Apoptosis and colorectal cancer. Studies on pathogenesis and potential therapeutic targets Koornstra, Jan

Transcription

1 University of Groningen Apoptosis and colorectal cancer. Studies on pathogenesis and potential therapeutic targets Koornstra, Jan IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2006 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Koornstra, J. J. (2006). Apoptosis and colorectal cancer. Studies on pathogenesis and potential therapeutic targets s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Apoptosis and colorectal cancer studies on pathogenesis and potential therapeutic targets

3 The studies presented in this thesis were supported by the Dutch Cancer Society, grant RUG (Chapters 4, 5, 6 and 8), grant RUG (Chapters 2, 3, 4, 6 and 8), a personal fellowship (Chapter 2) and by the Dutch Digestive Disease Foundation, grant WS (Chapters 3, 4, 5, 6 and 8). The printing of this thesis was supported by the Dutch Digestive Disease Foundation, AstraZeneca, AltanaPharma, Janssen-Cilag, Roche, Schering-Plough and Tramedico. CIP GEGEVENS KONINKLIJKE BIBLIOTHEEK DEN HAAG Koornstra, Jan Jacob Apoptosis and colorectal cancer studies on pathogenesis and potential therapeutic targets Proefschrift Groningen. ISBN: Copyright 2006, Jan Jacob Koornstra All rights reserved. No part of this publication may be reproduced, or transmitted in any form or by any means, without permission of the author. Schilderij cover: Jentsje Popma, verkaveld landschap, regio Oldambt Layout: Helga de Graaf, Groningen Printed by: Print Partners Ipskamp Enschede

4 RIJKSUNIVERSITEIT GRONINGEN Apoptosis and colorectal cancer studies on pathogenesis and potential therapeutic targets Proefschrift ter verkrijging van het doctoraat in de Medische Wetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op woensdag 11 januari 2006 om uur door Jan Jacob Koornstra geboren op 25 juli 1969 te Alkmaar

5 Promotores: Copromotor: Prof. dr. J.H. Kleibeuker Prof. dr. E.G.E. de Vries Prof. dr. H. Hollema Dr. S. de Jong Beoordelingscommissie: Prof. dr. P.L.M. Jansen Prof. dr. G.J.A. Offerhaus Prof. dr. T. Wiggers

6 Paranimfen: Drs. Bart W. Schot Drs. Sjoerd Hovenga

7

8 voor mijn vader en mijn moeder

9

10 CONTENTS Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Chapter 10 Introduction and outline of the thesis Changes in apoptosis during the development of colorectal cancer: a systematic review of the literature. Crit Rev Oncol Hematol 2003;45:37-53 Assessment of apoptosis by M30 immunoreactivity and the correlation with morphological criteria in normal colorectal mucosa, adenomas and carcinomas. Histopathology 2004;44:9-17 Expression of TRAIL (TNF-related apoptosis-inducing ligand) and its receptors in normal colonic mucosa, adenomas, and carcinomas. J Pathol 2003;200: Expression of TRAIL receptors DR4 and DR5 in sporadic and hereditary colorectal tumours: potential targets for apoptosis induction. Eur J Cancer 2005;41: Prognostic significance of TRAIL, DR4 and DR5 expression in adjuvant treated stage III colon cancer patients. submitted Non-steroidal anti-inflammatory drugs and molecular carcinogenesis of colorectal carcinomas. Lancet 2003;362: Sulindac inhibits β-catenin expression in normalappearing colon of hereditary nonpolyposis colorectal cancer and familial adenomatous polyposis patients. Cancer Epidemiol Biomarkers Prev 2005;14: Summary, conclusions and future perspectives Samenvatting, conclusies en toekomstperspectieven List of publications Curriculum vitae Dankwoord Appendix

11

12 Chapter 1 Introduction and outline of the thesis.

13 Chapter 1 Introduction and outline of the thesis Colorectal cancer is the second cause of cancer related deaths in the western world. The development of colorectal cancer is characterised by a sequence of events during which normal colonic epithelium gradually transforms to carcinoma tissue, in most cases via the development of colorectal adenomas. This sequence of events is driven by an accumulation of molecular (epi)genetic alterations causing progressive disorders in cell growth, differentiation and apoptosis. Apoptosis, or programmed cell death, plays an important role in the development and maintenance of tissue homeostasis but also represents an effective mechanism by which abnormal cells, such as tumour cells, can be eliminated. Abnormalities in apoptotic function or resistance to apoptosis have been identified as important events in the pathogenesis of colorectal cancer and its resistance to chemotherapeutic drugs and radiotherapy. During apoptosis, a complex death program is initiated that ultimately leads to the fragmentation of the cell. The death program can be either initiated by the cell itself or by certain external stimuli. These external stimuli may induce apoptosis by targeting one of two pathways. The extrinsic pathway is initiated by triggering cell death receptors on the cell surface, leading to activation of the intracellular apoptotic machinery. The intrinsic pathway of apoptosis is initiated via the mitochondria by cellular stress, such as chemotherapeutic drugs and radiation. The elucidation of the molecular mechanisms regulating these processes is of primary interest. Tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a transmembrane protein belonging to the TNF superfamily. TRAIL triggers the extrinsic apoptotic pathway by binding to its membrane-bound death receptors DR4 and DR5, which transmit an apoptotic signal via their intracellular death domains. TRAIL can also bind to decoy receptors DcR1 and DcR2 that are unable to initiate an apoptotic signal. In vitro, TRAIL induces apoptosis in malignant cells but not in normal cells. Preclinical data in mice and nonhuman primates have shown that TRAIL inhibits tumour growth without serious systemic toxicity. Therefore, TRAIL is considered a promising new anti-cancer agent. This thesis aims to gain insight into changes in apoptosis occurring during the development of colorectal cancer. In addition, the potential of TRAIL-mediated apoptosis as a therapeutic target in colorectal neoplasms is investigated. Finally, possible mechanisms behind apoptosis induction by non-steroidal anti-inflammatory drugs (NSAIDs), in particular sulindac, in colonic epithelial cells are explored, with emphasis on TRAIL mediated apoptosis. Chapter 2 summarises current knowledge of mechanisms of apoptosis and the role of dysregulation of apoptosis in the pathogenesis of colorectal cancer. Although apoptosis has been extensively investigated in colon tissue at different stages of colorectal cancer development, results are conflicting and controversial. These studies were systematically reviewed. Over the last years, new techniques have been developed to investigate apoptosis in human tissue. Chapter 3 describes a new technique of identification of apoptotic cells in colon cells by M30 immunoreactivity. The M30 antibody recognises caspase-cleaved cytokeratin 18 and has shown to be an early marker of apoptosis in epithelial cells. The technique was 12

14 Introduction compared to identification of apoptotic cells using morphological criteria, the method that is considered the gold standard. To investigate the potential use of TRAIL or TRAIL receptor agonists as therapeutic agents in colon neoplasms, the expression of the four membrane bound TRAIL receptors was studied in colon tissue sections in different stages of colon cancer development, i.e. normal colon, adenomas and carcinomas. The results are described in Chapter 4. In the same study, correlations between the degree of apoptosis and expression of TRAIL and its receptors were explored. Finally, possible differences in apoptosis were studied between truly normal colon tissue and normal colon obtained from resection margins from surgical specimens of colorectal cancer patients. Up to 10 % of all colorectal cancers occur in the setting of defined cancer predisposition syndromes. The remainder of cases are named sporadic disease. The two major entities known to be associated with a highly increased risk of developing colorectal cancer are familial adenomatous polyposis (FAP) and hereditary non-polyposis colorectal cancer (HNPCC). Tumours from HNPCC patients are characterised by length alterations in repetitive sequences distributed throughout the genome, so-called microsatellite instability (MSI). The MSI phenotype is also found in % of sporadic colorectal cancer cases. Expanding on the results described in Chapter 4, where patients with sporadic disease had been studied, the same experiments were performed in material obtained from patients with FAP and HNPCC and the results are described in Chapter 5. In addition, sporadic tumours with MSI were studied. Up to half of colorectal tumours with MSI contain mutations in the BAX gene. BAX, a pro-apoptotic member of the Bcl-2 family, has been reported to play a role in sensitivity to TRAIL-mediated apoptosis in vitro. The presence of BAX mutations in MSI positive tumours could potentially limit the use of TRAIL or TRAIL receptor agonists. Therefore, the relationship was investigated between the presence or absence of BAX mutations and the level of apoptosis and expression of TRAIL and its receptors in MSI positive tumours. Although the majority of colorectal carcinomas is macroscopically resectable, 50 % of patients subsequently relapse. In patients with established lymph-node metastasis (stage III), adjuvant chemotherapy following surgical resection has proven benefit. However, intrinsic as well as acquired resistance to chemotherapeutic drugs is a major problem in the treatment of this disease. It is therefore of importance to identify prognostic factors that can help to develop new patient-tailored treatment strategies for colorectal cancer patients. In Chapter 6, the prognostic impact of TRAIL, DR4 and DR5 expression was investigated in tumours of 376 stage III colon cancer patients, who had received adjuvant chemotherapy in a previous nation-wide randomised trial. An important strategy to prevent colorectal cancer is chemoprevention, i.e. the use of drugs that inhibit the development of colorectal adenomas or carcinomas. Chemoprevention is particularly important in patients who have an increased risk of developing colorectal neoplasia. The most promising chemopreventive agents are aspirin and other NSAIDs. Numerous studies have established the efficacy of the NSAIDs sulindac and celecoxib in patients with FAP, while chemoprevention studies in HNPCC patients are ongoing. The mechanisms by which NSAIDs inhibit neoplastic growth are not fully known. In Chapter 7, potential mechanisms behind the chemopreventive action of NSAIDs are briefly reviewed. 13

15 Chapter 1 In the same chapter, two important recent randomised placebo-controlled trials showing a chemopreventive effect of aspirin in patients with previous colorectal adenomas or carcinomas are discussed. The mechanisms behind the chemopreventive efficacy of NSAIDs partly involve the induction of apoptosis in neoplastic cells. Apoptosis induction by sulindac in colon cancer cells seems to involve the intrinsic, mitochondrial pathway, as well as the extrinsic, death receptor pathway. With respect to the TRAIL pathway, sulindac induced in vitro upregulation of DR5 mrna and protein levels, but not of DR4. Some studies have suggested that sulindac may mediate its chemopreventive effect through modulation of the APC-βcatenin-Wnt pathway, possibly mediated by the cell-cycle regulating protein p21. Activation of the APC-β-catenin-Wnt is considered the initial event in the neoplastic transformation of normal colon epithelium. In Chapter 8, the effects of sulindac were studied on apoptosis and expression of DR4 and DR5, β-catenin and p21 in normal appearing colon mucosa. For this purpose, tissue obtained before and after sulindac treatment during two chemoprevention trials in HNPCC and FAP patients respectively was investigated. Finally, the results of the studies described in this thesis are summarised and future perspectives are discussed (Chapter 9). 14

16

17

18 Chapter 2 Changes in apoptosis during the development of colorectal cancer: a systematic review of the literature. J.J. Koornstra 1,2, S. de Jong 1, H. Hollema 3, E.G.E. de Vries 1, J.H. Kleibeuker 2 Departments of 1 Medical Oncology; 2 Gastroenterology and Hepatology and 3 Pathology. University of Groningen Medical Centre, the Netherlands. Critical Reviews in Oncology/Hematology 2003;45:37-53

19 Chapter 2 Abstract The development of colorectal cancer is characterised by an accumulation of molecular genetic alterations causing disorders in cell growth, differentiation and apoptosis. Although changes in apoptosis with colorectal cancer development have been studied extensively, a clear consensus of opinion has not yet emerged. In this review, the literature about changes in the frequency and distribution of apoptosis in tissue sections of normal and neoplastic colorectal tissues was systematically reviewed. Using a PUBMED search, 53 relevant articles were identified. Data from these studies are discussed with respect to the following aspects: methods used to detect apoptotic cell death; frequency and locoregional distribution of apoptosis in normal mucosa, adenomas and carcinomas; the correlation between levels of apoptosis and proliferation and the prognostic significance of the degree of apoptosis in colorectal cancer. Possible underlying mechanisms of dysregulation of apoptosis are briefly discussed. Finally, possible therapeutic implications of knowledge of the molecular regulation of apoptosis are discussed and potential options for further research are suggested. Contents 1. Introduction 2. Methods 3. Results and discussion 3.1 Methods of apoptosis detection 3.2 Changes in frequency of apoptosis in the adenoma-carcinoma sequence 3.3 Changes in locoregional distribution of apoptosis in the adenoma-carcinoma sequence 3.4 The correlation between apoptosis and proliferation 3.5 Apoptosis and prognosis of colorectal cancer 4. Potential mechanisms of dysregulation of apoptosis 4.1 Abnormal expression of apoptosis-regulating genes in colorectal cancer APC P Bcl The role of apoptosis induction by the TNF ligand family 5. Possibilities for intervention: Apoptosis and chemoprevention 6. Conclusions References 18

20 Review: Apoptosis and colorectal cancer development 1. Introduction Colorectal cancer is a major cause of cancer death and, in most cases, develops from a pre-existing adenoma: the adenoma-carcinoma sequence 1. This sequence is characterised by an accumulation of molecular genetic alterations causing disorders in cell growth, differentiation and apoptosis 2-4. It is generally believed that the balance between the rates of cell growth and apoptosis maintains intestinal epithelial cell homeostasis 5, and that during cancer development this balance gets progressively disturbed 6,7. Changes in cell proliferation have been well studied in the adenoma-carcinoma sequence and it is generally accepted that the number of proliferating cells increases in proportion to the severity of dysplasia 8,9. However, less is known about changes in the incidence and regulation of apoptosis during colorectal cancer development. Apoptosis, or programmed cell death, plays an important role in many physiologic and pathologic processes 10. Amongst others, an important function of apoptosis lies in the elimination of damaged cells. For example, cells with genetic damage caused by exposure to carcinogens may be deleted by undergoing apoptosis, thereby preventing their replication and the accumulation of clones of abnormal cells. There is increasing evidence to support the hypothesis that failure of apoptosis may be an important factor in the evolution of colorectal cancer and its poor response to chemotherapy and radiation (reviewed by Watson 11 ). Inhibition of apoptosis causes an imbalance in normal tissue homeostasis promoting cell growth and it also allows the survival of genetically damaged cells, both contributing to tumour development and progression 12. However, considerable controversy exists as to whether the frequency of apoptosis increases or decreases during the adenoma-carcinoma sequence 13. A large number of studies have assessed the proportion of cells undergoing apoptosis, both in vitro and in human colorectal tissue sections. Several in vitro studies have shown changes in colorectal carcinogenesis affecting apoptosis (reviewed by Williams et al. 14 ). From these studies it appears that a gradual resistance to induction of apoptosis emerges during cancer development. Other evidence suggestive of the emergence of such a resistance comes from studies in which it was shown that bile salt-induced apoptosis was diminished in individuals with a history of colon cancer as compared to normal individuals 15,16. It was also found that adenoma cell lines were more sensitive than carcinoma cell lines to apoptosis induction by vitamin D3, its analogue EB1089 and sodium butyrate 17,18. Moreover, a decrease in the rate of spontaneous apoptosis was observed in the course of the adenoma-carcinoma sequence in colorectal epithelial cells in an ex-vivo culture experiment 6. Contradictory to the previous experiments, it was shown that colorectal carcinoma cells were more sensitive than colorectal adenoma cells to induction of apoptosis by the aspirin metabolite salicylate 19 and the non-steroidal anti-inflammatory drug NS Apoptotic cell death has been extensively investigated in tissue sections of human colorectal adenomas and carcinomas. However, these studies have yielded confusing and controversial results and have hitherto not been systematically reviewed. In an attempt to clarify this subject, we performed a systematic search of the literature, focusing on the detection and quantification of apoptosis in benign and malignant colorectal tissue sections. 19

21 Chapter 2 2. Methods We searched PUBMED ( from inception of the database to July 2001, with the MeSH headings apoptosis, colorectal or colon, adenomas, carcinomas or cancer including all subheadings. The resulting citations were examined on screen to identify possibly relevant studies and those thus identified were obtained in full. We included all studies in which apoptosis was studied in tissue sections of normal colonic mucosa, colorectal adenomas or carcinomas. Only papers published in English were included. The bibliographies of all included studies were carefully examined. In some studies, data on the frequency of apoptosis were not explicitly mentioned in the text but could be inferred from tables or figures. In some other studies, apoptosis was quantified and discussed in the article, without actual presentation of exact numbers. 3. Results and discussion In total, 53 studies were considered eligible. Individual studies are summarised in table I. The first part of the table summarises those studies that assessed and compared apoptotic frequencies in different colonic tissue types (normal mucosa and/or hyperplastic polyps and/or adenomas and/or carcinomas). In the second section of the table, studies are depicted in which only one colonic tissue type was studied. Similarities and differences between the studies will be discussed below with respect to the following aspects: the methods used to detect apoptotic cells; the frequency of apoptosis in normal mucosa, adenomas and carcinomas; the locoregional distribution of apoptosis in benign and malignant colorectal tissues; the correlation between levels of apoptosis and proliferation and the correlation between levels of apoptosis and prognosis of colorectal cancer. 3.1 Methods of apoptosis detection Several methods exist to examine apoptosis in tissue sections. The methods used in the reviewed studies were terminal deoxyribonucleotidyl transferase-mediated nick end labelling (TUNEL, n=29), in situ end labelling (ISEL, n=7), morphology by light microscopy (n=7), TUNEL or ISEL in conjunction with morphology by light microscopy (n=9), ISEL in conjunction with morphology by electron microscopy (n=1), morphology by light microscopy in conjunction with electron microscopy (n=2) and M30 immunoreactivity (n=2). All methods and their advantages and disadvantages will be briefly discussed below. Apoptotic cells can be recognised by careful light-microscopic examination of hematoxylineosin-stained sections, showing morphological criteria defined by Kerr et al. 5,21. These criteria include the identification of the formation of apoptotic bodies with nuclear condensation or nuclear fragmentation, with cytoplasmic shrinkage and loss of contact from surrounding cells. A more definitive method of morphologic identification of apoptotic cells is electron microscopy 5,22. Because of the brief duration of such morphologic changes and their seemingly low incidence, apoptosis has been difficult to detect in routine histologic sections 23. Therefore, several biochemical procedures have been developed to identify apoptotic cells in vivo; the methods most widely used are TUNEL and ISEL. These methods are based 20

22 Review: Apoptosis and colorectal cancer development Table 1. Overview of studies on the detection of apoptosis in colorectal tissues. Apoptosis assessed in different colon tissue types Normal Normal adjacent Hyperplastic polyps Adenomas Carcinomas First author [ref] Method(s) used AI # n AI n AI n AI n AI n Sinicrope [4] TUNEL & morphology (HE) Bedi [6] TUNEL Hao [7] ISEL Hawkins [12] ISEL 4.2 * * * 83 Carr [23] ISEL M30 reactivity Ward [35] ISEL 13.1 * * 58 Aotake [36] TUNEL ** Ikenaga [37] TUNEL Ikenaga [38] TUNEL ** Koike [39] TUNEL < Moss [41] TUNEL & morphology (HE) Moss [42] TUNEL Partik [43] ISEL & EM Regitnig [44] TUNEL Takano [45] TUNEL Kawasaki [47] TUNEL ** Nakamura [48] TUNEL Kikuchi [51] TUNEL Nomura [52] TUNEL Yamamoto [53] TUNEL Morphology (HE) Baretton [54] TUNEL Valentini [55] ISEL Sinicrope [56] TUNEL & morphology (HE) Sträter [71] TUNEL Kobayashi [128] TUNEL Miki [195] TUNEL Hao [196] ISEL & morphology (HE) Schölzel [197] TUNEL

23 Chapter 2 Table 1. (continued) Apoptosis assessed in one colon tissue type Normal Adenomas Carcinomas Liu [40] TUNEL Anti [46] TUNEL & morphology (HE) Barkla [75] Morphology (HE & EM) Keller [186] Morphology (HE) - 21 Arai [49] Morphology (HE & EM) Fazeli [50] TUNEL & morphology (HE) - 14 Hawkins [22] ISEL & morphology (HE) - 15 Michael-Robinson [33] M30 reactivity Tsujitani [57] TUNEL Ozawa [60] TUNEL Tanimoto [61] ISEL Tatebe [81] TUNEL & morphology (HE) Evertsson [82] TUNEL Sinicrope [83] Morphology (HE) Hashimoto [89] ISEL Kawasaki [90] TUNEL & morphology (HE) Paradiso [91] TUNEL & morphology (HE) Tenjo [92] TUNEL Sugamura [93] TUNEL Langlois [94] Morphology (HE) Schwandner [95] TUNEL Dolcetti [106] TUNEL Morphology (HE) De Angelis [129] TUNEL Sinicrope [144] Morphology (HE) Seong [198] TUNEL Table legend: Normal adjacent: normal mucosa adjacent to adenoma or carcinoma; # AI: Apoptotic index (expressed as mean or median), defined as the percentage of apoptotic cells of the total number of epithelial cells counted or defined as arbitrary units (*); n: number of samples investigated; TUNEL: terminal deoxynucleotidyl transferase using nick end labelling; ISEL: in situ end labelling; EM: electron microscopy; HE: hematoxylin-eosin; -: missing data; ** according to Japanese criteria 58 on the assumption that genomic DNA is fragmented in a dying cell, producing fragments of consistent length in apoptotic cell death, as opposed to necrotic cell death where DNA is believed to be randomly degraded. Both the TUNEL and ISEL method detect the DNA strand breaks by adding nucleotides, including deoxyuridine triphosphate labelled with biotin, to the ends of DNA fragments. The only essential difference between the two is that TUNEL uses the enzyme terminal deoxyribonucleotidyl transferase, whereas ISEL uses DNA polymerase I or its Klenow fragment Both methods (TUNEL/ISEL) have disadvantages. Most importantly, they both mark necrotic and autolytic cells in addition 22

24 Review: Apoptosis and colorectal cancer development to apoptotic cells 24, Furthermore, they require pre-treatment steps that need careful optimisation, and the results depend on how these steps are performed 23. There is also evidence that the physical act of cutting tissue to create sections produces TUNEL activity 30. This resulted in serious reservations concerning the applicability of the TUNEL method to evaluate apoptotic cell death in the gastrointestinal tract 31. Results derived solely from TUNEL staining should therefore be considered with caution, unless supported by light microscopic analysis of cell morphology. Cell morphology by light microscopy alone is considered by some as superior to TUNEL or ISEL for the detection of apoptotic cells 22,32. The vast majority of the studies discussed in this review have utilised TUNEL or ISEL staining as the primary measure of apoptosis, and only a few confirmed their findings by morphological identification of apoptotic cells. When only these studies are considered, ranges in reported apoptotic indices, which will be discussed below, vary considerably, underlining the limitations of the currently used methods to identify apoptotic cell death. In two studies, a new method of identifying apoptotic cells using the monoclonal antibody M30 was used 23,33. This method is based on the detection of a neo-epitope of cytokeratin 18, which is produced by caspase cleavage of cytokeratin 18 during apoptosis 34. As it remains immunoreactive in paraffin-embedded tissue and is not present in nonapoptotic cells, it has potential in the study of apoptosis in clinical and experimental material. A clear advantage over TUNEL and ISEL is that M30 is negative in necrotic cells. One disadvantage of the method might be that the time scale over which TUNEL or ISEL detects DNA fragmentation in the cell is considerably longer than the time during which cleaved cytokeratin 18 is expressed 23. Recently, results of M30 immunoexpression were compared with the ISEL method. In colorectal neoplasms, a strong positive correlation was found between in situ end labelling and M30 immunoreactivity 23. Immunoexpression of M30 was generally easier to interpret than in situ end labelling, and the procedure for M30 immunohistochemistry was technically simpler. It can be expected that in the near future new techniques with higher specificity for the detection of apoptotic cell death will be developed based on the detection of target proteins of caspase cleavage. 3.2 Changes in frequency of apoptosis in the adenoma-carcinoma sequence In all but two studies, apoptosis was quantified in tissue sections as an apoptotic index, defined by the number of apoptotic cells as a percentage of the total number of epithelial cells counted. In the two exceptions 12,35, numbers of apoptotic cells were determined in 10 high power fields (light microscope; field diameter 490 µm), corrected for the fraction of neoplastic epithelium in those fields. Normal mucosa. The proportion of epithelial cells undergoing spontaneous apoptosis in normal colonic mucosa appears to be highly variable and apoptotic indices of % have been reported 4, Two issues have to be considered here. First, in most studies sections of normal colonic mucosa were in fact retrieved from macroscopic normal tissue adjacent to adenomas or carcinomas. One could question whether this represents truly normal epithelium. For example, Moss et al. showed considerable differences in the degree of apoptosis between histologically normal adjacent epithelium in comparison to truly normal tissue, it being generally lower in the latter 41. Such alterations in normal tissue adjacent to colorectal neoplasms may reflect an early stage in the process of colorectal 23

