy with subcutaneous GM-CSF, low dose IL-2 and IFNa in progressive metastatic melanoma and renal cell carcinoma. Clin Cancer Res 1999 (in press). Groenewegen GC, De Gast GC. Sequential chemo-immunotherapy (SCIT) for metastatic melanoma: outpatient treatment with dacarbazine, molgramostim, interleukin-2, and interferon-alpha. Clin Cancer Res 1999 (in press). Nuyen B, Bouma M, Henrar REC, Jimeno JM, Manada C, Bult A, Beijnen JH. Compatibility and stability of aplidine in infusion devices and its hemolytic and precipitation potential upon intravenous administration. Anti-Cancer Drugs 1999 (in press). Paridaens R, Biganzoli L, Bruning PF, Klijn JGM, Gamucci T, Houston S, Coleman R, Schachter J, Van Vreckem A, Sylvester R, Awada A, Wildiers J, Piccart M. Taxol versus doxorubicin as firstline single agent chemotherapy for metastatic breast cancer - an EORTC randomized study with crossover. J Clin Oncol 1999 (in press). Pluim D, Maliepaard M, Van Waardenburg RCAM, Beijnen JH, Schellens JHM. 32 P-postlabeling assay for the quantification of the major platinum-dna adducts. Anal Biochem 1999 (in press). Reijers MHE, Weverling GJ, Hart AAM, Weigel HM, Ten Kate RW, Mulder JW, Reiss P, Schuitemaker H, Hoetelmans RMW, Lange JMA.Toxicity and drug exposure in a quadruple drug regimen in HIV 1 infected patients. AIDS 1999 (in press). Rodenhuis S. High-dose chemotherapy in solid tumors [editorial]. Revista de Oncología 1999 (in press). 123 MEDICAL ONCOLOGY Hoitink MA, Beijnen JH, Bult A, Damen MA, Van der Houwen OAGJ, Kruijtzer JAW, Tibben MM, Wiese G, Underberg WJM. Degradation of azaglycinamido residues in model tripeptides derived from goserelin. J Pharm Sci 1999 (in press).. Koks CHW, van Heeswijk RPG, Veldkamp AI, Meenhorst PL, Mulder JW, Van der Meer JTM, Beijnen JH, Hoetelmans RMW. Itraconazole as an alternative for ritonavir liquid formulation when combined with saquinavir. AIDS 1999 (in press). Rosing H, Man WY, Doyle E, Bult A, Beijnen JH. Bio-analytical chromatographic method validation: a review of current practices and procedures. J Liq Chromatogr Rel Techn 1999 (in press). Rosing H, Van Warmerdam LJC, Ten Bokkel Huinink WW, Dubbelman AC, Rodenhuis S, Beijnen JH. Pharmacokinetics and metabolism of docetaxel administered as an one hour intravenous infusion during a phase II clinical trial. Cancer Chemother Pharmacol 1999 (in press). Kuiper RAJ, Malingré MM, Beijnen JH, Schellens JHMS. Anaphylactic reaction after first ingestion of oral cyclosporin (Neoral ) capules. Ann Pharmacother 1999 (in press). Schellens JHM, Beijnen JH. Cisplatin. In: Ratain MJ, Tempero M, Skosey C, editors. Oncology therapeutics: a quick reference guide. Orlando: Saunders, 1999 (in press). Malingré MM, Meerum Terwogt JM, Beijnen JH, Rosing H, Koopman FJ, Van Tellingen O, Duchin KKL, Ten Bokkel Huinink WW, Swart M, Schellens JHM. Clinical pharmacology of oral paclitaxel in a dose escalating study. Clin Cancer Res 1999 (in press). Meerum Terwogt JM, Tibben MM, Schellens JHM, McGahen L, Beijnen JH. Validated method for the determination of platinum from a liposomal source (SPI-77) in plasma using graphite furnace atomic absorption spectrometry with Zeeman correction. Fresenius J Anal Chem 1999 (in press). Newell DR, McLeod H, Schellens JHM. The pharmacology of anticancer drugs. In: Souhami RL, Tannock I, Hohenberger P, Horiot JC, editors. Oxford Textbook of Oncology. Oxford: Oxford University Press, 1999 (in press). Nuyen B, Bouma M, Henrar REC, Bette P, Bult A, Beijnen JH. In vitro compatibility studies with the experimental anticancer agent BIBX1382BS. Int J Pharm 1999 (in press). Schellens JHM, Beijnen JH. Topotecan. In: Ratain MJ, Tempero M, Skosey C, editors. Oncology therapeutics: a quick reference guide. Orlando: Saunders, 1999 (in press). Sparidans RW, Veldkamp AI, Hoetelmans RMW, Beijnen JH. An improved and simplified liquid chromatographic assay for adefovir in plasma using derivatization with chloroacetaldehyde. J Chromatogr B 1999 (in press). Theuws JCM, Seppenwoolde Y, Kwa SLS, Boersma LJ, Damen EMF, Baas P, Muller SH, Lebesque JV. Recovery of local pulmonary injury 18 to 48 months after irradiation for lymphoma or breast cancer. Int J Radiat Oncol Biol Phys 1999 (in press). Van Dijk AMC, Kessler FL, Stadhouders-Keet SAE, Verdonck LF, De Gast GC, Otten HG. Selective depletion of major and minor histocompatibility antigen reactive T cells; Towards prevention of acute graft-versus-host disease. Br J Hematol 1999 (in press).

124 124 MEDICAL ONCOLOGY Veldkamp AI, Hoetelmans RMW, Lange JMA, Beijnen JH, Meenhorst PL. Ritonavir enables combined therapy with rifampin and saquinavir. Clin Infect Dis 1999 (in press). Veldkamp AI, Mulder JW, Meenhorst PL, Lange JMA, Beijnen JH, Hoetelmans RMW. Determination of the novel antiretroviral drug efavirenz in human plasma using high-performance liquid chromatography with ultraviolet detection. J Chromatogr B 1999 (in press). Rodenhuis S. Hoge dosis chemotherapie met stamcelondersteuning bij gemetastaseerd mammacarcinoom. Tijdschr Kanker 1999; 23: Schellens JHM, Meerum Terwogt JM, Malingré MM, Beijnen JH. Orale therapie met taxanen. In: Schellens JHM, Meinders AE, Vermeij P, redactie. Medicamenteuze therapie: condolidatie en innovatie. Leiden: Boerhaave Commissie voor Postacademisch Onderwijs in de Geneeskunde, 1999: Veldkamp AI, Sparidans RW, Hoetelmans RMW, Beijnen JH. Rapid determination of abacavir in human plasma using high-performance liquid chromatography with ultraviolet detection. J Chromatogr B 1999 (in press). Vielvoye-Kerkmeer APE, Bergman W. Letter to the editor. The Pain Clinic 1999 (in press). Vielvoye-Kerkmeer APE, Mattern C, Uiterdaal M. Transdermal fentanyl in opioid-naive cancer patients. J Pain Symptom Manage 1999 (in press). Vielvoye-Kerkmeer APE, Van der Weide M, Mattern C. Letter to the Editor. J Pain Symptom Manage 1999 (in press). Vulcan TG, Zhu TC, Rodrigues C, His A, Fraker DL, Baas P, Murrer L, Star WM, Yodh AG, Hahn SM. In vivo light measurement during intraperitoneal photodynamic therapy: a comparison of two dosimetry systems. Lasers Surg Med 1999 (in press). Schellens JHM. Variabiliteit in gevoeligheid voor farmaca. In: Van Ree JM, Breimer DD, redactie. Algemene farmacologie. Maarssen: Elsevier/Bunge, 1999: Taal BG, Hoefnagel C, Boot H, De Jong D, Rutgers. Carcinoide tumoren van de darm: ontwikkelingen binnen Nederland in diagnostiek en palliatieve behandeling. Ned Tijdschr Geneeskd 1999; 143: Taal BG, Van Loon HJ, Kahn N, De Jong D, Vasen HFA, Van t Veer. De rol van genetische factoren bij het ontstaan van maagcarcinoom. Ned Tijdschr Geneeskd 1999; 143: Van der Weide M, Vielvoye-Kerkmeer APE. Pijnbestrijding bij terminale kankerpatiënten. Mod Med 1999; 5: Van Dijk AMC. T cell reactivity against keratinocytes in acute graftversus-host disease after bone marrow transplantation [dissertation]. Utrecht: University of Utrecht, Local papers and theses Boogerd W. Neuro-oncologie: als de prognose nog maar een paar maanden bedraagt. Neurologie Actueel 1999; 2: 1-5. Van Zandwijk N, Smit EF, Vansteenkiste J. Oncogenese. In: Demedts M, Dijkman JH, Hilversing C, Postma DS, redactie. Longziekten. Assen: Van Gorcum/ Leuven: Universitaire Pers, 1999: Bronner GM, Koks CHW, Beijnen JH. Thalidomide opnieuw in de belangstelling. [ingezonden] Ned Tijdschr Geneeskd 1999; 143: 122. Koning CCE, Belderbos JSA, Van Zandwijk N. Radiotherapie. In: Demedts M, Dijkman JH, Hilversing C, Postma DS, redactie. Longziekten. Assen/Leuven: Van Gorcum/Universitaire Pers, 1999: Kuiper RAJ, Schellens JHM, Beijnen JH, Blijham GH, Voest EE. Grote verwachtingen van nieuwe aanpak: tumorangiogenese als doelwit voor kankertherapie. Pharm Weekbl 1999; 134: Meerum Terwogt JM. Clinical pharmacology of anticancer agents in relation to formulations and administration routes [disseration]. Utrecht: University of Utrecht, Rodenhuis S, De Vries EGE: Hoge dosis chemotherapie met stamcelondersteuning bij solide tumoren van volwassenen [capita selecta]. Ned Tijdschr Geneeskd 1999; 143: Van Zandwijk N, Vansteenkiste J, Festen J. Algemeen. In: Demedts M, Dijkman JH, Hilversing C, Postma DS, redactie. Longziekten. Assen: Van Gorcum/ Leuven:Universitaire Pers, 1999: Vanuytsel LJ, Van Zandwijk N. Longafwijkingen door ioniserende straling. In: Demedts M, Dijkman JH, Hilversing C, Postma DS, redactie. Longziekten. Assen: Van Gorcum/ Leuven: Universitaire Pers, 1999: Vasen HFA, Nagengast FM, Griffioen G, Kleibeuker JH, Menko FH, Taal BG. Periodiek onderzoek bij personen met een belaste familie anamnese voor colorectaal carcinoom. Ned Tijdschr Geneeskd 1999; 143: Vielvoye-Kerkmeer APE, Meijler WJ. Praktische aspecten van de pijnbestrijding. In: Wesseling H, Neef C, De Graeff PA, redactie. Algemene farmacotherapie. Houten/Diemen: Bohn Stafleu Van Loghum, 1999: Vielvoye-Kerkmeer APE, Van der Weide M. Behandeling van pijn door kanker. Tijdschr Kanker 1999; 23: 20-2.

125 125 Vielvoye-Kerkmeer APE, Van Luyt PA. Zin en onzin over dystrofie. Mod Med 1999; 6: Westermann AM. Novel applications of growth factors in solid tumors [dissertation]. Amsterdam: University of Amsterdam, Zoetmulder FAN, Van der Vange N, Witkamp AJ, Kaag MM, Boot H, Beijnen JH. Hypertherme intraperitoneale chemotherapie (HIPEC) bij patiënten met pseudomyxoma peritonei of peritoneummetastasen van colorectaal carcinoom; gunstige eerste ervaringen in het Nederlands Kanker Instituut. Ned Tijdschr Geneeskd 1999; 143: MEDICAL ONCOLOGY Local papers in press Jonkers DAP, Kerbusch T, Schellens JHM, Beijnen JH. Ifosfamide encephalopathie. Mechanisme, behandeling en preventie. Pharm Weekbl 1999 (in press). Meijler WJ, Vielvoye-Kerkmeer APE. Farmacotherapie van benigne pijn. Handboek Pijnbestrijding 1999 (in press). Rodenhuis S, Beijnen JH. Principes van medicamenteuze therapie. In: Voûte PA, Veenhof CHN, redactie. Behandeling van kanker. Houten: Bohn Stafleu Van Loghum, 1999 (in press). Vielvoye-Kerkmeer APE, Bergman W. Help, mijn benen doen zo zeer; een patiënte met een idiopathische erythromelalgie. Tijdschr Huisartsgeneeskd 1999 (in press). Vielvoye-Kerkmeer APE, Van Luijt PA. Complex regional pain syndrome. Tijdschr Huisartsgeneeskd 1999 (in press).

126 126 SURGICAL ONCOLOGY XI Division of Surgical Oncology Surgical Oncology Board: Bin Kroon and Fons Balm Board BBR Kroon MD PhD (Head) AJM Balm MD PhD FRCS FACS Permanent academic staff General Surgery BBR Kroon MD PhD (Head) OE Nieweg MD PhD, EJTh Rutgers MD PhD, F Van Coevorden MD PhD, FAN Zoetmulder MD PhD Head and Neck Surgery FJM Hilgers MD PhD (Head) AJM Balm MD PhD FRCS FACS, S Gonggrijp DDS, FHM Kroon DDS PhD, IB Tan MD PhD, AP Timmers Gynecology ACM Van Lindert MD PhD, (Head) M Van Beurden MD PhD, N Van der Vange MD PhD Urology S Horenblas MD PhD (Head) W Meinhardt MD PhD Plastic and Reconstructive Surgery KE Bos MD PhD (Head) JB De Boer MD (until May), RP Noordanus MD (until October) Anesthesiology PFE Schutte MD (Head) CL Blackburn MMBS FRCA (until February), D Buitelaar MD, MM Kaag MD, JM Ronday MD (from February), J Visscher MD Other academic staff involved in research activities AH Ackerstaff PhD, N Bijker MD, RBJ De Bondt MD, IF Faneyte MD, M Govaert MD, M Heetveld MD, L Jansen MD, R Kaas MD, N Kimmings MD, WMC Klop MD, R Lewicz MD, M Piek-Den Hartog MD, TA Roeleveld MD, LJA Strobbe MD, PJ Tanis MD, CJ Van As MSc, SAJM Van der Veen MD, AJ Witkamp MD, CL Zuur MD Undergraduate students W Deenik, CMG Dekker BSc, JDH De Vries, ME Grunstra, E Hensen, RG Houtkamp, M Koppe, ELAR Mutsaerts, M Raming, GK Roozendaal, M Smit, D Van Poll Fellows A Bex MD, MP Copper MD PhD, PHP Davids MD PhD, E De Bree MD, W IJzerman MD, JWS Merkus MD, KDHM Keuning MD DDS, HSA Oldenburg MD, IGM Tjakra MD, LAE Woerdeman MD Secretary E Van Damme, GMM Van Zuilen Research funded positions (full-time equivalents) Clinical permanent: 2.2 Clinical project: 4.0 Introduction The Division of Surgical Oncology focused on its main themes of preoperative staging, organ preservation and reconstructive surgery. The unique institutional experience with sexuality preserving cystectomy and neobladder reconstruction resulted in successful preservation of continence and sexual function for 25 patients. The significant improvement of disease-free survival in peritoneal pseudomyxoma by HIPEC (Hyperthermic Intraperitoneal Chemotherapy) was consolidated and is now recognized as the treatment of choice. The value of this method in colorectal cancer is being studied in the framework of a prospective randomized trial, supported by the Health Insurance Council Fund for Investigative Medicine. Hyperthermic intrapleural chemotherapy (HIPLEC), a variant of this treatment modality, started in patients with pleural metastases of limited mesothelioma or thymoma. A major achievement in the development of new surgical treatments is the sentinel node procedure which has now made the practice of elective axillary dissection obsolete in patients with breast cancer. The division was actively involved in organizing the First International Congress on the Sentinel Node in Diagnosis and Treatment of Cancer, held in Amsterdam in cooperation

127 with the Free University. This successful event was attended by more than 700 delegates. The organ preserving treatment of advanced stage IV head and neck squamous cell carcinoma by RADPLAT (super-selective intra-arterial infusion of high dose cisplatin and concomitant radiotherapy) was successfully continued in close collaboration with Divisions IX (Radiotherapy), X (Medical Oncology) and XIII (Diagnostic Oncology). A phase II study was completed in 76 evaluable patients, demonstrating a high loco-regional complete response rate of 88%. A randomized study comparing intravenous (RADPLAT i.v.) and intra-arterial infusion of cisplatin (RADPLAT i.a.) plus concomitant radiotherapy started at the end of In cooperation with Divisions IV (Immunology), X (Medical Oncology) and XIII (Diagnostic Oncology), perioperative immunotherapy and immuno-monitoring commenced in renal cancer patients and head and neck cancer patients. Continued collaboration with the Divisions VIII (Experimental Therapy), IX (Radiotherapy) and Departments of Radiotherapy and Otolaryngology/ Head and Neck Surgery of the Catholic University Leuven resulted in the first hypoxia measurements in fresh operation specimens of head and neck cancer. In addition to their traditional clinical program of organ preservation and preoperative staging, the surgical oncological specialties will concentrate on further development of translational research, with emphasis on prognostic markers and the development of experimental animal tumor model systems, in close collaboration with Divisions VI (Tumor Biology), VIII (Experimental Therapy) and XIII (Diagnostic Oncology). Melanoma AJM Balm, D Batchelor, JMG Bonfrer 2, GC De Gast 1, JDH De Vries, CA Hoefnagel 2, SP Israëls 1, L Jansen, MJ Kersten 1, SH Muller 2, GK Roozendaal, LJA Strobbe 3, PJ Tanis, RA Valdés Olmos 2, D Van Poll, BBR Kroon, OE Nieweg Lymphatic mapping with selective lymphadenectomy Lymphatic mapping with sentinel node biopsy can be used to stage the regional lymph node field and to select patients for therapeutic regional lymph node dissection. Lymphoscintigraphy is one of the three techniques used to aid in the identification of a sentinel node. The accuracy of this technique to determine the number of sentinel nodes was investigated in 199 patients and was found to be 77%. In 23% of the patients scintigraphy suggested too few or too many sentinel nodes. The principle reasons for discrepancies were non-visualization of a lymphatic duct on the dynamic (early) images and the limited resolution of the gamma camera. Drainage from melanomas in the head and neck is the most difficult to map. The sentinel node could be identified in 90% of the patients with a melanoma in this region. The need for a two-tracer ( 99m Tc-nanocolloid, Patent Blue Dye) approach was established because 47% of the nodes were missed with one of the tracers. The sensitivity for finding lymph node metastases was 80%. A study was undertaken to analyze sentinel nodes outside regular lymphatic fields. In the 379 patients studied, this phenomenon was encountered 25 times (7% incidence). The majority of these nodes was retrieved. A number of previously unknown drainage patterns were discovered. In four patients, such an abnormally situated node was involved with metastatic disease. The relevance of these findings extends to the follow-up of melanoma patients in general. The latter two studies were done in cooperation with the University Hospital Groningen. The division is one of the two European participants in the international randomized trial to determine potential benefits of lymphatic mapping in local-regional control and survival. That trial has now accrued well over 1,400 patients and will be closed in the year Analysis of 551 patients in this trial revealed that lymphatic mapping performed with lymphoscintigraphy, a gamma ray detection probe plus a blue dye, was more successful (99.1%) than lymphatic mapping performed with blue dye alone (95.2%) (p = 0.014). After a new participating center had completed the 30-case learning phase, the success of sentinel node identification was independent of the center s case volume or experience in the trial. Notes 1 Division X. 2 Division XIII. 3 Canisius Wilhelmina Hospital, Nijmegen. Isolated Limb Perfusion (ILP) for malignant tumors of the limbs J Clarke, AMM Eggermont 1, P Gustafson 6, JM Klausner 4, FJ Lejeune 2, D Liénard 2, PM Schlag 5, H Schraffordt Koops 3, G Steinmann 7, GW Van Slooten, BBR Kroon, OE Nieweg Limb salvage by ILP with Tumor Necrosis Factor alpha and Melphalan in patients with locally advanced soft tissue sarcomas Limb salvage in patients with irresectable soft tissue sarcomas (STS) conventionally treated by anatomic or functional amputation was investigated in a large international multi-institutional study. ILP with Tumor Necrosis Factor alpha (TNFα) + Melphalan (M) was given as induction biochemotherapy to obtain local control and/or make limb sparing surgery possible. A total of 246 patients with locally advanced STS were offered an ILP with TNF + M. There were 55% primary tumors, 45% recurrent tumors, 22% multiple tumors and 15% concurrent known metastatic disease. Thirteen percent of 127 SURGICAL ONCOLOGY