25 Chapter 2 carcinogenesis or a reaction to the adjacent tumour 41. Secondly, Liu et al. demonstrated significant differences between apoptotic indices in different parts of the colon 40. While the percentage of proliferating colonic epithelial cells was constant throughout the colon, apoptotic indices were found to be lower in the right colon in comparison to the left colon and rectum 40. However, Anti et al. could not show segmental differences in normal control subjects whereas in patients with a history of adenomas apoptotic indices were higher in the right colon compared to the left colon and rectosigmoid segment 46. The cause of these discrepancies is unclear and may be a result of the relatively small numbers of subjects studied. In most studies reviewed, the location from which normal mucosa was obtained was not stated. Adenomas. There is virtually no debate as to whether the proportion of apoptotic cells increases or decreases during the transition from normal colonic mucosa to adenomas. Most studies report increased apoptotic indices in adenomatous tissue compared to normal mucosa 4,12,36,38,39,43-45,47, whereas only a few studies found the opposite 6,41,42. Reported apoptotic indices in adenomas vary between %. Several studies have assessed the degree of apoptotic cell death in different types of adenomas, i.e. low or high grade dysplasia and/or tubular or villous architecture of the adenoma 7,12,36,38,39,43, In reports where adenomas with different grades of dysplasia were studied, apoptotic indices increased in proportion with the severity of dysplasia in most studies 7,38,49,51,53, whereas others have found no such relationship 48,52, or the opposite 36,39,47,50. With respect to tubular and villous adenomas, some found higher apoptotic indices in tubulovillous and villous adenomas than in tubular ones 12,43, whereas others found the opposite 38,49. Taken together, these studies do not allow general conclusions of differences in apoptotic cell death between different types of adenomas as the numbers of adenomas studied were generally low and considerable differences in the methods to identify apoptotic cell death were present. Carcinomas. Concerning the transition from adenoma to carcinoma, considerable controversy exists. Most researchers found that the frequency of apoptotic cell death in carcinomas was higher than in adenomas 7,12,23,35,38,39,45, Others observed a decrease 6,36,43,55 and some showed no differences 41,47,51,56. Tsujitani et al. compared carcinomatous with adenomatous components in the same tumours and found lower apoptotic rates in the carcinomatous components 57. Reported apoptotic indices in carcinomas vary between %. When considering only those studies in which TUNEL or ISEL staining was confirmed by morphological identification of apoptotic cells, and both adenomas and carcinomas were studied, no differences were found between adenomas and carcinomas 41,56. The one study not using TUNEL, ISEL or morphology, but the supposedly more specific method of M30 immunoreactivity, showed a higher apoptotic index in carcinomas than in adenomas 23. In summary, most studies indicate an increase in apoptotic frequency in the course of the adenoma-carcinoma sequence, in parallel with increasing dysplasia. The transition from normal mucosa to adenoma is more clearly accompanied by an increase in apoptotic frequency than that from adenoma to carcinoma. How can these differences in apoptotic indices be explained? First, discrepancies in conclusions may reflect methodological differences in the detection of apoptosis and selection and preparation of tissue. Second, diagnostic criteria for carcinoma of the colorectum differ between Japanese and Western pathologists 58,59. Since intramucosal carcinomas in Japan 24

26 Review: Apoptosis and colorectal cancer development are not accepted as carcinomas in the West, results from some Japanese studies 36,38,47,57,60,61 cannot readily be compared with the others. Many authors of the reviewed articles refer to the study by Bedi et al. 6, in which a progressive decrease was reported of the proportion of apoptotic cells during the transformation of normal colonic epithelium to adenomas and adenocarcinomas, associated with enhanced ex vivo survival. As many other studies have shown a gradual increase in frequency of apoptotic cell death, the study has given rise to confusion and controversy, which, although being recognised by most authors, is usually not further addressed. We feel that a few remarks concerning this report should be made. In the study of Bedi, the TUNEL method was used to identify apoptotic cell death in different colonic tissue sections. In ex vivo cultures of normal, adenomatous and carcinomatous epithelial cells, apoptotic cell death was determined by flow cytometry analysis. It must first be noted that the TUNEL studies were performed on frozen tissue sections, unlike all of the other studies discussed in this review. Second, using this method, the authors observed no apoptosis in colorectal carcinomas, which is most unusual. As all other studies have shown apoptotic cells in carcinomas, the reported absence of apoptosis in the Bedi report is likely to reflect deficiencies in the TUNEL technique. Furthermore, it must be realised that the analysis by flow cytometry in the ex vivo cultures of colonic epithelial cells represents a rather artificial approach, with several limitations. For example, the ex vivo cell survival studies were assessed after collagenase digestion and the preparation of single-cell suspensions. By thus interfering with normal cell-cell contact, which is known to affect the degree of apoptosis 62,63, results from the ex vivo cultures cannot readily be compared with those of other studies. As stated before, most studies show that the proportion of epithelial cells undergoing apoptotic cell death increases in the course of the adenoma-carcinoma sequence. At first glance, this seems contradictory to the general idea, mostly based on in vitro experiments, that tumour progression is associated with decreased sensitivity to apoptosis induction. This has led to some confusion in the literature, possibly because decreased sensitivity to apoptosis induction is confused with a decrease in apoptotic cell death. One should realize that all 53 reviewed studies concern spontaneous apoptotic cell death and not sensitivity to chemotherapy- or otherwise induced apoptosis. An increase in spontaneous apoptotic cell death in the course of the adenoma-carcinoma sequence is not necessarily contradictory to decreased sensitivity to apoptosis induction. First, higher rather than lower levels of apoptosis would provide a mechanism to keep the size of adenomatous polyps stationary for prolonged periods of time in spite of continuing proliferation and would explain the slow growth of adenomatous polyps 64. Second, it has become clear that cell proliferation and cell death are coupled processes in several respects 65. For example, there is evidence that the increased proliferative activity in tumours might directly activate the program of apoptosis due to lack of nutrients, competition for growth factors, or oxygen starvation 65. Furthermore, the oncogene c-myc, which is overexpressed early in the adenoma-carcinoma sequence, has been shown to be a potent inducer of not only cell proliferation but also cell death 66. Moreover, the higher the amount of c-myc expressed in cells, the greater the tendency to undergo apoptosis 65,67. The regulation of the balance between cell proliferation, differentiation and apoptosis is particularly important in tissues with high cell turnover rates, such as the colon. Thus, it might well be that increased cell proliferation paralleled by an increased proportion of 25

27 Chapter 2 apoptotic cell death during the adenoma-carcinoma sequence eventually may select for the survival of mutant cells or cells with a survival phenotype that are resistant to the induction of apoptosis. The selection of advantageous mutations may be as important as an increased mutation rate in carcinogenesis 68. Such a selective pressure to override apoptosis has indeed been suggested by others 15, Changes in locoregional distribution of apoptosis in the adenoma-carcinoma sequence Normal mucosa. The growth and differentiation of colonic epithelial cells proceeds through distinct phases, depending on the position of a particular cell in the crypt. Stem cells, located at the bottom of the crypts, divide to form daughter cells, which proliferate rapidly and migrate up the crypts, differentiating into columnar epithelium, goblet cells and enteroendocrine cells 70. In normal colonic mucosa the proliferative compartment is confined to the lower half to two thirds of the crypt. As cells move towards the upper segments, they differentiate, lose the capacity to divide, and die within several days. Until recently, it was assumed that the cells were then passively detached into the colonic lumen or mechanically shed by an exfoliating action of passage of the gut contents. While this may be the case for a proportion of cells, the main physiological mechanism by which enterocytes die is by apoptosis 71. Apoptosis of colon epithelial cells has clearly been demonstrated predominantly in the upper levels of the crypts 4,6,25,37,39,41,42,46,51,71,72. Apoptotic rates, calculated from the number of cells undergoing apoptosis in each colonic crypt, approximately balance the rate of cell renewal which supports the idea that apoptosis is responsible for the homeostasis of the epithelial cell population 72. At the luminal surface, part of the apoptotic cells are actively extruded from the epithelial layer into the lumen, while others are engulfed by surrounding epithelial cells or lymphocytes 71. This engulfment pattern is predominantly observed in the basal parts of the crypt 71. The occasional apoptotic events that are found in the lower, proliferative zones of the crypts are thought to occur in response to genetic damage 73,74. A recent morphological study suggested that epithelial cells undergoing apoptosis may also pass through fenestrations in the basement membrane to the lamina propria where they are taken up by macrophages 75. Adenomas. Several authors have observed that in adenomatous polyps proliferation predominates at the luminal surface and apoptosis at the base of the crypt, a complete reversal of the normal pattern 4,36,41-44,49,71. In all these studies, the lower portion of the adenomatous crypts showed more apoptotic cell death than normal tissue. This pattern was seen both in sporadic adenomas as well as those associated with familial adenomatous polyposis (FAP). Others have observed apoptotic cells along the whole length of the adenoma crypts 39,51,52. Thus, it is generally agreed that in adenomas apoptosis takes place mainly in the lower part of the crypt, and to a greater extent than in normal crypts 76. These findings can be interpreted in several ways. One view is that in adenomas genetic alterations occur preferentially in the cells at the base of the crypt and that these cells are eliminated by apoptosis 74. Diminished apoptosis at the luminal end of the crypt may mean that the effort to control the damage has failed there, so the adenoma grows. It supports the concept of a defensive function of apoptosis to eliminate immature cells carrying mutations before they have the chance to mature and migrate up the crypt 74. Alternatively, the adenomatous polyp may grow inward, i.e. proliferating cells migrate towards the base of the crypt, a concept first suggested by 26

28 Review: Apoptosis and colorectal cancer development Moss et al. 42. Support for the inward growth model comes from a kinetic study of colonocytic migration in a patient with Gardner syndrome 77. Distribution of thymidine labelled cells showed an abnormal retrograde migration away from the adenoma surface towards the crypt base 77. More support for an inward growth model comes from a recent study, showing that dysplastic cells in early adenomas were located at the luminal surface of the crypts whereas the cells at the base of the crypts appeared morphologically normal 78. Moreover, these dysplastic cells were found to be mutated in contrast to the cells at the base. The study strongly suggests that the dysplastic process is initiated at the luminal side of the crypt and that genetically altered cells spread laterally and downward from the surface to the base of the crypt rather than the other way around. From these data, it seems unlikely that the observed increased apoptotic rates at the base of the crypts in adenomas reflect the deletion of cells that have gained genetic damage, as in the normal mucosa. Finally, in carcinomas, there is probably no specific distribution of apoptotic cells, as they were usually randomly spread throughout the tumour according to the reviewed studies 4,12,33,36,39,41,43,51, The correlation between apoptosis and proliferation The balance between cell production through proliferation and cell loss through apoptosis determines how fast a tumour grows and is an important determinant of tumour behaviour. Most colorectal adenomas are stable for a long time before, if ever, transforming to malignancies. For example, 3-5 years after the initial diagnosis of adenomas, 70% showed no change in size 79. In animal studies, adenoma and carcinoma cells had cell cycle times about half of that of normal epithelium, reflecting higher growth rates 80. As increased cell death may be an attempt to limit the expansion of the tumour cell population, several studies have linked the rate of apoptosis with the proliferative rate of adenomas and carcinomas, yielding conflicting results. Levels of proliferation were associated with these of apoptotic cell death in adenomas and carcinomas in a number of studies 7,23,38,49,81-83, whereas others found no correlation 39,42,43,48,51,54,61. The finding that apoptosis and proliferation are not correlated is often used as an argument to support the assumption that an imbalance between these processes emerges in the course of the adenoma-carcinoma sequence. Supportive for this assumption are results from a study in which proliferative activity was shown to be correlated with apoptotic index in adenomas with low-grade dysplasia whereas this correlation was lost in adenomas with high-grade dysplasia and carcinomas 36. Some studies have focussed on the ratio of apoptotic cells to proliferating cells as a proposed measure of susceptibility to apoptosis 39,54. Koike et al. found that this ratio was higher in adenomas of small size and with low- and middle-grade dysplasia than in those of larger size and with high-grade dysplasia 39. Baretton et al. found that the ratio was higher in adenomatous components than in carcinomatous tissue components 54. Both studies thus support the hypothesis that in the course of the adenoma-carcinoma sequence the susceptibility to apoptosis induction gradually diminishes. In summary, most studies show a considerable increase in both apoptotic cell death and proliferation in colonic neoplasms, illustrating abnormally high cell turnover rates. However, the two processes do not seem to be correlated in individual tumours, providing further evidence for the concept of a progressively disturbed balance between apoptosis and proliferation in the course of the adenoma-carcinoma sequence. 27

29 Chapter Apoptosis and prognosis of colorectal cancer Currently, the most important prognostic variables in colorectal cancer are tumour stage and preoperative elevation of carcinoembryonic antigen 84. Several reports indicate that tumour cell kinetics may be an important prognostic variable in colorectal cancer Traditionally, an increase in cell proliferation, rather than a change in apoptosis, has been used to predict later tumour development and prognosis. However, in an animal model, the best predictor of tumour development was the degree of apoptosis 88. While there is general agreement that deregulation of apoptosis contributes to malignant transformation, the potential predictive or prognostic value of the degree of apoptosis in colorectal cancer is controversial. Several studies have examined the prognostic value of the apoptotic index in colorectal cancer, producing conflicting results 12,45,82,83, Schwandner et al. showed that the apoptotic index was not predictive of prognosis in a series of 160 cases of rectal cancer 95. In three other studies, no prognostic significance was found in large groups of carcinomas 83,91,92. However, stratification by tumour site revealed that the apoptotic index was an independent predictor of survival in a series of 82 distal tumours (distal to splenic flexure) 83. In two studies, it was shown that a low apoptotic index in the tumour was associated with poor survival 90,94. Two reports showed that apoptotic indices were higher in tumours that were more highly differentiated and had not invaded or metastasised than in those that were poorly differentiated and invasive or metastasising 89,93. Tanako et al. also found higher apoptotic indices in tumours without lymph node or distant metastases in comparison to tumours that had metastasised, but they found no correlation with the degree of tumour differentiation 45. On the other hand, Hawkins et al. demonstrated that Dukes A carcinomas had lower apoptotic indices than Dukes B to D carcinomas 12. Metastatic dissemination may depend upon the resistance of metastatic cells to apoptosis 96. Indeed, in several murine and human cancer cell lines a more aggressive metastatic phenotype was associated with increased resistance to apoptosis 97. Two recent interesting studies should be mentioned in this respect. Most colorectal carcinomas in hereditary non-polyposis colorectal cancer (HNPCC) and a proportion of sporadic colorectal carcinomas show a particular form of genetic instability, termed microsatellite instability (MSI), which means an accumulation of deletion and insertion mutations in simple repeated sequences 98,99. Patients with colorectal cancers with the MSI phenotype are known for better survival rates than patients with tumours without this phenotype It has been proposed that colorectal tumours with the MSI phenotype develop along a different pathway compared to colorectal tumours without this phenotype, and that these tumours have an increased mutation rate as a result of inactivation of a mismatch repair gene 104. It has also been shown that apoptosis can be induced by overexpression of the mismatch repair genes hmsh2 or hmlh1 105, suggesting that tumours with the MSI phenotype lose the ability to undergo efficient apoptosis 99. Surprisingly however, it was demonstrated by two groups that apoptotic cell death was actually more frequent in colorectal tumours with MSI than in those without MSI 33,106. Taken together, although most studies suggest that colorectal tumours with high apoptotic indices are associated with better prognosis and survival, we believe that more work is required before the apoptotic index can be considered as a potential prognostic marker or indicator for the choice of therapy in colorectal neoplasms. 28

30 Review: Apoptosis and colorectal cancer development 4 Potential mechanisms of dysregulation of apoptosis 4.1 Abnormal expression of apoptosis-regulating genes in colorectal cancer It has been firmly established that colorectal carcinogenesis is characterised by a stepwise accumulation of genetic alterations (for extensive recent reviews, see Chung 107 or Ilyas et al. 108 ). The accumulation of genetic alterations during the adenoma-carcinoma sequence may provide the basis for changes in the degree of apoptotic cell death, as many of the genes involved have been shown to regulate apoptosis. Most studies have focused on the possible roles of the tumour suppressor genes adenomatous polyposis coli (APC) and p53 and the proto-oncogene bcl-2, which will be briefly discussed below APC Mutations in the APC gene have been implicated in both sporadic and familial colorectal neoplasia 2,109. The frequency of APC mutations is similar in colonic adenomas and carcinomas (approximately 60%), suggesting that APC mutations may be an early or even the initiating event in the process of colorectal carcinogenesis 110. Immunohistochemical studies have shown that the APC protein is expressed in normal epithelial cells as they migrate toward the top of the crypt 111,112. Disruption of normal APC function presumably disturbs the equilibrium between new cell formation at the base of the crypts and cell death at the top of the crypts, leading to a relative expansion of the progeny of APC-mutant cells. An important role of the APC gene in regulating apoptosis is suggested by an experiment in which the expression of APC in human colorectal cancer cells containing endogenous inactive APC alleles resulted in the induction of cell death through apoptosis 113. The functional significance of the APC gene probably lies not only in the regulation of apoptosis, but also in control of cell cycle progression, migration and differentiation 114. It has been shown that in normal cells, the APC protein resides in a large complex with axin, glycogen synthase kinase 3ß (GSK3b) and ß-catenin 114,115. Loss of APC protein function leads to ß- catenin accumulation in the nucleus where it binds to TCF/Lef transcription factors 116,117. This complex then activates transcription of TCF target genes 118. Thus far, genes including c-myc, cyclin-d1, matrilysin, peroxisome proliferator-activated receptor delta and MDR1 (multidrug resistance) have been reported to be the target genes of this complex, influencing cell proliferation and apoptosis 119,120. In the herein-reviewed studies, APC protein function was never assessed by either immunohistochemistry or by mutation analysis P53 Mutations of the p53 gene occur in various human tumours, including colorectal cancer 121,122. Deletions and mutations of the p53 gene can be detected in up to 85% of colorectal tumours and usually occur during the transition from adenoma to adenocarcinoma 123. Several functions have been ascribed to the p53 tumour suppressor gene, reviewed by Levine 122 and Sigal 124. Its product, the p53 protein, may respond to DNA damage by triggering either growth arrest during G1 or G2 phase of the cell cycle or programmed cell death. In this manner, p53 may protect the normal cell from proceeding to replicate damaged DNA. The wild-type p53 protein, but not the mutant, can initiate apoptosis. Introduction of the wild-type p53 gene mediated apoptosis in human colon carcinoma cell lines and tumours in nude mice 29

31 Chapter 2 underwent regression if wild-type p53 expression was induced 125. The mutated p53 protein may block the function of the wild-type p53 protein and thereby inhibit induction of apoptosis. It has also been found that p53 has profound effects on responses to chemotherapeutic drugs used in colorectal cancer, and that these effects vary considerably depending on the drug 126. Several studies have assessed the correlation between p53 protein expression and apoptosis in colorectal neoplasms. In a murine model, levels of spontaneous apoptosis were similar in both normal and p53 knockout mice in normal small intestine and colon epithelium 127. In addition, homozygosity for an inactivating germ-line mutation of p53 did not affect the frequency of apoptosis in mouse adenomas 50. Due to a germline APC mutation, these mice were strongly predisposed to adenoma formation. Moreover, p53 inactivation did not additionally enhance adenoma formation or adenoma progression. In the same study, the degree of apoptosis was not different in human colon adenomas between areas with highgrade dysplasia, associated with loss of p53 function, and areas with low-grade dysplasia, associated with intact p53 function. Some of the herein reviewed studies have indicated that adenomas and/or carcinomas with a high percentage of cells expressing the p53 protein were more likely to have a low apoptotic index 4,61,94,95,128,129 whereas most studies did not show such a relationship 7,12,33,35,36,39,45,48,51,52,54,57,81,91,92,130,131. In summary, the fact that neither immunohistochemical overexpression of p53 nor p53 mutations correlated to the frequency of apoptotic cell death in the majority of studies does not support a major role for mutant p53 protein as an inhibitor of apoptosis in the development of colorectal cancer Bcl-2 At least 15 bcl-2 family member proteins have been identified in mammalian cells, including proteins that promote apoptosis and those that prevent it 132. The oncogene bcl-2 promotes cell survival by blocking apoptosis 133. Bcl-2 protein is normally expressed only in the lower half of the crypts of the colon, corresponding to the stem cell compartment, where bcl-2 is believed to protect stem cells from apoptosis 134. Most colonic adenomas express bcl-2 protein at high levels throughout the neoplastic epithelium 4,48,54,131, while non-neoplastic polyps have a normal pattern of bcl-2 expression 48,136,137. Overexpression of bcl-2 may therefore contribute to the transition between hyperplastic epithelium and adenomas. Bcl-2 protein expression in colorectal carcinomas is higher than in normal mucosa, but lower than in adenomas 12,35,48,51,54,70,131, In a murine model, bcl-2-null mice have increased levels of spontaneous apoptosis in colonic crypts compared with wild-type mice 134. An inverse correlation has indeed been reported between bcl-2 expression and the apoptotic index of colonic tissues 4,7,12,36,54,90, whereas others have found no such correlation 33,35,43,48,51,52,61,83,91,94,95. With respect to the correlation between bcl-2 expression and prognosis in colorectal cancer, reports are conflicting 91,135, Considering the relationship between bcl-2 and p53 overexpression, some have found an inverse correlation in adenomas 7,55,131,139. Taken together, bcl-2 expression seems to be gradually reduced in the course of the adenoma-carcinoma sequence and inversely related to p53 overexpression. As most studies show a gradual increase in frequency of apoptotic cell death, a possible relationship with downregulation of bcl-2 can be hypothesised. However, bcl-2 is probably only one of the genes that determine the incidence of apoptotic cell death in colorectal neoplasms. Indeed, 30