128 128 SURGICAL ONCOLOGY the patients had received previous radiotherapy and 15% chemotherapy. In 46% the diameter measured >10 cm. Grade III tumors were diagnosed in 66%. The majority of patients (222) underwent one ILP of 90 minutes at C with 2 4 mg TNF + Melphalan (10 13 mg/l limb volume). The first 62 patients also received Interferon-γ. Two to four months later a delayed marginal resection of the tumor remnant was done in 75% of cases. All cases were reviewed by two independent committees which agreed that for 80% (196 patients) ILP offered the only chance of limb salvage: 67% were otherwise clear cut amputation cases and 13% were resectable cases but with severe loss of function. Major tumor responses occurred in 75%, rendering tumors resectable in most cases. The following response rates were achieved: complete response in 28%, partial response in 47%, no change (17%), progressive disease (6%), missing values (1%). At median follow up of >3 years, limb salvage was achieved in 71% of the 196 patients considered justified ILP-candidates by independent reviews. Regional toxicity was limited. Systemic toxicity was minimal/moderate, easily managed with no toxic deaths. Matched-pair analysis with cases from the Scandinavian STS-database showed TNF-ILP had no negative effect on survival (p = 0.96). This study demonstrates that TNF based ILP is a new and effective option for the limb sparing management of locally advanced extremity STS. TNF destroys tumor vessels and increases the Melphalan uptake by the tumor four-fold as has recently been shown by the Rotterdam-perfusion program explaining the synergistic anti-tumor effects. TNF is now approved and registered in Europe for ILP of locally advanced grade II III extremity soft tissue sarcomas. Notes 1 Dr Daniel den Hoed Cancer Center, Rotterdam. 2 Centre Pluridisciplinaire d Oncologie, Lausanne. 3 University Hospital, Groningen. 4 University Hospital, Tel-Aviv. 5 Robert-Roessle University Clinic, Berlin. 6 University Hospital, Lund. 7 Boehringer Ingelheim Company, Germany. Figure XI.I The nasal airflow inducing manoeuvre (NAIM) with water manometer connected to one nostril through a tube with an olive and the other nostril closed with a finger, to demonstrate the negative pressure in the nasal cavity by displacement of the water column towards the nose. Head & Neck Oncology AH Ackerstaff, AC Begg 1, KE Bos, JM Coco Martin 1, MP Copper 2, LDA Dorresteijn 3, 12, S Gonggrijp, AAM Hart 4, K Haustermans 5, KDHM Keuning 2, 6, RB Keus 4, S Keijzers 7, WMC Klop, BBR Kroon, FHM Kroon 6, RJAM Michalides 8, RP Noordanus, BMR Op de Coul 9, FA Pameijer 10, CRN Rasch 4, JH Schornagel 3, L Schultze Kool 10, AP Timmers, RA Valdés Olmos 10, CJ Van As, FSAM Van Dam 7, VLM Vander Poorten 11, SAJM Van der Veen, MLF Van Velthuysen 10, LAE Woerdeman 13, CL Zuur, AJM Balm, FJM Hilgers, IB Tan Rehabilitation In the ongoing research program on post-laryngectomy rehabilitation, several topics have been expanded. A high incidence of complete loss of olfactory acuity is one of the disturbing consequences of total laryngectomy due to the absence of a nasal airflow. The key to rehabilitation of olfaction appears to be the development of a Nasal Airflow Inducing Manoeuver (NAIM), which can be described as polite yawning, i.e. yawning with closed lips. This technique is based on our earlier observations in a series of 63 patients and has been refined during an intervention study. The effecti- Table XI.I. Objective measurement of the width of the neoglottis at rest and during phonation on digitized videofluoroscopic images. Measurement Near normal voice Abnormal voice Difference Maximal width at rest 4.2 mm 7.3 mm p = Minimal width during phonation 0.5 mm 2.7 mm p = 0.029

129 veness of the manoeuver could be demonstrated with a water- and/or digital manometer (Figure XI.I). The intervention consisted of one half-hour training session in the majority of the patients. Using an odor detection test and a structured questionnaire in a series of 44 patients, the percentage of patients with functional olfactory acuity could be increased from 25% to 57% (p<0.001). Evaluation of the pharyngo-esophageal segment or neoglottis as the sound source of prosthetic tracheoesophageal voice after total laryngectomy was further evaluated with video-fluoroscopy and correlated with perceptual analysis. Relevant data were collected in 39 patients. Perceptual analysis was done by four experienced listeners. A dichotomy in near normal (13 patients) and abnormal (25 patients) voice was used. Video-fluoroscopy was analyzed objectively by measurements on digitized video images, using the Drawer software for image analysis, developed in Division IX. Results show that objective measurements (maximal width of the neoglottis at rest and the minimal width during phonation) are significantly different between the two patient categories (Table XI.I). It can be concluded that image analysis is a new, promising tool for objective evaluation of the neoglottis during prosthetic vocalization in laryngectomized patients. Long-term results of prosthetic vocal rehabilitation after total laryngectomy were assessed in a cohort of 318 patients, using an indwelling (Provox) voice prosthesis. In this period 3,008 events with or without a voice prosthesis were retrospectively analyzed. The median follow-up was 68 months. The long-term success rate with respect to (fair to excellent) voice quality was 88%. The main indication for replacement, in approximately three-quarters of the patients, was leakage of fluids through the prosthesis. The median device life was 11 days. There was a significant negative association between radiotherapy and device life (p<0.05), but prostheses in patients older than 70 years had a significantly longer life span (p<0.05). Leakage around the prosthesis was seen in 13% of the patients. This problem was solved by simple down sizing in more than half of the cases. Short-time removal to allow for spontaneous shrinkage or temporary closure of the fistula was necessary in 18% of the patients, resulting in only one definitive closure because of untreatable leakage around the device. Hypertrophy and/or infection of the fistula (19% of the patients) were successfully treated by upsizing the prosthesis, antibiotics or a temporary closure, which was definitive in one patient. Radiotherapy was again shown to have a significantly negative influence on device life (p<0.05). Evaluation of treatment strategies Analysis of a cohort of 1,960 patients ( ) referred for diagnosis and treatment of primary head and neck squamous cell carcinoma demonstrated a 2% incidence (40 patients) with neck node metastasis of an unknown primary. The number of non-protocol radiodiagnostic measures carried out before admission to our institute (30), demonstrates the unfamiliarity of referring physicians with this type of neck disease. The development of two primaries at non-irradiated sites proves the validity of including the laryngo-pharyngeal axis in the radiation portals. A group of 95 patients with a recurrent head and neck carcinoma and a control group, matched for age, sex, primary tumor and duration of follow-up, were analyzed for pain complaints as first symptom of recurrent disease. Of the patients with proven recurrent disease, 69% reported pain complaints as an initial symptom, whereas in the control group this was only reported by 2%. This confirms that each pain complaint after intentional curative treatment should be regarded as a warning sign. Recurrence of disease without preceding pain complaints (31%) emphasizes the importance of a thorough follow-up. The incidence of stroke was analyzed in 255 patients irradiated for T1 or T2 larynx carcinoma (n=120), pleiomorphic adenoma (n=65), and parotid carcinoma (n=70). The relative risk of stroke was determined by comparison with population stroke-incidence register, adjusted for sex and age. The following rates were calculated: incidence of stroke related to larynx carcinoma 8 (expected 0.923; OR 8.67; 95% CI ), pleiomorphic adenoma 3 (expected 0.330; OR 9.09; 95% CI ), parotid carcinoma 0 (exp ). These findings demonstrate that radiation on the neck increases the risk of stroke significantly (p<0.000). Development of function-preserving treatment modalities Photodynamic therapy (PDT) using the photosensitizer Foscan was continued in a group of 33 patients with 47 superficially spreading oral and oropharyngeal carcinoma. A complete prolonged tumor regression was achieved in 93% of tumors <2cm and 29% in tumors >3cm, with excellent preservation of function. Side effects include generalized photosensitivity during 2-3 weeks. Foscan PDT is effective in primary oral and oropharyngeal carcinoma, is repeatable, does not compromise future treatments and is free of major toxicity. Super-selective intra-arterial infusion of high dose cisplatin (150 mg/m 2, days 2, 9, 16, 23) with concomitant radiotherapy (RADPLAT) has been studied in 76 patients with advanced head and neck squamous cell carcinoma (stage IV). A complete locoregional response was achieved in 88% with a median follow-up period of 11 months (range 2-30 months). No severe toxicity was encountered. Limited ototoxicity was found in 53% with a median audiometric hearing loss of 5 db at 1-2 KHz. These results consolidate our earlier experience of function-preserving therapy in patients with a poor prognosis. A randomized study comparing intra-arterial (RADPLAT i.a.) and intravenous (RADPLAT i.v.) administration has started. Notes 1 Division VIII. 2 Fellow Head and Neck Oncology. 129 SURGICAL ONCOLOGY

130 130 SURGICAL ONCOLOGY 3 Division X. 4 Division IX. 5 Department of Radiotherapy, Catholic University, Leuven, Belgium. 6 Department of Oral/Maxillofacial Surgery, Academic Medical Center, Amsterdam. 7 Division XIII. 8 Divsion VI. 9 Department of Otolaryngology/Head and Neck Surgery, University Hospital Nijmegen. 10 Division XIII. 11 Department of Otolaryngology/Head&Neck Surgery, Catholic University, Leuven, Belgium. 12 Department of Neurology, University Hospital Nijmegen. 13 Department of Plastic and Reconstructive Sugery, Academic Medical Center, Amsterdam. Breast Cancer H Bartelink 1, APE Besnard 2, N Bijker, JH Borger 1, JB De Boer, IF Faneyte, ME Grunstra, CA Hoefnagel 2, L Jansen, R Kaas, H Klaren 2, BBR Kroon, OE Nieweg, JL Peterse 2, BR Pieters 1, R Valdes Olmos 2, MJ Van de Vijver 2, FE Van Leeuwen 3, EJTh Rutgers Prophylactic mastectomy Files of 69 women who underwent prophylactic mastectomy between were analyzed. Thirty eight women were asymptomatic (mean age 37, 29 BRCA-1/2 mutations) and 31 underwent contralateral mastectomy after previous treatment for breast cancer (mean age 42, 20 BRCA-1/2 mutations). A total of 107 breasts were surgically removed. Two patients were diagnosed with breast cancer by mammography/mri screening before surgery. At surgery, one patient had a T1N0 breast cancer and three had ductal carcinoma in situ (DCIS) /lobular carcinoma in situ (LCIS). Fifty-six patients underwent reconstruction of whom 50 were reconstructed in the same surgical session. There were 58 tissue expanders (TE) placed, of which twelve were lost due to infection. All 13 transpositions of rectoabdominal muscle (TRAM) were successful. During follow-up (142 women years at risk), no breast cancer was encountered. The clinical outcome seems to confirm the estimated risk reduction by prophylactic mastectomy. Immediate TE reconstruction has a substantial failure rate. Ductal Carcinoma In Situ The effects of selection of patients for the EORTC trial, investigating the role of radiotherapy in breast conserving therapy, were studied in five participating centers by analyzing all patients treated for DCIS. Reasons for non-entry were evaluated and treatment results of entered and non-entered patients were compared. Of the eligible patients 64% were entered; the main reason for non-entry were doctors preference for one of the treatment arms, or patients refusal to trial participation (respectively 26% and 9%). These percentages varied significantly among the centers. At four years, non-entered patients treated with local excision and radiotherapy had higher local recurrence rates than randomized patients (17% versus 2% respectively, p=0.003). Patients treated with local excision only had equal recurrence rates (11% in both groups). These results suggest that the trial results may not be applicable to all patients with DCIS. Clinical relevance of sentinel node outside the axilla Unexpected lymphatic drainage may be found during the sentinel node procedure. The relevance of sentinel nodes (SN) outside the lower axilla was studied in 113 patients with breast cancer. In 21 (19%) SN s outside the lower axilla were found using pre-operative lymphoscintigraphy, a gamma ray detection probe and intra-operative patent blue dye. SN biopsy was performed in the internal mammary chain (17), the breast parenchyma (3), the interpectoral fossa (2) and the infraclavicular fossa (2). All but 3 patients also had a SN in the lower axilla. One patient had a metastasis only in the SN outside the axilla. In three patients SN s within the axilla and elsewhere contained tumor. In three patients the adjuvant treatment policy was modified based on the extra-axillar SN status only. No postoperative complications were seen due to the extra-axillar SN biopsies. Based on these results, biopsy of extra-axillar SN is recommended. Uptake of radiocolloid in sentinel nodes The chance of intra-operatively SN retrieval with the gamma ray detection probe depends on the amount of radioactivity in the node in relation to the uptake in the surrounding normal tissue. In a series of 60 patients with breast cancer and 22 patients with melanoma, uptake in SN s was measured and compared to possibly related factors in a multivariate analysis. Median uptake was 6.5 kbq (0.16% of the injected dose; 95% range kbq) 24 hrs after injection of 64 MBq in breast cancer patients, and 33 kbq (0.82%; kbq) in patients with melanoma. The radioactivity level depends mainly on the time interval between injection and surgery, and not on age sex, nodal status, or dose. Notes 1 Division IX. 2 Division XIII. 3 Division XII. Gastrointestinal Tract Cancer JH Beijnen 1, H Boot 2, E De Bree, MM Kaag, SH Muller 3, N Van der Vange, GW Van Slooten, AJ Witkamp, F Van Coevorden, FAN Zoetmulder

131 Figure XI.2 Actuarial survival curve (Kaplan-Meier) after intra-operative hyperthermic intraperitoneal chemotherapy (HIPEC) in patients with carcinosis peritonei of colorectal origin. Colorectal cancer Since November 1995, patients with peritoneal carcinosis of colorectal origin without distant metastases are treated by extensive surgical cytoreduction in combination with intra-operative hyperthermic intraperitoneal chemotherapy (HIPEC). Twenty-nine patients were included in a phase II study with a median follow-up of 28 months (19-39 months). At the time of evaluation, eleven patients were still alive, six of them with proven recurrence (two local, four distant). The actuarial survival at one and two years is 79% and 48% respectively (Figure XI.2 ). A total of fifty-three patients have now been accrued in the ongoing randomized phase III trial (financed by the Health Insurance Council Fund for Investigative Medicine), comparing extensive surgical cytoreduction + HIPEC and systemic 5-FU to systemic 5-FU alone. Pseudomyxoma peritonei Forty patients with pseudomyxoma peritonei were treated by HIPEC. In most patients the tumor was widely spread throughout the abdomen needing extensive surgical cytoreduction (mean operating time 11 hours). Consequently, a significant postoperative morbidity was encountered. Four patients had long-term nutritional problems related to bowel fistula. Eleven patients needed a re-laparotomy. After a median follow-up of 14 months, 34 patients are still alive (25 of them disease-free). Four patients died from treatment related causes and one patient died from recurrent disease. These data are promising and support previously described results (Sugarbaker PH. Adv Surg 1996; 30: ). Notes 1 Department of Pharmacology, Slotervaart Hospital, Amsterdam. 2 Division X. 3 Division XIII. Urologic Oncology CA Hoefnagel 1, D De Jong 1, L Jansen, OE Nieweg, WW Ten Bokkel Huinink 2, S Horenblas, W Meinhardt Lymphatic mapping and sentinel node biopsy in penile cancer With the aim of reducing the number of unnecessary lymph node dissections in penile cancer, we prospectively assessed the value of a dynamic sentinel node procedure based on the individual drainage pattern of the tumor. From 1994 to 1998, 55 consecutive patients with squamous cell carcinoma of the penis were prospectively entered in this study. A positive sentinel node was found in 11 patients who underwent a subsequent regional lymph node dissection. To date, one false negative sentinel lymph node was observed. We conclude that this technique is minimally invasive and seems promising in identifying patients with lymphatic metastases at an early stage of the disease. Salvage prostatectomy From , 20 patients with residual prostate cancer after 125 I or external beam radiotherapy were treated with a radical prostatectomy. This type of surgery is seldom performed because of the assumed surgical risks and the presumed low potential for local control and cure. Analysis of our results shows an excellent local control with an overall survival of 90% (median follow-up 78 months) and a five-years disease-free survival of 72% (prostate specific antigen (PSA) progression-free survival and/or absence of local recurrence). All patients suffered from erectile dysfunction before surgery and the majority of patients (70%) did not need to wear any external appliance for incontinence. We conclude that in selected cases salvage prostatectomy is a feasible procedure with an acceptable morbidity and excellent results regarding local control and overall survival. Non-seminoma From , 90 patients were entered in a waitand-see-protocol after orchidectomy and 26% developed a recurrence. The eight years disease-specific survival was 98%. Prognostic factors for recurrence are: tumor size, vascular invasion and percentage of embryonal carcinoma (p< 0.03). By using these prognostic factors, a subgroup of patients with high risk for recurrence can be identified for adjuvant treatment. Bladder cancer From 1994 to date, three females and 17 males underwent a modified sexuality- and continencepreserving cystectomy for removal of muscle invasive bladder cancer. One patient developed a urinary fistula which needed surgical closure. Storage function was excellent in 18 patients who are fully continent. Sixteen patients empty their bladder without residual urine. Three patients need to catheterize themselves and one patient 131 SURGICAL ONCOLOGY

132 132 SURGICAL ONCOLOGY is not yet fully evaluable. In 19 patients, normal sexual function could be established. Maximum organ and tissue conservation is achieved through modified cystectomy, leaving a normal sexual function with good urinary tract reconstruction. Testicular tissue preservation Patients with seminoma stage I treated with one-sided orchidectomy and radiotherapy were studied prospectively for serum testosterone levels (LH, FSH and SHBG). Within a normal spread of serum levels (grouped around P30), pathologically low testosterone levels were found in 17% of the patients, indicating that hormonal recovery is at risk from one-sided orchidectomy. In selected cases testicular tissue preservation is therefore strongly recommended. Notes 1 Division XIII. 2 Division X. Gynecologic Oncology MCG Aalders 1, MG Klein 2, HJCM Sterenborg 3, FA Stewart 4, WW Ten Bokkel Huinink 5, MJ Van de Vijver 6, FAN Zoetmulder, M Van Beurden, N Van der Vange, ACM Van Lindert Photodetection with 5-aminolevulinic acid induced protoporphyrin IX in the rat abdominal cavity: Drug dose dependent fluorescence kinetics Macroscopically non-visible peritoneal metastases of ovarian tumors can be detected by fluorescence diagnostics as an aid to diagnostic laparoscopy or laparotomy. The technique of photodetection of tumor metastases is based on differential concentration of photoactive dyes/fluorophores between tumor and normal tissue. The higher the differential, the greater the sensitivity and specificty of photodetection. In this study we performed fluorescence pharmacokinetics on 5- aminolevuline acid (ALA) induced protoporphyrin IX (PpIX) on abdominal organs and tumor metastases in the peritoneal cavity of rats. Three different drug doses (100, 25 and 5 mg/kg) were used and PpIX fluorescence profiles were followed up to 24 hours after i.v. administration. Maximum tumor to normal tissue ratios were found two to three hours after administration of ALA with all drug doses. Three hours after administration of 5 mg/kg, a significant tumor to normal tissue contrast ratio was obtained with all included organs (Figure XI.3). This indicates that low dosages of ALA are optimal for photodetection and are also safe with respect to unwanted white light phototoxicity. Notes 1 Laser center, Academic Medical Center, Amsterdam. 2 Department of Experimental Surgery, Academic Medical Center, Amsterdam. 3 Department of Radiotherapy, Daniel den Hoed Cancer Center, Rotterdam. 4 Division VIII. 5 Division X. 6 Division XIII. Publications: Division XI Full papers Aalders MCG, Van der Vange N, Stewart FA, Klein MG, Van de Vijver MJ, Sterenborg HJCM. White-light toxicity, resulting from systemically administered 5-aminolevulinic acid, under normal operating conditions. J Photochem Photobiol B Biol 1999; 50: Ackerstaff AH, Hilgers FJM, Meeuwis CA, Van der Velden LA, Van den Hoogen FJA, Marres HAM, Manni JJ, Vreeburg GCM. Multiinstitutional assessment of the Provox 2 voice prosthesis. Arch Otolaryngol Head Neck Surg 1999; 125: Figure XI.3 Tumor/normal ratio of fluorescence contrast as function of time after administration of 5mg/kg 5 - aminolevulinic acid (ALA) Bijker N, Rutgers EJT, Peterse JL, Van Dongen JA, Hart AAM, Borger JH, Kroon BBR. Low risk of loco-regional recurrence of primary breast carcinoma after treatment with a modification of the Halsted radical mastectomy and selective use of radiotherapy. Cancer 1999; 85:

133 De Gruijl TD, Bontkes HJ, Peccatori F, Gallee MP, Helmerhorst TJ, Verheijen RH, Aarbiou J, Mulder WM, Walboomers JM, Meijer CJ, Van de Vange N, Scheper RJ. Expression of CD3-zeta on T-cells in primary cervical carcinoma and in metastasis-positive and negative pelvic lymph nodes. Br J Cancer 1999; 79: De Gruijl TD, Bontkes HJ, Van den Muysenberg AJ, Van Oostveen JW, Stukart MJ, Verheijen RH, Van der Vange N, Snijders PJ, Meijer CJ, Walboomers JM, Scheper RJ. Differences in cytokine mrna profiles between premalignant and malignant lesions of the uterine cervix. Eur J Cancer 1999; 35: Meinhardt W, Horenblas S. Time to normalisation of serum testosteroneafter 3-months LHRH agonist administration [letter]. J Urol 1999; 162: Morton DL, Thompson JF, Essner R, Elashoff R, Stern SL, Nieweg OE, Roses DF, Karakousis CP, Mozillo N, Reintgen D, Wang H, Glass EC, Cochran AJ, and the Multicenter Selective Lymphadenectomy Trial Group. Validation of the accuracy in a multicenter trial of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma. Ann Surg 1999; 230: SURGICAL ONCOLOGY Deenik W, Mooi WJ, Rutgers EJT, Peterse JL, Hart AAM, Kroon BBR. Clear cell sarcoma (malignant melanoma) of soft parts. Cancer 1999; 86: Herben VM, Panday VR, Richel DJ, Schellens JH, Van der Vange N, Rosing H, Beusenberg FD, Hearn S, Doyle E, Beijnen JH, Ten Bokkel Huinink WW. Phase I and pharmocologic study of the combination of paclitaxel, cisplatin, and topotecan administered intravenously every 21 days as first-line therapy in patients with advanced ovarian cancer. J Clin Oncol 1999; 17: Nieweg OE, Jansen L, Valdés Olmos RA, Rutgers EJT, Peterse JL, Hoefnagel CA, Kroon BBR. Lymphatic mapping and sentinel lymph node biopsy in breast cancer. Eur J Nucl Med 1999; 26 (suppl): Olieman AFT, Eggermont AMM, Lejeune FJ, Kroon BBR, Hoekstra HJ, Schraffordt Koops H. Hyperthermic isolated limb perfusion with tumor necrosis factor-alpha, interferon-gamma and melphalan for locally advanced extremity non-melanoma skin tumors; a multicenter study. Arch Surg 1999; 134: Hilgers FJM, Ackerstaff AH. Comprehensive rehabilitation after total laryngectomy is more than voice alone. Folia Phoniatrica Logopaedica 2000; 52: Pameijer FA, Mukherji SK, Balm AJM, Van der Laan BFAM. Imaging of squamous cell carcinoma of the hypopharynx [review]. Semin Ultrasound CT MRI 1998; 19: Hilgers FJM, Ackerstaff AH, Van As CJ. Tracheoesophageal puncture: prosthetic voice management. Curr Opin Otolaryngol Head Neck Surg 1999; 7: Strobbe LJA, Jonk A, Hart AAM, Kroon BBR. Positive iliac and obturator nodes in melanoma: survival and prognostic factors. Ann Surg Oncol 1999; 6: Hoefnagel CA, Moonen LMF, Haustermans KMG, Balm AJM. Imaging of tumour hypoxia in nuclear medicine. In: Oxygen and onco therapy 98; Vol 3, Van der Kleij AJ, Voûte PA, Sminia P, editors. Archimedes Hyperbarmedizin und Verlags-GmbH, 1998: Tan IB, Oppelaar H, Ruevekamp MC, Veenhuizen RB, Timmers AP, Stewart FA. The importance of in situ light dosimetry for photodynamic therapy of oral cavity tumors. Head Neck 1999; 21: Horenblas S. Surgical management, Carcinoma of the penis and scrotum. In: Petrovich Z, Baert L, Brady LW, editors. Medical radiology: diagnostic imaging and radiation oncology; Volume: Carcinoma of the kidney, testis and uncommon tumors of the genitourinary tract. Heidelberg: Springer, 1999: Kerrebijn JDF, Balm AJM, Freeman JL, Dosch HM, Drexhage HA. Who is in control of the immune system in head and neck cancer? Crit Rev Oncol Hematol 1999; 31: Tjiong MY, Van der Vange N, Ten Kate FJ, Tjong-A-Hung SP, Ter Schegget J, Burger MP, Out TA. Increased IL-6 and IL-8 levels in cervicovaginal secretions of patients with cervical cancer. Gynecol Oncol 1999; 73: Valdés Olmos RA, Hoefnagel CA, Nieweg OE, Jansen L, Rutgers EJT, Borger J, Horenblas S, Kroon BBR. Lymphoscintigraphy in oncology: a rediscovered challenge. Eur J Nucl Med 1999; 26 (suppl): Kroon BBR, Bergman W, Coebergh JWW, Ruiter DJ on behalf of the Dutch Melanoma Working Party. Consensus on the management of malignant melanoma of the skin in the Netherlands: a consensus report. Melanoma Res 1999; 9: Valdés Olmos RA, Jansen L, Muller SA, Hoefnagel CA, Nieweg OE. Contribution of nuclear medicine to lymphatic mapping and sentinel node identification in oncology. Rev Esp Med Nucl 1999; 18: Ligtenberg MJL, Hogervorst FBL, Willems HW, Arts PJW, Brink G, Hageman S, Bosgoed EAJ, Van der Looij E, Rookus MA, Devilee P, Vos EMAW, Wigbout G, Struycken PM, Menko FH, Rutgers EJT, Hoefsloot EH, Mariman ECM, Brunner HG, Van t Veer LJ. Characteristics of small breast and/or ovarian cancer families with germline mutations in BRCA1 and BRCA2. Br J Cancer 1999; 79: Valdés Olmos RA, Koops W, Loftus BM, Liem IH, Gregor RT, Hoefnagel CA, Hilgers FJM, Balm AJM. Correlative 201 TI SPECT, MRI, and ex-vivo 201 TI uptake in detecting and characterizing cervical lymphadenopathy in head and neck squamous-cell carcinoma. J Nucl Med 1999; 40:

134 134 SURGICAL ONCOLOGY Van As CJ, Tigges M, Wittenberg T, Op de Coul BMR, Eysholdt U, Hilgers FJM. High-speed digital imaging of neoglottic vibration after total laryngectomy. Arch Otolaryngol Head Neck Surg 1999; 125: Van Dam FSAM, Hilgers FJM, Emsbroek G, Touw FI, Van As CJ, De Jong N. Deterioration of olfaction and gustation as a consequence of total laryngectomy. Laryngoscope 1999; 109: Van der Poel HG, Roukema JA, Horenblas S, Van Geel AN, Debruyne FMJ. Metastasectomy in renal cell carcinoma: a multicenter retrospective analysis. Eur J Urol 1999: Vander Poorten VLM, Balm AJM, Hilgers FJM, Tan IB, Loftus-Coll BM, Keus RB, Hart AAM. Prognostic factors for long term results of the treatment of patients with malignant submandibular gland tumors. Cancer 1999; 85: Vrouenraets BC, Kroon BBR, Ogilvie AC, Van Geel AN, Nieweg OE, Swaak AJG, Eggermont AMM. Absence of severe systemic toxicity after leakage-controlled isolated limb perfusion with tumor necrosis factor-alpha and melphalan. Ann Surg Oncol 1999; 6: Full papers in press Ackerstaff AH, Hilgers FJM, Meeuwis CA, Knegt PPM, Weenink C. Pulmonary function pre and post total laryngectomy. Clin Otolaryngol 1999 (in press). Hilgers FJM, Van Dam FSAM, Keyzers S, Koster MN, Van As CJ, Muller MJ. Rehabilitation of olfaction in laryngectomized patients by means of a nasal airflow inducing maneuver: the polite yawning technique. Arch Otolaryngol Head Neck Surg 1999 (in press). Vander Poorten VLM, Balm AJM, Hilgers FJM, Tan IB, Loftus-Coll BM, Van Leeuwen FE, Hart AAM. The development of a prognostic score for patients with parotid carcinoma. Cancer 1999; 85: Horenblas S. Dynamic sentinel node procedure in squamous cell carcinoma of the penis. In: Nieweg OE, Reitgen DS, Thompson FF, Essner R, editors. Lymphatic mapping in cancer. New York: Marcel Dekker, 1999 (in press). Van Ginkel RJ, Kole AC, Nieweg OE, Molenaar WM, Pruim J, Schraffordt Koops H, Vaalburg W, Hoekstra HJ. L-[1-11C]-tyrosine PET to evaluate response to hyperthermic isolated limb perfusion for locally advanced soft-tissue sarcoma and skin cancer. J Nucl Med 1999; 40: Verdonck-de Leeuw IM, Hilgers FJM, Keus RB, Koopmans-Van Beinum FJ, Greven AJ, De Jong A, Bartelink H. Multidimensional assessment of voice characteristics following radiotherapy for early glottic cancer. Laryngoscope 1999; 109: Verdonck-de Leeuw IM, Keus RB, Hilgers FJM, Koopmans-van Beinum FJ, Greven AJ, De Jong A. Bartelink H. Consequences of voice impairment in daily-life for patients following radiotherapy for early glottic cancer: voice quality, vocal function, and vocal performance. Int J Radiat Oncol Biol Phys 1999; 44: Voogd AC, Van Tienhoven G, Peterse JL, Crommelin MA, Rutgers EJT, Van de Velde CJH, Van Geel AN, Slot A, Rodrigus PTR, Jobsen JJ, Von Meyenfeldt MF, Coebergh JWW, for the Dutch Study Group on Local Recurrence after Breast Conservation (BORST). Local recurrence after breast conservation therapy for early breast cancer: detection, treatment and outcome in 266 patients. Cancer 1999; 85: Vrouenraets BC, Hart AAM, Eggermont AMM, Klaase JM, Van Geel BN, Nieweg OE, Kroon BBR. Relation between limb toxicity and treatment outcome after isolated limb perfusion for recurrent melanoma. J Am Coll Surg 1999; Horenblas S, Jansen L, Meinhardt W, Hoefnagel CA, De Jong D, Nieweg OE. Detection of occult metastasis in squamous cell carcinoma of the penis using a dynamic sentinel node procedure. J Urol 1999 (in press). Jansen L, Nieweg OE, Peterse JL, Hoefnagel CA, Valdés Olmos RA, Kroon BBR. Reliability of sentinel node biopsy for staging melanoma. Br J Surg 1999 (in press). Jansen L, Schraffordt Koops H, Nieweg OE, Doting MHE, Kapteijn BAE, Balm AJM, Vermey A, Plukker JT, Hoefnagel CA, Piers DA, Kroon BBR. Sentinel node biopsy for melanoma in the head and neck region. Head Neck 1999 (in press). Kroon BBR, Jansen L, Rutgers EJT, Nieweg OE. Future of lymphatic mapping and sentinel node biopsy. In: Lymphatic mapping in cancer. Nieweg OE, Reintgen D, Thompson J, Essner R, editors. New York: Marcel Dekker, 1999 (in press). Liénard D, Eggermont AMM, Schraffordt Koops H, Kroon BBR, Towse G, Hiemstra S, Schmitz P, Clarke J, Steinmann G, Rosenkaimer F, Lejeune FJ. Isolated limb perfusion with TNFalpha and Melphalan with or without Interferon-y for treatment of in-transit melanoma metastases: a multicenter randomised phase II study. Melanoma Res 1999 (in press). Nieweg OE, Rutgers EJT, Jansen L, Valdés Olmos RA, Peterse JL, Hoefnagel CA, Kroon BBR. Is lymphatic mapping in breast cancer adequate and safe? World J Surg 1999 (in press). Vrouenraets BC, In t Veld GJ, Nieweg OE, Van Slooten GW, Van Dongen JA, Kroon BBR. Long-term functional morbidity after mild hyperthermic isolated limb perfusion with melphalan. Eur J Surg Oncol 1999; 25:

135 135 Local papers and theses Horenblas S, Efthymiou KM, Meinhardt W, Moonen LMF, Van Tinteren H. Totale prostatectomie goede mogelijkheid na Jodium- 125-behandeling of uitwendige radiotherapie. Ned Tijdschr Geneeskd 1999; 143: Nieweg OE, Sosef MN, Van Coevorden F. Wekedelensarcoom, diagnostiek: self assessment quiz. Ned Tijdschr Heelkd 1999; 8: 26, 29. SURGICAL ONCOLOGY Kroon BBR. Is een melanoom in de voorgeschiedenis een contraindicatie voor hormonale substitutie in verband met osteoporose? Vademecum 1998; 50. Kroon BBR, Bergman W, Coebergh JWW, Ruiter DJ. Consensus melanoom van de huid. Tijdschr Huisartsgeneeskd 1999; 6: Pameijer FA. Pre- and post-radiotherapy computed tomography in laryngeal and hypopharyngeal cancer. Imaging-based prediction of local control [dissertation]. Utrecht: University of Utrecht, Roeleveld T, Horenblas S, Meinhardt W, Van de Vijver M, De Vries F, Ten Bokkel Huinink WW. Wait-and-see-beleid bij non-seminoma testis stadium I. Ned Tijdschr Urol 1999; 7: 6-9. Rutgers EJT. Minder mutileren bij borstkanker: de schildwachtklier: de derde revolutie? In: Liber Amoricum M.A. Crommelin. Tilburg: Het inventief, 1999: Sosef MN, Nieweg OE, Van Coevorden F. Wekedelensarcoom, therapie: self assessment quiz. Ned Tijdschr Heelkd 1999; 8: 26, 29. Zoetmulder FAN. Wat te doen als de appendix bij onderzoek een carcinoïd van 2 cm bevat? Vademecum 1999; 17: 2a. Zoetmulder FAN, Van der Vange N, Witkamp AJ, Kaag MM, Boot H, Beijnen JH. Hypertherme intraperitoneale chemotherapie (HIPEC) bij patiënten met pseudomyxoma peritonei of peritoneummetastasen van colorectaal carcinoom; gunstige eerste ervaringen in het Nederlands Kanker Instituut. Ned Tijdschr Geneeskd 1999; 143: Local papers in press Kroon BBR. Van Lanschot JJB. Principes van kankerchirurgie. In: Veenhof CHN, Voûte PA, redactie. Behandeling van kanker. Houten/Antwerpen: Bohn/Stafleu/Van Loghum,1999 (in press). Neering H, Kroon BBR. Huidtumoren. In: Van de Velde CHJ Bosman FT, Wagner DJT, redactie. Oncologie, 6e druk. Houten/Antwerpen: Bohn/Stafleu/Van Loghum, 1999 (in press).

136 136 XII Division of Psychosocial Research and Epidemiology Division head Neil Aaronson Introduction The psychosocial research group is pursuing three primary research lines: 1) the development, testing and application of quality of life (QL) measures in clinical oncology research and practice; 2) assessment of the psychosocial consequences of genetic counseling and testing for hereditary forms of cancer; and 3) the development of methods for monitoring and managing symptoms related to cancer and its treatment. Preliminary results of a prospective, randomized study indicated that the routine provision of information derived from standardized QL assessments increased significantly (by approximately 20%) the frequency with which physicians discussed their patients physical and psychosocial functioning, and their symptom experience during outpatient palliative chemotherapy visits. Additionally, physicians provided with the QL summary data were more successful in identifying patients with moderate to severe functional impairments and symptoms. Our on-going research into the feasibility of employing proxy respondents for assessing cancer patients QL indicated that the spouses of patients with metastatic cancer are able to evaluate, with a relatively high degree of accuracy, patients physical and psychosocial functioning, symptom burden, and overall quality of life. Patient accrual continued for two companion studies on the impact of low-grade and high-grade glioma and its treatment on patients QL and neuropsychological functioning. Preliminary results indicated significant postsurgical QL and neuropsychological deficits among highgrade glioma in almost all QL domains assessed. More disturbing were the preliminary findings indicating significant QL and neuropsychological deficits among the mid-term to long-term survivors of low-grade glioma, compared with age- and gender-matched controls from the general population. In a retrospective, multicenter study we found that approximately 1 in 6 individuals at increased risk of developing colorectal cancer underwent preventive screening (colonoscopy or sigmoidoscopy) less frequently than had been advised. The most important reasons for non-adherence were the painful nature of the screening, and the belief that screening was not necessary due to the absence of symptoms. Data collection is continuing for a prospective, multicenter study of the effect of genetic counseling and testing for colorectal cancer on risk perception, levels of distress, family relationships, work, family and financial planning, and preventive health behavior. We have also initiated a cross-sectional, multicenter study of the physical and psychosocial impact of gynecological screening versus prophylactic oophorectomy among women from hereditary breast/ovarian (HBOC) families. Finally, in the area of symptom management and control, work continued on a prospective, multicenter study of the neuropsychological sequelae of high-dose, adjuvant chemotherapy in patients with high-risk breast cancer and high-grade lymphoma. Neurophysiological (qeeg) testing of a small sample of patients who had received high-dose chemotherapy in the past, indicated pathological asymmetry of the alpha rhythm. Finally, in a laboratory study carried out in collaboration with the University of Groningen, mice injected with a single, highdose of chemotherapy (CTC) were found to perform significantly less well on a maze test than a non-treated control group. These results suggest a diminished consolidation of information, possibly indicative of long-term memory deficits. The cancer epidemiology group is currently concentrating on two principal research lines: 1) the etiology of hormone-related cancers; 2) the long-term health consequences of cancer treatment, particularly in terms of the risk of developing a second cancer. Data collection is being carried out for a large case-control study which compares risk factors for ductal carcinoma in situ of the breast and unilateral (invasive) breast cancer. Further, we are conducting a nation-wide cohort study to examine whether women who underwent in vitro fertilization are at increased risk of ovarian cancer and other hormonerelated cancers. Using data from a large populationbased case-control study of breast cancer and oral contraceptives, we examined the association between physical activity and risk of breast cancer. We found that physically active women had a 30 percent lower risk of breast cancer than inactive women. In collaboration with the Molecular Pathology laboratory (LJ Van t Veer) and the Netherlands Collaborative Group on Hereditary Breast Cancer, the epidemiology group is also conducting a nation-wide prospective study of gene-environment inter-