32 Review: Apoptosis and colorectal cancer development changes in the expression of other members of the bcl-2 family have been shown during progression of colorectal tumors, such as the anti-apoptotic proteins bcl-x L, mcl-1 and the pro-apoptotic protein bak, which may be more important than bcl The role of apoptosis induction by the TNF ligand family Apoptotic cell death can be initiated by endogenous stimuli, such as the absence of oxygen, nutrients or growth/survival factors, the presence of DNA damage or the action of cytokines as well as by exogenous stimuli such as ionising radiation and chemotherapeutic drugs 69. It has been suggested that alterations in frequency of apoptosis and the development of resistance to apoptosis induction may be explained by changes in apoptosis-regulating cytokines 146,147. The tumour necrosis factor (TNF) ligand family comprises cytokines that serve important functions as mediators of immune regulation, inflammation and apoptosis 148. Members of the TNF ligand family bind and interact with specific receptors of the TNF receptor (TNF-R) family. Binding of the ligands to their respective receptors triggers diverse intracellular signalling pathways, promoting cell differentiation, apoptosis, but also cell proliferation 149. Three major apoptosis-inducing cytokines have been identified so far 150. These are TNF-α, Fas ligand (FasL, also known as CD95L) and tumour necrosis factor-related apoptosis-inducing ligand (TRAIL), all of which induce apoptosis by binding to their respective receptors 151. Current knowledge on the possible roles of these cytokines in the development of colorectal cancer is briefly discussed below. The most important physiological function of TNF-α probably lies in the regulation of inflammation rather than apoptosis-induction and its role in colorectal cancer development seems limited 152. Fas mediates apoptotic cell death upon engagement by its natural ligand, FasL 153. Fas-mediated apoptosis is involved in several regulatory functions within the immune system (for recent reviews see Walczak 148 or Krammer 154 ). Human colonic epithelial cells constitutively express Fas and an intact Fas signalling pathway has been shown to be capable of regulating apoptosis in human colon carcinoma cell lines 146,147,155. It is known that many tumour cells escape from immune cytolysis and become resistant to FasL from anti-tumour immune effector cells 156. The mechanisms responsible for this resistance are complex, and several possibilities have been proposed, including downregulation of the Fas receptor and upregulation of FasL in tumour cells p53 mutations in colon carcinomas may also directly inhibit Fas signalling 156,160. A concept that has become known as the Fascounterattack has gained wide attention 156,161,162. It is based on the observations that tumour cells from various origins, including colon cancer, expressed FasL and induced apoptosis in Fas-expressing tumour-infiltrating lymphocytes and/or peripheral T cells 163. However, recent experiments could not confirm these results in several colon cancer cell lines 163. In summary, although the Fas/FasL pathway has been extensively studied in colorectal cancer and several experiments have suggested a possible role for disturbances in Fas/FasLmediated apoptosis in colorectal carcinogenesis, its precise impact still has to be defined and the counterattack hypothesis is seriously debated 164. TRAIL has been shown to induce apoptosis in a wide variety of tumour cells in vitro. Approximately 80% of human cancer cell lines, representing colon, lung, breast, skin, kidney and brain tumours were sensitive at least to some extent to TRAIL, whereas most normal cell types were relatively resistant Yet, little is known about the physiological 31

33 Chapter 2 role of TRAIL mediated apoptosis. In a murine model, it was recently shown that TRAIL was partly responsible for surveillance of tumour metastasis by liver natural killer cells, suggesting a physiological role as a tumour suppressor 168. For TRAIL, five different receptors have been identified: two cell death-inducing receptors (DR4/TRAIL-R1 and DR5/ TRAIL-R2), two non-cell-death-inducing receptors (DcR1/TRAIL-R3 and DcR2/TRAIL-R4) and osteoprotegerin 169,170. Until recently, levels of TRAIL-receptors were known only from mrna expression. TRAIL mrna and DR4 and DR5 transcripts are expressed in several tissues, including normal colon epithelium and colon adenocarcinomas 171. Future studies aimed at elucidating the expression of TRAIL and its receptors in the adenoma-carcinoma sequence will not only contribute to our general understanding of colorectal carcinogenesis but may also serve to design new therapeutic approaches aimed at stimulating apoptosis. TRAIL specifically appears to be a promising new anti-cancer agent which, unlike TNF-α and FasL, seems to induce tumour-selective apoptosis, while sparing normal cells 165,172. With respect to differences in sensitivity towards TRAIL-induced apoptosis between normal and tumour cells, several theories exist, including the decoy hypothesis. This hypothesis is based on the assumption that cellular sensitivity or resistance to TRAIL-induced apoptosis is dependent on the presence or absence of specific receptors 173. It was supposed that TRAIL does not induce apoptosis in normal cells because these are protected by the presence of the non-signaling receptors DcR1 and DcR2. However, examination of TRAIL receptors in large panels of cultured tumor cells showed no correlation between the expression of DcR1 and DcR2 mrna and the resistance or sensitivity to TRAIL treatment 174. In addition, it has been shown that the expression and localisation of TRAIL receptors varies between different cells and that resistance to TRAIL is mediated by different mechanisms such as the differential subcellular localisation of decoy receptors and intracellular inhibitors of apoptosis 175. Although the precise mechanism of TRAIL-mediated apoptosis in normal and tumour tissues as well as the clinical significance and the molecular biologic regulatory mechanisms have yet to be elucidated, the cell killing properties of this cytokine has made it an exciting target for drug development 176. Concerns of hepatotoxicity or brain toxicity associated with the use of TRAIL have been expressed following experiments where polyhistidine- and FLAG-tagged recombinant versions of human TRAIL were found to induce apoptosis in vitro in isolated human hepatocytes and brain tissue slices, respectively 177,178. However, in contrast to polyhistidine-tagged TRAIL, native rhtrail was nontoxic to cultured hepatocytes 179. In addition, two relevant non-human primate models have indicated that systemic administration of native rhtrail is unlikely to cause major toxicity to the liver or other organs 179,

34 Review: Apoptosis and colorectal cancer development 5. Possibilities for intervention: Apoptosis and chemoprevention Therapies designed to stimulate apoptosis in target cells play an increasingly central role in the prevention and treatment of both hereditary and non-hereditary colorectal cancer 181. Chemoprevention of colorectal cancer is defined as the use of pharmacological agents to prevent or reverse the development of adenomatous polyps and subsequent progression to colorectal cancer 181. The importance to identify drugs that can prevent disease development and progression, particularly in high-risk populations is generally recognised. Several agents have been studied for chemoprevention for colorectal cancer (for recent reviews, see Sharma 182 or Krishnan 183 ). Of these agents, non-steroidal anti-inflammatory drugs (NSAIDs) are currently most widely used 184,185. It has been shown that NSAIDs inhibit cell growth in culture, decrease the incidence of colon tumours in carcinogen induced murine models and decrease the relative risk of incidence and mortality of colorectal cancer in human epidemiological studies 186. Although NSAIDs may exert effects through a number of potential mechanisms, accumulating evidence suggests an effect of these drugs on apoptotic cell death 187. For example, numerous reports have shown that the NSAID sulindac induces apoptosis in epithelial cells in normal colonic mucosa and adenomas in familial adenomatous polyposis (FAP) patients and in a murine model of FAP 186,188,189. Several animal studies and cell culture experiments have confirmed the ability of sulindac to induce apoptosis Unfortunately, there is significant toxicity associated with long-term NSAID use, in particular gastrointestinal bleeding 193. Such toxicity seriously compromises the overall value of NSAID-mediated chemoprevention in at-risk individuals and necessitates the development of safer agents. Agents that directly induce apoptosis may reduce the risk of toxicity and reduce the opportunity for acquired drug resistance. A critical issue in chemoprevention is the relative sensitivity of cells at different stages of neoplastic progression. The molecular elements determining the sensitivity remain to be determined. Nevertheless, the identification of other biochemical targets for chemoprevention that specifically induces apoptosis would provide new strategies to prevent colonic neoplasms. As death receptor-mediated apoptosis is predicted to be independent of p53 status, application in treatment or prevention of colorectal cancer seems promising 194. In the coming years, it seems likely that rational strategies to manipulate apoptotic programs will produce new therapies that are less toxic than current chemoprevention strategies. 6. Conclusions The ability to undergo apoptosis is an important mechanism for maintaining control over a population of cells under continuous renewal, as in the colonic epithelium. Furthermore, apoptosis is particularly important for the elimination of cells with unrepaired DNA damage. A reduction of this ability would result in the retention of cells with DNA damage and a consequent increased risk of mutations, including those that are carcinogenic. Several genes that regulate apoptosis are inappropriately expressed or mutated in colorectal cancer. Although the idea of decreased apoptosis during colorectal cancer development is conceptually attractive, the majority of studies show that the proportion of cells undergoing apoptosis increases in the course of the adenoma-carcinoma sequence, possibly as a 33

35 Chapter 2 mechanism to delete cells with sustained DNA damage. As most in vitro studies indicate the emergence of gradual resistance to apoptosis-induction towards tumour progression, it might be that the increased cell turnover rates together with increase in apoptotic cell death selects for the survival of cells resistant to induction of apoptosis. Considering changes in the locoregional distribution of apoptotic cells, consensus emerges from the reviewed studies that in adenomas a reversal of the normal pattern of proliferation and apoptosis in colonic crypts is observed. Evidence is accumulating supporting an inward or top-down growth model of adenomas. The degree of apoptosis in colorectal tumours is not related to the degree of proliferation in most studies. The use of the apoptotic index in a tumour as a potential prognostic marker or indicator for the choice of therapy, as suggested by some, is currently not supported by the available data. In the reviewed studies, the methods used to identify apoptotic cell death varied widely, with in situ end labelling methods prevailing. As a consequence, reported apoptotic indices varied considerably. Considering the limitations of the methods of in situ end labelling of DNA fragments, used in the vast majority of studies, improved diagnostic criteria for the identification of apoptotic cells are needed and can be expected in the near future. Further identification of the pathways controlling apoptosis including the role of TNF ligand family members will provide opportunities to understand the homeostatic control of normal colon tissue and reveal new targets for prevention and therapy of colorectal cancer. 34

36 Review: Apoptosis and colorectal cancer development References 1. Midgley R, Kerr D. Colorectal cancer. Lancet 1999;353: Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61: Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. Cell 1996;87: Sinicrope FA, Roddey G, McDonnell TJ, Shen Y, Cleary KR, Stephens LC. Increased apoptosis accompanies neoplastic development in the human colorectum. Clin Cancer Res 1996;2: Kerr JF, Winterford CM, Harmon BV. Apoptosis. Its significance in cancer and cancer therapy. Cancer 1994;73: Bedi A, Pasricha PJ, Akhtar AJ, et al. Inhibition of apoptosis during development of colorectal cancer. Cancer Res 1995;55: Hao X, Du M, Bishop AE, Talbot IC. Imbalance between proliferation and apoptosis in the development of colorectal carcinoma. Virchows Arch 1998;433: Diebold J, Lai MD, Lohrs U. Analysis of proliferative activity in colorectal mucosa by immunohistochemical detection of proliferating cell nuclear antigen (PCNA). Methodological aspects and application to routine diagnostic material. Virchows Arch B Cell Pathol Incl Mol Pathol 1992;62: Johnston PG, O Brien MJ, Dervan PA, Carney DN. Immunohistochemical analysis of cell kinetic parameters in colonic adenocarcinomas, adenomas, and normal mucosa. Hum Pathol 1989;20: Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Science 1995;267: Watson AJ. Review article: manipulation of cell death-the development of novel strategies for the treatment of gastrointestinal disease. Aliment Pharmacol Ther 1995;9: Hawkins N, Lees J, Hargrave R, O Connor T, Meagher A, Ward R. Pathological and genetic correlates of apoptosis in the progression of colorectal neoplasia. Tumour Biol 1997;18: Leonardos L, Butler LM, Hewett PJ, Zalewski PD, Cowled PA. The activity of caspase-3-like proteases is elevated during the development of colorectal carcinoma. Cancer Lett 1999;143: Williams AC, Hague A, Elder DJ, Paraskeva C. In vitro models for studying colorectal carcinogenesis: cellular and molecular events including APC and Rb cleavage in the control of proliferation, differentiation and apoptosis. Biochim Biophys Acta 1996;1288:F Bernstein C, Bernstein H, Garewal H, et al. A bile acid-induced apoptosis assay for colon cancer risk and associated quality control studies. Cancer Res 1999;59: Garewal H, Bernstein H, Bernstein C, Sampliner R, Payne C. Reduced bile acid-induced apoptosis in normal colorectal mucosa: a potential biological marker for cancer risk. Cancer Res 1996;56: Diaz GD, Paraskeva C, Thomas MG, Binderup L, Hague A. Apoptosis is induced by the active metabolite of vitamin D3 and its analogue EB1089 in colorectal adenoma and carcinoma cells: possible implications for prevention and therapy. Cancer Res 2000;60: Hague A, Elder DJ, Hicks DJ, Paraskeva C. Apoptosis in colorectal tumour cells: induction by the short chain fatty acids butyrate, propionate and acetate and by the bile salt deoxycholate. Int J Cancer 1995;60: Elder DJ, Hague A, Hicks DJ, Paraskeva C. Differential growth inhibition by the aspirin metabolite 35

37 Chapter 2 salicylate in human colorectal tumor cell lines: enhanced apoptosis in carcinoma and in vitrotransformed adenoma relative to adenoma relative to adenoma cell lines. Cancer Res 1996;56: Elder DJ, Halton DE, Crew TE, Paraskeva C. Apoptosis induction and cyclooxygenase-2 regulation in human colorectal adenoma and carcinoma cell lines by the cyclooxygenase-2-selective non- steroidal anti-inflammatory drug NS-398. Int J Cancer 2000;86: Kerr JF, Wyllie AH, Currie AR. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 1972;26: Hawkins NJ, Lees J, Ward RL. Detection of apoptosis in colorectal carcinoma by light microscopy and in situ end labelling. Anal Quant Cytol Histol 1997;19: Carr NJ. M30 expression demonstrates apoptotic cells, correlates with In situ end-labeling, and is associated with Ki-67 expression in large intestinal neoplasms. Arch Pathol Lab Med 2000;124: Ansari B, Coates PJ, Greenstein BD, Hall PA. In situ end-labelling detects DNA strand breaks in apoptosis and other physiological and pathological states. J Pathol 1993;170: Gavrieli Y, Sherman Y, Ben Sasson SA. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992;119: Cummings MC, Winterford CM, Walker NI. Apoptosis. Am J Surg Pathol 1997;21: Grasl-Kraupp B, Ruttkay-Nedecky B, Koudelka H, Bukowska K, Bursch W, Schulte-Hermann R. In situ detection of fragmented DNA (TUNEL assay) fails to discriminate among apoptosis, necrosis, and autolytic cell death: a cautionary note. Hepatology 1995;21: Hayashi R, Ito Y, Matsumoto K, Fujino Y, Otsuki Y. Quantitative differentiation of both free 3 -OH and 5 -OH DNA ends between heat-induced apoptosis and necrosis. J Histochem Cytochem 1998;46: Wijsman JH, Jonker RR, Keijzer R, van de Velde CJ, Cornelisse CJ, van Dierendonck JH. A new method to detect apoptosis in paraffin sections: in situ end- labeling of fragmented DNA. J Histochem Cytochem 1993;41: Sloop GD, Roa JC, Delgado AG, Balart JT, Hines MO, III, Hill JM. Histologic sectioning produces TUNEL reactivity. A potential cause of false-positive staining. Arch Pathol Lab Med 1999;123: Potten CS. What is an apoptotic index measuring? A commentary. Br J Cancer 1996;74: Butler LM, Hewett PJ, Fitridge RA, Cowled PA. Deregulation of apoptosis in colorectal carcinoma: theoretical and therapeutic implications. Aust N Z J Surg 1999;69: Michael-Robinson JM, Biemer-Huttmann AE, Purdie DM, et al. Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status. Gut 2001;48: Leers MP, Kolgen W, Bjorklund V, et al. Immunocytochemical detection and mapping of a cytokeratin 18 neo- epitope exposed during early apoptosis. J Pathol 1999;187: Ward RL, Todd AV, Santiago F, O Connor T, Hawkins NJ. Activation of the K-ras oncogene in colorectal neoplasms is associated with decreased apoptosis. Cancer 1997;79: Aotake T, Lu CD, Chiba Y, Muraoka R, Tanigawa N. Changes of angiogenesis and tumor cell apoptosis during colorectal carcinogenesis. Clin Cancer Res 1999;5: Ikenaga M, Takano Y, Saegusa M, et al. Apoptosis of colon cancers assessed by in situ DNA nick endlabeling method. Pathol Int 1996;46:

38 Review: Apoptosis and colorectal cancer development 38. Ikenaga M, Takano Y, Ohtani Y, et al. Low levels of apoptosis and proliferative activity in colorectal villous tumors: comparison with tubular tumors. Pathol Int 1998;48: Koike M. Significance of spontaneous apoptosis during colorectal tumorigenesis. J Surg Oncol 1996;62: Liu LU, Holt PR, Krivosheyev V, Moss SF. Human right and left colon differ in epithelial cell apoptosis and in expression of Bak, a pro-apoptotic Bcl-2 homologue. Gut 1999;45: Moss SF, Scholes JV, Holt PR. Abnormalities of epithelial apoptosis in multistep colorectal neoplasia demonstrated by terminal deoxyuridine nick end labeling. Dig Dis Sci 1996;41: Moss SF, Liu TC, Petrotos A, Hsu TM, Gold LI, Holt PR. Inward growth of colonic adenomatous polyps. Gastroenterology 1996;111: Partik G, Kahl-Rainer P, Sedivy R, Ellinger A, Bursch W, Marian B. Apoptosis in human colorectal tumours: ultrastructure and quantitative studies on tissue localization and association with bak expression. Virchows Arch 1998;432: Regitnig P, Denk H. Lack of pseudomelanosis coli in colonic adenomas suggests different pathways of apoptotic bodies in normal and neoplastic colonic mucosa. Virchows Arch 2000;436: Takano Y, Saegusa M, Ikenaga M, Mitomi H, Okayasu I. Apoptosis of colon cancer: comparison with Ki-67 proliferative activity and expression of p53. J Cancer Res Clin Oncol 1996;122: Anti M, Armuzzi A, Morini S, et al. Severe imbalance of cell proliferation and apoptosis in the left colon and in the rectosigmoid tract in subjects with a history of large adenomas. Gut 2001;48: Kawasaki H, Toyoda M, Shinohara H, et al. Expression of survivin correlates with apoptosis, proliferation, and angiogenesis during human colorectal tumorigenesis. Cancer 2001; 91: Nakamura T, Sakai T, Nariya S. Cell death in colorectal polyps as evaluated by in situ 3 -tailing reaction and its relationship to BCL-2 expression. Pathol Int 1995;45: Arai T, Kino I. Role of apoptosis in modulation of the growth of human colorectal tubular and villous adenomas. J Pathol 1995;176: Fazeli A, Steen RG, Dickinson SL, et al. Effects of p53 mutations on apoptosis in mouse intestinal and human colonic adenomas. Proc Natl Acad Sci U S A 1997;94: Kikuchi Y, Dinjens WN, Bosman FT. Proliferation and apoptosis in proliferative lesions of the colon and rectum. Virchows Arch 1997;431: Nomura M, Watari J, Yokota K, Saitoh Y, Obara T, Kohgo Y. Morphogenesis of nonpolypoid colorectal adenomas and early carcinomas assessed by cell proliferation and apoptosis. Virchows Arch 2000;437: Yamamoto T, Igarashi N, Kato Y, Kobayashi M, Kawakami M. Apoptosis in adenoma and early adenocarcinoma of the colon. Histol Histopathol 1998;13: Baretton GB, Diebold J, Christoforis G, et al. Apoptosis and immunohistochemical bcl-2 expression in colorectal adenomas and carcinomas. Aspects of carcinogenesis and prognostic significance. Cancer 1996;77: Valentini AM, Caruso ML, Armentano R, et al. Programmed cell death in colorectal carcinogenesis. Anticancer Res 1999;19: Sinicrope FA, Roddey G, Lemoine M, et al. Loss of p21waf1/cip1 protein expression accompanies progression of sporadic colorectal neoplasms but not hereditary nonpolyposis colorectal cancers. Clin Cancer Res 1998;4:

39 Chapter Tsujitani S, Shirai H, Tatebe S, et al. Apoptotic cell death and its relationship to carcinogenesis in colorectal carcinoma. Cancer 1996;77: Japanese Research Society for Cancer of the Colon and Rectum. General rules for clinical and pathological studies on cancer of the colon, rectum and anus. Jpn J Surg 1993;13: Schlemper RJ, Itabashi M, Kato Y, et al. Differences in the diagnostic criteria used by Japanese and Western pathologists to diagnose colorectal carcinoma. Cancer 1998;82: Ozawa A, Konishi F, Fukayama M, Kanazawa K. Apoptosis and its regulation in flat-type early colorectal carcinoma: comparison with that in polypoid-type early colorectal carcinoma. Dis Colon Rectum 2000;43:S23-S Tanimoto T, Tanaka S, Haruma K, et al. Growth patterns in various macroscopic types of noninvasive intramucosal colorectal carcinoma with special reference to apoptosis and cell proliferation. Dis Colon Rectum 1998;41: Raff MC. Social controls on cell survival and cell death. Nature 1992;356: Hague A, Hicks DJ, Bracey TS, Paraskeva C. Cell-cell contact and specific cytokines inhibit apoptosis of colonic epithelial cells: growth factors protect against c-myc-independent apoptosis. Br J Cancer 1997;75: Muto T, Bussey HJ, Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975;36: Evan G, Littlewood T. A matter of life and cell death. Science 1998;281: Evan G. Cancer a matter of life and cell death. Int J Cancer 1997;71: Juin P, Hueber AO, Littlewood T, Evan G. c-myc-induced sensitization to apoptosis is mediated through cytochrome c release. Genes Dev 1999;13 : Tomlinson IP, Novelli MR, Bodmer WF. The mutation rate and cancer. Proc Natl Acad Sci USA 1996;93: Lowe SW, Lin AW. Apoptosis in cancer. Carcinogenesis 2000;21: Bach S, Renehan A, Potten C. Stem cells: the intestinal stem cell as a paradigm. Carcinogenesis 2000;21: Strater J, Koretz K, Gunthert AR, Moller P. In situ detection of enterocytic apoptosis in normal colonic mucosa and in familial adenomatous polyposis. Gut 1995;37: Hall PA, Coates PJ, Ansari B, Hopwood D. Regulation of cell number in the mammalian gastrointestinal tract: the importance of apoptosis. J Cell Sci 1994;107: Potten CS, Loeffler M. Stem cells: attributes, cycles, spirals, pitfalls and uncertainties. Lessons for and from the crypt. Development 1990;110: Potten CS, Wilson JW, Booth C. Regulation and significance of apoptosis in the stem cells of the gastrointestinal epithelium. Stem Cells 1997;15: Barkla DH, Gibson PR. The fate of epithelial cells in the human large intestine. Pathology 1999;31: Shiff SJ, Rigas B. Colon adenomatous polyps-do they grow inward? Lancet 1997;349: Lightdale C, Lipkin M, Deschner E. In vivo measurements in familial polyposis: kinetics and location of proliferating cells in colonic adenomas. Cancer Res 1982;42:

40 Review: Apoptosis and colorectal cancer development 78. Shih IM, Wang TL, Traverso G, et al. Top-down morphogenesis of colorectal tumors. Proc Natl Acad Sci U S A 2001;98: Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL. Natural history of untreated colonic polyps. Gastroenterology 1987;93: Sunter JP, Hull DL, Appleton DR, Watson AJ. Cell proliferation of colonic neoplasms in dimethylhydrazine-treated rats. Br J Cancer 1980;42: Tatebe S, Ishida M, Kasagi N, Tsujitani S, Kaibara N, Ito H. Apoptosis occurs more frequently in metastatic foci than in primary lesions of human colorectal carcinomas: analysis by terminaldeoxynucleotidyl-transferase-mediated dutp-biotin nick end labeling. Int J Cancer 1996;65: Evertsson S, Bartik Z, Zhang H, Jansson A, Sun XF. Apoptosis in relation to proliferating cell nuclear antigen and Dukes stage in colorectal adenocarcinoma. Int J Oncol 1999;15: Sinicrope FA, Hart J, Hsu HA, Lemoine M, Michelassi F, Stephens LC. Apoptotic and mitotic indices predict survival rates in lymph node- negative colon carcinomas. Clin Cancer Res 1999;5: Compton CC, Fielding LP, Burgart LJ, et al. Prognostic factors in colorectal cancer. College of American Pathologists Consensus Statement Arch Pathol Lab Med 2000;124: Bauer KD, Lincoln ST, Vera-Roman JM, et al. Prognostic implications of proliferative activity and DNA aneuploidy in colonic adenocarcinomas. Lab Invest 1987;57: Schutte B, Reynders MM, Wiggers T, et al. Retrospective analysis of the prognostic significance of DNA content and proliferative activity in large bowel carcinoma. Cancer Res 1987;47: Witzig TE, Loprinzi CL, Gonchoroff NJ, et al. DNA ploidy and cell kinetic measurements as predictors of recurrence and survival in stages B2 and C colorectal adenocarcinoma. Cancer 1991;68: Chang WC, Chapkin RS, Lupton JR. Predictive value of proliferation, differentiation and apoptosis as intermediate markers for colon tumorigenesis. Carcinogenesis 1997;18: Hashimoto S, Koji T, Kohara N, Kanematsu T, Nakane PK. Frequency of apoptosis relates inversely to invasiveness and metastatic activity in human colorectal cancer. Virchows Arch 1997;431: Kawasaki H, Altieri DC, Lu CD, Toyoda M, Tenjo T, Tanigawa N. Inhibition of apoptosis by survivin predicts shorter survival rates in colorectal cancer. Cancer Res 1998;58: Paradiso A, Simone G, Lena MD, et al. Expression of apoptosis-related markers and clinical outcome in patients with advanced colorectal cancer. Br J Cancer 2001;84: Tenjo T, Toyoda M, Okuda J, et al. Prognostic significance of p27(kip1) protein expression and spontaneous apoptosis in patients with colorectal adenocarcinomas. Oncology 2000;58: Sugamura K, Makino M, Kaibara N. Apoptosis as a prognostic factor in colorectal carcinoma. Surg Today 1998;28: Langlois NE, Lamb J, Eremin O, Heys SD. Apoptosis in colorectal carcinoma occurring in patients aged 45 years and under: relationship to prognosis, mitosis, and immunohistochemical demonstration of p53, c-myc and bcl-2 protein products. J Pathol 1997;182: Schwandner O, Schiedeck TH, Bruch HP, Duchrow M, Windhoevel U, Broll R. Apoptosis in rectal cancer: prognostic significance in comparison with clinical histopathologic, and immunohistochemical variables. Dis Colon Rectum 2000;43: Glinsky GV, Glinsky VV, Ivanova AB, Hueser CJ. Apoptosis and metastasis: increased apoptosis resistance of metastatic cancer cells is associated with the profound deficiency of apoptosis execution mechanisms. Cancer Lett 1997;115:

41 Chapter Glinsky GV. Apoptosis in metastatic cancer cells. Crit Rev Oncol Hematol 1997;25: Boland CR, Thibodeau SN, Hamilton SR, et al. A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998;58: Wheeler JM, Bodmer WF, Mortensen NJ. DNA mismatch repair genes and colorectal cancer. Gut 2000;47: Lothe RA, Peltomaki P, Meling GI, et al. Genomic instability in colorectal cancer: relationship to clinicopathological variables and family history. Cancer Res 1993;53: Sankila R, Aaltonen LA, Jarvinen HJ, Mecklin JP. Better survival rates in patients with MLH1-associated hereditary colorectal cancer. Gastroenterology 1996;110: Gryfe R, Kim H, Hsieh ET, et al. Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer. N Engl J Med 2000;342: Thibodeau SN, Bren G, Schaid D. Microsatellite instability in cancer of the proximal colon. Science 1993;260: Lengauer C, Kinzler KW, Vogelstein B. Genetic instability in colorectal cancers. Nature 1997;386: Zhang H, Richards B, Wilson T, et al. Apoptosis induced by overexpression of hmsh2 or hmlh1. Cancer Res 1999;59: Dolcetti R, Viel A, Doglioni C, et al. High prevalence of activated intraepithelial cytotoxic T lymphocytes and increased neoplastic cell apoptosis in colorectal carcinomas with microsatellite instability. Am J Pathol 1999;154: Chung DC. The genetic basis of colorectal cancer: insights into critical pathways of tumorigenesis. Gastroenterology 2000;119: Ilyas M, Straub J, Tomlinson IP, Bodmer WF. Genetic pathways in colorectal and other cancers. Eur J Cancer 1999;35: Groden J, Thliveris A, Samowitz W, et al. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 1991;66: Powell SM, Zilz N, Beazer-Barclay Y, et al. APC mutations occur early during colorectal tumorigenesis. Nature 1992;359: Miyashiro I, Senda T, Matsumine A, et al. Subcellular localization of the APC protein: immunoelectron microscopic study of the association of the APC protein with catenin. Oncogene 1995;11: Smith KJ, Johnson KA, Bryan TM, et al. The APC gene product in normal and tumor cells. Proc Natl Acad Sci U S A 1993;90: Morin PJ, Vogelstein B, Kinzler KW. Apoptosis and APC in colorectal tumorigenesis. Proc Natl Acad Sci U S A 1996;93: Goss KH, Groden J. Biology of the adenomatous polyposis coli tumor suppressor. J Clin Oncol 2000;18: Ilyas M, Tomlinson IP. The interactions of APC, E-cadherin and beta-catenin in tumour development and progression. J Pathol 1997;182: Korinek V, Barker N, Morin PJ, et al. Constitutive transcriptional activation by a beta-catenin-tcf complex in APC-/- colon carcinoma. Science 1997;275:

42 Review: Apoptosis and colorectal cancer development 117. Morin PJ, Sparks AB, Korinek V, et al. Activation of beta-catenin-tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 1997;275: Polakis P. Wnt signaling and cancer. Genes Dev 2000;14: Yamada T, Takaoka AS, Naishiro Y, et al. Transactivation of the multidrug resistance 1 gene by T-cell factor 4/beta-catenin complex in early colorectal carcinogenesis. Cancer Res 2000;60: Watson SA. Oncogenic targets of beta-catenin-mediated transcription in molecular pathogenesis of intestinal polyposis. Lancet 2001;357: Hollstein M, Sidransky D, Vogelstein B, Harris CC. p53 mutations in human cancers. Science 1991;253: Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997;88: Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal-tumor development. N Engl J Med 1988;319: Sigal A, Rotter V. Oncogenic mutations of the p53 tumor suppressor: the demons of the guardian of the genome. Cancer Res 2000;60: Shaw P, Bovey R, Tardy S, Sahli R, Sordat B, Costa J. Induction of apoptosis by wild-type p53 in a human colon tumor-derived cell line. Proc Natl Acad Sci U S A 1992;89: Bunz F, Hwang PM, Torrance C, et al. Disruption of p53 in human cancer cells alters the responses to therapeutic agents. J Clin Invest 1999;104: Merritt AJ, Potten CS, Kemp CJ, et al. The role of p53 in spontaneous and radiation-induced apoptosis in the gastrointestinal tract of normal and p53-deficient mice. Cancer Res 1994;54: Kobayashi M, Watanabe H, Ajioka Y, Yoshida M, Hitomi J, Asakura H. Correlation of p53 protein expression with apoptotic incidence in colorectal neoplasia. Virchows Arch 1995;427: De Angelis PM, Stokke T, Thorstensen L, Lothe RA, Clausen OP. Apoptosis and expression of Bax, Bclx, and Bcl-2 apoptotic regulatory proteins in colorectal carcinomas, and association with p53 genotype/ phenotype. Mol Pathol 1998;51: Paradiso A, Simone G, Petroni S, et al. Thymidilate synthase and p53 primary tumour expression as predictive factors for advanced colorectal cancer patients. Br J Cancer 2000;82: Sinicrope FA, Ruan SB, Cleary KR, Stephens LC, Lee JJ, Levin B. bcl-2 and p53 oncoprotein expression during colorectal tumorigenesis. Cancer Res 1995;55: Gross A, McDonnell JM, Korsmeyer SJ. BCL-2 family members and the mitochondria in apoptosis. Genes Dev 1999;13: Hockenbery D, Nunez G, Milliman C, Schreiber RD, Korsmeyer SJ. Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature 1990;348: Merritt AJ, Potten CS, Watson AJ, et al. Differential expression of bcl-2 in intestinal epithelia. Correlation with attenuation of apoptosis in colonic crypts and the incidence of colonic neoplasia. J Cell Sci 1995;108: Bosari S, Moneghini L, Graziani D, et al. bcl-2 oncoprotein in colorectal hyperplastic polyps, adenomas, and adenocarcinomas. Hum Pathol 1995;26: Bronner MP, Culin C, Reed JC, Furth EE. The bcl-2 proto-oncogene and the gastrointestinal epithelial tumor progression model. Am J Pathol 1995;146: Flohil CC, Janssen PA, Bosman FT. Expression of Bcl-2 protein in hyperplastic polyps, adenomas, and 41

43 Chapter 2 carcinomas of the colon. J Pathol 1996;178: Mosnier JF, Perret AG, Vindimian M, et al. An immunohistochemical study of the simultaneous expression of bcl-2 and p53 oncoproteins in epithelial tumors of the colon and rectum. Arch Pathol Lab Med 1996;120: Watson AJ, Merritt AJ, Jones LS, et al. Evidence of reciprocity of bcl-2 and p53 expression in human colorectal adenomas and carcinomas. Br J Cancer 1996;73: Ofner D, Riehemann K, Maier H, et al. Immunohistochemically detectable bcl-2 expression in colorectal carcinoma: correlation with tumour stage and patient survival. Br J Cancer 1995;72: Pereira H, Silva S, Juliao R, Garcia P, Perpetua F. Prognostic markers for colorectal cancer: expression of P53 and BCL2. World J Surg 1997;21: Manne U, Myers RB, Moron C, et al. Prognostic significance of Bcl-2 expression and p53 nuclear accumulation in colorectal adenocarcinoma. Int J Cancer 1997;74: Leahy DT, Mulcahy HE, O Donoghue DP, Parfrey NA. bcl-2 protein expression is associated with better prognosis in colorectal cancer. Histopathology 1999;35: Sinicrope FA, Hart J, Michelassi F, Lee JJ. Prognostic value of bcl-2 oncoprotein expression in stage II colon carcinoma. Clin Cancer Res 1995;1: Krajewska M, Moss SF, Krajewski S, Song K, Holt PR, Reed JC. Elevated expression of Bcl-X and reduced Bak in primary colorectal adenocarcinomas. Cancer Res 1996;56: Möller P, Koretz K, Leithauser F, et al. Expression of APO-1 (CD95), a member of the NGF/TNF receptor superfamily, in normal and neoplastic colon epithelium. Int J Cancer 1994;57: Von Reyher U, Sträter J, Kittstein W, Gschwendt M, Krammer PH, Moller P. Colon carcinoma cells use different mechanisms to escape CD95-mediated apoptosis. Cancer Res 1998;58 : Walczak H, Krammer PH. The CD95 (APO-1/Fas) and the TRAIL (APO-2L) apoptosis systems. Exp Cell Res 2000;256: Tschopp J, Martinon F, Hofmann K. Apoptosis: Silencing the death receptors. Curr Biol 1999;9:R381- R Salvesen GS, Dixit VM. Caspases: intracellular signaling by proteolysis. Cell 1997;91: MacFarlane M, Ahmad M, Srinivasula SM, Fernandes-Alnemri T, Cohen GM, Alnemri ES. Identification and molecular cloning of two novel receptors for the cytotoxic ligand TRAIL. J Biol Chem 1997;272: Remacle-Bonnet MM, Garrouste FL, Heller S, Andre F, Marvaldi JL, Pommier GJ. Insulin-like growth factor-i protects colon cancer cells from death factor-induced apoptosis by potentiating tumor necrosis factor alpha- induced mitogen-activated protein kinase and nuclear factor kappab signaling pathways. Cancer Res 2000;60: Nagata S. Apoptosis by death factor. Cell 1997;88: Krammer PH. CD95 s deadly mission in the immune system. Nature 2000;407: Houghton JA, Harwood FG, Gibson AA, Tillman DM. The fas signaling pathway is functional in colon carcinoma cells and induces apoptosis. Clin Cancer Res 1997;3: O Connell J, O Sullivan GC, Collins JK, Shanahan F. The Fas counterattack: Fas-mediated T cell killing by colon cancer cells expressing Fas ligand. J Exp Med 1996;184: Butler LM, Dobrovic A, Bianco T, Cowled PA. Promoter region methylation does not account for the 42

44 Review: Apoptosis and colorectal cancer development frequent loss of expression of the Fas gene in colorectal carcinoma. Br J Cancer 2000;82: Butler LM, Hewett PJ, Butler WJ, Cowled PA. Down-regulation of Fas gene expression in colon cancer is not a result of allelic loss or gene rearrangement. Br J Cancer 1998;77: Peduto EL, Guillou L, Saraga E, et al. Fas and Fas ligand expression in tumor cells and in vascular smooth- muscle cells of colonic and renal carcinomas. Int J Cancer 1999;81: O Connell J, Bennett MW, Nally K, Houston A, O Sullivan GC, Shanahan F. Altered mechanisms of apoptosis in colon cancer: Fas resistance and counterattack in the tumor-immune conflict. Ann N Y Acad Sci 2000;910: O Connell J, Bennett MW, O Sullivan GC, Collins JK, Shanahan F. Fas counter-attack-the best form of tumor defense? Nat Med 1999;5: O Connell J, Houston A, Bennett MW, O Sullivan GC, Shanahan F. Immune privilege or inflammation? Insights into the Fas ligand enigma. Nat Med 2001;7: Favre-Felix N, Fromentin A, Hammann A, Solary E, Martin F, Bonnotte B. Cutting edge: the tumor counterattack hypothesis revisited: colon cancer cells do not induce T cell apoptosis via the Fas (CD95, APO-1) pathway. J Immunol 2000;164: Restifo NP. Countering the counterattack hypothesis. Nat Med 2001;7: Walczak H, Miller RE, Ariail K, et al. Tumoricidal activity of tumor necrosis factor-related apoptosisinducing ligand in vivo. Nat Med 1999;5: Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A, Ashkenazi A. Induction of apoptosis by Apo- 2 ligand, a new member of the tumor necrosis factor cytokine family. J Biol Chem 1996;271: Wiley SR, Schooley K, Smolak PJ, et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 1995;3: Hayakawa Y, Smyth MJ, Kayagaki N, et al. Involvement of tumor necrosis factor-related apoptosisinducing ligand in surveillance of tumor metastasis by liver natural killer cells. Nat Med 2001;7: Ashkenazi A, Dixit VM. Apoptosis control by death and decoy receptors. Curr Opin Cell Biol 1999;11: Degli-Esposti M. To die or not to die-the quest of the TRAIL receptors. J Leukoc Biol 1999;65: Ashkenazi A, Dixit VM. Death receptors: signaling and modulation. Science 1998;281: French LE, Tschopp J. The TRAIL to selective tumor death. Nat Med 1999;5: Gura T. How TRAIL kills cancer cells, but not normal cells. Science 1997;277: Griffith TS, Rauch CT, Smolak PJ, et al. Functional analysis of TRAIL receptors using monoclonal antibodies. J Immunol 1999;162: Zhang XD, Nguyen T, Thomas WD, Sanders JE, Hersey P. Mechanisms of resistance of normal cells to TRAIL induced apoptosis vary between different cell types. FEBS Lett 2000;482: de Vries EG, van der Graaf WT, Heijenbrok FJ, Hoekstra HJ, de Jong S. Don t blaze the trailblazer TRAIL too early. Lancet 2000;356: Jo M, Kim TH, Seol DW, et al. Apoptosis induced in normal human hepatocytes by tumor necrosis factor- related apoptosis-inducing ligand. Nat Med 2000;6:

45 Chapter Nitsch R, Bechmann I, Deisz RA, et al. Human brain-cell death induced by tumour-necrosis-factorrelated apoptosis-inducing ligand (TRAIL). Lancet 2000;356: Lawrence D, Shahrokh Z, Marsters S, et al. Differential hepatocyte toxicity of recombinant Apo2L/ TRAIL versions. Nat Med 2001;7: Ashkenazi A, Pai RC, Fong S, et al. Safety and antitumor activity of recombinant soluble Apo2 ligand. J Clin Invest 1999;104: Hawk E, Lubet R, Limburg P. Chemoprevention in hereditary colorectal cancer syndromes. Cancer 1999;86: Sharma RA, Manson MM, Gescher A, Steward WP. Colorectal cancer chemoprevention. biochemical targets and clinical development of promising agents. Eur J Cancer 2001;37: Krishnan K, Ruffin MT, Brenner DE. Chemoprevention for colorectal cancer. Crit Rev Oncol Hematol 2000;33: Kubba AK. Non steroidal anti-inflammatory drugs and colorectal cancer: is there a way forward? Eur J Cancer 1999;35: Janne PA, Mayer RJ. Chemoprevention of colorectal cancer. N Engl J Med 2000;342: Keller JJ, Offerhaus GJ, Polak M, et al. Rectal epithelial apoptosis in familial adenomatous polyposis patients treated with sulindac. Gut 1999;45: Marx J. Cancer research. Anti-inflammatories inhibit cancer growth--but how? Science 2001;291: Mahmoud NN, Boolbol SK, Dannenberg AJ, et al. The sulfide metabolite of sulindac prevents tumors and restores enterocyte apoptosis in a murine model of familial adenomatous polyposis. Carcinogenesis 1998;19: Pasricha PJ, Bedi A, O Connor K, et al. The effects of sulindac on colorectal proliferation and apoptosis in familial adenomatous polyposis. Gastroenterology 1995;109: Piazza GA, Rahm AK, Finn TS, et al. Apoptosis primarily accounts for the growth-inhibitory properties of sulindac metabolites and involves a mechanism that is independent of cyclooxygenase inhibition, cell cycle arrest, and p53 induction. Cancer Res 1997;57: Rice PL, Goldberg RJ, Ray EC, Driggers LJ, Ahnen DJ. Inhibition of extracellular signal-regulated kinase 1/2 phosphorylation and induction of apoptosis by sulindac metabolites. Cancer Res 2001;61: Chan TA, Morin PJ, Vogelstein B, Kinzler KW. Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis. Proc Natl Acad Sci U S A 1998;95: Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of nonsteroidal antiinflammatory drugs. N Engl J Med 1999;340: Nicholson DW. From bench to clinic with apoptosis-based therapeutic agents. Nature 2000;407: Miki Y, Nishisho I, Miyoshi Y, Utsunomiya J, Nakamura Y. Interstitial loss of the same region of 5q in multiple adenomas and a carcinoma derived from an adenomatous polyposis coli (APC) patient. Genes Chromosomes Cancer 1992;4: Hao X, Bishop AE, Wallace M, et al. Early expression of cyclo-oxygenase-2 during sporadic colorectal carcinogenesis. J Pathol 1999;187: Scholzel S, Zimmermann W, Schwarzkopf G, Grunert F, Rogaczewski B, Thompson J. 44

46 Review: Apoptosis and colorectal cancer development Carcinoembryonic antigen family members CEACAM6 and CEACAM7 are differentially expressed in normal tissues and oppositely deregulated in hyperplastic colorectal polyps and early adenomas. Am J Pathol 2000;156: Seong J, Chung EJ, Kim H, et al. Assessment of biomarkers in paired primary and recurrent colorectal adenocarcinomas. Int J Radiat Oncol Biol Phys 1999;45:

47

48 Chapter 3 Assessment of apoptosis by M30 immunoreactivity and the correlation with morphological criteria in normal colorectal mucosa, adenomas and carcinomas. J.J. Koornstra 1,2, F.E.M. Rijcken 1, S. de Jong 2, H. Hollema 3, E.G.E. de Vries 2, J.H. Kleibeuker 1. Departments of 1 Gastroenterology and Hepatology; 2 Medical Oncology and 3 Pathology. University of Groningen Medical Centre, the Netherlands. Histopathology 2004;44:9-17

49 Chapter 3 Abstract Background: Although TUNEL and ISEL are the methods most often used to demonstrate and quantify apoptosis in histological tissue sections, the interpretation and specificity of these techniques have been controversial. Immunohistochemistry using the monoclonal antibody M30 that recognises caspase-cleaved cytokeratin 18 is considered a promising alternative but has yet to be validated against a generally accepted standard. Methods: Paraffin sections of normal colonic mucosa (n=30), normal mucosa obtained from resection margins from carcinomas (n=30), colorectal adenomas (n=84) and carcinomas (n=40) were studied. Apoptosis of epithelial cells was assessed by M30 immunoreactivity and morphological criteria and expressed as a proportion of the total number of cells counted (apoptotic index). Results: Mean apoptotic indices assessed by M30 immunoreactivity were 0.18 ± 0.04 % in normal mucosa, 0.42 ± 0.04 % in adenomas and 1.97 ± 0.24 % in carcinomas. Apoptotic indices using morphological criteria were 0.23 ± 0.03 %, 0.62 ± 0.06 % and 1.78 ± 0.19 %, respectively. Apoptotic counts were higher in normal mucosa obtained from resection margins than in genuinely normal mucosa using the M30 antibody. Apoptotic indices obtained by M30 immunoreactivity and morphological criteria were positively correlated (r = 0.71, p < 0.01). Conclusion: Assessment of apoptotic cells by M30 immunoreactivity correlates well with morphological criteria. Apoptotic indices increase in the course of the adenoma-carcinoma sequence. Apoptosis in normal mucosa obtained from resection margins differs from genuinely normal mucosa that warrants caution when interpreting studies on apoptosis in normal colonic mucosa. Our findings support the use of M30 immunoreactivity in the study of apoptosis in colorectal tissues. Introduction The adenoma-carcinoma sequence of the colon represents one of the most well studied and characterised models of tumour progression 1. The gradual progression from normal mucosa through adenoma to colorectal cancer is accompanied by genetic alterations that affect normal tissue homeostasis 2. Alterations in apoptosis and proliferation contribute to carcinogenesis 3. Indeed, many studies indicate that both proliferative activity and apoptosis are different in various stages of colorectal cancer development 2. Although it is generally accepted that proliferative activity gradually increases with tumour progression, considerable controversy exists about whether the frequency of apoptosis increases or decreases in the course of the adenoma-carcinoma sequence 4. In most studies, mainly in-situ end-labelling techniques of DNA fragments (TUNEL and ISEL) have been used to obtain data concerning apoptotic cell death in human colorectal mucosa 4. However, serious reservations have been expressed concerning the applicability of these methods to evaluate apoptosis in the gastrointestinal tract 5-7. Recently, several antibodies have been introduced that specifically identify apoptotic cells in different human cell types 8,9. One of these is the murine monoclonal antibody 48

50 Detection of apoptosis by M30 immunoreactivity M30, which reacts with a caspase-cleaved product of cytokeratin Immunoreactivity of M30 is confined to the cytoplasm of apoptotic epithelial cells and is expressed during early apoptosis 11. It has been suggested that this antibody is a promising alternative for TUNEL to detect apoptotic cells 7. The identification of apoptotic cells by morphological criteria with light-microscopic examination of hematoxylin-eosin stained sections is generally considered as the reference standard 3,4,6, However, no studies are available that have validated the M30 method against this reference. Another limitation of previous studies on apoptosis in normal colon epithelium is that in most studies normal mucosa had been obtained from resection margins from surgically removed carcinomas 4. One could argue whether this tissue represents genuinely normal mucosa. The primary aim of this study was to assess the frequency of apoptosis in different stages in colon cancer development by M30 immunoreactivity and compare it with morphological identification of apoptotic cells. Second, possible correlations between the degree of apoptosis, proliferative activity and histopathological characteristics were explored. Finally, we compared the frequency and localisation of apoptotic cells in normal mucosa obtained from resection margins from surgical specimens with that in genuinely normal mucosa. Materials and methods Patients and tissue samples Paraffin-embedded specimens from 84 colorectal adenomas and 40 adenocarcinomas were retrieved from the files of the Department of Pathology. The adenomas were of at least 3-mm circumferential size and had been consecutively removed by endoscopy in 1997 at the Department of Gastroenterology. Cases of primary adenocarcinomas had been consecutively removed in 1999 at the Department of Surgery. Samples from patients with familial adenomatous polyposis or hereditary non-polyposis colorectal cancer were excluded. None of the patients had received chemo- or radiotherapy prior to resection or removal of the tumour. In addition, tissue sections were selected from the 40 patients with carcinomas from the non-involved resection margins with normal macroscopic appearance. From these cases, 30 samples of resection margins were available. For comparison, 30 samples of genuinely normal colonic mucosa were selected from archival materials, which had been consecutively obtained in These samples were from patients with normal macroscopic findings at colonoscopy and from whom random biopsies had been taken which had not shown any abnormalities at histologic examination. If the biopsies had been taken from normal mucosa from patients with a history of adenomas or carcinomas, they were excluded. Histologic classifications were carried out on hematoxylin-eosin (HE) stained slides. The morphological classification of the carcinomas and adenomas was conducted according to the World Health Organisation (WHO) criteria 16. From adenomas, the circumferential size was measured, and the severity of dysplasia expressed as low- or high-grade dysplasia. Adenomas were classified as tubular, tubulovillous or villous. For statistical purposes, adenomas with tubulovillous and villous architecture were taken together. Adenocarcinomas 49