137 actions in familial and hereditary breast and ovarian cancer. Since penetrance and age at onset vary between and within BRCA1 and BRCA2 families, the development of cancer may also be influenced by hormonal or lifestyle factors or other distinct genetic loci. With the increased survival for several malignancies, it has become exceedingly important to evaluate the potential carcinogenic effects of cancer treatment. The epidemiology group is therefore examining the long-term risk of second cancers and cardiovascular disease following treatment of Hodgkin s disease, breast cancer and testicular cancer. In 1999, analyses in a large group of Hodgkin s disease survivors demonstrated that the relative risk of solid tumors increased strongly with younger age at first treatment of Hodgkin s disease. This was not only observed for breast cancer but also for all other solid tumors. Reassuringly, the strongly increased solid tumor risks in patients who were very young ( 20) at first treatment appeared to decrease as these patients grow older. We are also examining, in collaboration with the Molecular Pathology laboratory (LJ Van t Veer), whether AT heterozygotes have an increased risk of developing radiation-induced second malignancies, especially breast cancer. In view of the large interest for tamoxifen as a chemopreventive agent, it is of great importance to better quantify its effect on endometrial cancer. In a series of 300 endometrial cancers following breast cancer, we found that the endometrial tumors of tamoxifen users had less favorable clinicopathological characteristics, and were more often p53-positive and estrogen-receptor negative. Psychosocial Research Assessing health-related quality of life (QL) in clinical research and clinical practice NK Aaronson PhD JH Schornagel MD PhD 1 M Hagedoorn PhD 2 M Klein PhD 3 SB Detmar MSc 4 R Hoopman MSc 5 KCA Sneeuw MSc 6 MJ Muller MSc WHJJ Cleyne 3 J Grit 3 WJM Oomen 3 LDV Wever 4 Group leader Academic staff Graduate student Graduate student Graduate student Statistical analyst Research assistant Research assistant Research assistant Research assistant The use of QL assessments in daily clinical oncology practice: an intervention study to facilitate doctor-patient communication in outpatient palliative care settings 4 This prospective, randomized (cross-over) study is evaluating the efficacy of incorporating standardized QL assessments as a routine part of the outpatient palliative treatment of cancer patients in terms of two primary outcomes: 1) facilitating doctor-patient communication; and 2) increasing physicians awareness of patients physical and psychosocial health problems. The study sample includes 10 medical oncologists and 214 of their patients treated with palliative chemotherapy. In the intervention group, graphic summaries of the patients QL (as assessed by the EORTC QLQ-C30) were generated and shared with both the patients and their physicians immediately prior to outpatient clinic visits. To evaluate doctor-patient communication, four consecutive medical consultations were audiotaped and content-analyzed using standardized computer software (the RIAS system). To evaluate physicians awareness, comparisons were made between physicians and patients ratings on the COOP/WONCA charts. Preliminary results indicate that patients level of physical and psychosocial functioning, and their symptom experience, was discussed significantly more frequently in the intervention than in the control group (on average, a 20% increase in QL topics discussed). Physicians in the intervention condition were more successful at identifying patients with moderate to severe problems in several QL domains (e.g. social functioning, pain, fatigue). All of the physicians and 75% of the patients in the intervention condition believed that the intervention facilitated communication and expressed interest in continued use of the procedure. The role of proxies in evaluating the QL of cancer patients 6 In this study, carried out in collaboration with the University of Connecticut Health Center (P Albertsen), patients with metastatic prostate cancer and their spouses independently completed the EORTC QLQ-C30 and a prostate cancer-specific questionnaire module assessing, in total, 21 QL outcomes. Systematic, but relatively minor differences in mean scale scores were noted for 5 of the 21 patient-proxy comparisons, with spouses rating the patients as having more QL impairments than the patients themselves. Moderate to good patient-proxy agreement was observed at the individual level (ICC s ranging from 0.40 to 0.75). The reliability and validity of spouses ratings were similar to that of the patients. These results support our earlier findings that proxies are able to evaluate, with a relatively high degree of accuracy, patients physical and psychosocial functioning, symptom burden, and overall quality of life. The QL and neuropsychological status of patients with high-grade and low-grade glioma 3, 7 In collaboration with the Departments of Neurology (JJ Heimans) and Medical Psychology (HM Van der Ploeg) of the Vrije Universiteit, Amsterdam, two companion studies are being conducted to assess: 1) the impact of low-grade and high-grade glioma and its treatment on the QL (as measured by the SF-36) of patients and their partners; and 2) the relationship between the patients neuropsychological status and their perceived QL. Newly diagnosed patients with high-grade 137 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY

138 138 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY glioma are assessed prospectively at 4-monthly intervals, along with a comparison group of patients with non-small cell lung cancer (NSCLC). Mid-term to long-term survivors of low-grade glioma are assessed at a single point in time, together with a comparison group of patients with low-grade hematologic malignancies. To date, 68 patients and 50 controls have been entered into the high-grade glioma study, and 185 patients and 23 controls in the low-grade glioma study. Normative data from the general population are also available for both the SF-36 and the neuropsychological test battery. Preliminary analyses of the baseline data from the high-grade glioma study indicate that the QL of glioma and NSCLC patients is similar, with both groups deviating significantly from normative data from the general population. Cognitive impairment, based on objective performance tests, was found in 49% of the glioma patients and in 26% of the NSCLC controls. Preliminary analyses comparing 145 low-grade glioma survivors with 145 age- and gender-matched general population controls revealed significant impairment in QL in the survivor group (on 7 of the 8 SF-36 scales). Additionally, significant neuropsychological deficits suggestive of a reduced information processing capacity were found in this survivor group. These findings indicate that survivors of low-grade glioma continue to experience significant QL and cognitive deficits for many years after their primary diagnosis and treatment. QL assessment among ethnic minority cancer patients in the Netherlands 5 In 1999, preparatory work was carried out for a large scale study whose primary aims are to translate, adapt and validate two widely used generic QL questionnaires (the SF-36 and the COOP/WONCA charts) and two cancer-specific questionnaires (the EORTC QLQ-C30 and the Rotterdam Symptom Checklist) for use among Turkish and Moroccan cancer patients. The study instruments have been translated into the requisite languages, and procedures have been developed for recruiting patients from regional hospitals. This project will facilitate inclusion of ethnic minorities in QL investigations, and will provide a wealth of descriptive information on the impact of cancer and its treatment on the functional health and well-being of these populations. Notes 1 Division X. 2 Fellow of the Dutch Cancer Society. 3 Funding: Dutch Cancer Society, Project VU/NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI External collaborators: JJ Heimans MD PhD, Academic Hospital Vrije Universiteit, Amsterdam; HM Van der Ploeg PhD, Department of Medical Psychology, Vrije Universiteit, Amsterdam. Publications: Quality of life Aaronson NK, Fayers P. Quality of life. In: Souhami RL, Tannock I, Hohenberger P, Horiot JC, editors. Oxford Textbook of Oncology, 2nd ed. Oxford: Oxford University Press (in press). Fosså SD, Curran D, Aaronson NK, Keuppens F, Kliment J, Robinson MRG, De Rijke TM, Hetherington J, Kil PJM, Rea LA. Quality of life of patients with newly diagnosed poor prognosis M1 prostate cancer undergoing orchiectomy without or with Mitomycin C. Eur J Urol (in press). Groenvold M, Fayers PM, Sprangers MAG, Bjorner JB, Klee MC, Aaronson NK, Bech P, Mouridsen HT. Anxiety and depression in breast cancer patients at low risk of recurrence compared with the general population - a valid comparison? J Clin Epidemiol 1999; 52 : Langendijk JA, Ten Velde GPM, Aaronson NK, De Jong JMA, Muller MJ, Wouters EFM. Quality of life after palliative radiotherapy in non-small cell lung cancer: a prospective study. Int J Radiat Oncol Biol Phys (in press). Middleton MR, Grob JJ, Aaronson NK, Fierlbeck G, Tilgen W, Seiter S, Goreo M, Amadal S, Cebon J, Coates A, Dreno B, Henz M, Schadendorf D, Kapp A, Weiss J, Fraass U, Statkevich P, Muller M, Thatcher N. Randomized phase III study of temozolomide versus dacarbazine in the treatment of patients with advanced, metastatic malignant melanoma. J Clin Oncol (in press). Sneeuw KCA, Aaronson NK, Sprangers MAG, Detmar SB, Wever LDV, Schornagel JH. Evaluating the quality of life of cancer patients: assessments by patients, significant others and nurses. Br J Cancer 1999; 81: Sprangers MAG, Aaronson NK, De Haes JCJM. Onderzoek naar de kwaliteit van leven van kankerpatiënten. In: Van de Velde CJH, Bosman FT, Wagener DJ, redactie. Oncologie, 6e dr. Houten: Bohn Stafleu Van Loghum (in press). Sprangers MAG, Te Velde A, Aaronson NK. The construction and testing of the EORTC colorectal cancer-specific quality of life questionnaire module (the QLQ-CR38). Eur J Cancer 1999; 35: Van der Zouwe N, Van Dam FSAM, Aaronson NK. Het gebruik van alternatieve kankertherapieën naast de reguliere behandeling. In: Van de Velde CJH, Bosman FT, Wagener DJ, redactie. Oncologie, 6e dr. Houten: Bohn Stafleu Van Loghum (in press).

139 Psychosocial issues in genetic counseling and testing for hereditary cancer NK Aaronson PhD I Kluijt MD 1 FH Menko MD PhD 1 BG Taal MD PhD 2 M Van Beurden MD PhD 3 S Verhoef MD 1 EMA Bleiker PhD 4 RM Rosenbrand MSc MA Gerritsma 5 LDV Wever 4 ALE Van Rens 1 G Wigbout 1 J Hollenstein Group leader Academic staff Academic staff Academic staff Academic staff Academic staff Graduate student Research assistant Research assistant Genetic nurse Genetic nurse Undergraduate student Screening adherence among individuals at high risk of developing colorectal cancer This retrospective, multicenter study is investigating adherence to screening advice among individuals at an increased risk of developing colorectal cancer who were counseled at one of three familial cancer clinics in Amsterdam (AvL, AZVU, or AMC) in the period Questionnaires were completed by 178 individuals (83% response rate). Preliminary results based on selfreport data indicate that approximately 16% of the sample underwent screening (colonoscopy or sigmoidoscopy) less frequently than had been advised. The most important reasons for non-adherence were the painful, unpleasant, and embarrassing nature of the screening, and the belief that screening was not necessary due to the absence of symptoms. No statistically significant differences between compliers and non-compliers were found as a function of sociodemographics, risk perception, or worries about cancer. Self-reported adherence behavior will be confirmed, where possible, by objective data obtained from medical chart audits. The psychosocial and behavioral impact of genetic counseling for colorectal cancer (CRC): a prospective, multicenter study 4, 6 The primary objectives of this prospective, multicenter (AvL, AZVU, AMC, RUG, and LUMC) study are to: 1) assess the effect of genetic counseling and testing on risk perception, levels of distress, family relationships, work, and family and financial planning; 2) identify risk factors for poor psychological adjustment to the counseling and testing outcome and for early withdrawal from the genetic counseling process; and 3) establish rates of short-term compliance with recommended screening practices. The study will include approximately 375 individuals. To date, 45 individuals have been entered into the study. The physical and psychosocial impact of gynecological screening or prophylactic oophorectomy among women from hereditary breast/ovarian (HBOC) cancer families 7 This cross-sectional, retrospective, multicenter (AvL, AZVU, AMC) study is investigating: 1) the decisionmaking process surrounding the choice of preventive health actions among women at increased risk of developing ovarian cancer; 2) the impact of screening versus prophylactic surgery on psychosocial well-being; 3) compliance with screening advice for those who opt for this form of prevention; and 4) the prevalence and severity of menopausal symptoms among women who opt for surgery, as well as the use and perceived benefit of hormone-replacement therapy. In 1999, preparatory activities were carried out, including identification of the study population, and selection and pilot-testing of the questionnaires and interview schedules. In total, approximately 375 women will be invited to participate in the study. Client satisfaction with family cancer clinics 5 This study is evaluating client satisfaction with the quality of care provided at three family cancer clinics in Amsterdam (AvL, AZVU, and AMC). All individuals who complete the genetic counseling process are asked to complete a questionnaire assessing experiences with and the perceived quality of the services provided, the role of the general practitioner in the genetic counseling process, and the impact of genetic testing on psychosocial well-being, family interactions, work, family planning, and insurance eligibility. In total, 300 individuals will be included in the study. To date, 118 individuals have been surveyed. Notes 1 Family Cancer Clinic, NKI/AvL. 2 Division X. 3 Division XI. 4 Funding: Dutch Cancer Society, Project NKI Funding: Clinical Genetics Foundation, Amsterdam (SKGA). 6 External collaborators: A Bröcker-Vriends MD PhD, Leiden University Medical Center; RH Sijmons MD, Academic Hospital Groningen; H Valdimarsdottir PhD, Mount Sinai School of Medicine, New York, USA. 7 External collaborators: J Van der Velden MD PhD, Academic Medical Center, University of Amsterdam; RHM Verheijen MD PhD, Academic Hospital Vrije Universiteit, Amsterdam; H Valdimarsdottir PhD, Mount Sinai School of Medicine, New York, USA. 139 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY

140 140 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY Publications: Genetic counseling Bleiker EMA, Aaronson NK. Genetic counseling for cancer: a family issue. In: Cooper CL, Baider l, editors. Cancer and the Family, 2nd ed. New York: John Wiley & Sons (in press). Bleiker EMA, Pouwer, F, Van der Ploeg HM, Leer JWH, Adèr HJ. Psychological distress two years after diagnosis of breast cancer: frequency and prediction. Patient Educ Couns (in press). Bleiker, EMA, Van der Linden MHM. Erfelijkheidsonderzoek bij kanker (2). MGV Maandbl Geest Volksgez 1999; 3: Bleiker, EMA, Van der Ploeg HM. Psychosocial factors in the etiology of breast cancer: a review of a popular link. Patient Educ Couns 1999; 37: Valdimarsdottir HB, Bovbjerg D, Brown K, Jacobsen P, Schwartz M, Bleiker EMA, Offit K, Borgen P, Heerdt A, Van Zee K. Cancerspecific distress is related to women s decisions to undergo BRCA1 testing. Cancer Res Ther Control 1999; 8: FSAM Van Dam PhD W Boogerd MD PhD 1 FJM Hilgers MD PhD 2 JF Jongkind MSc 3 S Rodenhuis MD PhD 1 CJ Van As MSc 2 SB Schagen MSc 4 MJ Muller MSc MEJ De Rond MSc AS Biervliet 4 BPC Kreukels 4 RM Rosenbrand 4 MCH Weevers 4 W De Ruyter 3 H Andersson HA Jansen ART Kalksma TA Overbeek DCM Van Rhijn A Vlietstra Symptom perception and management in cancer patients Group leader Academic staff Academic staff Academic staff Academic staff Academic staff Graduate student Statistical analyst Nursing scientist Research assistant Research assistant Research assistant Research assistant Research nurse Undergraduate student Undergraduate student Undergraduate student Undergraduate student Undergraduate student Undergraduate student Cognitive deficits as a consequence of high-dose chemotherapy among patients with high-risk breast cancer and high-grade lymphoma 4, 5 The primary objective of this study is to investigate cognitive function prior to and following (adjuvant) high-dose chemotherapy in patients with high-risk breast cancer and high-grade lymphoma. The neuropsychological status of the patients is being assessed at three points in time with a standard battery of tests. Patient accrual started in April To date, 428 test assessments have been carried out. In collaboration with the Department of Clinical Neurophysiology of the Slotervaart Hospital, neurophysiological tests were performed on a subgroup of 33 highrisk breast cancer patients who had participated in a randomized study comparing adjuvant high-dose (HD) chemotherapy with standard-dose (SD) chemotherapy. An additional control group included 13 high-risk breast cancer patients who had not received chemotherapy. The neurophysiological examination consisted of P300 and quantitative EEG including analysis of asymmetry and blocking of alpha-rhythm. Patients were tested two years after completion of treatment. Asymmetry of the alpha rhythm of 0.5 Hz was found in 7 patients of the HD group, in 2 patients of the SD group and in none of the control patients (p=0.02). An asymmetry of the alpha rhythm of >0.2 Hz is considered as pathological. The neurophysiological results did not correlate with the neuropsychological test results, nor with patients subjective complaints about their cognitive functioning. In collaboration with the Department of Clinical Chemistry and the Laboratory for Animal Physiology of the University of Groningen, the effects of CTC chemotherapy in FVB mice were studied. A significant difference was found between the treatment group and control group with regard to long-term learning strategies. In the treatment group, a diminished consolidation of information was observed, which could be indicative of longterm memory deficits. Empowering the cancer patient with chronic pain 6 Chronic pain is a major problem in cancer patients. Based on previous research in the AvL, an instruction manual for nurses is being developed and problem areas for pain management in the inpatient setting are being identified. Central to the program is instructing nurses in how they can train patients to monitor their pain complaints and to use pharmacological and non-pharmacological pain interventions effectively. This project is being carried out at the University Hospital Rotterdam and at the Antoni van Leeuwenhoek Hospital. Following initial evaluation, the program will be offered to other regional hospitals. Olfactory rehabilitation after total laryngectomy In cooperation with Division XI, a method to rehabilitate olfaction in laryngectomized patients has been developed. This Nasal Airflow Inducing Maneuver (NAIM), can be learned with 30 minutes of training. This technique could

141 be successfully learned by 57% of patients who could not smell due to a total laryngectomy. Further work is being carried out to investigate the impact of NAIM on olfaction in daily life, and to better understand why some patients are unable to learn and apply the technique. Notes 1 Division X. 2 Division XI. 3 Nursing department, NKI/AvL. 4 Funding: Dutch Cancer Society, Project NKI External collaborators: J Snel PhD, KR Ridderinkhof PhD, Department of Psychology, University of Amsterdam. 6 External collaborators: R De Wit PhD, Department of Medical Psychology and Psychiatry, Erasmus University Rotterdam; FE Witkamp, University Hospital Rotterdam. Publications: Symptom perception De Rond MEJ, De Wit R, Van Dam FSAM, Van Campen BTM, Den Hartog YM, Klievink RMA. A Pain Monitoring Program for nurses: effects on nurses pain knowledge and attitude. J Pain Symptom Manage (in press). Van Dam FSAM. Alternatieve behandelingen bestrijden de angst, niet de kanker, Oncologica 1999; 14: Van Dam FSAM. Meer gebruik van alternatieve diëten en van andere alternatieve behandelingen door kankerpatiënten: Houtsmuller is in, Moerman is uit. Ned Tijdschr Geneeskd 1999; 143: Van Dam FSAM. Psychosociale aspecten van kankerbehandeling: over kwaliteit van leven. In: CHN Veenhof, PA Voute, redactie. Behandeling van Kanker. Houten: Bohn Stafleu Van Loghum (in press). Van Dam FSAM. Psychosociale zorg van de kankerpatiënt. In: Van de Velde CJH, Bosman FT, Wagener DJTh, redactie. Oncologie, 6e ed. Houten: Bohn Stafleu Van Loghum (in press). Van Dam FSAM, Hilgers FJM, Emsbroek G, Touw FI, Van As CJ, De Jong N. Deterioration of olfaction and gustation as a consequence of total laryngectomy. Laryngoscope 1999; 109: Van Dam FSAM, Roodbergen R, Jongkind H, Loonstra S. Evaluatie van de kwaliteit van de verpleegkundige zorgverlening à la carte: het Amsterdams Modulair Kwaliteitsbewaking Systeem voor de Verpleegkundige Zorgverlening (AMKS-VZ). In: Handboek Verpleegkundig Konsult. Houten: Bohn Stafleu Van Loghum, PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY De Rond M, De Wit R, Van Dam F, Van Campen B, Den Hartog Y, Klievink R, Nieweg R, Noort J, Wagenaar M, Van Campen B. Daily pain assessment: value for nurses and patients. J Adv Nurs 1999; 29: Van der Zouwe N, Van Dam FSAM, Aaronson NK. Het gebruik van alternatieve kankertherapienn naast de reguliere behandeling. In: Van de Velde CJH, Bosman FT, Wagener DJTh, redactie. Oncologie, 6e ed. Houten: Bohn Stafleu Van Loghum (in press). De Rond MEJ, Van Dam FSAM, De Wit R. De kennis en attitude van verpleegkundigen en artsen ten aanzien van pijn en pijnbestrijding. Tijdschr Gezondheidswetenschappen 1999; 77: De Wit R, Van Dam F, Hanneman M, Zandbelt L, Van Buuren A, Van der Heijden K, Leenhouts G, Loonstra S, Huijer Abu-Saad H. Evaluation of the use of a pain diary in chronic cancer patients at home. Pain 1999; 79: De Wit R, Van Dam F, Huijer Abu-Saad H, Loonstra S, Zandbelt L, Van Buuren A, Van der Heijden K, Leenhouts G. Empirical comparison of commonly used measures to evaluate pain treatment in cancer patients with chronic pain. J Clin Oncol 1999; 17: De Wit R, Van Dam F, Vielvoye-Kerkmeer A, Mattern C, Huijer Abu-Saad H. The treatment of chronic cancer pain in a cancer hospital in the Netherlands. J Pain Symptom Manage 1999; 17: Renckens CNM, Van Dam FSAM. Het Koningin Wilhelmina Fonds en de Houtsmuller-therapie bij kanker. Ned Tijdschr Geneeskd 1999; 143: Schagen SB, Van Dam FSAM, Muller MJ, Boogerd W, Lindeboom J, Bruning PF. Cognitive deficits after postoperative adjuvant chemotherapy for breast cancer. Cancer 1999; 85:

142 142 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY Epidemiology Risk factors for hormone-related cancer FE Van Leeuwen PhD Group leader AAM Hart MSc 1 Academic staff I Kluyt MD 2 Academic staff FH Menko MD PhD 2 Academic staff JL Peterse MD 3 Academic staff MA Rookus PhD Academic staff EJT Rutgers MD PhD 4 Academic staff L Van t Veer PhD 3 Academic staff S Verhoef MD 2 Academic staff K Van der Kooy PhD 5 RM Brohet MSc 6 Graduate student EJ De Boer MSc 7 Graduate student H Klip MPH 8, 9 Graduate student J Verloop MSc 10 Graduate student S Braak 9 Research assistant PAM Braas 6 Research assistant AC Buitelaar 6 Research assistant BPG Maertzdorf 11 Research assistant MCJ Schippers 9 Research assistant AW Van den Belt-Dusebout 6, 9 Research assistant CM Versteegden 9 Research assistant MMPGM Woordes-van Baalen 6 Research assistant C Spronk 5 ALE Van Rens 2 Genetic nurse G Wigbout 2 Genetic nurse M Hollmann Undergraduate student J Tomas Undergraduate student M Van Vliet Undergraduate student Risk factors for breast cancer5, 6, 10, 11, 12 In 1999, we stopped the accrual of patients for our population-based case-control studies of risk factors for ductal carcinoma in situ of the breast (n=546), and invasive unilateral (n=1233) and contralateral breast cancer (n=348). An important aim of the study is to compare risk factor profiles for different forms of breast cancer, as defined by histopathologic characteristics and genetic alterations. For this purpose, we collected 1558 (85%) of the 1832 tumor blocks from 43 pathologylabatories. Immunohistochemical staining and scoring for C-ErbB2, p53, estrogen and progestagen receptor, Cyclin D1 and bcl2 is in progress. In 1999 we conducted further analyses with data from a large, population-based case-control study of oral contraceptive use and breast cancer. Breast cancer cases, aged at diagnosis (n=918), were pair-matched on age and region with controls randomly selected from municipal registries. We investigated the association between lifetime physical activity and risk of breast cancer. Women who were more active than their peers at ages had a 32% decreased risk of breast cancer. Women who had ever been engaged in recreational physical activity had a reduced risk of breast cancer compared to inactive women (relative risk (RR)=0.70, 95% CI= ). Neither very early recreational activity (before age 20) nor recent activity (last five years) was associated with a greater risk reduction than recreational activity in the intermediate period. Lean women appeared to benefit more from the protective effect of recreational physical activity than women with a higher body mass index. We continued the accrual of participants for our national genetic epidemiologic cohort study in breast and/or ovarian cancer families with the following aims: 1) to examine whether hormonal/life-style factors modify cancer risks in BRCA1/2 and non-brca1/2 families; 2) to assess in non-brca1/2 families the age-specific cumulative risks of breast, ovarian and other cancers based on full pedigree information. So far, data collection started in five out of the ten genetic centers in the Netherlands. Pedigree information of 348 families has been collected from clinical genetic records and 563 women out of 31 families have been approached to complete a mailed questionnaire on hormonal/life-style factors. So far, 445 questionnaires have been received. In total, the study will include 10,000 family members. Preliminary analyses were conducted to examine which pedigree characteristics were most predictive for finding a BRCA1/2 mutation. In families with both breast and ovarian cancer cases (n=67), BRCA1/2 mutations were found in 64.2% of the families, whereas this proportion was 17.4% in breast cancer-only families (n=167). In these families selected for breast and/or ovarian cancer, brain tumors were diagnosed more often than expected from age-, calendar period-specific incidence data of the general population. We examined cancer prevalence in women exposed to diethylstilbestrol (DES) in utero and registered at the Netherlands DES Information Center (NDESIC). Selfreported cancers were checked in the medical records. Compared to the general population, a more than fourfold increased prevalence rate of invasive cervical cancer was found among DES-daughters. Hyperstimulation of the ovary in relation to the risk of hormone-related cancers7, 8, 9, 13 We are conducting a large-scale cohort study to examine whether women who have undergone hyperstimulation for in vitro fertilization (IVF) are at increased risk of ovarian cancer and other hormone-related cancers. All IVF centers in the Netherlands are collaborating in this nationwide effort. A historical cohort of 20,000 subfertile women who received hyperstimulation for IVF between 1983 and 1995 was identified. As a control group 7,500 women were recruited whose subfertility was diagnosed

143 in the same hospitals, but who did not receive IVF treatment. In 1999, address information was updated for all cohort members and 25,000 of those received a risk factor questionnaire. The response rate is approximately 70%. For 15,000 participating women, medical data (cause of infertility, details of fertility treatment) have been collected from the medical records, including information on 45,000 IVF cycles. Every five years, the incidence of hormone-related cancers in the IVF group and the control group will be compared by linkage with the Netherlands Cancer Registry. Within this large cohort of subfertile women, we initiated a study examining determinants of age at menopause. Factors of interest are cause of subfertility, number of retrieved oocytes at IVF and IVF treatment. Notes 1 Division IX. 2 Family Cancer Clinic, NKI/AvL. 3 Division VIII. 4 Division XI. 5 Funding: Dutch Cancer Society, Project NKI Funding: Dutch Cancer Society, Project NKI Funding: International Health Foundation, Utrecht. 8 Funding: Prevention Fund, Project Funding: Ministry of Health. 10 Funding: Prevention Fund, Project Funding: International Agency for Research on Cancer. 12 External collaborator: DE Goldgar PhD, International Agency for Research on Cancer, Lyon, France. 13 External collaborators: CW Burger MD PhD, University Hospital Rotterdam; I Den Tonkelaar PhD, International Health Foundation Utrecht; ER Te Velde MD PhD, Academic Hospital Utrecht. Publications: Hormone related cancer Ligtenberg MJL, Hogervorst FBL, Willems HW, Arts PJW, Brink G, Hageman S, Bosgoed EAJ, Van der Looij E, Rookus MA, Devilee P, Vos EMAW, Wigbout G, Struycken PM, Menko FH, Rutgers EJ, Hoefsloot EH, Mariman ECM, Brunner HG, Van t Veer LJ. Characteristics of small breast and/or ovarian cancer families with germline mutations in BRCA1 and BRCA2. Br J Cancer 1999; 79: Rookus MA. De epidemiologie van mammacarcinoom, in het bijzonder in de menopauze. Bijblijven 1999; 15: Rookus MA. Is screening op borstkanker nodig bij oestrogeensuppletie? Huisartsen Vademecum (in press). Rookus MA. Invited commentary: reporting bias in case-control studies on induced abortion and breast cancer. Am J Epidemiol (in press). Verloop J, Rookus MA, Van der Kooy K, Van Leeuwen FE. Physical activity and breast cancer risk in women aged years. J Natl Cancer Inst (in press). Cancer risk following medical treatment FE van Leeuwen PhD BMP Aleman MD 1 MPW Gallee MD PhD 2 JL Peterse MD 2 NS Russell MD 1 L Van t Veer PhD 2 EAE Welp MPH 3 WJ Klokman MD MSc EK Adriaans 3 MLR Beelen 3 VB Hartog 3 AW Van den Belt-Dusebout Group leader Academic staff Academic staff Academic staff Academic staff Academic staff Graduate student Statistical analyst Research assistant Research assistant Research assistant Research assistant Long-term risk of second cancers following treatment for Hodgkin s disease, testicular cancer and breast cancer 3, 4 The specific aims of this research project are: 1) to evaluate the risk of second cancers (SC) following Hodgkin s disease (HD), breast cancer and testicular cancer over a period of up to 30 years, and to determine for which SC types risk is increased compared with the incidence in the general population; 2) to similarly evaluate the risk of cardiac death and acute myocardial infarction; 3) to quantify the separate and combined effects of different treatments on SC risk. We are conducting three cohort studies which include 2,650 5-year survivors of HD treated between , 2,200 5-year survivors of testicular cancer treated between 1966 and 1990, and 6,050 5-year survivors of breast cancer treated between 1970 and Patients included in these cohorts were treated in the Netherlands Cancer Institute, the Dr Daniel den Hoed Cancer Center/EUR, the University hospitals of Leiden and Groningen and hospitals participating in the Eindhoven Cancer Registry. In 1999 we collected treatment data on 1,600 breast cancer patients treated between 1970 and Further, we conducted a more extensive analysis as to the effect of treatment on the risk of second malignancies in 1,253 survivors of HD who were less than 40 years old at diagnosis of HD. In all, 137 patients developed a SC, compared with 19.4 cases expected on the basis of incidence rates in the general population (RR=7.0). Significantly increased RRs were observed for leukemia (RR=37.5), NHL (RR=21.5) and solid tumor overall (RR=6.1). The 25-year actuarial risk of SC overall was 27.7% and of breast cancer (in females) 16.3%. The RR of solid malignancies rose with longer follow-up duration. Among patients with 20 years of follow-up, the RR was 7.6-fold increased, slightly less than the RR of 9.1 observed in the year follow up period. The RR of solid tumors increased strongly with younger age at first treatment of HD. This was not only observed for breast cancer but also for all other solid tumors, with RRs of 4.9, 6.9 and 12.7 for women first treated at ages 31-39, 21-30, and 20, respectively. Reassuringly, the strongly increased solid tumor risks in patients who were very young (20) at first treatment appear to decrease as these 143 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY

144 144 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY patients grow older. Among patients treated at age 20 or before, the RR of developing a solid tumor before age 40 was significantly greater than the RR of solid tumor development at ages (RR=27.9 versus RR=4.2). In a related project we are examining to what extent ATM heterozygosity increases the risk of radiationassociated breast cancer. Two case control studies have been conducted on patients who developed breast cancer following radiotherapy. Patients and matched controls donated blood for analysis of germline mutations in the ATM gene using the PTT test and subsequent sequencing when there were abnormal results. We found that classical mutations in the ATM gene do not contribute to the strongly increased risk of breast cancer following HD. Detailed results are discussed in the report of Division XIII. Tamoxifen and second cancer risk 5 In the 1998 Scientific Report we described the first results of our nationwide case-control study (n=309) investigating the effects of tamoxifen on the risk and prognosis of endometrial cancer following breast cancer. We reported that the three-year endometrial cancerspecific survival was significantly worse for long-term tamoxifen users than for non-users. In 1999 we found that long-term tamoxifen users more often developed endometrial tumors with unfavorable clinicopathological characteristics (malignant mixed mesodermal tumors and sarcomas of the endometrium). Furthermore, the endometrial tumors of long-term tamoxifen users were more often estrogen-receptor negative than those of non-users (61% versus 26%), and were more often p53-positive than those of non-users (31% versus 18%). Late health effects of nasopharyngeal radium irradiation in childhood 6 In close collaboration with the Reinaert Clinic in Maastricht and the US National Cancer Institute we are conducting a nationwide cohort study among children who have undergone nasopharyngeal radium irradiation. The aim of the study is to examine the very long-term risks of head and neck cancer and disorders related to radiation-exposure of the thyroid and pituitary glands. From nine Dutch hospitals a historic cohort was constructed of 5,392 children who were treated with radium at an ear-nose-throat (ENT) ward between 1945 and The control group consists of 5,291 children who were treated in the same hospital for ENT-related disorders, but without radium. After completion of extensive follow-up procedures, a total of 624 (8.5%) deaths were observed. Cause of death information could be obtained for all but 2 deceased at Statistics Netherlands. Due to incomplete treatment data the final number of deceased persons eligible for analysis was 616, 302 among the irradiated group and 314 among controls. Among the irradiated group we observed 97 deaths due to malignant disease whereas 84 were expected based on gender-, age- and calendar periodspecific reference data, rendering an observed/expected (O/E) ratio of 1.15 (95% CI, ). Among controls, the O/E ratio was 1.01 (87/86.3). Analyses of tumorspecific mortality as well as direct comparisons of the exposed and non-exposed groups by survival analysis are underway. Furthermore, cancer incidence data were collected and will be analyzed similarly. Dosimetric data will soon be available for incorporation in further analyses. Notes 1 Division IX. 2 Division XIII. 3 Funding: Dutch Cancer Society, Project NKI External collaborators: MA Crommelin MD, Catharina Hospital, Eindhoven; JGM Klijn MD PhD, ADG Krol MD, MB Van t Veer MD PhD, Daniel den Hoed Cancer Center/Erasmus University Rotterdam; EM Noordijk MD PhD, Leiden University Medical Center; CL Land PhD, LB Travis MD PhD, Radiation Epidemiology Branch, National Cancer Institute, USA: M Stovall PhD, The University of Texas MD Anderson Cancer Center, USA. 5 External collaborator: J Benraadt MD, Comprehensive Cancer Center Amsterdam. 6 External collaborators: PG Verduijn MD PhD, CM Ronckers MSc, Reinaert Clinic, Maastricht. Publications: Cancer risk Broeks A, Russell NS, Floore AN, Urbanus JHM, Dahler EC, Van t Veer MB, Hagenbeek A, Noordijk EM, Crommelin MA, Van Leeuwen FE, Van t Veer LJ. Increased risk of breast cancer following irradiation for Hodgkin s disease is not a result of ATM germline mutations. Int J Radiat Biol (in press). Broeks A, Urbanus JHM, Floore AN, Dahler EC, Klijn JGM, Rutgers EJT, Devilee P, Russell NS, Van Leeuwen FE, Van t Veer LJ. Classical ATM germline mutations contribute to breast cancer susceptibility. Am J Hum Genet (in press). Metayer C, Clarke EA, Glimelius B, Storm H, Pukkala E, Van Leeuwen FE, Lynch CF, Curtis RE, Holowaty EJ, Wikstrand G, Andersson M, Wiklund T, Gospodarowicz M, Travis LB. Second cancer in long-term survivors of Hodgkin s disease diagnosed in childhood and adolescence. J Natl Cancer Inst (in press). Van Leeuwen FE. Does risk of endometrial cancer increase with longer duration of tamoxifen use? Eur J Cancer 1998; 34: S44-5. Van Leeuwen FE, Klokman WJ, Van t Veer MB, Hagenbeek A, Krol ADG, Vetter UAO, Schaapveld M, Van Heerde P, Burgers JMV, Somers R, Aleman BMP. Long-term risk of second malignancy in survivors of Hodgkin s disease treated during adolescence or young adulthood. J Clin Oncol (in press). Van Leeuwen FE, Swerdlow AJ, Valagussa P, Tucker MA. Second cancers. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, editors. Hodgkin s disease. Philadelphia: Lippincott Williams & Wilkins, 1999:

145 145 Secretary Psychosocial Research and Epidemiology M Van den Hoorn Research staff positions (full time equivalents) Scientific permanent: 3.6 Scientific project: 9.9 Technical permanent: 6.0 Technical project: 9.4 PSYCHOSOCIAL RESEARCH AND EPIDEMIOLOGY

146 146 XIII Division of Diagnostic Oncology Division head M Van de Vijver Head MJ Van de Vijver MD PhD Permanent academic staff APE Besnard MD, JMG Bonfrer PhD, D De Jong MD PhD, MPW Gallee MD PhD, CA Hoefnagel MD PhD, FBL Hogervorst PhD, W Koops MD, R Kröger MD, SH Muller PhD, PM Nederlof PhD, WJ Nooijen PhD, FA Pameijer MD PhD, JL Peterse MD, L Schultze Kool MD PhD, MJA Smid-Geirnaerdt MD PhD, RA Valdés Olmos MD PhD, LJ Van t Veer PhD, MLF Van Velthuysen MD PhD, S Verhoef MD Other academic staff N Bijker, E Deurloo, IF Faneyte, KGH Gilhuys, H Klaren, W Klein Zeggelink, I Kluijt MD PhD, FH Menko MD PhD, LMTh Sterk MD, N Verra, M Van Vliet Permanent technical staff D Atsma, LH Boerrigter-Barendsen, MS Boutmy-De Lange, G Brink, CPE Ottenheim, J Poodt, R Pruntel, R Regnerus, C Schippers-Gillissen, O Van Tellingen PhD, P Wisman Other technical staff T Van Welsem Guest C Hari Aparici MD Secretary SG Massink Research funded positions (full time equivalent) Clinical permanent: 16.6 Scientific Permanent: 3.1 Clinical projects: 5.6 Scientific projects: 0.8 Introduction The Division of Diagnostic Oncology comprises the Departments of Clinical Chemistry, Nuclear Medicine, Pathology and Radiology. The Department of Radiology is involved in an increasing number of collaborative research projects. There is collaborative work with the universities of Malmo and Leiden to develop a new percutaneous method of isolated liver perfusion as a treatment for liver metastases for colorectal cancer. National collaboration projects include: evaluation of the value of MRI (and dynamic MRI) in the diagnosis and follow-up of soft tissue tumors; dynamic MRI screening in women with an increased risk to develop breast cancer; assessment of the diagnostic value of stereotactic biopsy of occult breast lesions. In-house collaborative projects include the treatment of head and neck malignancies with intra-arterial chemotherapy and radiotherapy (RADPLAT); quantitative assessment of treatment margins and implementation of multimodality imaging in breast-conserving therapy. All these projects were started over the past two years, and the first results can be expected next year. Within the department of Pathology, much effort is being put into building a strong translational research program. As part of the Markerstudy project, the collection of DNA and mrna from well characterized tumours has continued. N Bijker, working on the analysis of a large clinical trial together with the EORTC Breast Cancer Cooperative Group, has found that radiotherapy improves local recurrence free survival in breast conserving treatment of ductal carcinoma in situ (DCIS) of the breast. There have been site visits to the centers participating in this clinical trial and the histology of the DCIS cases has been reviewed. This has resulted in very high quality of the clinical and pathological data for this trial. Ongoing research on 131 I-MIBG in the diagnosis and treatment of various tumor types continues to result in refinements to diagnostic and therapeutic regimens involving this compound. The sentinel node procedure, an important clinical research area at our institute, is very rapidly turning into a standard clinical procedure. This will result in great benefit for patients with breast, vulva and penis cancer and malignant melanoma; standard lymph node dissection can now be reserved for only those patients with metastatic disease.