51 Chapter 3 were graded into well, moderately and poorly differentiated. For statistical purposes, tumours with Dukes stages A and B were compared with those with stages C and D. The location of adenomas and carcinomas was retrieved from endoscopy, surgical or pathology reports. The caecum, ascending colon and transverse colon were regarded as the proximal colon, while the descending colon and sigmoid were referred to as the distal colon. Rectal localisation was recorded separately. Immunohistochemistry For immunohistochemical staining, 3 μm-thick sections were cut from paraffin blocks. After deparaffinisation, endogenous peroxidase was blocked with 0.3 % hydrogen peroxide for 30 min. The primary antibodies M30 (Boehringer Mannheim, GmbH, Mannhein, Germany) and MIB-1 (against Ki-67, Immunotech, Marseilles, France) were applied in a 1:50 and 1:400 solution, respectively. Antigen retrieval was performed, for MIB-1 in a pressure cooker, for M30 in a microwave oven for 8 min in 0.01 M citrate buffer (ph 6.0). After 1 h incubation with the primary antibody, samples were incubated with a secondary rabbit-anti-mouse antibody conjugated with peroxidase (DAKO, Glostrup, Denmark) and a tertiary goat-anti-rabbit peroxidase-conjugated antibody (DAKO) to intensify the staining reaction. Counterstaining was performed with hematoxylin. As negative controls, slides were immunostained in the absence of the primary antibody. Evaluation of staining results Evaluation of staining was performed by two investigators using light microscopy in serial tissue sections. Quantitative analysis was performed in complete crypts in normal mucosa and adenomas, and at random in carcinomas. M30 positivity was identified as brown cytoplasmic staining. Morphological characteristics including the presence of apoptotic bodies, nuclear condensation, cytoplasmic shrinkage and membrane blebbing were assessed in HE stained tissue sections. In all cases, at least 1000 epithelial cells were counted and morphological apoptotic and M30 positive cells were expressed as a percentage of the total number of cells counted (apoptotic index). In normal mucosa and adenomas, the localisation of M30 positive cells was recorded in the upper and lower half of the crypt separately as positive or negative. Apoptotic cells located in the lumen, assessed either by morphological criteria or M30 positivity, were excluded. For evaluation of MIB-1 staining in normal mucosa and adenomas, complete crypts were counted. In carcinomas, counts were made randomly. In all samples, at least 1000 cells were counted and MIB-1 positive cells were expressed as a percentage of the total number of epithelial cells counted. Statistical analysis The correlation between the identification of apoptotic cells by morphological criteria and the M30 method was assessed using the non-parametric Spearman test. Mean M30 and MIB- 1 counts were compared across different types of colon tissue and different histopathological characteristics, using the Mann-Whitney test. Multiple linear regression analysis was used to compare apoptotic and proliferative indices across different types of colon tissue and different histopathological characteristics after adjusting for potential confounding by other variables such as size, tumour type, stage and grade. Differences in distribution of apoptotic cells 50

52 Detection of apoptosis by M30 immunoreactivity Table 1. Frequency of apoptosis, assessed by M30 immunoreactivity and proliferative activity in adenomas and carcinomas in relation to histopathological characteristics. Apoptotic index (AI) and proliferative index (PI) are expressed as mean ± SEM, n represents number of cases. n AI p p adjusted # PI p p adjusted # Adenomas Size < 10 mm ± ± mm ± ± 3.3 ns ns Growth Tubular ± ± 2.4 ns ns Villous ± ± 3.2 ns ns Dysplasia Low-grade ± ± 2.1 ns ns High-grade ± ± Location Proximal ± ± 3.7 Distal ± * ns 32.5 ± 2.9 ns ns Rectum ± ± 3.2 Carcinomas Stage Dukes A/B ± ± Dukes C/D ± ± 4.4 ns ns Grade Well ± ± 12.6 moderate ± 0.31 ns ns 57.1 ± 4.4 ns ns Poor ± ± 7.8 Location Proximal ± ± 6.5 Distal ± ** ns 55.0 ± 8.0 ns ns Rectum ± ± 4.9 # after multiple linear regression analysis; * for the difference between proximal and rectal adenomas; ** for the difference between distal and rectal carcinomas. between genuinely normal mucosa and normal mucosa obtained from resection margins were assessed using chi-square tests. Correlations between apoptosis and proliferative activity and continuous histopathological variables were assessed using the Spearman test. P-values below 0.05 were considered significant. SPSS for Windows software (SPSS Inc., Chicago, IL, USA) was used in all statistical analyses. Results Patient and histological characteristics The mean age of patients at the time of adenoma removal was 64.2 ± 13.4 years (range 28-89), and the male/female ratio was 1.8. The mean age at time of cancer diagnosis was 65.2 ± 12.4 years (range 40-88), with a male/female ratio of 1.1. Histopathological characteristics of the adenomas and carcinomas are depicted in table I. Mean adenoma size was 9.6 mm (range 3-42). Apoptosis and proliferative activity in the adenoma-carcinoma sequence Figure 1 summarises the apoptotic indices as assessed by M30 immunoreactivity (A) and proliferative indices (B) in different colon tissues. The mean apoptotic index in genuinely normal epithelial mucosa was 0.18 ± 0.04 %, which was lower than in normal mucosa adjacent to carcinomas: 0.41 ± 0.03 %. The mean apoptotic index in adenomas 51

53 Chapter 3 1A 4 p< P< P< ns Apoptotic index (%) 2 P< B normal adjacent adenomas carcinomas P< P< P< P<0.001 Proliferative index (%) ns P < normal adjacent adenomas carcinomas Figure 1. Apoptotic indices assessed by M30 immunoreactivity (A) and proliferative indices (B) in normal mucosa, normal mucosa from resection margins ( adjacent ), adenomas and carcinomas. Box plots show median value, interquartile ranges and 90% ranges (extremes not shown). ns: not significant.. 52

54 Detection of apoptosis by M30 immunoreactivity Figure 2. Examples of M30 immunoreactivity (A-B; magnification x 100), Ki-67 staining (C-D; magnification x 100) and HE-staining (E; magnification x 400) in normal mucosa. (A) Genuinely normal colon mucosa, M30 labelling is predominantly limited to cells at the mucosal surface (arrow). (B) Normal mucosa obtained from a resection margin, M30 labelling is seen along the crypt axis (arrows). (C) Normal mucosa and (D) normal mucosa obtained from resection margins show Ki-67 positive cells limited to the proliferation zone in both cases with comparable proliferative indices. (E) High power view of normal mucosa obtained from a resection margin, showing the presence of apoptotic cells (arrows) in the crypt base. See appendix for colour pictures. was 0.42 ± 0.04 %, lower than the apoptotic index of 1.97 ± 0.24 % in carcinomas. Using morphological criteria, these indices were 0.23 ± 0.03 %, 0.58 ± 0.05 %, 0.62 ± 0.06 % and 1.78 ± 0.19 %, respectively. There was a positive correlation between the frequency of apoptosis as assessed by M30 immunoreactivity and morphological criteria (Spearman correlation coefficient 0.71, p < 0.01). In genuinely normal epithelial mucosa, the mean percentage of proliferating cells was 27.4 ± 1.6 %, comparable with that in normal mucosa from resection margins with a mean of 27.6 ± 2.3 %. Mean proliferative activity in adenomas was 35.3 ± 1.93 %, higher than in normal mucosa and normal mucosa from resection margins but lower than that in carcinomas: 56.5 ± 3.7 %. 53

55 Chapter 3 Apoptosis assessed by M30 immunoreactivity in relation to histopathological characteristics and proliferative activity Results are summarised in table 1. In adenomas, the degree of apoptosis was higher in adenomas 10 mm than in smaller ones and higher in proximal than rectal adenomas. There was no difference between adenomas with different grades of dysplasia or growth type. When multiple regression analysis was performed adjusting for all variables (size, localisation, dysplasia and growth type), only size larger than 10 mm remained significant (p=0.04), indicating that size was the only variable independently associated with a high apoptotic index. There was also a significant but weak linear correlation between the apoptotic index and adenoma size (r = 0.235, p < 0.05) but no correlation with proliferative activity (r = -0.15, p = 0.89). Proliferative activity in adenomas was higher in lesions with high-grade dysplasia compared to those with low-grade dysplasia. In carcinomas, the degree of apoptosis was higher in Dukes stages C and D compared to A and B and higher in rectal than in distal tumours. However, after multiple regression analysis only the difference between stages remained significant (p = 0.02), indicating that a higher Dukes stage was the only variable independently associated with a high apoptotic index. As in adenomas, the degree of apoptosis was not correlated with proliferative activity (r = 0.22, p = 0.16) in carcinomas. Distribution of apoptotic cells assessed by M30 immunoreactivity in normal mucosa, adenomas and carcinomas Examples of M30 immunostaining are shown in figures 2 and 3. In genuinely normal mucosa, apoptotic cells were exclusively seen in the upper (luminal) half of the crypt and the luminal surface (figure 2A). In contrast, in normal mucosa obtained from resection margins, apoptotic cells were not only observed in these regions but also in the lower half of the crypts in 18 of 30 cases (figure 2B; p < 0.01 for the difference with genuinely normal mucosa). Similar results were obtained when morphological criteria were used (figure 2E). Proliferating cells in both types of normal mucosa were exclusively located in the lower half of the crypt (figure 2 C-D). In adenomas, apoptotic cells were seen in both the upper and lower half of the crypts in all cases (figure 3, A-B). In carcinomas, apoptotic cells were randomly distributed throughout the tumour (figure 3 C-D). Discussion The main physiological mechanism by which colon epithelial cells die is by apoptosis 17. Apoptosis has been studied extensively in the course of the adenoma-carcinoma sequence. In a recent systematic review of these studies, a wide variety in reported apoptotic indices was found, differing up to a 100-fold 4. This variety can be partly attributed to drawbacks of TUNEL and ISEL, the methods that were most often used to demonstrate apoptosis 6. Although the morphological changes that are seen in hematoxylin-eosin stained tissue sections are considered the reference standard for recognising apoptotic cells, neither TUNEL nor morphological investigation discriminates between various cell types, e.g. apoptotic epithelial cells cannot be differentiated from apoptotic lymphocytes or apoptotic 54

56 Detection of apoptosis by M30 immunoreactivity Figure 3. Examples of M30 immunostaining in an adenoma (A-B) and a carcinoma (C-D). (A) M30 positive cells are seen along the entire crypt (magnification x 100). (B) Detail of a crypt containing numerous M30 positive cells with strong cytoplasmic staining (magnification x 400). (C) M30 positive cells are heterogeneously distributed throughout the carcinoma (magnification x 100). (D) High power view of the carcinoma shown in C (magnification x 400). See appendix for colour pictures. mesenchymal cells. A clear advantage of the monoclonal antibody M30 over in situ endlabelling techniques and morphological methods is that M30 is exclusively expressed in apoptotic epithelial cells 10,11. M30 immunoreactivity was shown to be positive in epithelial cells with apoptotic characteristics such as chromatin condensation, nuclear fragmentation and detachment of cytoplasm from the environment, whereas cells with normal morphology and end-stage apoptotic cells were negative 11. In the present study, M30 staining and morphological identification of apoptotic epithelial cells were well correlated. In most cases, apoptotic counts were slightly higher using morphological criteria compared to M30 immunostaining. This may suggest false-negative staining results with the M30 method. However, it must be noted that cells with morphological criteria of apoptosis are of various, including epithelial, origin, whereas M30 positivity is limited exclusively to epithelial cells. Second, M30 immunoreactivity eventually disappears in end-stage apoptotic cells 11. In a previous study, M30 immunoexpression was compared with in situ end-labelling in colorectal neoplasms, showing a strong positive correlation between the two methods 18. Hence, the current validation of M30 immunostaining against two reference methods supports the application of this technique in the study of apoptosis in colorectal epithelial cells. Using M30 immunohistochemistry, we found a higher apoptotic index in carcinomas than in adenomas and the lowest counts in normal mucosa. The mean apoptotic index of 1.97 % in our series of carcinomas is in accordance with the other studies in which apoptotic cell death was assessed using M30 immunoreactivity. In these studies, apoptotic indices of 1.50 % and 2.46 % were found respectively Although this may suggest some 55

57 Chapter 3 inconsistency in the results of M30 immunostaining, the variation in apoptotic indices with M30 is considerably less compared to the range of % reported using in-situ end-labelling methods 4. We found a lower degree of apoptosis in adenomas using M30 immunoreactivity as compared to a study in which 31 adenomas were included showing a mean apoptotic index of 0.98 % 18. This difference may be attributable to differences in adenoma characteristics. Using two different methods, we found that the proportion of epithelial cells undergoing apoptotic cell death increases in the course of the adenoma-carcinoma sequence, in accordance with the majority of studies on this subject 4. At first glance, this may seem contradictory to the general idea, mostly based on in vitro experiments, that tumour progression is associated with a decrease in apoptotic cell death 19. On the other hand, it is well known that colorectal epithelial cells accumulate genetic alterations in the course of the adenoma-carcinoma sequence 20. As DNA damage can be a trigger for the induction of apoptosis 3, one could well expect higher rates of spontaneous apoptotic cell death towards carcinoma development. It has been shown that the presence of aneuploidy in adenomas is associated with adenoma size, architecture and grade of dysplasia, but that only adenoma size is correlated to the presence of specific gene mutations 21,22. Interestingly, the only variable independently associated with a high apoptotic index in adenomas in our study was size larger than 10 mm. The absence of a correlation between apoptosis and proliferation in colorectal neoplasms, as found in this and other studies 4, suggests that apoptotic cell death is not primarily driven by increasing proliferative activity. Alternatively, apoptotic cell death may be induced by accumulating genetic alterations. It must be noted that in the vast majority of studies on apoptosis in normal mucosa, histologically normal mucosa was investigated adjacent to carcinomas, usually obtained from resection margins 4. One could argue whether this represents genuinely normal mucosa. Indeed, we found a higher degree of apoptosis in normal mucosa obtained from resection margins than in genuinely normal mucosa. This is in accordance with the only other study in which this was investigated 23. In that study, in which TUNEL was used, apoptotic indices were 11% and 3% in adjacent normal mucosa and genuinely normal mucosa, respectively. The discrepancy in apoptotic indices between this study as compared to ours may be explained by the fact that the time scale over which TUNEL detects DNA fragmentation in the cell can be assumed to be longer than the time during which cleaved cytokeratin 18 is expressed in an apoptotic cell 18. Not only the degree of apoptosis, but also the distribution of apoptotic cells was found in our study to be different between genuinely normal mucosa and mucosa obtained from resection margins. Our finding of apoptotic cells in the lower half of the crypts in normal mucosa from resection margins, a pattern similar to that in adenomas, supports the assumption that this type of mucosa does not represent genuinely normal mucosa. We found no difference in proliferative activity between genuinely normal mucosa and normal mucosa obtained from resection margins, suggesting that the difference in degree of apoptosis is probably not driven by different proliferative activities. The higher degree of apoptosis may be explained on technical grounds, for example hypoxia-induced apoptosis, occurring during the surgical procedure or delayed fixation procedures. Irrespective of the underlying mechanism, the fact that the degree and distribution of apoptotic cells differ 56

58 Detection of apoptosis by M30 immunoreactivity between normal mucosa from resection margins and genuinely normal mucosa must be kept in mind when assessing studies on apoptosis in normal colon mucosa. In conclusion, M30 immunohistochemistry is well suited for the demonstration of apoptotic cells in benign and malignant colorectal tissues. Our results confirm previous studies that the degree of apoptosis increases in the course of the adenoma-carcinoma sequence. 57

59 Chapter 3 References 1. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100: Chung DC. The genetic basis of colorectal cancer: insights into critical pathways of tumorigenesis. Gastroenterology 2000;119 : Kerr JF, Winterford CM, Harmon BV. Apoptosis. Its significance in cancer and cancer therapy. Cancer 1994;73: Koornstra JJ, de Jong S, Hollema H, de Vries EG, Kleibeuker JH. Changes in apoptosis during the development of colorectal cancer: a systematic review of the literature. Crit Rev Oncol Hematol 2003;45: Michael-Robinson JM, Biemer-Huttmann AE, Purdie DM, et al. Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status. Gut 2001;48: Potten CS. What is an apoptotic index measuring? A commentary. Br J Cancer 1996;74: Walker JA, Quirke P. Viewing apoptosis through a TUNEL. J Pathol 2001;195: Marshman E, Ottewell PD, Potten CS, Watson AJ. Caspase activation during spontaneous and radiation-induced apoptosis in the murine intestine. J Pathol 2001;195: Suurmeijer AJ, van der WJ, van Veldhuisen DJ, Yang F, Cole GM. Fractin immunostaining for the detection of apoptotic cells and apoptotic bodies in formalin-fixed and paraffin-embedded tissue. Lab Invest 1999;79: Caulin C, Salvesen GS, Oshima RG. Caspase cleavage of keratin 18 and reorganization of intermediate filaments during epithelial cell apoptosis. J Cell Biol 1997;138: Leers MP, Kolgen W, Bjorklund V, et al. Immunocytochemical detection and mapping of a cytokeratin 18 neo- epitope exposed during early apoptosis. J Pathol 1999;187: Arai T, Kino I. Role of apoptosis in modulation of the growth of human colorectal tubular and villous adenomas. J Pathol 1995;176: Langlois NE, Lamb J, Eremin O, Heys SD. Apoptosis in colorectal carcinoma occurring in patients aged 45 years and under: relationship to prognosis, mitosis, and immunohistochemical demonstration of p53, c-myc and bcl-2 protein products. J Pathol 1997;182: Sinicrope FA, Roddey G, McDonnell TJ, Shen Y, Cleary KR, Stephens LC. Increased apoptosis accompanies neoplastic development in the human colorectum. Clin Cancer Res 1996;2: Hawkins NJ, Lees J, Ward RL. Detection of apoptosis in colorectal carcinoma by light microscopy and in situ end labelling. Anal Quant Cytol Histol 1997;19: Jass JR, Sobin LH, Watanabe H. The World Health Organization s histologic classification of gastrointestinal tumors. A commentary on the second edition. Cancer 1990;66: Sträter J, Koretz K, Gunthert AR, Moller P. In situ detection of enterocytic apoptosis in normal colonic mucosa and in familial adenomatous polyposis. Gut 1995;37: Carr NJ. M30 expression demonstrates apoptotic cells, correlates with In situ end-labeling, and is associated with Ki-67 expression in large intestinal neoplasms. Arch Pathol Lab Med 2000;124: Bedi A, Pasricha PJ, Akhtar AJ, et al. Inhibition of apoptosis during development of colorectal cancer. Cancer Res 1995;55:

60 Detection of apoptosis by M30 immunoreactivity 20. Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. Cell 1996;87: Brabletz T, Herrmann K, Jung A, Faller G, Kirchner T. Expression of nuclear beta-catenin and c-myc is correlated with tumor size but not with proliferative activity of colorectal adenomas. Am J Pathol 2000; 156: Saraga E, Bautista D, Dorta G, et al. Genetic heterogeneity in sporadic colorectal adenomas. J Pathol 1997;181: Moss SF, Scholes JV, Holt PR. Abnormalities of epithelial apoptosis in multistep colorectal neoplasia demonstrated by terminal deoxyuridine nick end labeling. Dig Dis Sci 1996;41:

61

62 Chapter 4 Expression of TRAIL (TNF-related apoptosisinducing ligand) and its receptors in normal colon mucosa, adenomas and carcinomas. J.J. Koornstra 1,2, J.H. Kleibeuker 2, C.M.M. van Geelen 1, F.E.M. Rijcken 2, H. Hollema 3, E.G.E. de Vries 1, S. de Jong 1. Departments of 1 Medical Oncology; 2 Gastroenterology and Hepatology; 3 Pathology, University of Groningen Medical Centre, the Netherlands. Journal of Pathology 2003;200:

63 Chapter 4 Abstract Background: Tumour necrosis factor-related apoptosis-inducing ligand (TRAIL) induces apoptosis in tumour cell lines. Four membrane-bound receptors for TRAIL have been identified, two apoptosis-mediating receptors DR4 and DR5 and two apoptosis-inhibiting receptors DcR1 and DcR2. The aim of this study was to examine the role of TRAIL and its receptors in colorectal cancer development. Methods: The immunohistochemical expression and localisation of TRAIL and its receptors were investigated in normal mucosa (n = 10), adenomas (n = 19) and carcinomas (n = 21). Correlations between expression of TRAIL and its receptors and the degree of apoptosis (assessed by M30 immunoreactivity) and histopathological characteristics were explored. Results: TRAIL and its receptors were expressed in normal mucosal epithelium. Expression of the receptors was seen in adenomas and carcinomas. TRAIL expression was lost in a subset of colorectal tumours, more frequently in carcinomas than in adenomas (p < 0.05). DR4 and DR5 staining was stronger in neoplastic cells compared to normal cells, and was accompanied by a higher degree of apoptosis. No differences were found between tumour and normal cells regarding DcR1 and DcR2 expression. No correlations were found between TRAIL or TRAIL receptor expression and histopathological characteristics. Conclusion: Marked changes were seen in the course of the adenoma-carcinoma sequence with respect to the expression of TRAIL and TRAIL receptors DR4 and DR5. The stronger expression of DR4 and DR5 in neoplastic cells as opposed to normal cells, together with a higher degree of apoptosis suggests a possible functional role for these receptors in apoptosis induction in colorectal neoplastic cells. Introduction Apoptosis plays a critical role in the normal development of multicellular organisms and in maintaining tissue homeostasis 1. Apoptosis also represents an effective mechanism by which cells with DNA damage can be eliminated while dysregulation of normal apoptotic mechanisms can provide a growth advantage to cancer cells 2,3. Colorectal carcinogenesis is characterised by an accumulation of molecular genetic alterations causing progressive disorders in cell growth, differentiation and apoptosis 4,5. Elucidation of the molecular mechanisms regulating these processes is therefore of primary interest. Apoptosis is controlled through a variety of intracellular and extracellular signals 2. Cytokines from the tumour-necrosis factor (TNF) family have been identified as participants in the regulation of apoptosis 6. These cytokines are TNF-α, Fas ligand (FasL) and TNFrelated apoptosis-inducing ligand (TRAIL), all of which induce apoptosis by binding to their respective receptors 2. The role of TNF-α in the development of colorectal cancer has not been studied extensively but seems to be limited 7. The FasL/Fas apoptosis-induction pathway in colorectal cancer has been studied to a greater extent. Although several experiments have suggested a possible role for disturbances in this pathway in colorectal carcinogenesis, its precise impact still has to be defined 8. TRAIL was isolated in TRAIL induces rapid apoptosis in a wide variety of transformed cell lines, but seems to have little or no 62