147 Department of Clinical Chemistry P Baas 1, H Bardelmeijer, CGJ Cleypool, HM De Feij -de Graaf, BM Jansen, L Jansen 2, CM Korse, TC Linders, OE Nieweg 3, M Ouwehand, M Rozenhart, JH Schouwink 1, Y Souwer, HR Van der Woude, O Van Tellingen, E Van Zandbergen, JH Beijnen 1, JMG Bonfrer, WJ Nooijen Pharmacological studies in mice We have shown that the nonlinear pharmacokinetic behavior of paclitaxel in mice and patients is a pseudononlinearity caused by Cremophor EL. This probably occurs through the formation of micelles which entrap paclitaxel in the plasma compartment. The linear pharmacokinetic behavior of docetaxel can be explained by the fact that the excipient polysorbate 80 is unable to form micelles because of rapid degradation in plasma. In line with studies in mice, the use of the P-glycoprotein blocker cyclosporin A results in an enhanced oral bioavailability of paclitaxel in patients. Cyclosporin A is also able to enhance the oral bioavailability of docetaxel in mice and clinical studies have been initiated. In spite of its apparent activity in other studies the compound R was unable to increase the oral bioavailability of paclitaxel in mice and consequently further clinical testing was not attempted. In contrast, GF was able to enhance the oral bioavailability of paclitaxel to the level observed in P-glycoprotein deficient mice. Moreover, this compound appears to be a very selective blocker of P-glycoprotein, since its co-administration to P-glycoprotein deficient mice did not change the pharmacokinetic handling of paclitaxel. Clinical studies with GF have been initiated. The clinical studies with oral paclitaxel and cyclosporin A indicate a reduced oral bioavailability of paclitaxel at higher dose levels. Studies in mice suggest that this may depend on the amount of Cremophor EL in the oral formulation. In P-glycoprotein deficient mice, only about 5% of an oral dose of 10 mg/kg of paclitaxel was recovered in the feces indicating almost complete gastrointestinal uptake. However, the fecal recovery of paclitaxel increased substantially when the amount of Cremophor EL increased by 6-fold. Similar to the findings in plasma, Cremophor EL is thought to form micelles, which entrap paclitaxel, thus reducing the fraction of paclitaxel available for absorption. P-glycoprotein in the blood barrier barrier protects the brain against the penetration of many toxic substances. The usefulness of various (presumed) P-glycoprotein blockers for increasing the brain penetration of paclitaxel has been tested. Blocking of P-glycoprotein in the BBB has proven to be very difficult. Only very effective blockers, such as GF120918, were able to partially block P-glycoprotein in the BBB. Tumor markers Using immunohistochemistry, S-100 protein can be demonstrated in malignant melanoma cells. It was later demonstrated that the S-100 protein was also present in serum and a S-100B specific commercial test has been developed. Previous studies have shown that the serum level of S-100B is an independent prognostic factor for survival in patients with malignant melanoma. Increasing marker levels can predict recurrence of disease before clinical symptoms have developed. The evaluation of response to therapy in lung carcinoma may sometimes be difficult. Cyfra 21.1 and Tissue Polypeptide Antigen (TPA) were previously evaluated. A newly developed assay, using mabs to detect specific epitopes of cytokeratin 19 fragments, was tested for its sensitivity and specificity in non small cell lung cancer. It was demonstrated that the effectiveness of this marker, in terms of sensitivity/specificity, was no greater than TPA and Cyfra Notes 1 Division X. 2 Division IX. 3 Division XI. Department of Nuclear Medicine E Bais 2, H Boot 2, JH Borger 1, O Dalesio 4, J De Kraker 5, L Jansen 3, M Kooi 4, BBR Kroon 3, C Mari Aparici, OE Nieweg 3, BR Pieters 1, EJTh Rutgers 3, BG Taal 2, P Tanis 3, WW Ten Bokkel Huinink 2, PA Voûte 5, CA Hoefnagel, SH Muller, RA Valdés Olmos 131I-MIBG therapy of neural crest and other tumors For the purpose of pooling European results of 131I-MIBG (meta-iodobenzylguanidine) therapy in neural crest tumors, the data of 208 patients treated with 131I-MIBG at the Netherlands Cancer Institute since 1984 were reviewed. Objective response rates in malignant pheochromocytoma and paraganglioma were 43% and 57%, respectively. 140 children with neuroblastoma were treated with 131I-MIBG and 3 clinical situations can be distinguished. 1) In 59 patients receiving 131I-MIBG therapy after conventional treatment had failed, the objective response rate was 52.7% and palliation was provided to most patients; 5 of these patients survived >5 years; hematological toxicity occurred in 82.5%. 2) 27 patients with residual tumor or first recurrence after chemotherapy received 131 I-MIBG in combination with oxygen under hyperbaric conditions. Partial remission was attained in 41.7% and hematological toxicity occurred in 90%. 3) In contrast, 54 patients with inoperable stage III or IV neuroblastoma treated with 131 I-MIBG at first diagnosis instead of combination chemotherapy, had an objective response rate of 81.5% with considerably less hematological toxicity (37.3%); 11 of 30 patients with >5 years follow up are alive (longest survival >9 years). In 48 patients with symptomatic, metastatic carcinoid 147 DIAGNOSTIC ONCOLOGY

148 148 DIAGNOSTIC ONCOLOGY tumors, no objective responses and minimal toxicity was recorded; however palliation with a mean duration of 8 months was attained in 60% of the patients. In addition, 2 of 5 patients with medullary thyroid carcinoma had a partial remission, while 4 had a palliative response to 131I-MIBG therapy. In patients with non-neuroendocrine tumors who received escalating doses of unlabeled MIBG in the context of a phase I study, scintigraphy using 131 I-MIBG did not reveal tumor uptake in any of these tumours, confirming the specificity of this agent for neural crest tumours. The influence of unlabeled MIBG on the biodistribution of 131 I-MIBG was studied. Radioimmunoscintigraphy of radioimmunotherapy In a pilot study of radioimmunoscintigraphy of neuroblastoma using 131 I-labeled chimeric chce7 antibodies, encouraging results were obtained in 7 patients, demonstrating the complementarity of this targeting modality to 131I-MIBG imaging. In collaboration with the Paul Scherrer Institute (PA Schubiger and I Novak-Hofer) a protocol was submitted for a comparative study of combined diagnostic scintigraphy using 131 I-chCE7 and 131 I-MIBG in selected cases of neuroblastoma. The multicenter phase I clinical study on radioimmunotherapy of medullary thyroid cancer using the two-step targeting technique with bispecific immunoconjugate and 131 I- labelled bivalent hapten (M96MTC) was concluded and the promising results in 24 patients were published. A phase II study is in preparation. A working party was created to explore the possibilities of starting up radioimmunotherapy of recurrent non-hodgkin lymphoma using radiolabeled anti-cd20 antibodies. Dosimetry of 131 I-therapy in differentiated thyroid carcinoma The EU-project (F14CCT960009) on improvement of dosimetry for 131 I-therapy of intrathoracic tumors was concluded in June In patients with differentiated thyroid carcinoma and neural crest tumors, treated with 131I-iodide or 131 I-MIBG respectively, absorbed radiation in pulmonary metastases was calculated, at the University of Wuerzburg, after calibration of the gamma cameras of participating centers. A follow up study is in preparation. Sentinel Node Lymphoscintigraphy The findings of mammary lymphoscintigraphy were evaluated in 150 consecutive patients with breast carcinoma: 100 patients (group A) investigated in the validation phase of the study, and 50 (group B) after the dose of 99m Tc nanocolloid was optimized. In group A (mean dose 61.6 MBq, range MBq) scintigraphy visualized lymph nodes in 83 patients (83%) with an increase in the visualization rate from 72% for the first 40 patients to 90% for the last 60. Only patient age (p=0.01) and administered tracer dose (p=0.04) were found to be significant factors for visualization in a multifactorial analysis. In group B (mean dose 90.8 MBq, range MBq) the visualization rate was 94%. In combination with intraoperative blue dye mapping and gamma probe the overall identification rate was 90% for group A and 98% in group B. A phantom study, performed to evaluate scatter suppression in sentinel node detection, showed that for superficial nodes the gamma probe should be fitted with a shield and for deep nodes with a collimator. 111In antimyosin in anthracycline related cardiotoxicity A prospective longitudinal study was started in cooperation with the University Hospital Sant Pau of Barcelona (N16 AMBAR), to determine the value of 111 In antimyosin in the detection of early subclinical anthracycline cardiotoxicity and the prediction of late cardiac events. In 24 patients who completed treatment, 111 In antimyosin cardiac uptake, as measured by heart-to-lung ratios, was significantly increased (p=0.0001) at low cumulative dose ( mg/m 2 ) and subsequent dose levels, whereas left ventricle ejection fraction and peak filling rate were both only significantly changed at high cummulative dose levels ( mg/m 2 ). These results suggest that anthracycline related myocyte damage precedes systolic and diastolic cardiac function alterations even at low cumulative doses. Internal mammary chain A computer assisted method to measure the lateralization of internal mammary lymph nodes was developed. In 100 consecutive patients both analog and digital measurements of two 57 cobalt point sources positioned on the jugular notch and xiphoid process correlated well (r=0.9). In approximately 70% of the patients the most lateral lymph node was located less than 4 cm from the sternal midline. In cooperation with Division IX, this computer assisted method will be used routinely in patients to adjust the irradiation field. Diagnosis of pulmonary embolism In cooperation with the Slotervaart Hospital, the Department of Nuclear Medicine has participated in a multicentre study on optimizing diagnostic strategies for pulmonary embolism (COG nr. D94-90). A total of 627 consecutive patients clinically suspected for pulmonary embolism were included and in 517 of them a diagnostic conclusion was reached. With respect to patients with segmental or larger embolism and a high probability perfusion/ventilation scan, the sensitivity of spiral CT was almost 90%, whereas in subsegmental embolism this decreased to 21%. For lung ventilation, 99m Tc-technegas was compared with Krypton-81, showing a relative low concordance (kappa 0.69) and limited clinical value. Notes 1 Division IX. 2 Division X. 3 Division XI. 4 Biometrics Department. 5 Emma Kinderziekenhuis/AMC.

149 Department of Pathology Genetic alterations in tumors D Atsma, N Bijker, O Dalesio 2, L Duchateau 8, IS Fentiman 6, M Gallee, FBL Hogervorst, J-P Julien 7, J Poodt, EJTh Ruthers 5, C Schippers-Gillissen, B Taal 3, N Ter Haar 1, M Van Beurden 4, JA Van Dongen 5, ACM Van Lindert 4, L Van Velthuysen, T Van Welsem, CJ Vos 1, D De Jong, PM Nederlof JL Peterse, MJ Van de Vijver, LJ Van t Veer 149 DIAGNOSTIC ONCOLOGY Marker study The purpose of this project is to establish well-characterized series of various tumor types, for which DNA, mrna and extensive clinicopathological information is available. These series are available for testing of expression pattern or mutational status of novel genes in association with pathological and clinical parameters. Seven series have been prepared: breast carcinoma stage I and II; breast carcinoma stage III and IV; malignant melanoma (mainly lymph node metastases); renal cell carcinoma; malignant mesothelioma; gastric carcinoma; malignant lymphoma. Future series will include squamous cell carcinoma of the head and neck; colon carcinoma; lung carcinoma; and gynecological malignancies. Genetic alterations specific for breast and ovarian carcinomas The purpose of this project is to characterize the (somatic) genetic alterations in breast and ovarian carcinomas of patients with BRCA1 germline mutations and compare these with genetic alterations in sporadic breast carcinomas. We have used comparative genomic hybridization (CGH) for the detection of genetic alterations in the following set of tumors: 20 breast carcinomas from BRCA1 germline mutation carriers; 29 breast carcinomas from patients with bilateral breast cancer (for 12 patients Figure XIII.1 Deletion of several regions on chromosome 4 in breast cancer. 49 Breast carcinomas were analyzed by CGH; the frequency of loss of various regions on chromosome 4 is depicted. It appears that there are four distinct regions that are likely to harbor a tumor suppressor gene. Figure XIII. 2 Model for the role of genetic alterations in the development of invasive breast cancer. There are at least four distinct pathways leading to different histologic types of breast cancer. Each of these pathways is characterized by specific genetic alterations. ILC = invasive lobular carcinoma; IDC = invasive ductal carcinoma; LCIS = lobular carcinoma in situ; DCIS = ductal carcinoma in situ. the tumors from both breasts were analyzed); 10 breast carcinomas from patients that also developed ovarian carcinoma. Our first analysis indicates that the profile of genetic alterations in breast carcinomas from patients with BRCA1 germline mutations differ, with respect to the frequency with which gains and losses are observed, from patients with sporadic breast carcinoma. It has been reported that losses on chromosome 4 are more frequent in carcinomas from patients with BRCA1 germline mutations. We have focused on losses on chromosome 4 and identified four regions with frequent loss (Figure XIII.1). We do not find differences between sporadic and BRCA1 associated carcinomas with respect to losses on chromosome 4. In collaboration with Leiden University Medical Center we have analyzed genetic alterations in a large series of ductal carcinomas in situ (DCIS) of the breast without an invasive component. By comparing these with the genetic alterations known to be present in invasive breast cancer, we have designed the model for breast cancer development (Figure XIII.2). In brief: inactivation of E- cadherin is involved in the development of all cases of lobular carcinoma in situ and is never found in the development of cancers of ductal type. Inactivation of a, not yet identified, tumor suppressor gene on chromosome 16q is involved in the development of well differentiated DCIS; amplification of the HER2/neu gene and mutational inactivation of the p53 gene are involved in the development of poorly differentiated DCIS. It is not known which genetic alterations play a role in the progression of carcinoma in situ to invasive breast cancer. The identification of these genetic alterations is subject of ongoing studies presented elsewhere (Division VIII).

150 150 DIAGNOSTIC ONCOLOGY Ductal carcinoma in situ of the breast: conserving treatment: first results of EORTC trial A randomized phase III clinical trial has been conducted by the EORTC Breast Cancer Cooperative Group, to investigate the role of radiotherapy in breast conserving treatment of DCIS. Women (n=1,010) with ductal carcinoma in situ (DCIS) of the breast were treated by complete local excision of the lesion and then randomly assigned to either no further treatment or to radiotherapy. At a median follow-up of 51 months, the 4- year local relapse-free interval was 84% in the group treated with local excision alone, compared with 91% in the women treated by local excision plus radiotherapy (p=0.005, HR 0.62). Long-term follow-up will be needed to determine whether improved local control leads to a reduction in distant metastases and deaths from breast cancer. Of the cases randomized, histologic sections from 845 (84%) patients have been reviewed; in 694 cases the diagnosis DCIS was confirmed. In 44 cases a benign lesion was found; in 62 patients the lesion was considered atypical ductal hyperplasia rather than DCIS, in 26 cases a (micro)invasive lesion was found, and in another 12 cases there was suspicion of invasion. At a median follow-up of 51 months only one event was seen in the group of ADH. The DCIS subtype was not significantly related to the risk of local recurrence. However, poorly differentiated DCIS had a significantly higher rate of distant metastases than did the other subtypes (p=0.01). In a separate study, the diagnostic and therapeutic procedures which were followed in the trial were evaluated. It emerged that large variations occurred, particularly with respect to the surgical procedures and histopathological work-up. Important risk factors like tumor size and margin status were poorly quantified in the medical files. These findings emphasize the need for establishing uniform guidelines for diagnostic and therapeutic procedures for DCIS, and for clearly defined risk factors for recurrence after BCT for DCIS. Notes 1 Leiden University Medical Center, Funding: Dutch Cancer Society, Project RUL Biometrics Department. 3 Division X. 4 Division XI (Gynecology). 5 Division XI (Surgery). 6 Guy s Hospital, London. 7 Centre Henri Becquerel, Rouen. 8 EORTC Data Center, Brussels. Lymphoma LH Boerrigter-Barendsen, MS Boutmy-de Lange, T Dellemijn, FA Vyth-Dreese, H Boot, D De Jong H.pylori eradication is generally accepted as the first choice of treatment for stage I low-grade gastric MALT lymphoma. Treatment failure may be attributed to the extent of the disease and to progression into an antigenindependent phase. Morphological and functional immunohistochemical markers were investigated to identify this transition in 23 consecutive low-grade MALT lymphoma patients treated with H.pylori eradication in the NKI/AvL. Complete regression after H.pylori eradication was achieved in 13/23 patients (56%), partial regression in 2 patients (9%) and no response in 8 patients (35%). Histological grading was highly predictive of clinical response, especially in stage I patients, with complete remissions in 10/12 tumors with purely low-grade (type A) morphology and 1/8 tumors with increased numbers of blasts (type B) (p=0.0046). Histological grading was also related to the expression of co-stimulatory markers (p=0.0061). CD86 as a single marker proved to be of predictive value for treatment outcome (p=0.0086). These tumor-related factors should be taken into account in the decision to treat gastric MALT-NHL patients with H.pylori eradication alone or to add classical anti-cancer modalities early in the course of follow-up. Since variations in diagnostic criteria and staging procedures in gastric MALT have important consequences for patient selection and strongly bias the outcome of clinical trials, we performed a survey on the management of this disease among 19 leading institutes with a special interest in this field in Europe, the United States and Japan. Minimum histological criteria varied among pathologists with a notable influence of the classification system used in the different countries. Detailed evaluation of the lymphoma distribution in the gastric wall and routine staging of the GI-tract differed between groups lead by medical oncologists and gastro-enterologists. Similar effects were recorded for the role of gastric resection and radiotherapy. The Netherlands Cancer Institute Family Cancer Clinic LH Boerrigter-Barendsen, G Brink, D Hahn, FBL Hogervorst, E Kaats, H Klaren, I Kluijt 2, FH Menko 3, CPE Ottenheim, R Pruntel, R Regnerus, EJTh Rutgers4, M Van Beurden 1, A Van Rens, S Verhoef, G Wigbout, LJ Van t Veer Since the start of the Family Cancer Clinic in 1995, over 500 families (approximately 1,400 individuals) have sought genetic advice in our hospital. DNA testing was performed in 310 breast cancer families (BRCA1/BRCA2) and in 130 colon cancer families (microsattellite instability,

151 mismatch repair genes: hmlh1/hmsh2/hmsh6). Moreover, DNA testing is carried out for approximately 200 families counseled at the Academic Medical Hospital, Amsterdam. Germline mutations in one of the cancer susceptibility genes were found in 82 breast cancer families and in 8 colon cancer families. In addition, for 40 colon cancer families genomic instability in the tumor tissue (indicating the presence of mismatch repair deficiency) was determined and used as a marker for further genetic investigation and counseling. Our evaluation of probabilities of finding a DNA mutation revealed that positive test results can be found in 25% of the families tested. Specific family characteristics raise this chance. In 1999 special attention was given to the psychosocial need of families in counseling. As part of a pilot study, all families counseled were offered standard psychosocial care and over 80% of the participants were in favor of this support in addition to the genetic counseling itself. It is well known that cancer susceptibility varies between and within families with a hereditary predisposition. In a nationwide cohort study (study-coordinators from the Department of Epidemiology and the Family Cancer Clinic, see Division XII), gene-environment interactions, including hormonal and life-style factors, will be examined in 1,400 high-risk breast and/or ovarian cancer families. More specific risk estimates will also be assessed for families stratified according to their extent of genetic predisposition. Clinical outcome of treatment of BRCA1/BRCA2 breast cancer patients will be evaluated in a case-control study for breast conserving therapy and in a case-control study for contralateral breast cancer development after radiotherapy. Results will be essential for designing optimal therapeutic strategies for BRCA1/BRCA2 carriers. For pre-symptomatic BRCA1 and BRCA2 carriers, important information might be generated that can be used in decision making between surveillance by mammography or bilateral prophylactic mastectomy. carcinoma [n = 30] and T1 - T4 supraglottic carcinoma [n = 29]) treated with definitive RT. On the pretreatment CT study, each tumor was assigned a high- or low-risk profile for local failure after RT. The post-rt CT examinations were evaluated for post-treatment changes using a three-point post-rt CT-score: 1 = expected post-rt changes (Figure XIII.3); 2 = focal mass with a maximal diameter of < 1cm and/or asymmetric obliteration of laryngeal tissue planes; 3 = focal mass with a maximal diameter of > 1cm, or < 50% estimated tumor volume reduction (Figure XIII.4). The local control rates at 2 years post-rt based on pre-treatment CT evaluation were 88% for low pre-treatment risk profile patients and 34% for high pretreatment risk profile patients (p = ). 151 DIAGNOSTIC ONCOLOGY Notes 1 Division XI. 2 Department of Clinical Genetics, Academic Medical Center, UvA-Amsterdam. 3 Department of Clinical Genetics, Free University Hospital, AZVU-Amsterdam. 4 Division XI. Department of Radiology R Hermans, PS Kubilis, AA Mancuso, WM Mendenhall, JT Parsons, SP Stringer, H Van Tinteren, FA Pameijer 1 Pre- and post-radiotherapy (RT) computed tomographic (CT) studies were used to predict local control in 59 patients with laryngeal carcinoma (T3 glottic Figure XIII. 3 CT images of a patient with a T3 carcinoma of the right true vocal cord and a low pretreatment risk profile. A. Pretreatment CT image at the true cord level shows involvement of the entire right cord, which is a paramedian position. The ipsilateral paraglottic space is obliterated (large arrow), when compared to the contralateral side (small arrow). B. CT image obtained 4 months post-rt at the same level. There is complete resolution of the tumor. The laryngeal tissues appear symmetric, and the paraglottic space on the right side has reappeared (arrow); post-rt CT-score 1. The patient is without evidence of disease 31 months after completion of radiotherapy.

152 152 DIAGNOSTIC ONCOLOGY Based on post-treatment CT, the local control rates at 2 years post-rt were 94% for score 1, 67% for score 2 and 10% for score 3 (p= ). Post-RT CT-scores added significant information to the pre-treatment risk profiles on prognosis. We conclude that pre-treatment CT risk profiles, as well as post-rt CT evaluation, can identify patients irradiated for laryngeal carcinomas at high risk for developing local failure. When the post-rt CT-score is available, it proves to be an even better prognosticator than pretreatment CT risk profile. Note 1 Funding: Dutch Cancer Society Fellowship. Publications: Division XIII Full papers Awada A, Punt CJA, Piccart MJ, Van Tellingen O, Van Manen L, Kerger J, Groot Y, Wanders J, Verweij J, Wagener DJT. Phase I study of Carzelesin (U-80,244) given (4-weekly) by intravenous bolus schedule. Br J Cancer 1999; 79: Bonfrer JMG, Korse CM. TPA and CA 15.3 measurements for breast cancer monitoring in a routine setting. Int J Biol Markers 1999; 14: Boven E, Jansen WJM, Hulscher TM, Beijnen JH, Van Tellingen O. The influence of P170-glycoprotein modulators on the efficacy and the distribution of vincristine as well as on MDR1 expression in BRO/mdr1.1 human melanoma xenografts. Eur J Cancer 1999; 35: Bijker N, Rutgers EJT, Peterse JL, Van Dongen JA, Hart AAM, Borger JH, Kroon BBR Low risk of locoregional recurrence of primary breast carcinoma after treatment with a modification of the Halsted radical mastectomy and selective use of radiotherapy. Cancer 1999; 85: Chatal JF, Hoefnagel CA. Radionuclide therapy. Lancet 1999; 354: Clahsen PC, Van de Velde CJ, Duval C, Pallud C, Mandard AM, Delobelle-Deroide A, Van den Broek L, Van de Vijver MJ. The utility of mitotic index, oestrogen receptor and Ki-67 measurements in the creation of novel prognostic indices for node-negative breast cancer. Eur J Surg Oncol 1999; 25: De Jong D, Aleman BMP, Taal BG, Boot H. Controversies and consensus in the diagnosis, work-up and treatment of gastric lymphoma: an international survey. Ann Oncol 1999; 10: Baars JW, De Jong D, Willemse EM, Gras L, Dalesio O, Van Heerde P, Huygens PC, Van der Lelie H, Von dem Borne AEGK. Diffuse large B-cell non-hodgkin lymphomas: the clinical relevance of histological subclassification. Br J Cancer 1999; 79: Figure XIII. 4 CT images of a patient with a T3 carcinoma of the left true vocal cord and a high pretreatment risk profile. C. Pretreatment CT image just above the false cord level demonstrates involvement of the left aryepiglottic fold and obliteration of the paraglottic space (arrow). D. CT image at the same level, obtained at 3 months post-rt. A focal mass, in the left aryepiglottic fold, with a maximal diameter slightly over 1 cm (arrowheads) has developed; post-rt CT-score 3. Salvage laryngectomy confirmed the presence of local recurrence. The patient is alive without evidence of disease 2 years, 3 months after the operation. Elkhuizen PH, Voogd AC, Van den Broek LC, Tan IT, Van Houwelingen HC, Leer JW, Van de Vijver MJ. Risk factors for local recurrence after breast-conserving therapy for invasive carcinomas: a case-control study of histological factors and alterations in oncogene expression. Int J Radiat Oncol Biol Phys 1999; 45: Hoefnagel CA, Clarke SEM, Fischer M, Chatal JF, Lewington VJ, Nilsson S, Troncone L, Vieira MR. Radionuclide therapy practice and facilities in Europe. Eur J Nucl Med 1999; 26: Hoefnagel CA, Moonen LMF, Haustermans KMG, Balm AJM. Imaging of tumour hypoxia in nuclear medicine. In: Van der Kleij AJ, Voûte PA, Sminia P, editors. Oxygen and onco therapy 98; Vol

153 Archimedes Hyperbarmedizin und Verlags-GmbH, 1998, Hoefnagel CA. Therapy of neuroblastoma. The Update 1999; 6: Kraeber-Bodéré F, Bardet S, Hoefnagel CA, Vieira MR, Vuillez JP, Murat A, Ferreira TC, Bardiés M, Ferrer L, Resche I, Gautherot E, Rouvier E, Barbet J, Chatal JF. Radioimmunotherapy in medullary thyroid cancer using bispecific antibody and iodine-131-labeled bivalent hapten: papreliminary results of a phase I/II clinical trial. Clin Cancer Res 1999; 5: 3190s-8s. Kremer LCM, Tiel-van Buul MMC, Offringa M, Ottenkamp J, Valdés Olmos RA, Voûte PA. Indium-111-antimyosin scintigraphy in the early detection of heart damage after anthracycline therapy in children. J Clin Oncol 1999; 17: Ligtenberg MJL, Hogervorst FBL, Willems HW, Arts PJW, Brink G, Hageman S, Bosgoed EAJ, Van der Looij E, Rookus MA, Devilee P, Vos EMAW, Wigbout G, Struijcken PM, Menko FH, Rutgers EJT, Hoefsloot EH, Mariman ECM, Brunner HG, Van t Veer LJ. Characteristics of small breast and/or ovarian cancer families with germline mutations in BRCA1 and BRCA2. Br J Cancer 1999; 79: Mancuso AA, Mukherji SK, Schmalfuss I, Mendenhall W, Parsons J, Pameijer F, Hermans R, Kubilis P. Preradiotherapy computed tomography as a predictor of local control in supraglottic carcinoma. J Clin Oncol : Nieweg OE, Jansen L, Valdés Olmos RA, Rutgers EJT, Peterse JL, Hoefnagel CA, Kroon BBR. Lymphatic mapping and sentinel lymph node biopsy in breast cancer. Eur J Nucl Med 1999; 26: S11-6. Pameijer FA, Hermans R, Mancuso AA, Mendenhall WM, Parsons JT, Stringer SP, Kubilis PS, Van Tinteren H. Pre- and post-radiotherapy computed tomography in laryngeal cancer: imagingbased prediction of local failure. Int J Radiat Oncol Biol Phys 1999; 45: Taal BG, Hoefnagel CA, Rutgers M. Carcinoid tumors [letter to the editor]. N Engl J Med 1999; 341:54. Ten Berge RL, Dukers DF, Oudejans JJ, Pulford K, Ossenkoppele GJ, De Jong D, Misere JF, Meijer CJ. Adverse effects of activated cytotoxic T lymphocytes on the clinical outcome of nodal anaplastic large cell lymphoma. Blood 1999; 93: Valdés Olmos RA, Koops W, Loftus BM, Liem IH, Gregor RT, Hoefnagel CA, Hilgers FJM, Balm AJM. Correlative Tl-201 SPECT, MRI and ex-vivo Tl-201 uptake for detecting and characterizing cervical lymphadenopathy in head and neck squamous cell carcinoma. J Nucl Med 1999; 40: Valdés Olmos RA, Hoefnagel CA, Nieweg OE, Jansen L, Rutgers EJTh, Borger J, Horenblas S, Kroon BBR. Lymphoscintigraphy in oncology: a rediscovered challenge. Eur J Nucl Med 1999; 26: S2-S10. Valdés Olmos RA, Jansen L, Muller SH, Hoefnagel CA, Nieweg OE. Aportacion de la medicina nuclear al estudio del drenaje linfático y la identificacion de ganglio linfático centinela en oncologia. Rev Esp Med Nucl 1999; 18: Van Asperen J, Van Tellingen O, Schinkel AH, Beijnen JH. Comparative pharmacokinetics of vinblastine after a 96-hour continuous infusion in wild-type mice and mice lacking mdr1a P-glycoprotein. J Pharmacol Exp Ther 1999; 289: Van Asperen J, Van Tellingen O, Tijssen F, Schinkel AH, Beijnen JH. Increased accumulation of doxorubicin and doxorubicinol in cardiac tissue of mice lacking mdr1a P-glycoprotein. Br J Cancer 1999; 79: Van de Vijver MJ. The pathology of familial breast cancer: the pre- BRCA1/BRCA2 era: historical perspectives. Breast Cancer Res 1999; 1: Van der Sande JJ, Boogerd W, Kröger R, Kapelle AC. Recurrent spinal epidural metastases: a prospective study with a complete follow up. J Neurol Neurosurg Psychiatry 1999; 66: DIAGNOSTIC ONCOLOGY Sloane JP, Amendoeira I, Apostolikas N, Bellocq JP, Bianchi S, Boecker W, Bussolati G, Coleman D, Connolly CE, Eusebi V, De Miguel C, Dervan P, Drijkoningen R, Elston CW, Faverly D, Gad A, Jacquemier J, Lacerda M, Martinez-Penuela J, Munt C, Peterse JL, Rank F, Sylvan M, Tsakraklides V, Zafrani B. Consistency achieved by 23 European pathologists from 12 countries in diagnosing breast disease and reporting prognostic features of carcinomas. European Commission Working Group on Breast Screening Pathology. Virchows Arch 1999; 434: Szymanska H, Sitarz M, Krysiak E, Piskorowska J, Czarnomska A, Skurzak H, Hart AA, De Jong D, Demant P. Genetics of susceptibility to radiation-induced lymphomas, leukemias and lung tumors studied in recombinant congenic strains. Int J Cancer Van Tellingen O, Huizing MT, Nannan Panday VR, Schellens JHM, Nooijen WJ and Beijnen JH. Cremophor EL causes (pseudo) nonlinear pharmacokinetics of paclitaxel in patients. Br J Cancer 1999; 81: Van Tienhoven G, Voogd AC, Peterse JL, Nielsen M, Andersen KW, Mignolet F, Sylvester R, Fentiman IS, Van der Schueren E, Van Zijl K, Blichert-Toft M, Bartelink H, Van Dongen JA Prognosis after treatment for loco-regional recurrence after mastectomy or breast conserving therapy in two randomised trials (EORTC and DBCG-82TM): EORTC Breast Cancer Cooperative Group and the Danish Breast Cancer Cooperative Group. Eur J Cancer 1999; 35: 32-8.

154 154 DIAGNOSTIC ONCOLOGY Van Zandwijk N, Van t Veer LJ, Antioxidants and the chemoprevention of lung cancer. In: Brambilla C, Brambilla E, editors. Lung tumors, fundamental biology and clinical management. New York: Marcel Dekker, 1999: Vos CB, Ter Haar NT, Peterse JL, Cornelisse CJ, Van de Vijver MJ. Cyclin D1 gene amplification and overexpression are present in ductal carcinoma in situ of the breast. J Pathol 1999; 187: Voogd AC, Van der Horst F, Crommelin MA, Peterse JL, Van Beek MW, Repelaer Van Driel OJ, Van der Heijden LH, Coebergh JW. The relationship between findings on pre-treatment mammograms and local recurrence after breast-conserving therapy for invasive breast cancer. Eur J Surg Oncol 1999; 25: Zuetenhorst H, Taal BG, Boot H, Valdés Olmos RA, Hoefnagel CA. Longterm palliation in metastatic carcinoid tumours with various applications of meta-idobenzylguanidine: pharmacological MIBG, 131I-labeled MIBG and the combination. Eur J Gastroenterol Hepatol 1999; 11: Full papers in press Bardelmeijer HA, Van Tellingen O, Schellens JHM, Beijnen JH. The oral route for administering cytotoxic drugs: strategies to increase the efficiency and consistency of drug delivery. Invest New Drugs 1999 (in press). De Jong D, Boot H, Taal B. Histological grading with clinical relevance in gastric mucosa-associated lymphoid tissue (MALT) lymphoma. Recent Results Cancer Res 1999 (in press). Elkhuizen PHM, Van Slooten HJ, Clahsen PC, Hermans J, Van de Velde CJH, Van den Broek LCJM, Van de Vijver MJ and cooperating investigators. The high local recurrence risk after breast conserving therapy in node negative premenopausal patients is greatly reduced by one cycle of perioperative chemotherapy, an EORTC Breast Cancer Cooperative Group study. J Clin Oncol 1999 (in press). Hermans R, Pameijer FA, MancusoAA, ParsonsJT, et al.. Can follow-up computed tomography after definitive radiotherapy for laryngeal or hypopharyngeal carcinoma detect failure earlier than clinical examination alone? Radiology 1999 (in press) Hoefnagel CA. Therapy of neuroblastoma. J Nucl Med 1999 (in press). Hoefnagel CA. Targeted radiotherapy of neuroblastoma. In: Abrams PG, Fritzberg AR, editors. Radioimmunotherapy of cancer. New York: Marcel Dekker, 1999 (in press). Hoefnagel CA, Taal BG, Sivro F, Boot H, Valdés Olmos RA. Enhancement of I-131 MIBG uptake in carcinoid tumours by administration of unlabelled meta-iodobenzylguanidine. Nucl Med Commun (in press). Horenblas S, Jansen L, Meinhardt W, Hoefnagel CA, De Jong D, Nieweg OE. Detection of occult metastasis in squamous cell carcinoma of the penis using a dynamic sentinel node procedure. J Urol 1999 (in press). Julien JP, Bijker N, Fentiman IS, Peterse JL, Delledonne V, Rouanet P, Avril A, Sylvester R, Mignolet F, Bartelink H, Van Dongen JA on behalf of the EORTC Breast Cancer Cooperative Group and EORTC Radiotherapy group. The role of radiotherapy in breast conserving treatment for ductal carcinoma in situ (DCIS): first results of EORTC randomized phase III trial Lancet 1999 (in press). Staalman CR and Hoefnagel CA. Imaging of neuroblastoma and metastasis. In: Brodeur GM, Sawada T, Tsuchida Y, Voûte PA, editors. Neuroblastoma. Amsterdam: Elsevier, 1999 (in press). Theuws JCM, Muller SH, Seppenwoolde Y, Kwa SLA, Boersma LJ, Hart GAM, Baas P, Lebesque JV. The effect of radiotherapy and chemotherapy on pulmonary function following treatment for breast cancer and lymphoma. J Clin Oncol 1999 (in press) Van Asperen J, Van Tellingen O, Beijnen JH. The role of mdr1a P-glycoprotein in the biliary and intestinal secretion of doxorubicin and vinblastine in mice. Drug Metab Dispos 1999 (in press). Van de Vaart PJM, Belderbos J, De Jong D, Sneeuw KCA, Majoor D, Bartelink H, Begg AC. DNA-adduct levels as a predictor of outcome for NSCLC patients receiving daily cisplatin and radiotherpay. Int. J Cancer 1999 (in press). Van Tellingen O, Beijnen JH, Verweij J, Scherrenburg EJ, Nooijen WJ, Sparreboom A. Rapid esterase sensitive breakdown of polysorbate 80 and its impact on the plasma pharmacokinetics of docetaxel and metabolites in mice. Clin Cancer Res 1999 (in press). Theuws JCM, Muller SH, Seppenwoolde Y, Kwa SLA, Boersma LJ, Hart GAM, Baas P, Lebesque JV. The effect of radiotherapy and chemotherapy on pulmonary function following treatment for breast cancer and lymphoma. J Clin Oncol 1999 (in press). Valdés Olmos RA, Jansen L, Hoefnagel CA, Nieweg OE, Muller SH, Rutgers EJTh, Kroon BBR. Evaluation of mammary lymphoscintigraphy by single intratumoral injection for sentinel node identification. J Nucl Med 1999 (in press). Hoefnagel CA. Therapy of Neuroendocrine Tumors with MIBG. In: Salvajoli JV, editor: Radiotherapy. Rio de Janeiro: Editora MEDSI, 1999 (in press). Vos CBJ, Ter Haar NT, Rosenberg C, Peterse JL, Cleton-Jansen A-M, Cornelisse CJ, Van de Vijver MJ. Genetic alterations on chromosome 16 and 17 are important features of ductal carcinoma in situ of the breast and are associated with histologic type. Br J Cancer 1999 (in press).

155 155 Local papers De Klerk JMH, Goslings BM, Koppeschaar HPF, Huysmans DAKCJM, Hoekstra OS, Oyen WJG, Van Eck-Smit BLF, Hoefnagel CA, Jager PL, Lips P, Arndt JW, Krenning EP. Behandeling van het gedifferentieerde schildkliercarcinoom met jodium-131: op weg naar een landelijke consensus. Tijdschr Nucl Geneeskd 1999; 21: DIAGNOSTIC ONCOLOGY Hoefnagel CA. MIBG therapie. In: Nascholingscursus medisch nucleair werkers therapie, SOANG. Rotterdam, 1999: Hoefnagel CA. Radionuclide therapy. In: Wiarda KS, redactie. Leerboek nucleaire geneeskunde SOANG. Maarssen: Elsevier/De Tijdstroom, 1999; pp Jansen EPM, Valdés Olmos RA, Dewit LGH, Muller SH, Hamburger HL, Hoefnagel CA, Bartelink H. Residual pituitary adenoma visualized by functional 99mTc-HMPAO SPECT. Tijdschr Nucl Geneeskd 1999; 21: Kroon BBR, Jansen L, Valdés Olmos RA, Peterse JL, Hoefnagel CA, Nieweg OE. Sentinel node biopsie bij het melanoom: pro? In: Syllabus thema-avond de Sentinel node biopsy bij het melanoom. Leiden: Werkgroep Huidtumoren IKW: Taal BG, Hoefnagel CA, De Jong D, Rutgers M, Boot H. Carcinoïde tumoren: recente ontwikkelingen binnen Nederland in diagnostiek en palliatieve behandeling. Ned Tijdschr Geneeskd 1999; 143: Taal BG, Van Loon H, Kahn N, De Jong D, Vasen HFA, Van t Veer LJ. De rol van genetische factoren bij het ontstaan van maagcarcinoom. Ned Tijdschr Geneeskd 1999; 143: Valdés Olmos RA, Van Dongen A, Hoefnagel CA. Oncologie. In: Wiarda KS, redactie. Leerboek nucleaire geneeskunde SOANG. Maarssen: Elsevier/De Tijdstroom: 1999; pp

156 156 Biometrics Department Division head Otilia Dalesio Head OB Dalesio MSc Academic staff M Kooi MSc, H Van Tinteren MSc Permanent technical staff A Hiemstra, K Hogema, I Jansen, J Lieverst, I Mandjes BSc, C Modder, M Mahn-Schaefers, A Reinders-Som, L Valkenet MD, T Van der Velde, W Van Waardenberg, A Wals, E Willemse Other technical staff D Baars, R Bakx, A Boucher, F De Lange, W Deurloo, F De Vries, L Frenken, R Hakvoort, J Maaskant, A Mathy, A Ndoye, D Roberts, R Tougha, E Van der Donk, E Van Soest, E Weeda, L Ziblat Guest J Oosting PhD Secretary L Rentenaar (until September) Research funded positions (full time equivalents) Scientific permanent: 1.5 Scientific project: 1.0 Technical permanent: 8.9 Technical project: 8.8 Introduction The Biometrics Department provides statistical services, data management and central information on clinical trials to the institute through different functional units: the statistical and data handling group, the tumor register, the trial bureau and the trial secretariat. The supportive tasks include different degrees of collaboration in projects and clinical trials initiated within the clinic and the research sections. In addition, the department has developed collaborations with external groups and functions as the data center for several studies of different national and international groups. There are two areas of special interest in which the department has initiated research and has developed projects during recent years: 1) the use of telematics in providing information services; 2) meta-analyses of treatment in cancer. At the end of 1999, the HORIZON project (EC DG XIII) was successfully completed, EC support for the activities in the area of telematics was increased and the department was involved in several existing projects. The development of the Extranet of Centres of Excellence in Oncology (Onconnet) continued. New services were added to the TRION Intranet server offering information on clinical trials. The different networks in the hospital were linked and protected by a router and a series of Proxy servers to allow access to a unique TRION server from the different places in the hospital. A new system (year 2000 compatible) for online patient randomization and registration in clinical trials (ALEA) was developed to replace PARADIGM, and will be in use in the year 2000, not only at the Biometrics Department but also at the Medical Research Council in UK and Institute Gustave Roussy in France. The 2nd cycle of meta-analysis in prostate cancer of randomized trials comparing standard hormonal treatment with and without the addition of an antiandrogen was finalized.