64 TRAIL and its receptors in normal and neoplastic colon detectable cytotoxic effect on normal cells in vitro and in vivo 6,9-11. It has been suggested that TRAIL may be involved in endogenous tumour surveillance 12,13. At present, five different receptors for TRAIL have been identified: two cell death-inducing receptors (DR4/TRAIL- R1 and DR5/TRAIL-R2), two non-cell-death-inducing receptors (DcR1/TRAIL-R3 and DcR2/TRAIL-R4) and osteoprotegerin 14,15. DR4 and DR5 are membrane bound receptors that contain a so-called death domain in their intracellular segment. This death domain is required for TRAIL-mediated apoptotic cell death in a manner similar to that of other death receptors in the TNF-receptor family 14, In contrast, DcR1 lacks the intracellular death domain and DcR2 contains a truncated death domain. Therefore, neither receptor can transduce the death signals 16. Antibodies against TRAIL and TRAIL-receptors were not available until recently. Therefore, TRAIL and TRAIL-receptor expression were only reported at the level of mrna expression. TRAIL mrna was detected in a wide range of tissues, including the small intestine and colon 9,19. DR4, DR5 and DcR2 transcripts are expressed in normal colonic epithelium and colon adenocarcinomas 14,16. DcR1 mrna levels were high in various tumours of the gastrointestinal tract, including colorectal cancer, in comparison to normal mucosa 20. Protein expression and localisation of TRAIL and its receptors in normal colonic epithelium have recently been described 21. There are no data available on the protein expression and localisation of TRAIL and its receptors in colorectal neoplasms. In order to examine the possible role of TRAIL mediated apoptosis, we analysed the immunohistochemical expression and localisation of TRAIL and its receptors in different stages of colorectal cancer development. In addition, correlations between TRAIL and TRAIL receptor expression and the degree of apoptosis and histopathological characteristics were explored. Materials and methods Tissue collection Paraffin-embedded tissue specimens from 10 samples of normal mucosa, 19 colorectal adenomas and 21 carcinomas were retrieved from the files of the Department of Pathology. Samples of normal mucosa were randomly selected from archival materials obtained from patients without macroscopic abnormalities at colonoscopy where random biopsies had been taken. The adenomas were the first 19 sporadic adenomas endoscopically removed at the Department of Gastroenterology in Adenomas from patients with previous or simultaneous colorectal cancer were excluded. Primary sporadic adenocarcinomas had been removed at the Department of Surgery in Mucinous carcinomas were excluded. Tumours from patients that had received chemo- or radiotherapy prior to removal of the tumour were excluded. Histologic classifications were carried out on haematoxylin and eosin (H&E) stained slides. Adenocarcinomas were staged according to the modified Dukes classification and graded into well, moderately and poorly differentiated 22. For statistical purposes, tumours with Dukes stages A/B were compared with Dukes stages C/D and tumours with good/moderate differentiation with tumours with poor differentiation. For adenomas, circumferential size was measured and dysplasia expressed as low- or high- 63

65 Chapter 4 grade. Adenomas were classified as tubular or villous when the villous component exceeded 25 % of the adenoma. In all cases of adenomas and carcinomas, slides were selected in which adjacent or surrounding normal epithelial cells were included. Immunohistochemistry Expression of TRAIL and receptors DR4, DR5, DcR1 and DcR2 For immunohistochemical staining, serial 3 μm-thick-sections were cut from paraffin blocks. After deparaffinisation in xylene, endogenous peroxidase was blocked with 0.3 % hydrogen peroxide for 30 min. For TRAIL, DR5 and DcR2, antigen retrieval was carried out by microwave treatment, for DR5 8 min at 700 W in 0.01 M citrate buffer (ph 6.0), for TRAIL and DcR2 in Tris/HCl buffer (ph 9.0). For DR4 and DcR1 staining, no antigen retrieval was required. After incubation with avidin and biotin blocking solutions (Vector Laboratories, Burlingame, CA, USA) primary antibodies were applied for 1 h at room temperature. Slides were stained with a goat polyclonal IgG for TRAIL (1:25; clone C-20, Santa Cruz, CA, USA), a goat polyclonal IgG for DR4 (1:100; clone C-20, Santa Cruz), a rabbit polyclonal IgG for DR5 (1:100; Oncogene Research, Cambridge, MA, USA), a goat polyclonal IgG for DcR1 (1:50; Calbiochem, San Diego, CA, USA) and a rabbit polyclonal IgG for DcR2 (1:75; Oncogene Research). After washing with PBS, slides were incubated with a 1:300 dilution of a biotinylated rabbit-anti-goat or swine-anti-rabbit antibody (DAKO, Glostrup, Denmark) respectively, followed by addition of streptavidin-conjugated peroxidase. Peroxidase activity was visualised with diaminobenzidine. Slides were counterstained with haematoxylin. Controls To ensure specificity of the primary antibodies several experiments were conducted. First, immunoblotting of the antibodies on several colon cancer cell lines was performed. For this study, the colon carcinoma cell lines Colo320, Caco-2 and SW948 were obtained from the ATCC (Rockville, MD). Immunoblotting of the antibodies for TRAIL, DR4 and DR5 on three cell lines and cases of colorectal cancer yielded protein products of expected size (see figure 1). The specificity of the antibodies for DcR1 and DcR2 was confirmed at the RNA and protein level on the cell lines Colo320, Caco-2 and SW948. Second, a number of tissue sections were immunostained using non-immunised goat or rabbit IgG antibody respectively as a substitute for the primary antibodies, as described by Moller et al. 23. In these cases, no immunostaining was detected. Third, for DR4, TRAIL and DcR1, immunohistochemical staining was performed in the presence or absence of a 10-fold excess of the corresponding blocking peptides (Santa Cruz for TRAIL and DR4, Calbiochem for DcR1), as described by Mitsiades et al. 24. Again, in these cases no immunostaining was detected. Finally, slides were immunostained in the absence of the primary antibody and in these cases no immunostaining was detected. As positive controls, sections of normal human liver (DR4 25, DR5 26,27 ) or kidney (TRAIL 19, DcR1 26, DcR2 28 ) were included to check the reliability and reproducibility of the staining procedures. 64

66 TRAIL and its receptors in normal and neoplastic colon Caco-2 Colo320 SW948 TRAIL DR4 DR5 Figure 1. Western-blot analysis for TRAIL, DR4 and DR5 in representative tissue homogenates of frozen human colon carcinomas(1266, 5695, 7380, 15250) and colon carcinoma cell lines Caco-2, Colo320 and SW948. TRAIL was readily detected in two of four carcinoma samples but not observed in the three cell lines. DR4 and DR5 protein was observed in all carcinoma samples and cell lines. Apoptosis Apoptotic cells were determined by immunohistochemistry with the murine monoclonal antibody M30 (Boehringer Mannheim, Mannhein, Germany). M30 reacts with a cleavage product of cytokeratin 18, released by activated caspase 29, and the immunoreactivity is present during early apoptosis 30. M30 staining has been shown to be markedly more sensitive than the morphological identification of apoptosis by H&E staining in colorectal tissues 31. In addition, a good correlation has been found between in situ end labelling (ISEL) and expression of M30 in colorectal neoplasms 32. After deparaffinisation and blocking endogenous peroxidase activity, sections were immersed in 0.01 M citrate buffer (ph 6.0) and heated in a microwave oven for 8 min. M30 was applied to the sections for 60 min at room temperature in a 1:50 solution, followed by incubation with a secondary rabbitanti-mouse antibody conjugated with peroxidase (DAKO, Glostrup, Denmark) and a tertiary goat-anti-rabbit peroxidase-conjugated antibody (DAKO). Peroxidase activity was visualised with diaminobenzidine. Slides were counterstained with haematoxylin. As negative controls, slides were immunostained in the absence of the primary antibody. Evaluation of staining results Slides were evaluated by light microscopy by at least two independent investigators without knowledge of the histopathological data. Immunostaining was evaluated in serial tissue sections. The percentage of positive cells was estimated semiquantitatively and sorted into four categories (negative, < 10 %, % and > 50 %). When the observers scores differed more than 10 %, cases were re-evaluated using a multiheaded microscope and the final grade was reached by consensus. For statistical analysis, tumours with positive staining in >10 % of tumour cells were considered positive, while tumours without staining or staining in less than 10 % of cells were considered negative. The pattern of staining was recorded as membranous, cytoplasmic or nuclear. To exclude any effects of variation in intensity of staining among different samples, the immunoreactivity of positive tumour cells was 65

67 Chapter 4 Figure 2. Representative examples of expression of TRAIL and TRAIL receptors in adenomas. A: TRAIL; B: DR4; C: DR5; D: DcR1; E: DcR2. Epithelial cells express cytoplasmic immunoreactivity for DR4, DR5, DcR1, DcR2 and TRAIL. The intensity of staining for DR4 and DR5 is generally stronger in adenoma cells as compared to normal cells. Magnification 100x. See appendix for colour pictures. assessed relative to adjacent normal epithelial cells. M30 positivity was identified as brown cytoplasmic staining. M30 positive cells were determined within whole crypts in normal mucosa and adenomas and randomly in carcinomas. In all cases, at least 500 epithelial cells were counted and M30 positive cells were expressed as a percentage of the total number of cells counted. Statistical analysis For statistical assessment of differences in the expression of TRAIL, DR4, DR5, DcR1 and DcR2 between normal tissue, adenomas and carcinomas, the chi-square test was used. Differences in expression of TRAIL, DR4, and DR5 with respect to histopathological characteristics of adenomas and carcinomas were assessed using the chi-square test. Mean percentages of M30 positive cells were compared across categories of DR4, DR5 and TRAIL positive cells using the Wilcoxon test. To assess differences in the percentage of apoptotic cells between different tissue types, the Mann-Whitney test was used. P-values < 0.05 were considered significant. SPSS for Windows software (SPSS Inc., Chicago, IL) was used for all statistical analyses. 66

68 TRAIL and its receptors in normal and neoplastic colon Table 1. Staining results for TRAIL, DR4, DR5, DcR1 and DcR2 in normal tissue, adenomas and carcinomas, expressed as number of samples with staining in > 10 % of epithelial cells. TRAIL DR4 DR5 DcR1 DcR2 n n (%) n (%) N (%) n (%) N (%) Normal (100) 10 (100) 10 (100) 10 (100) 10 (100) Adenomas (84) 19 (100) 19 (100) 14 (74) 19 (100) Carcinomas (52)* 21 (100) 21 (100) 19 (90) 21 (100) * p < 0.05 TRAIL expression in carcinomas versus adenomas. Results Expression of TRAIL, DR4, DR5, DcR1 and DcR2 in normal colon epithelium Immunohistochemical staining showed expression of TRAIL and all four TRAIL-receptors in all cases of normal epithelium, including both goblet and columnar epithelial cells, along the entire length of the crypt. There were no appreciable differences in staining intensities between individuals. For TRAIL, DR4 and DcR1, the immunoreactivity of epithelial cells increased gradually from the base of the crypt to the crypt mouth and the luminal surface. For DR5 and DcR2, there was no apparent variability in staining intensity between the basal part of the crypt and the luminal surface. The expression was cytoplasmic for TRAIL, DR4, DR5, DcR1 and DcR2. In addition, nuclear stained cells with DcR2 were seen, predominantly located in the basal part of the crypts. At the mucosal surface, TRAIL staining was seen throughout the cytoplasm whereas for DR4, DR5, DcR1 and DcR2 cytoplasmic staining was strongest at the basolateral surface of the cell. Expression of TRAIL, DR4, DR5, DcR1 and DcR2 in colorectal adenomas and carcinomas Staining results are summarised in Tables 1-3. Both adenomas and carcinomas exhibited cytoplasmic staining for TRAIL, DR4, DR5, DcR1 and DcR2 (figures 2, A-E: adenomas; figure 3, A-E: carcinomas). Loss of TRAIL expression was more frequently observed in carcinomas than in adenomas (Figure 3E). Loss of TRAIL expression was seen more often in adenomas larger than 9 mm than those smaller than 9 mm (mean circumferential adenoma size, table 2). There was no relationship between loss of TRAIL expression in carcinomas and tumour stage or grade. Nuclear TRAIL staining was seen in one carcinoma. In the adenomas there were no significant correlations between the presence of TRAIL expression and growth type or degree of dysplasia. DR4 staining was seen in all adenomas and carcinomas. Strikingly, in all adenomas and carcinomas, staining intensity was stronger in neoplastic cells relative to adjacent normal cells. The percentage of DR4 positive cells was higher in carcinomas than in adenomas. 67

69 Chapter 4 Figure 3. Representative examples of expression of TRAIL receptors and TRAIL in adenocarcinomas. A: TRAIL; B: DR4; C: DR5; D: DcR1; E: DcR2. Tumour cells express cytoplasmic immunoreactivity for DR4, DR5, DcR1 and DcR2. Loss of TRAIL expression is seen in tumour cells compared to adjacent normal cells. The intensity of staining for DR4 and DR5 is stronger in carcinoma cells as compared to normal cells. Magnification 100x. See appendix for colour pictures. The percentage of DR4 positive cells was not correlated to size, grade of dysplasia or growth type (tubular or villous) of adenomas and was also not correlated to stage or grade of carcinomas. DR5 staining was detected in all adenomas and carcinomas. As for DR4, staining intensity in all adenomas and carcinomas was stronger in neoplastic cells than in adjacent normal epithelial cells. The majority of adenomas and carcinomas harboured more than 50 % positive cells. The percentage of positively staining cells did not differ between adenomas or carcinomas with different histopathological characteristics. DcR1 staining was observed in the majority of adenomas and carcinomas. Staining intensities were similar in neoplastic cells compared to adjacent normal epithelial cells. Strong staining was seen in stromal cells. DcR2 staining was observed in all adenomas and carcinomas, with staining intensities in neoplastic cells comparable to adjacent normal cells. Nuclear staining was seen in one adenocarcinoma. 68

70 TRAIL and its receptors in normal and neoplastic colon Table 2. Immunohistochemical staining for TRAIL, DR4, DR5 and M30 immunoreactivity in adenomas and carcinomas in relation to histopathological characteristics. Results are expressed as number of samples with positive staining cells (TRAIL), the number of samples with more than 50 % staining cells (DR4, DR5) and the mean (± SEM) percentage of M30 positive cells. n TRAIL + DR4 > 50 % DR5 > 50 % M30 Adenomas Size < 9 mm (0.10) > 9 mm 7 4 * (0.16) Growth Tubular (0.12) Villous (0.14) Dysplasia Low-grade (0.09) High-grade (0.18) Carcinomas Stage Dukes A/B (0.16) Dukes C/D (0.52) Grade Well/moderate (0.36) Poor (0.28) * p < 0.05 TRAIL expression in adenomas > 9 mm versus adenomas < 9 mm. Degree of apoptosis in adenoma-carcinoma sequence In normal mucosa, the mean percentage of M30 positive cells was 0.11 (range ) in truly normal mucosa, 0.12 (range ) in normal mucosa adjacent to adenomas and 0.56 (range ) in normal mucosa adjacent to carcinomas. The mean percentage of M30 positive cells in adenomas was 0.51 (range ). This was higher than in truly normal mucosa and normal mucosa adjacent to adenomas (p < 0.001). In carcinomas, the mean percentage of M30 positive cells was 1.66 (range ). This was higher than in adenomas and normal mucosa adjacent to carcinomas (p < 0.001). TRAIL and TRAIL receptor expression in correlation to histopathological characteristics and the degree of apoptosis Possible relations were investigated between histopathological characteristics and the expression of TRAIL, DR4 and DR5, and the degree of apoptosis (table 2). No significant differences were found in the degree of apoptosis between either adenomas or carcinomas with different histopathological characteristics. Table 3 depicts the degree of positively staining cells for TRAIL, DR4 and DR5 in adenomas and carcinomas in relation to the degree of apoptotic cell death. No significant correlations were found between the degree of apoptosis and the extent of positive staining cells in adenomas and carcinomas. There was a trend towards a positive correlation between DR4 expression and the degree of apoptosis (p = 0.07). Within individual tumours, no clear patterns of immunostaining were observed with varying grades of dysplasia (adenomas) or varying degrees of differentiation (carcinomas). 69

71 Chapter 4 Table 3. Distribution of positively staining cells in categories (negative or < 10 %, % and > 50 %) for TRAIL, DR4 and DR5 in adenomas and carcinomas and the mean percentage (± SEM) of M30 positive cells. TRAIL DR4 DR5 Adenomas Carcinomas n M30 n M30 < (0.36) (0.62) (0.10) 0 - > (0.16) (0.13) < (0.04) (0.15) > (0.21) 20 * 1.63 (0.31) < (0.05) 0 - > (0.11) (0.30) * p < for difference in categories in percentage of positive cells for DR4 expression in adenomas versus carcinomas. Discussion In the present study, marked changes were shown in the course of the adenoma-carcinoma sequence with respect to the expression of TRAIL and TRAIL receptors, in particular DR4 and DR5. Our results in normal colonic epithelium are largely in accordance with the recent study by Sträter et al. 21. They also found TRAIL, DR4, DR5 and DcR2 expression in normal colonic epithelial cells. Furthermore, as in the Sträter study, we observed TRAIL and DR4 staining predominantly in the upper third of the crypts and the surface epithelium. We found DcR1 expression in almost all samples of normal mucosa, adenomas and carcinomas whereas Sträter et al. were not able to detect DcR1 in normal colonic epithelium, neither by rt-pcr nor by immunohistochemistry. Their rt-pcr results are in contrast to a study by Sheikh et al., who demonstrated low, yet detectable, DcR1 mrna expression in normal colonic epithelium 20. The contradictory results may be explained by the use of different antibodies for immunohistochemistry and the probable low level of DcR1 protein expression in normal colon. Similar to Sträter et al, we found predominantly cytoplasmic staining of the TRAIL receptors where one would rather expect membrane staining. However, comparable results have been obtained with immunohistochemical staining of other TNF-receptor family members like Fas 23 in the human colon and TNF-R1 and R2 in mouse colon 33. A possible explanation may be that TRAIL receptors, like Fas, exist in both membrane-bound and soluble forms. In the present study, stronger immunohistochemical DR4 and DR5 expression was observed in the majority of adenomas and carcinomas compared to adjacent normal 70

72 TRAIL and its receptors in normal and neoplastic colon epithelial cells. These results suggest that both DR4 and DR5 expression are upregulated in adenomas and carcinomas. This suggestion is supported by experiments showing that both DR4 and DR5 expression are enhanced by DNA damage, induced by ionising radiation as well as by chemotherapeutic agents 34,35. The accumulation of molecular genetic alterations that characterises the development of colorectal cancer leads to increasing DNA damage and, consequently, may lead to increased expression of DR4 and DR5. Similar observations have been made in cervical and pancreatic cancer 36,37. In cervical cancer, mrna levels of DR4 and DR5 were higher in tumour cells than in normal cells which was confirmed by stronger immunohistochemical expressions 36. In pancreatic cancer, mrna levels of DR4 and DR5 but also of TRAIL were elevated in most pancreatic cancer cases, compared with normal pancreatic tissue 37. The DR4 and DR5 receptors can induce apoptosis after interacting with their ligand, TRAIL, and are therefore a marker for cells that are predisposed to apoptosis. We found a higher degree of apoptosis in adenomatous and carcinomatous areas compared to adjacent normal tissue. Our results with M30 immunostaining in adenomas and carcinomas are in accordance with other studies using this method to assess apoptotic cell death 32,38. Percentages of DR4 or DR5 positively staining cells were however not correlated to the degree of apoptosis in our series. In contrast to the study in pancreatic tissue, we found decreased or lost TRAIL expression in 16 % of adenomas and 48 % of carcinomas, while staining was intact in adjacent normal cells. The degree of apoptosis did not differ between TRAIL positive and TRAIL negative carcinomas. Expression patterns of TRAIL have recently been reported in other tumour types. In brain tumours, TRAIL was expressed in astrocytomas and glioblastomas, but not in medulloblastomas, oligodendrogliomas, meningiomas, neurocytomas and schwannomas 39,40. In these brain tumours, TRAIL expression did not correlate with the degree of apoptosis. In breast cancer tissues TRAIL expression was found in 21/40 breast cancers as opposed to none of five normal breast tissue samples 41. The loss of TRAIL expression in a subset of colorectal adenomas and carcinomas suggest that these tumours could have evaded induction of apoptosis due to downregulation of TRAIL. The hypothesis of loss of TRAIL expression in the course of the adenoma-carcinoma sequence is supported by a recent small study. In this study, expression profiles of eight colorectal cancer tumours were compared with corresponding non-cancerous colonic cells using a DNA micro-array technique, consisting of 9216 genes. It was shown that among other genes, TRAIL was downregulated in colorectal cancer 42. So far, mutations in the TRAIL gene have not been studied in colorectal cancer. No clear differences were observed in the expression of the apoptosis-inhibiting receptors DcR1 and DcR2 between colorectal neoplasms and normal mucosa. In contrast, others have shown that DcR1 mrna was expressed at higher levels in four of six colon carcinomas than in matched control tissues 20. The authors suggested that colorectal tumours may gain a growth advantage by overexpressing DcR1 in order to protect themselves against TRAIL-mediated apoptosis. However, others have shown that DR4 and DR5, and not the decoy receptors, were expressed on the plasma membrane of four human colon carcinoma cell lines, and that sensitivity to TRAIL correlated with the level of expression of DR4 and DR

73 Chapter 4 It is generally accepted that the main physiological mechanism by which colon epithelial cells die is apoptosis, yet the precise mechanisms behind the induction of apoptosis still have to be identified 21. Changes in the degree of apoptosis in the course of the adenomacarcinoma sequence have been studied extensively and were recently reviewed 44. From these studies, it appears that the proportion of cells undergoing apoptosis increases with tumour progression, possibly as a mechanism to delete cells with sustained DNA damage. Alterations in the degree of apoptosis may be explained by changes in apoptosis-regulating death-receptor pathways. From the currently known death receptor pathways, both FasLas well as TNF-α-mediated apoptosis are not thought to be involved in colorectal cancer development 7,21. In this regard, the expression of TRAIL and its receptors in the course of the adenoma-carcinoma sequence may be relevant. Our results of increased expression of the pro-apoptotic TRAIL receptors DR4 and DR5 in neoplastic cells compared to normal cells, and a high degree of apoptotic cell death in neoplastic cells, suggests a possible functional role for these receptors in the induction of apoptosis of colorectal epithelial cells with DNA damage. The functionality of TRAIL-mediated apoptosis in normal and tumour tissues as well as the precise clinical significance and the molecular biologic regulatory mechanisms have yet to be elucidated. The cell killing properties of this cytokine have made it an exciting target for drug development. Several studies have shown that recombinant human (rh) TRAIL induces apoptosis in a variety of cancer cell lines, including colon carcinoma cell lines 11. Combining chemotherapy and rhtrail resulted in potentiation of antitumour activity in colon carcinoma cell lines as well as in mice carrying subcutaneous tumours from colorectal cell lines 43,45,46. Based on preclinical toxicity and activity profiling, TRAIL is considered to be of interest for clinical use 47. In colorectal cancer, rhtrail may provide a useful drug for killing tumour cells as our data suggest not only increased levels of DR4 and DR5 receptor expression but also frequent loss of TRAIL expression. In conclusion, our study demonstrates that TRAIL and its receptors are expressed in normal colon mucosa and that several changes occur during the course of the adenomacarcinoma sequence. Most strikingly, tumour progression is associated with an increase in expression of the pro-apoptotic receptors DR4 and DR5 and an increase in apoptotic cell death. Furthermore, the expression of TRAIL decreases in the adenoma-carcinoma sequence. More studies are needed to elucidate the precise role of TRAIL as a mediator of apoptosis in the colon, its role in the development of colorectal cancer and its potential as a therapeutic agent. 72

74 TRAIL and its receptors in normal and neoplastic colon References 1. Steller H. Mechanisms and genes of cellular suicide. Science 1995;267: Nagata S. Apoptosis by death factor. Cell 1997;88: Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Science 1995;267: Hao X, Du M, Bishop AE, Talbot IC. Imbalance between proliferation and apoptosis in the development of colorectal carcinoma. Virchows Arch 1998;433: Sinicrope FA, Roddey G, McDonnell TJ, Shen Y, Cleary KR, Stephens LC. Increased apoptosis accompanies neoplastic development in the human colorectum. Clin Cancer Res 1996;2: Walczak H, Krammer PH. The CD95 (APO-1/Fas) and the TRAIL (APO-2L) apoptosis systems. Exp Cell Res 2000;256: Remacle-Bonnet MM, Garrouste FL, Heller S, Andre F, Marvaldi JL, Pommier GJ. Insulin-like growth factor-i protects colon cancer cells from death factor-induced apoptosis by potentiating tumor necrosis factor alpha- induced mitogen-activated protein kinase and nuclear factor kappab signaling pathways. Cancer Res 2000;60: Restifo NP. Countering the counterattack hypothesis. Nat Med 2001;7: Wiley SR, Schooley K, Smolak PJ, et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 1995;3: Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A, Ashkenazi A. Induction of apoptosis by Apo- 2 ligand, a new member of the tumor necrosis factor cytokine family. J Biol Chem 1996;271: Walczak H, Miller RE, Ariail K, et al. Tumoricidal activity of tumor necrosis factor-related apoptosisinducing ligand in vivo. Nat Med 1999;5: Hayakawa Y, Smyth MJ, Kayagaki N, Yamaguchi N, Kakuta S, Iwakura Y, et al. Involvement of tumor necrosis factor-related apoptosis-inducing ligand in surveillance of tumor metastasis by liver natural killer cells. Nat Med 2001;7: Takeda K, Smyth MJ, Cretney E, et al. Critical role for tumor necrosis factor-related apoptosis-inducing ligand in immune surveillance against tumor development. J Exp Med 2002;195: Ashkenazi A, Dixit VM. Apoptosis control by death and decoy receptors. Curr Opin Cell Biol 1999;11: Degli-Esposti M. To die or not to die--the quest of the TRAIL receptors. J Leukoc Biol 1999;65: Ashkenazi A, Dixit VM. Death receptors: signaling and modulation. Science 1998;281: Pan G, Ni J, Wei YF, Yu G, Gentz R, Dixit VM. An antagonist decoy receptor and a death domaincontaining receptor for TRAIL. Science 1997;277: Pan G, O Rourke K, Chinnaiyan AM, et al. The receptor for the cytotoxic ligand TRAIL. Science 1997;276:

75 Chapter Tan KB, Harrop J, Reddy M, et al. Characterization of a novel TNF-like ligand and recently described TNF ligand and TNF receptor superfamily genes and their constitutive and inducible expression in hematopoietic and non-hematopoietic cells. Gene 1997;204: Sheikh MS, Huang Y, Fernandez-Salas EA, et al. The antiapoptotic decoy receptor TRID/TRAIL- R3 is a p53-regulated DNA damage-inducible gene that is overexpressed in primary tumors of the gastrointestinal tract. Oncogene 1999;18: Sträter J, Walczak H, Pukrop T, et al. TRAIL and its receptors in the colonic epithelium: a putative role in the defense of viral infections. Gastroenterology 2002;122: Jass JR, Sobin LH, Watanabe H. The World Health Organization s histologic classification of gastrointestinal tumors. A commentary on the second edition. Cancer 1990;66: Möller P, Koretz K, Leithauser F, et al. Expression of APO-1 (CD95), a member of the NGF/TNF receptor superfamily, in normal and neoplastic colon epithelium. Int J Cancer 1994;57: Mitsiades N, Poulaki V, Tseleni-Balafouta S, Koutras DA, Stamenkovic I. Thyroid carcinoma cells are resistant to FAS-mediated apoptossi but sensitive tumor necrossis factor-related apoptosis-inducing ligand. Cancer Res 2000;60: Schneider P, Bodmer JL, Thome M, Hofmann K, Holler N, Tschopp J. Characterization of two receptors for TRAIL. FEBS Lett 1997;416: MacFarlane M, Ahmad M, Srinivasula SM, Fernandes-Alnemri T, Cohen GM, Alnemri ES. Identification and molecular cloning of two novel receptors for the cytotoxic ligand TRAIL. J Biol Chem 1997;272: Sheridan JP, Marsters SA, Pitti RM, et al. Control of TRAIL-induced apoptosis by a family of signaling and decoy receptors. Science 1997;277: Pan G, Ni J, Yu G, Wei YF, Dixit VM. TRUNDD, a new member of the TRAIL receptor family that antagonizes TRAIL signalling. FEBS Lett 1998;424: Caulin C, Salvesen GS, Oshima RG. Caspase cleavage of keratin 18 and reorganization of intermediate filaments during epithelial cell apoptosis. J Cell Biol 1997;138: Leers MP, Kolgen W, Bjorklund V, et al. Immunocytochemical detection and mapping of a cytokeratin 18 neo- epitope exposed during early apoptosis. J Pathol 1999;187: Michael-Robinson JM, Biemer-Huttmann AE, Purdie DM, et al. Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status. Gut 2001;48: Carr NJ. M30 expression demonstrates apoptotic cells, correlates with In situ end-labeling, and is associated with Ki-67 expression in large intestinal neoplasms. Arch Pathol Lab Med 2000;124: Mizoguchi E, Mizoguchi A, Takedatsu H, et al. Role of tumor necrosis factor receptor 2 (TNFR2) in colonic epithelial hyperplasia and chronic intestinal inflammation in mice. Gastroenterology 2002;122: Guan B, Yue P, Clayman GL, Sun SY. Evidence that the death receptor DR4 is a DNA damage-inducible, p53- regulated gene. J Cell Physiol 2001;188:

76 TRAIL and its receptors in normal and neoplastic colon 35. Sheikh MS, Burns TF, Huang Y, et al. p53-dependent and -independent regulation of the death receptor KILLER/DR5 gene expression in response to genotoxic stress and tumor necrosis factor alpha. Cancer Res 1998;58: Ryu HS, Chang KH, Chang SJ, Kim MS, Joo HJ, Oh KS. Expression of TRAIL (TNF-related apoptosisinducing ligand) receptors in cervical cancer. Int J Gynecol Cancer 2000;10: Ozawa F, Friess H, Kleeff J, et al. Effects and expression of TRAIL and its apoptosis-promoting receptors in human pancreatic cancer. Cancer Lett 2001;163: Michael-Robinson JM, Reid LE, Purdie DM, et al. Proliferation, apoptosis, and survival in high-level microsatellite instability sporadic colorectal cancer. Clin Cancer Res 2001;7: Nakamura M, Rieger J, Weller M, Kim J, Kleihues P, Ohgaki H. APO2L/TRAIL expression in human brain tumors. Acta Neuropathol (Berl) 2000;99: Rieger J, Ohgaki H, Kleihues P, Weller M. Human astrocytic brain tumors express AP02L/TRAIL. Acta Neuropathol (Berl) 1999;97: Herrnring C, Reimer T, Jeschke U, et al. Expression of the apoptosis-inducing ligands FasL and TRAIL in malignant and benign human breast tumors. Histochem Cell Biol 2000;113: Kitahara O, Furukawa Y, Tanaka T, et al. Alterations of gene expression during colorectal carcinogenesis revealed by cdna microarrays after laser-capture microdissection of tumor tissues and normal epithelia. Cancer Res 2001;61: Lacour S, Hammann A, Wotawa A, Corcos L, Solary E, Dimanche-Boitrel MT. Anticancer agents sensitize tumor cells to tumor necrosis factor- related apoptosis-inducing ligand-mediated caspase-8 activation and apoptosis. Cancer Res 2001;61: Koornstra JJ, de Jong S, hollema H, de Vries EG, Kleibeuker JH. Changes in apoptosis during the development of colorectal cancer: a systematic review of the literature. Crit Rev Oncol Hematol 2003:45: Gliniak B, Le T. Tumor necrosis factor-related apoptosis-inducing ligand s antitumor activity in vivo is enhanced by the chemotherapeutic agent CPT-11. Cancer Res 1999;59: Ashkenazi A, Pai RC, Fong S, Leung S, et al. Safety and antitumor activity of recombinant soluble Apo2 ligand. J Clin Invest 1999;104: de Jong S, Timmer T, Heijenbrok FJ, de Vries EG. Death receptor ligands, in particular TRAIL, to overcome drug resistance. Cancer Metastasis Rev 2001;20:

77

78 Chapter 5 Expression of TRAIL receptors DR4 and DR5 in sporadic and hereditary colorectal tumours: potential targets for apoptosis induction. J.J. Koornstra 1, M. Jalving 1, 2, F.E.M. Rijcken 1, J. Westra 3, N. Zwart 2, H. Hollema 4, E.G.E. de Vries 2, R.W.M. Hofstra 3, J.T.M. Plukker 5, S. de Jong 2, J.H. Kleibeuker 1. Departments of 1 Gastroenterology and Hepatology; 2 Medical Oncology; 3 Medical Genetics; 4 Pathology, 5 Surgery, University of Groningen Medical Centre, the Netherlands. European Journal of Cancer 2005;41:

79 Chapter 5 Abstract Background: Tumour necrosis factor-related apoptosis-inducing ligand (TRAIL) and antibodies against TRAIL receptors death receptor 4 (DR4) and death receptor 5 (DR5) are under investigation for cancer therapy. To study potential application of these agents, the expression of DR4 and DR5 were studied in colorectal adenomas and carcinomas from patients with sporadic and hereditary disease. The role of the BAX gene, frequently mutated in tumours with high frequency microsatellite instability (MSI-H) and playing a role in sensitivity to TRAIL, was also studied. Methods: Adenomas and carcinomas from patients with sporadic disease (n = 74 and 56 respectively), familial adenomatous polyposis (FAP, n = 41 and 4 respectively) and hereditary nonpolyposis colorectal cancer (HNPCC, n = 50 and 21 respectively) were studied by immunohistochemistry. MSI-H carcinomas (n = 42, of which 27 sporadic and 15 HNPCC) were analysed for apoptotic activity, assessed by M30 immunoreactivity, and BAX mutations. Results: Most adenomas from all three patient groups expressed DR4 and DR5. Most carcinomas expressed DR4, except for 6 cases, all with mucinous histology. All carcinomas, including mucinous carcinomas, showed DR5 expression. BAX mutations were found in 6/42 MSI-H cancers with similar apoptotic indices and expression of DR4, DR5 and TRAIL in BAX mutant and wild-type cases. Conclusion: Since most sporadic and hereditary colorectal neoplasms express DR4 and DR5, targeting of these receptors may be a potential prevention or treatment strategy. No evidence was found supporting the concept of BAX inactivation as a critical mechanism to evade TRAIL-receptor-mediated apoptosis in vivo. Introduction Colorectal cancer (CRC) is the second leading cause of cancer-related mortality in the western world. The shortcomings of current treatment modalities for CRC call for novel strategies to treat or prevent the disease. Attention focuses on early detection of CRC or its precursor lesion, the adenoma, for example by endoscopic screening. An alternative approach to prevent the development of adenomas or carcinomas is the use of chemopreventive agents, especially in high-risk patients. Two major entities are known to carry a highly increased risk of developing CRC: familial adenomatous polyposis (FAP) and hereditary non-polyposis colorectal cancer (HNPCC) 1. FAP, caused by a germline mutation in the APC gene is clinically characterised by numerous adenomatous polyps in the colon. HNPCC is caused by a germline mutation in one of the DNA mismatch repair (MMR) genes, in particular hmsh2, hmlh1 and hmsh6. As a consequence of defective DNA mismatch repair, tumours from HNPCC patients are characterised by length alterations in repetitive sequences distributed throughout the genome, so-called microsatellite instability (MSI) 2. The MSI phenotype is also found in % of sporadic CRC cases, as a result of hypermethylation of the hmlh1 gene promoter region 3. In FAP patients, chemoprevention using nonsteroidal anti-inflammatory drugs (NSAIDs) reduced adenoma size and number in several studies 4. However, complete regression of 78

80 TRAIL and its receptors in hereditary colorectal neoplasms adenomas in FAP patients is unusual, CRC can develop during NSAID treatment and longterm NSAID use is associated with side effects 5. Chemoprevention studies in HNPCC patients are underway, however, it has been suggested that NSAIDs may be less effective in the setting of MSI 6. The development of other agents is therefore needed. TNF-related apoptosis inducing ligand (TRAIL) is a type II transmembrane protein which induces apoptosis in a variety of tumour cell lines but not in normal cells 7. Four membranebound receptors for TRAIL have been identified: death receptor 4 (DR4), death receptor 5 (DR5), decoy receptor 1 (DcR1) and decoy receptor 2 (DcR2) 7. TRAIL induces apoptosis by binding to DR4 or DR5, whereas DcR1 and DcR2 do not transduce apoptotic signals. Apoptosis induction through pro-apoptotic death receptors by recombinant human (rh) TRAIL or agonistic antibodies against these receptors is considered a promising approach for cancer therapy 8,9. Expression patterns of DR4 and DR5 have been described in normal colon, sporadic adenomas and sporadic carcinomas 10,11. The majority of tumours show DR4 and DR5 expression, with stronger intensity in neoplastic cells compared to normal tissue 10,11, suggesting preferential susceptibility of these cells to TRAIL-receptor mediated apoptosis. Whether the same expression patterns apply to hereditary cases is currently unknown. Recently, BAX, a pro-apoptotic member of the Bcl-2 family, has been shown to play a role in sensitivity to TRAIL-mediated apoptosis in vitro Since up to half of colorectal tumours with high frequency MSI (MSI-H) contain frameshift mutations in a polyg tract of the BAX gene 2, this could potentially limit the use of TRAIL or agonistic antibodies in MSI-H tumours. The aims of this study were twofold. First, DR4 and DR5 expression was investigated in colorectal tumours from patients with sporadic disease, FAP and HNPCC. Second, the relationship between BAX mutations, apoptosis and expression of DR4, DR5 and TRAIL was explored in MSI-H tumours. Materials and methods Patient and tissue selection Sporadic patients. Adenomas (n = 74), consecutively removed endoscopically at the Department of Gastroenterology in 1997 of which sufficient material was available to allow serial sectioning for immunohistochemical staining, were selected. MSI-H carcinomas were selected from a previously reported cohort of 500 patients with stage III colon cancer 16. Sufficient DNA for MSI analysis was available from 273/500 specimens. MSI analysis was performed using the ABI Prism 377 DNA sequencer (Promega, Madison WI, USA) when tumour and matching normal tissue were avaliable. In cases without normal tissue, the HNPCC MSI kit was employed (Roche, Basel, Switzerland), containing 5 previously described consensus markers 17. MSI-H, defined as instability in 3/9 or 2/5 markers respectively was detected in 44/273 samples (16 %). From 27 of these 44 samples, paraffin embedded tissue sections were available to allow serial sectioning for immunohistochemistry. As MSS controls, a series of sporadic carcinomas (n = 29) previously analysed 10 was studied. MSI was excluded in this series by immunohistochemical staining for MLH1 expression 18. FAP patients. Data from all patients (14 males, 18 females) with classical FAP treated at 79

81 Chapter 5 the University of Groningen Medical Centre from 1970 to December 2001 were reviewed. All available slides from colectomy specimens and endoscopically removed adenomas were revised. When adenomas showed similar growth patterns and degree of dysplasia in a single patient, one adenoma per patient was studied. Otherwise, additional adenomas were studied. In total, 4 carcinomas and 41 adenomas were included. HNPCC patients. HNPCC patients had a germline mutation in one of the MMR genes and/or fulfilled the Amsterdam II criteria. All colorectal adenomas and carcinomas removed between 1979 and 2002 with sufficient material available to allow serial sectioning for DNA extraction and immunohistochemistry were selected. In total, 21 carcinomas were examined, 18 from mutation carriers (10 hmsh2, 5 hmlh1, 3 hmsh6) and 3 from patients fulfilling the Amsterdam II criteria. Fifty adenomas, 21 from proven mutation carriers, were analysed. Only HNPCC carcinomas displaying the MSI-H phenotype (n = 15), assessed as described above, were subjected to BAX mutation analysis. Histopathological classification Histopathological classifications were performed with hematoxylin-eosin stained slides according to the WHO criteria 19. For adenomas, circumferential size was measured. Adenomas were classifed as tubular, tubulovillous, villous or serrated. As the group of serrated adenomas was relatively small (n = 4), they were not separately analysed. For statistical purposes, adenomas with tubulovillous or villous histology were combined. Data concerning tumour localisation were retrieved from pathology reports and/or endoscopy reports. Tumours were defined as right-sided (coecum, ascending or transverse colon) or left-sided (descending and sigmoid colon, rectum). Immunohistochemistry For immunohistochemical staining, serial 3 μm-thick-sections were cut from paraffin blocks. After deparaffinisation, blocking of endogenous peroxidase with 0.3 % hydrogen peroxide and incubation with avidin and biotin blocking solutions (Vector Laboratories, Burlingame, CA, USA) the primary antibodies were applied for 1 h at room temperature. Staining and control procedures for DR4, DR5 and TRAIL were performed as previously described 10. Apoptosis was assessed by M30 immunoreactivity, based on the detection of cleaved cytokeratin-18, which is expressed during early apoptosis of epithelial cells 20. The method has been validated against the gold standard of apoptosis detection by morphological criteria 20. Staining procedures for M30 were carried out as previously described 20. MLH1 staining was carried out with a mouse monoclonal antibody (1:100, PharMingen, San Diego, CA, USA). Staining was evaluated by light microscopy by two investigators, with re-evaluation under a multi-headed microscope if results did not agree. For DR4, DR5 and TRAIL, the percentage of staining cells was estimated semiquantitatively. Samples with staining in more than 10 % of cells were regarded as positive. Apoptosis was assessed in at least 1000 epithelial cells and expressed as a percentage of the total number of cells counted (apoptotic index). Intra and inter-observer variability were less than 10 %. 80

82 TRAIL and its receptors in hereditary colorectal neoplasms Table 1. Patient and tumour characteristics. Sporadic FAP HNPCC Ad * Ca MSS * Ca MSI-H * Ad Ca Ad Ca N Male (%) Age (yrs, median, range) 65 (28-89) 65 (40-88) 64 (26-76) 30 (11-52) 43 (21-52) 48 (30-67) 54 (31-77) Size (mm, median/range) 6.0 ( ) ( ) ( ) - Tubular (%) HGD (%) Localisation ** (%) # # 43 Tumour stage (%) I/II III/IV Differentiation (%) Poor Good/moderate * Ad = adenoma; Ca = carcinoma; MSS = microsatellite stable; MSI-H: microsatellite instability-high; HGD: high-grade dysplasia; ** localisation: 1 = right; 2 = left; # unknown in number of cases BAX mutation analysis From 27 sporadic MSI-H carcinomas and 15 HNPCC-associated MSI-H carcinomas, sufficient DNA could be extracted from microdissected sections of paraffin embedded samples with the Qiaquick PCR purification kit (Qiagen Inc, Chatsworth, CA, USA). PCR was performed using previously published primer sequences 21, amplifying a 94-base pair DNA fragment. PCR was carried out for 32 cycles, each cycle consisting of denaturation for 1 min at 94 C, annealing for 1 min at 55 C and extension for 1 min at 72 C. PCR products were visualised on a 1.5 % agarose gel and subsequently subjected to direct sequencing using the ABI Prism genetic analyser (Applied Biosystems Product, Foster City CA, USA). Statistical analysis Appropriate tests were used to assess differences in patient and tumour characteristics (chi-square test) and immunohistochemical findings (Mann-Whitney test for continuous variables, chi-square test for discontinuous variables). Correlations between percentages of positive staining and apoptotic indices were calculated with the Spearman test. P-values < 0.05 were considered significant. SPSS for Windows software was used in all statistical analyses. 81

83 Chapter 5 Table 2. DR4, DR5 and TRAIL expression patterns in sporadic, FAP and HNPCC adenomas. Sporadic FAP HNPCC DR4 pos (n) # 69/74 (93 %) 38/41 (93 %) 49/50 (98 %) % DR4 pos cells * 35 (15-100) 65 (15-100) 50 (20-100) DR5 pos (n) # 74/74 (100 %) 40/41 (98 %) 48/50 (96 %) % DR5 pos cells * 90 (20-100) 90 (40-100) 100 (20-100) TRAIL pos (n) # 58/74 (78 %) 32/41 (78 %) 35/50 (70 %) # number of cases with positive staining relative to the number of samples investigated * median (range) percentage of positively staining cells among positive samples Figure 1. DR4 and DR5 expression in mucinous adenocarcinoma. Hematoxylin-eosin (A), DR4 (B) and DR5 staining (C) showing absence of DR4 expression with intact DR5 expression. See appendix for colour pictures. 82

84 TRAIL and its receptors in hereditary colorectal neoplasms Table 3. DR4, DR5 and TRAIL expression patterns in sporadic, FAP and HNPCC carcinomas. Sporadic MSS Sporadic MSI-H FAP HNPCC DR4 pos (n) # 29/29 (100 %) 21/27 (78 %) 4/4 (100 %) 21/21 (100 %) % DR4 pos cells * 100 (50-100) 95 (20-100) 90 (75-100) 100 (20-100) DR5 pos (n) # 29/29 (100 %) 27/27 (100 %) 4/4 (100 %) 21/21 (100 %) % DR5 pos cells * 100 (60-100) 100 (20-100) 100 (95-100) 100 (20-100) TRAIL pos (n) # 21/29 (72 %) 10/27 (37 %) 3/4 (75 %) 17/21 (81 %) # number of cases with positive staining relative to the number of samples investigated * median (range) percentage of positively staining cells among positive samples Results Patient and tumour characteristics Patient and tumour characteristics are summarised in table 1. Inherent to the patient groups, several differences were observed between groups. Median age of patients with sporadic tumours was higher than FAP and HNPCC cases (p < 0.001). Median adenoma size was smaller in HNPCC compared to FAP and sporadic cases (p < 0.001). HNPCCassociated carcinomas and MSI-H sporadic carcinomas were more often localised in the proximal colon compared to FAP and sporadic MSS cases. Tumour stage was lower in HNPCC and FAP-associated cases than in their sporadic MSS counterparts (p < 0.001). Expression of DR4, DR5 and TRAIL and apoptosis in adenomas Immunohistochemical staining results are summarised in table 2. DR4 and DR5 expression was positive in virtually all adenomas. Among positive adenoma samples, median percentages of positively staining cells were similar in patient groups and independent of histopathological characteristics such as size, growth type or degree of dysplasia. Staining patterns of DR4, DR5 and TRAIL expression were generally heterogenous throughout adenoma tissue, with no consistent co-localisation for the different proteins. Adenomas with absence of DR4 or DR5 expression were not characterised by distinct histopathological features. There were no adenomas staining negative for both DR4 and DR5. TRAIL expression was positive in about 75 % of adenomas, independent of the patient group, and not associated with any histopathological parameter. Strikingly, 8/9 adenomas with absent DR4 expression also stained negative for TRAIL. Apoptotic indices, assessed by M30 immunoreactivity, were similar in patient groups. There was a positive correlation between apoptotic index and DR4 positivity (r = 0.23, p < 0.01) and TRAIL positivity (r = 0.18, p = 0.02). This correlation was not observed for DR5. 83