157 Study of chemotherapy with vitamin A and/or N-acetylcysteine (EUROSCAN) H Van Tinteren, OB Dalesio, N Van Zandwijk 1 The EUROSCAN study is a chemoprevention study, set up in 1987 as a joint venture of the EORTC Head and Neck Group and the EORTC Lung Group. The main objective is to investigate the effectiveness of vitamin A, N-acetylcysteine or the combination of both in preventing or delaying recurrences and second primary cancers in patients treated curatively for oral, laryngeal or lung cancer. Almost 2600 patients have been included in the 4 arms of the study and more than 1500 events were recorded (recurrences, second tumors and deaths). The Department was responsible for the statistical analysis. No benefit was observed either by the use of NAC or vitamin A in the incidence of 2nd tumors, recurrence rate or duration of survival. During 1999 a report of the results was submitted for publication. Notes 1 Division X. Funding: DGV-EC Europe against Cancer program. Collaboration with NDDO Study of adjuvant 5FU plus Levamisole in colorectal cancer (NACCP) B Taal, F Zoetmulder, OB Dalesio, H van Tinteren Although the American Consensus meeting adopted 5FU plus Levamisole as standard therapy in stage III colon cancer shortly after the publication of Moertel et al. (N Engl J Med 1990; 322:352-8), many clinicians in the Netherlands remained unconvinced about the value of adjuvant therapy in colorectal cancer. A cooperative group (the Netherlands Adjuvant Colorectal Cancer Project) was therefore created to investigate this. The group includes 52 hospitals in the Netherlands that treat approximately 60% of colorectal cancer cases in the country. The Biometrics department provided statistical support, central randomization and data handling and coordinated data management. From 1990 to 1996, 1,029 patients were randomized. Patients are followed for recurrence and survival. The overall survival showed a significant difference in favor of the adjuvant treatment with 68% 5-year survival compared to 58% in the control arm. This benefit can also be described as a 25% (SD 9%) reduction in the odds of dying (logrank p=0.007) (Figure 1). 157 BIOMETRICS D Baars, F De Vries, M Kooi, A Ndoye, H Van Tinteren, L Ziblat, OB Dalesio For the last 11 years the department has collaborated with the New Drug Development Office (NDDO) in the phase II studies of the Early Clinical Studies Group (ECSG) and the Biological Therapeutics Development Study Group (BTDSG) of the EORTC. Patients are registered centrally, through on-line randomization by modem connection or through telephone calls. Case record forms from more than 50 participating centers worldwide are sent to the department to be processed. Data entry, checking and querying takes place according to the European Good Clinical Practice guidelines. Reports of ongoing studies are produced for the ECSG investigators and the sponsoring companies, together with statistical analysis for presentations and publications. In 1999 four phase II studies were finalized: 1) GI147211; 2) Carzelesin a DNA-sequence-specific alkynating agent; 3) UFT+Leucovorin in patients with gastric cancer; 4) TA-HPV in patients with cervical cancer. Seven other studies continued with patient registration and data collection. These include studies with Docetaxel, Paclitaxel, LU103793, Mage-3, UFT/5- FU+Leucovorin, ISIS 3521 and ISIS The total number of patients in all studies is 724. From January 1999 until November 1999, 69 new patients were entered in the open studies. Figure 1 Mortality results of FU and levamisol as adjuvant treatment vs no adjuvant treatment in colo-rectal cancer. Telematics projects A Boucher, A Mathy, E Van der Donk, OB Dalesio The Biometrics Department has been involved in telematics projects for the last 11 years and has obtained support of the EC to develop them. In 1999, the department continued to play a key role in the AC:TION

158 158 BIOMETRICS cluster of oncology related EC funded projects in the 4th framework program of DGXIII Telematics for Healthcare. The common denominator of this cluster of projects emerged as what is referred to as Onconnect.net, an Extranet of European reference centers for cancer. The cluster develops telematics services in the areas of cancer treatment, research and education. The Global Horizon project, in which the Biometrics Department is a major partner, supports a G7 initiative to develop global healthcare networks. Key persons in healthcare policy in G7 countries and observer countries are involved in the development of a political, organizational and technical strategy for the implementation of global healthcare networks for oncology, cardiology and other medical disciplines. The Netherlands has been identified as an observer country in these developments. As such, the institute is represented at the meetings of this group. In 1999, the Biometrics Department received additional funding to develop the Onconnect.net concept. This resulted in a concerted implementation plan addressing different aspects of the implementation of a healthcare Extranet, such as legal issues (with Coopers and Lybrand), ISO 9000 certification (with IKO) and the technological options (with Origin). In addition, replication tools were developed in Java, a common helpdesk tool was developed and taken into use for both the onconnect and Intranet services provided by Biometrics in the institute. Extranets are a fast evolving area in the Information and Communication Technology (ICT) sector. In an extranet, a private network of collaborating partners is created based on the same technology as the Internet (TCP/IP). The only difference is the infrastructure: the Extranet uses direct links between participating institutes rather than through Internet Service Providers. The advantage of such a private infrastructure is that the performance is deterministic: the bandwidth is allocated on demand, which enables the use of services such as video, audio and conferencing tools. In addition, being a private infrastructure makes an Extranet less vulnerable to security breaches. A spin-off of these activities was the initiation of the development of a national component of the Onconnect.net. This national network (nl.onconnect.net) currently provides access to TRION (Trial Information Online), ALEA (Online patient randomization and registration in clinical trials) and some international content packages (EuroTransMed realvideo and realaudio), as well as generic IP communication tools such as Netmeeting. Note Funding: DGXIII-EC telematics program. TRION (Trials On line) J Lieverts, L Rentenaar, D Roberts, R Tougha, E Van der Donk, OB Dalesio TRION is a computer system developed by the Biometrics Department to support clinical research in the institute. In 1999 TRION was linked with the live database as it is being used in the Trial Bureau. In effect, any data entered modified or deleted by Biometrics Department staff is immediately effective on TRION. Further, it was extended with a number of extra facilities and web sites. One of the new services on TRION is a web based helpdesk. This site allows any TRION user to drop a support question or report an incident. The incidents or questions are automatically forwarded to the appropriate support staff. The previous situation of TRION being offered through four different servers for WA, RT, DZA and the research net was changed into a single TRION server serving all networks. In addition, TRION is made available to external users through a login mechanism and an encryption of data transmitted from TRION over the Internet. This service will be offered to a restricted number of users in selected hospitals. Maximum androgen blockade in advanced prostate cancer: an overview of the randomized trials W Deurloo, H Van Tinteren, OB Dalesio Since 1989 the Biometrics Department coordinates, in collaboration with the CTSU of the University of Oxford, an overview of all randomized trials of the treatment of prostate cancer. The Prostate Cancer Trialists Collaborative Group was created to make data available for this overview. In advanced prostate cancer androgen suppression (AS) by surgery or drugs controls testicular hormone secretion, and the further addition of an anti-androgen such as nilutamide, flutamide or cyproterone acetate (CPA) produces maximal androgen blockade (MAB). Several randomized trials have been carried out by different groups world-wide to compare AS alone versus MAB and a collaborative meta-analysis of these trials, involving central re-analysis of the data on each of 8,275 men (98% of those ever randomized) was organized. Follow-up was typically for about 5 years, and 71% had died. Five-year survival was 25.4% with MAB versus 23.6% with AS alone, a non-significant gain of 1.8% SD 1.3 (logrank 2p=0.11). There was no significant heterogeneity of benefit with respect to age or stage. The results for CPA, which accounted for only one fifth of the evidence, appeared slightly unfavorable (five-year survival 15.4% MAB versus 18.1% AS alone, difference 2.8% SD 2.4, logrank 2p=0.04 adverse), while those for

159 159 BIOMETRICS Figure 2 Mortality results of MAB vs standard hormonal treatment, by type of anti-androgen in prostate cancer nilutamide and flutamide appeared slightly favourable (five-year survival 27.6% MAB versus 24.7% AS alone, difference 2.9% SD 1.3, logrank 2p=0.005). Nonprostate-cancer deaths (although not significantly affected by treatment) accounted for some of the apparently adverse effects of CPA. Figure 2 presents overall survival curves of MAB compared to standard androgen suppression. Note Funding: DG XII-EC BIOMED II program.

160 160 Research Facilities The research divisions within the NKI/AvL are supported by a number of indispensable research service facilities, each comprising a group of experts who provide information, instruction and service assistance. Besides the facilities described below, the institute has a central cryogenic storage facility, a glassware kitchen, a (electro-)technical workshop and an audiovisual department. Biometrics OB Dalesio MSc, D Baars, R Bakx, A Boucher, F De Lange, W Deurloo, F De Vries, L Frenken, A Hiemstra, K Hogema, I Jansen, M Kooi MSc, J Lieverst, J Maaskant, I Mandjes BSc, A Mathy, C Modder, M Mahn-Schaefers, A Ndoye, A Reinders-Som, L Rentenaar (until September), D Roberts, R Tougha, L Valkenet MD, T Van der Velde, E Van der Donk, H Van Tinteren MSc, W Van Waardenberg, A Wals, E Willemse, E Weeda, L Ziblat In addition to its own research activities (described separately), the Biometrics Department provides the infrastructure for clinical and fundamental research concerning biostatistical support, centralized patient data collection and documentation, data processing and coordinated administration and monitoring of clinical trials. The statisticians collaborate in several research and clinical projects both within the institute and for national and international multicenter studies. The department maintains a Tumor Register database, with information on patients with benign tumors, premalignant and malignant tumors seen in the institute since The database contains almost 97,000 registrations, including second tumors. In excess of 30,000 patients are currently being followed up in the register. This database is a valuable resource for research. Support of the medical divisions for central administration, patient registration and data collection of clinical trials is provided through the Trial Office. Between July 1998 and June 1999, 689 patients were entered in one of the 92 trials open to accrual. One third of them were included in studies supported by the Dutch Cancer Society or by the Health Insurance, a quarter in studies sponsored by industry and the rest in studies sponsored by the institute. Registration and randomization services for 9 national and 9 international studies were provided. In 1999 a quality assurance officer was appointed to review, develop and standardize medical and administrative procedures related to clinical trials. During this year the department reviewed and documented all standard operating procedures (SOPs) according to ISO guidelines and a computer system was developed to maintain the SOPs updated and accessible. A course on Medical Statistics for researchers and physicians that was piloted the previous year was organized twice in 1999 due to the great interest generated. Biophysics GJF Blommestijn PhD, HHJ Brocks PhD, LCJM Oomen PhD, HJ Stoffers MSc, SA Turpijn-la Pas, IV Van de Pavert, GJ Van der Stroom, RO Van Drimmelen Figure 1 Biophysics department; computers and network.

161 161 Computers and Network The NKI/AvL research network facility provides general computer and network support for all research divisions as well as specific site wide services (such as facilities) for hospital personnel. Tasks include: the management of routers, ethernet switches, file servers, job-and-database servers, the configuration of network software on PCs and Apple Macintosh computers of end users and, if necessary, the development and maintenance of custom server and/or client side software. Special care is given to the integration of central data acquisition systems, e.g. cell sorters, phosphor imagers, and confocal laser scanning microscopes, into the network infrastructure. By this service heterogeneous local data and services, internet services and application programs are integrated into a unified scientific office environment and put on the researcher s desktop. RESEARCH FACILITIES Digital Microscopy The biophysics department also provides a digital image acquisition and analysis facility. There are two confocal laser scanning microscopes (one with continuously adjustable wavelength selection bands in all four emission channels) and a microscope setup with a cooled CCD-camera. A second microscope setup with a cooled CCD-camera has been bought to meet the increasing need for these systems. Electron Microscopy J Calafat PhD, JWRM Janssen The facility consists of a Philips CM10 transmission electron microscope and Leica (cryo)-ultramicrotomes. Next year a second electron microscope will be Figure 2 Biophysics department; digital microscopy. purchased. In addition to its own research activities (described in Division I), the facility assists researchers in their ultrastructural morphological analysis of macromolecules, viruses, cells and tissues. Immunoelectron microscopy of ultrathin cryosections reveals a great wealth of structural details in the 2 to 100 nm range providing information that can not be obtained in another way. This technique is used for the following studies: the exact localization of antigens/molecules in the cell organelles, co-localization of two different antigens using colloidal gold of different sizes, trafficking of molecules in the cell, and cell-cell and cell-matrix interactions. Flow cytometry E Noteboom, AS Pfauth Figure 3 Electron Microscopy. The FACS-facility harbors two sorters (one has been upgraded with a HeNe laser), three analyzers and three computers systems for data analysis. The two operators maintain the equipment, instruct users and assist with experiments. Every user is trained so that he or she can operate the analyzers independently. A wide variety of experiments are being performed, because every research group in the institute makes use of this facility. Examples include measurements of apoptosis, cell cycle progression and cell surface markers. Cell sorting experiments include those for transfected cells (co-transfection with surface marker gene), progenitor cells, green fluorescent protein (GFP) and LacZ positive cells, and for subcloning cells. Sorting of live as well as fixed cells is routinely carried out, including four way sorts, double laser experiments and sorting into 96-well plates.

162 162 RESEARCH FACILITIES Laboratory Animal Department RGM Ten Berg DVM PhD, K Ankema, NH Bosnie, NJ Bovenkerk, MT Bulale, C Friederich, HMG Grimminck, D Grund, PJ Handgraaf, TL Hetem-Maidment, M Hoffmann, JJ Janssen, WST Kenbeek, J Kleinsma, FC Krekko, M Lageweg-Beumkes, AJL Lagro, AJ Linden, FT Nicolaas, AJ Schrauwers, C Spaans, HP Starrevelt, EH Tanger, MAM Timpico, AJM Tolkamp, R Van den Berg, FH Van der Ahe, D Van der Pijl, HCA Van Hattum, N Van Roosmaelen, WK Van t Wout, W Wolff, A Zwerver A large part of the oncology research in the institute is carried out using rodent models (mice and rats). The laboratory animal facility is located in a separate wing in order to maintain a good microbiological barrier. The department provides the infrastructure for animal experiments: production and maintenance of mice and rats, renders biotechnical assistance, sanitation of strains, and production of gnotobiotic animals. The facility possesses a large isolator unit for gnotobiology, sanitation and quarantine, plus a separate laboratory for microbiology. Cryopreservation of murine embryos are routinely carried out using a vitrification method. Laboratory Animal Pathology and Histology MA Van der Valk MSc, JC Bulthuis, CCJ De Goeij, DA Hoogervorst, EW Van Muylwijk As cancer and tumorgenesis can be defined as progressive changes in the relational and biological nature of tissues, both clinical and animal (histo-) pathology are indispensable in cancer research. With the increasing production of transgenic and knock-out mice, animal pathology is an important supportive discipline within the institute. The animal pathology and histology department has a well-trained staff (including an animal pathologist) and modern equipment. The department provides researchers with basic histological techniques as well as full histopathology carried out by the pathologist. Advice and instruction are also given to researchers on histotechnical and pathological aspects of their work. In the scope of animal health control, the pathologist is involved in the Sick Animal Program. Apart from serology and bacteriology, routine screening on gastrointestinal endoparasites is performed histologically; this adds significantly to the reliability of conventional techniques. Library S Bakker MSc, IS Benne, L Broertjes, IN Goede, GJ Kroeskamp The library serves the research, clinical, nursing and paramedical departments of the institute and also functions as the Central Cancer Library in the Netherlands. Occasionally patients are referred to the library by the staff of the patient information center. The aim of the library is to maintain an up-to-date collection of both printed and electronic publications covering all aspects of clinical and experimental oncology and related research areas, and to play an active role as an intermediary in the process of information provision and document delivery. The library provides access to the major electronic literature databases and, increasingly, to the full-text electronic version of scientific journals. The role of the library as an information intermediary is extended with the responsibility to maintain the NKI/AvL Internet service consisting of a web-site with information about and from the institute and of an Intranet service with links to bibliographic and literature related databases of the library and access to in-house and externally available electronic resources and services. The library offers current awareness services both in electronic and printed format. On request literature searches, bibliographic instruction, help and advise about literature database management (systems) are offered. The library is responsible for the scientometric analyses (based on impact factors and citations) and the list of publications published in the annual scientific report. DNA Microarrays R Bernards PhD, RM Kerkhoven PhD In January of 1999 a DNA microarray facility was started at the Institute. This facility was made possible by a grant from the Center for Biomedical Genetics. The facility collaborates closely with a similar facility at Leiden University to implement the use of DNA microarray technology in basic and translational research programs. In its first year of operation, a set of 40,000 sequenceverified human cdnas was obtained and methods have been optimized to PCR amplify cdna inserts from library plasmids on a large scale. Equipment to produce microarrays (spotter) and to read arrays (scanner) have been installed and the first microarrays of some 2,000 genes have been produced. At present, technology to generate probes from small amounts of clinical material and to hybridize arrays with fluorescent probes are being optimized. Arrays with larger numbers of genes will become available in the next year.

163 Peptide synthesis LN Vernie MSc (until May), HAM Hilkmann This year 142 peptides were synthesized, mainly for members of our institute but also for other research institutions and universities collaborating within the framework of the Oncology Graduate School Amsterdam. As part of routine quality control each peptide is checked by HPLC. For elucidating the cause of a less successful synthesis and for a further control of the peptides, mass spectrometry is performed, in collaboration with the Free University of Amsterdam and the department of Pharmacy and Pharmacology of the Slotervaartziekenhuis of Amsterdam. A meeting for peptide chemists in the Netherlands has been organized at the institute to discus recent developments in peptide synthesis. Most of the peptides synthesized are used for raising antisera or for biological studies in tissue culture. However, a shift is noted towards the use of peptides in in vitro binding experiments. Peptides used for the production of antibodies are now routinely synthesized onto a branching lysine core. The results of the multiple antigenic peptides are satisfactory. There is always a need for peptides with special features. Examples include the synthesis of peptides which were biotinylated, acetylated or contained amino acids with defined side chains. Peptides are also being synthesized for clinical use; for this purpose the synthesizer has been adapted to increase the yield. In addition to normal routine controls, these peptides are extensively tested in the Department of Pharmacy and Pharmacology of the Slotervaartziekenhuis of Amsterdam. Small amounts of different peptides are synthesized on the recently installed Syro multi peptide synthesizer (in collaboration with Division IV). the staff of the department takes an active part in various research activities, including synthesis of [ 125 I]MIBG with low specific activity (Division VIII) and synthesis of radiolabeled peptides and proteins. Sequence Facility LJ Van t Veer PhD, FBL Hogervorst PhD, R Pruntel The rapid increase in the number of DNA sequence and fragment length analyses and the complex equipment required led to the establishment of a sequence facility in March of this year. The facility is equipped with a ABI 377XL slab-gel automated sequencer which can handle 64 samples simultaneously. The communication between the users, already more than 85, and the facility is through the computer network. In the start-up phase of the facility, users perform the sequence reactions themselves. The chemicals can be purchased from the facility. In addition, specific runs are performed on request. On average, 1,500 samples are analyzed by the sequence facility each month. Another 1,500 samples concern fragment length analysis. In the near future an automated 96-capillary sequencer will be purchased. This machine will enable us to carry out large sequence projects without delay. Furthermore, the facility will be moved to a dedicated laboratory and the service will be extended (including sequence reaction and sample preparation). Transgenic Mouse Service PJA Krimpenfort PhD, R Bobeldijk (until February), CJH Van Veen-Buurman, WK Van t Wout (until April) 163 RESEARCH FACILITIES Radionuclide Laboratory PAJ Jonkergouw MSc, TE Lamers, ALM Luts In addition to departmental laboratories licensed for radioactive work, which are present on each floor of the research building, there is a central radiochemistry facility (class B) available for specific and general experimental use. The staff of the Radionuclide Laboratory offer help and advice on all aspects of radioactive work. The department is equipped with up to date gamma and scintillation counters, gamma analyzers and HPLC apparatus for on line radiodetection. There are also separate facilities for animal experiments with radioactivity. The department provides regular courses on Radiation Protection, level 5B, for all new personnel (including students and guests) whose work entails the use of radioactive material. In addition to general aspects of education and safety, Figure 4 Transgenic mouse service.

164 164 RESEARCH FACILITIES In collaboration with the breeding unit of the laboratory animal department, the transgenic mouse service routinely provides researchers with genetically modified mouse strains. Two distinct techniques are being used: DNA injection into the pronuclei of zygotes and (modified) ES cell transfer into the blastocyst cavity. The unit also assists researchers who want to generate modified mice themselves or who want to perform specific embryological experiments. Pronuclear injections were performed using over 25 transgenic constructs, generated by investigators from almost all research divisions, resulting in a total of 120 independent transgenic mouse strains. The standard recipient zygotes are from the inbred FVB/N background but, upon request, zygotes from a different genetic background can be used (B10, B6, SJL) for specific (e.g. immunological) experiments. Genetically modified ES cell clones were generated by researchers and injected into blastocysts giving rise to chimeras which have been used either in experiments directly or for the establishment of mutant mouse strains. For chimeric studies a total of 12 clones were injected and for the generation of mutant mouse strains 25 clones. Currently, procedures are being developed to use ES cells from genetic backgrounds other than the standard 129/Ola (e.g. B6, Balb/C). Staff of other general research facilities: Cryogenic storage: MM Dijkstra, REJ Kambey Glassware kitchen: TAM Holman-van Doorn, M Aboutalib, EV Holman, REJ Kambey, HL Kempff Technical workshop: JA Marijnissen, WJ Kraan, CWJ Sier Audiovisual department: JM Lomecky, SL Bakker, SM Grentje, ARM Jagt, C Koedijk, ER Tragter, CM Walle Figure 5 Cryogenic storage.