85 Chapter 5 Table 4. Apoptosis, DR4, DR5 and TRAIL expression in MSI-H carcinomas with wild-type or mutant BAX. BAX mutant (n=6) Wild-type BAX (n=36) p Apoptotic index + (%) 1.4 ± ± 0.2 ns DR4 pos (n) # 4/6 32/36 ns % DR4 pos cells * 100 (50-100) 100 (20-100) ns DR5 pos (n) # 6/6 36/36 ns % DR5 pos cells * 95 (60-100) 100 (20-100) ns TRAIL pos (n) # 3/6 19/36 ns + expressed as mean ± SEM.# number of cases with positive staining relative to the number of samples investigated.* median (range) percentage of positively staining cells among positive samples. ns: not significant Expression of DR4, DR5 and TRAIL and apoptosis in carcinomas Results are summarised in table 3. DR4 expression was positive in almost all carcinomas, with similar percentages of positive cells among positive samples in the different patient groups. Staining was generally homogenous throughout tumour tissue. Six out of 27 sporadic MSI-H carcinomas, all with mucinous histology, were DR4 negative (figure 1). Four other mucinous carcinomas, also MSI-H, stained positive for DR4. DR5 expression was positive in all carcinomas, including those negative for DR4 (figure 1). Percentages of positively staining cells among positive samples in the various groups were similar. Staining results for DR4 or DR5 were independent of tumour stage, localisation or degree of differentiation. TRAIL expression was positive in % of the carcinomas, depending on the patient group. Similar to the observation in adenomas, 5/6 carcinomas with absence of DR4 expression also stained negative for TRAIL. As in adenomas, there was no apparent colocalization of DR4, DR5 and TRAIL expression. There was a positive correlation between apoptotic index and DR4 positivity (r = 0.39, p < 0.001) and DR5 positivity (r = 0.18, p = 0.01), but not with TRAIL positivity. BAX mutation analysis and the relationship with apoptosis, DR4, DR5 and TRAIL expression Frameshift mutations in the G(8) repeat of the BAX gene were detected in 2/15 (13%) MSI-H HNPCC-associated carcinomas and 4/27 (15 %) MSI-H sporadic carcinomas. Sequence analyses of BAX mutations showed a 1 bp deletion in 5 cases and in 1 case an insertion of 1 bp. Apoptosis and expression patterns of DR4, DR5 and TRAIL were compared between MSI-H tumours with (n=6) and without (n=36) BAX gene mutations, depicted in table 4. Apoptotic indices were comparable between both groups. DR4, DR5 and TRAIL was observed with similar frequencies in BAX mutant and wild-type BAX tumours. Median percentages of DR4 and DR5 positivity also did not differ between these groups. 84

86 TRAIL and its receptors in hereditary colorectal neoplasms Discussion The present study shows extensive expression of the TRAIL death receptors DR4 and DR5 in sporadic and hereditary colorectal neoplasms. Moreover, no support was found for a critical role for BAX gene mutations as an apoptosis-evading mechanism in MSI-H tumours. Expression patterns of DR4 and DR5 in sporadic neoplasms were in accordance with previous findings 10,11. A novel finding was that a subset of sporadic carcinomas, all MSI-H and with mucinous histology, showed no DR4 expression. The background for this finding is unclear. The gene for DR4 is located on chromosome 8p21, a region associated with frequent loss of heterozygosity (LOH) in CRC development 22,23. However, 8p LOH is not particularly related with mucinous histology or MSI 22,23. Mutations in the DR4 gene have been described in several tumour types 8,9 but have not yet been studied in CRC. Alternatively, it may be that absence of DR4 expression in MSI-H tumours occurs as a result of hypermethylation. Interestingly, in the present study, absent DR4 expression was associated with absent TRAIL expression in the majority of cases. On the other hand, absent TRAIL expression was more often encountered in our study than absent DR4 expression, a finding also reported by others 11. Future studies may be directed at determining the mechanisms behind absence of TRAIL and DR4 expression in certain colorectal tumours. It must be realised however that the technique of immunohistochemistry may not be sensitive enough to determine TRAIL and DR4 expression. The importance of absent or low levels of endogenous TRAIL expression in colon epithelial cells for induction of apoptosis is unknown. Other cells such as macrophages and NK cells are known to express TRAIL, which may play a role in TRAIL-induced apoptosis. For example, TRAIL-expressing macrophages isolated from pleural effusions of cancer patients induced apoptosis in colon cancer cells 24. It is therefore important to stress that DR4 and DR5 expression was observed in almost every colorectal neoplasm, with no tumour being negative for both receptors, suggesting that neoplastic cells are prone to TRAIL induced apoptosis. The fulfilment of this prerequisite offers hope for the future use of rhtrail, agonistic TRAIL-receptor antibodies or other, yet to be developed, TRAIL agonists. Data on preclinical activity profiling of these agents are promising 25. TRAIL induces apoptosis in a wide variety of cancer cell lines of diverse origins, including colon cancer, while having little or no detectable cytotoxic effect on normal cells in vitro and in vivo 8. In cynomolgus monkeys and chimpanzees, repeated intravenous injections of native rhtrail did not cause detectable toxicity 26,27. The combination of chemotherapy and rhtrail potentiated antitumour activity in human colon cancer cell lines as well as in xenografted mice 26,28,29. In addition, specific targeting of DR4 and DR5 with monoclonal agonistic antibodies exerted similar effects as TRAIL in vitro and in vivo 30,31. Phase I trials using rhtrail as well as monoclonal agonistic antibodies are underway. Targeting of death receptors may even be useful at a premalignant stage given our finding of pro-apoptotic death receptor expression in adenomas, for example in chemopreventive regimens for FAP or HNPCC patients. Several studies indicate that TRAIL act synergistically with NSAIDs in induction of apoptosis 15,32,33, which may be a potentially useful combination regimen, deserving further study. Recent studies in cell lines suggested that the presence of functional BAX is important 85

87 Chapter 5 in determining the outcome of TRAIL-based therapeutic regimens BAX is frequently mutated in mismatch repair deficient tumours 2. It was shown that BAX-deficient human colon carcinoma cells were resistant to death-receptor ligands, whereas BAX-expressing sister clones were sensitive In search for evidence for such a mechanism, we investigated whether apoptotic indices, DR4, DR5 and TRAIL expression were related to the presence or absence of BAX mutations in MMR-deficient, MSI-H tumours. We found similar apoptotic indices in BAX mutant and wild-type cases. Although the number of tumours investigated was quite small, the results are in accordance with recent data from others 34. The small number of BAX mutations in our study does not allow definite conclusions on the role of BAX. Nevertheless, our data suggest that BAX mutations in these tumours do not protect completely against apoptosis. One explanation may lie in the following. TRAIL can induce apoptosis by direct activation of caspases (extrinsic, type I pathway), but also by activation of the mitochondrial (intrinsic, type II) pathway with subsequent release of factors such as cytochrome c 15. It seems that in some cell types, both pathways are activated, while in others, only one pathway is preferentially activated 9. One could speculate that the mitochondrial pathway, which involves BAX, is less important than the extrinsic pathway in colorectal cancer cells. Alternatively, it may be that other ligands than TRAIL, e.g. FasLigand, may be more important mechanisms in apoptosis induction in these tumours. The finding of almost ubiquitous expression of DR4 and DR5 in MSI-H tumours together with a low prevalence of BAX mutations, does not suggest a survival benefit for BAX mutated cells in the setting of MMR deficiency. Taken together, we found no evidence supporting the concept of BAX inactivation as a critical mechanism to evade TRAIL-receptor-mediated apoptosis in vivo. Whether TRAIL-mediated apoptosis plays a role in colon cancer cells in vivo remains to be proven. There are however some observations supporting the functionality of this pathway. First, studies using TRAIL-deficient mice demonstrated that TRAIL is important in controlling tumour growth and metastasis 35,36. Second, sensitivity to TRAIL-induced apoptosis in colon cancer cell lines correlated with the relative expression of DR4 and DR5 on the cell membrane 37. Third, TRAIL expressing macrophages isolated from pleural effusions of cancer patients induced apoptosis in colon cancer cells in a concentration dependent fashion 24. Finally, in our study, apoptotic indices were positively correlated to DR4 positivity, both in adenomas and carcinomas, and to DR5 positivity in carcinomas. In conclusion, the widespread expression of pro-apoptotic death receptors for TRAIL in sporadic and hereditary colorectal neoplasms provides potential targets for apoptosis induction. In addition, we found no support for BAX inactivation as a mechanism to evade apoptosis in vivo in MSI-H tumours. These observations hold promise for targeting of DR4 or DR5 with TRAIL, agonistic antibodies or other TRAIL-receptor agonists to treat or prevent colorectal neoplasms. 86

88 TRAIL and its receptors in hereditary colorectal neoplasms References 1. Jass JR. Familial colorectal cancer: pathology and molecular characteristics. Lancet Oncol 2000;1: De la Chapelle A. Microsatellite instability. N Engl J Med 2003;349: Jass JR, Whitehall VL, Young J, et al. Emerging concepts in colorectal neoplasia. Gastroenterology 2002;123: Huls G, Koornstra JJ, Kleibeuker JH. Non-steroidal anti-inflammatory drugs and molecular carcinogenesis of colorectal carcinomas. Lancet 2003;362: Cruz-Correa M, Hylind LM, Romans KE, et al. Long-term treatment with sulindac in familial adenomatous polyposis: a prospective cohort study. Gastroenterology 2002;122: Sinicrope FA, Lemoine M, Xi L, et al. Reduced expression of cyclooxygenase 2 proteins in hereditary nonpolyposis colorectal cancers relative to sporadic cancers. Gastroenterology 1999;117: Walczak H, Krammer PH. The CD95 (APO-1/Fas) and the TRAIL (APO-2L) apoptosis systems. Exp Cell Res 2000;256: Wang S, El Deiry WS. TRAIL and apoptosis induction by TNF-family death receptors. Oncogene 2003;22: Younes A, Kadin ME. Emerging applications of the tumor necrosis factor family of ligands and receptors in cancer therapy. J Clin Oncol 2003;21: Koornstra JJ, Kleibeuker JH, van Geelen CMM, et al. Expression of TRAIL (TNF-related apoptosisinducing ligand) and its receptors in normal colonic mucosa, adenomas, and carcinomas. J Pathol 2003;200: Sträter J, Hinz U, Walczak H, et al. Expression of TRAIL and TRAIL receptors in colon carcinoma: TRAIL-R1 is an independent prognostic parameter. Clin Cancer Res 2002;8: Deng Y, Lin Y, Wu X. TRAIL-induced apoptosis requires Bax-dependent mitochondrial release of Smac/ DIABLO. Genes Dev 2002;16: Ravi R, Bedi A. Requirement of BAX for TRAIL/Apo2L-induced apoptosis of colorectal cancers: synergism with sulindac-mediated inhibition of Bcl-x(L). Cancer Res 2002;62: Theodorakis P, Lomonosova E, Chinnadurai G. Critical requirement of BAX for manifestation of apoptosis induced by multiple stimuli in human epithelial cancer cells. Cancer Res 2002;62: LeBlanc H, Lawrence D, Varfolomeev E, et al. Tumor-cell resistance to death receptor--induced apoptosis through mutational inactivation of the proapoptotic Bcl-2 homolog Bax. Nat Med 2002;8: Bleeker WA, Mulder NH, Hermans J, et al. The addition of low-dose leucovorin to the combination of 5- fluorouracil- levamisole does not improve survival in the adjuvant treatment of Dukes C colon cancer. IKN Colon Trial Group. Ann Oncol 2000;11: Boland CR, Thibodeau SN, Hamilton SR, et al. A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 1998;58: Lindor NM, Burgart LJ, Leontovich O, et al. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 2002;20: Jass JR, Sobin LH, Watanabe H. The World Health Organization s histologic classification of 87

89 Chapter 5 gastrointestinal tumors. A commentary on the second edition. Cancer 1990;66: Koornstra JJ, Rijcken FE, De Jong S, Hollema H, de Vries EG, Kleibeuker JH.Assessment of apoptosis by M30 immunoreactivity and the correlation with morphological criteria in normal colorectal mucosa, adenomas and carcinomas. Histopathology 2004;44: Rampino N, Yamamoto H, Ionov Y, et al. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 1997;275: Erebours F, Olschwang S, Thuille B, et al. Deletion mapping of the tumor suppressor locus involved in colorectal cancer on chromosome band 8p21. Genes Chromosomes Cancer 1999;25: Kelemen PR, Yaremko ML, Kim AH, et al. Loss of heterozygosity in 8p is associated with microinvasion in colorectal carcinoma. Genes Chromosomes Cancer 1994;11: Herbeuval JP, Lambert C, Sabido O, et al. Macrophages from cancer patients: analysis of TRAIL, TRAIL receptors, and colon tumor cell apoptosis. J Natl Cancer Inst 2003;95: de Jong S, Timmer T, Heijenbrok FJ, et al. Death receptor ligands, in particular TRAIL, to overcome drug resistance. Cancer Metastasis Rev 2001;20: Ashkenazi A, Pai RC, Fong S, et al. Safety and antitumor activity of recombinant soluble Apo2 ligand. J Clin Invest 1999;104: Kelley SK, Harris LA, Xie D, et al. Preclinical studies to predict the disposition of Apo2L/tumor necrosis factor-related apoptosis-inducing ligand in humans: characterization of in vivo efficacy, pharmacokinetics, and safety. J Pharmacol Exp Ther 2001;299: Gliniak B, Le T. Tumor necrosis factor-related apoptosis-inducing ligand s antitumor activity in vivo is enhanced by the chemotherapeutic agent CPT-11. Cancer Res 1999;59: Lacour S, Hammann A, Wotawa A, et al. Anticancer agents sensitize tumor cells to tumor necrosis factor-related apoptosis-inducing ligand-mediated caspase-8 activation and apoptosis. Cancer Res 2001;61: Chuntharapai A, Dodge K, Grimmer K, et al. Isotype-dependent inhibition of tumor growth in vivo by monoclonal antibodies to death receptor 4. J Immunol 2001;166: Ichikawa K, Liu W, Zhao L, et al. Tumoricidal activity of a novel anti-human DR5 monoclonal antibody without hepatocyte cytotoxicity. Nat Med 2001;7: He Q, Luo X, Huang Y, et al. Apo2L/TRAIL differentially modulates the apoptotic effects of sulindac and a COX-2 selective non-steroidal anti-inflammatory agent in Bax-deficient cells. Oncogene 2002;21: Huang Y, He Q, Hillman MJ, et al. Sulindac sulfide-induced apoptosis involves death receptor 5 and the caspase 8-dependent pathway in human colon and prostate cancer cells. Cancer Res 2001;61: Trojan J, Brieger A, Raedle J et al. BAX and caspase-5 frameshift mutations and spontaneous apoptosis in colorectal cancer with microsatellite instability. Int J Colorectal Dis 2004;19: Cretney E, Takeda K, Yagita H, et al. Increased susceptibility to tumor initiation and metastasis in TNFrelated apoptosis-inducing ligand-deficient mice. J Immunol 2002;168: Takeda K, Smyth MJ, Cretney E, et al. Critical role for tumor necrosis factor-related apoptosis-inducing ligand in immune surveillance against tumor development. J Exp Med 2002;195: Van Geelen CM, de Vries EG, Le TK, et al. Differential modulation of the TRAIL receptors and the CD95 receptor in colon carcinoma cell lines. Br J Cancer 2003;89:

90

91

92 Chapter 6 Prognostic significance of TRAIL, DR4 and DR5 expression in adjuvant treated stage III colon cancer patients C.M.M. van Geelen 1, J.L. Westra 2, E.G.E. de Vries 1, W. Boersma-van Ek 3, N. Zwart 1, H. Hollema 4, H.M. Boezen 5, N.H. Mulder 1, J.T.M. Plukker 6, S. de Jong 1, J.H. Kleibeuker 3, J.J. Koornstra 3 Departments of 1 Medical Oncology, 2 Medical Genetics, 3 Gastroenterology and Hepatology, 4 Pathology, 5 Epidemiology, and 6 Surgery. University of Groningen Medical Centre. submitted

93 Chapter 6 Abstract Background: In preclinical models there is a synergy between chemotherapy and recombinant human TNF-related apoptosis-inducing ligand (TRAIL), on apoptosis induction in tumour cells. Therefore, the prognostic relevance was analysed of the expression of TRAIL and its death receptors DR4 and DR5 on disease free survival and overall survival in stage III colon cancer patients treated with adjuvant chemotherapy. Methods: Tissue microarrays were constructed of tumour tissue from 376 stage III colon cancer patients previously treated in a randomised fluorouracil-based adjuvant chemotherapy study and stained immunohistochemically for TRAIL, DR4 and DR5. Kaplan-Meier survival analysis and Cox proportional hazard analysis (with adjustment for treatment arm, sex and age) were performed. Results: Median follow-up was 43 months. The majority of tumours showed high expression of TRAIL (82.8 % of cases), DR4 (91.8 %) and DR5 (87.1 %). High DR4 expression was associated with worse overall (OR (95 % CI) = 2.22 ( ); p = 0.04) and disease free survival (OR = 2.19 ( ); p = 0.03). Within patients who relapsed, time to recurrence was longer for those with high DR5 expression (p = 0.03) or high TRAIL expression (p = 0.007) compared to those with low DR5 or TRAIL expression. Conclusion: High DR4 expression is associated with worse disease free survival and overall survival while high DR5 and TRAIL expression are associated with a longer time to recurrence in stage III adjuvant-treated colon cancer patients. Given the high expression of DR4 and DR5 in the majority of tumours, rhtrail or agonistic TRAIL death receptor antibodies may be considered an interesting constituent of adjuvant treatment of stage III colon cancer patients. Introduction Colorectal cancer is one of the leading causes of cancer-related death worldwide. Although over 80 % of colorectal carcinomas are macroscopically resectable at the time of diagnosis, 50 % of patients subsequently relapse, due to the presence of micrometastases. Adjuvant treatment aims to eliminate residual tumour cells and increases the proportion of patients achieving long-term disease-free and overall survival. In patients with established lymph-node metastasis (stage III), adjuvant 5-fluorouracil based chemotherapy following surgical resection has proven benefit 1. However, intrinsic as well as acquired resistance to chemotherapeutic drugs is a major problem in the treatment of this disease. Therefore, it is of great importance to identify prognostic factors that can help to develop new patienttailored treatment strategies for colorectal cancer patients. Tumour suppressor gene p53 plays a key role in induction of the intrinsic apoptosis pathway in response to several chemotherapeutic agents. Most of the colorectal cancers have lost or inactivated p53 and this may limit the clinical efficacy of chemotherapeutic regimens 2-4. Data from our institution recently confirmed the importance of p53 mutations as a prognostic factor in adjuvant-treated stage III colon cancer patients 5. Tumour-necrosisfactor-related-apoptosis-inducing-ligand (TRAIL) is a so-called death ligand of the tumour necrosis factor (TNF) family that induces apoptosis in tumour cells 6,7. Interestingly, in 92

94 TRAIL and its receptors and colon cancer survival contrast to most chemotherapeutic agents, recombinant human (rh)trail can initiate apoptosis through a distinct (extrinsic) apoptosis pathway, independent of the p53 status of the tumour 4. Furthermore, chemotherapeutic agents cooperate with rhtrail and facilitate p53-independent apoptosis of colon cancer cells 4. RhTRAIL, alone or in combination with chemotherapy, has demonstrated an anti-tumour effect in pre-clinical studies without harming normal tissue Therefore it harbours potential as a therapeutic agent. TRAIL binds to four membrane receptors. Following binding to the death receptors DR4 and DR5, TRAIL can trigger apoptosis in tumour cells. The other two receptors, DcR1 and DcR2, do not transfer an apoptosis signal because they either lack an intracellular domain or have a truncated intracellular domain, respectively. Therefore DcR1 and DcR2 are thought to act as decoy receptors (reviewed in ). Expression patterns of DR4 and DR5 have already been described in normal colon, sporadic adenomas and sporadic carcinomas 17,18. Sträter et al suggested that DR4 expression might be an important prognostic factor for colorectal cancer 18. However, this study included both stage II and III patients. Interestingly, DR5 was recently shown to be important for sensitivity to 5-fluorouracil 19. No data are available on the role of TRAIL, DR4 and DR5 expression in tumours of patients who receive adjuvant chemotherapy for stage III colon cancer. This would be of major interest because of the observed synergy between chemotherapy and rhtrail on apoptosis induction in tumour cells and the potential clinical application of rhtrail. The aim of this study was to determine the DR4, DR5 and TRAIL expression pattern in stage III colon cancer and to correlate these expression patterns with disease-free survival and overall survival. Tissue microarrays of 376 tumours of patients with stage III colon cancer who participated in an adjuvant chemotherapy study were analysed. Materials and methods Patients material Formalin-fixed paraffin-embedded primary tumour tissue from 376 patients with stage III colon cancer included in a nation-wide randomised colon cancer trial (n = 500) was available. In the remaining 124 patients no primary tumour specimens or insufficient material were available. All patients had undergone surgical resection with histologically negative resection margins but positive lymph nodes and were treated adjuvantly with 5- fluorouracil-based chemotherapy (5-fluorouracil /levamisole or 5-fluorouracil/levamisole/ leucovorin as described earlier) 20. The medical ethical committees of all participating hospitals approved the study. All patients gave written informed consent. Clinical and tumor characteristics For all patients clinical and tumour characteristics were derived from the clinical database maintained at the Comprehensive Cancer Centre North Netherlands. The mean age of the 376 patients at time of randomisation was 58.8 years (ranging from years), 201 (53.5 %) were males. 173 tumours (46 %) were right-sided i.e. proximal of the splenic flexure. By histological classification according to the WHO criteria 21, the degree of differentiation was good in 16.2 %, moderate in 53.5 %, poor in 22.6 % and unknown in 7.7 % of tumours. Patients of the subset of 376 cases for this study were equally distributed between treatment 93

95 Chapter 6 with 5-fluorouracil/levamisole (50.8 %) and 5-fluorouracil/leucovorin/levamisole (49.2 %). The addition of leucovorin had no effect on overall survival or disease free survival. Therefore, no distinction was made between the treatment arms in further analysis. During a mean follow-up of 43 months, 160 deaths were observed in the cohort of 376 patients. Tissue microarrays The construction of tissue microarrays from this patient group was recently described 22. Briefly, representative tumour tissues were selected using haematoxylin and eosin stained slides and arrayed into 4 different tissue microarrays, according to standard protocols. 23 Three 0.6 mm tissue cores were taken from three discrete representative regions of each tumour specimen using the manual tissue arrayer from Beecher Instruments (Sun Prairie, WI, USA) and put into a standard-size recipient paraffin block, the tissue microarray block. Each array contained 322 tissue cores, representing 94 tumour samples in triplicate and 20 internal controls, i.e. the same on all arrays, in duplicate (10 normal tissues and 10 tumour tissues). Sections of 4 µm were taken from each array block using the paraffin tape-transfer system with adhesive-coated slides from Instrumedics (Hackensack, NJ, USA). Immunohistochemistry The tissue microarrays were stained for DR4, DR5 and TRAIL. After deparaffinisation in xylene, blocking of endogenous peroxidase with 0.3 % hydrogen peroxide and incubation with avidin and biotin blocking solutions (Vector Laboratories, Burlingame, CA, USA) the primary antibodies were applied for 1 h at room temperature. Staining procedures for DR4, DR5 and TRAIL were performed as described previously 17. Briefly, slides were stained with a goat polyclonal IgG for TRAIL (1:25; clone K-18, Santa Cruz Biotechnology, Santa Cruz, CA, USA), a goat polyclonal IgG for DR4 (1:50; clone C-20, Santa Cruz) and a rabbit polyclonal IgG for DR5 (1:100; Oncogene Research, Cambridge, MA, USA). After washing with PBS, slides were incubated with a 1:300 dilution of a biotinylated rabbitanti-goat or swine-anti-rabbit antibody (DAKO, Glostrup, Denmark) respectively, followed by addition of streptavidin-conjugated peroxidase. Peroxidase activity was visualised with diaminobenzidine. Slides were counterstained with haematoxylin. Evaluation of staining was performed by light microscopy by two investigators, without knowledge of the clinical data of the patients and with re-evaluation under a multi-headed microscope if results did not agree. Cores containing less than 10 % tumour cells were not evaluated. The score was determined by the degree of staining seen in the majority of the tumour tissue core. If cores from one tumour differed in staining intensity, the median score of the three related cores determined the score of the tumour. The immunoreactivity was assessed using a grading system of four categories based on the intensity (0 = no expression, 1 = low expression, 2 = moderate expression, 3 = strong expression). For statistical purposes, categories 0 and 1 were considered as low and categories 2 and 3 as high expression. Statistical analyses Differences in distributions of categorical variables between groups were tested using Chi-square tests. Differences between groups in distributions of continuous variables were tested using t-tests or Mann-Whitney U-tests, as appropriate. Patients were categorised into low and high expression levels of DR4, DR5 and TRAIL as described above. Differences 94

96 TRAIL and its receptors and colon cancer survival Table 1. Staining results for DR4, DR5 and TRAIL in colon carcinomas. DR4 expression n (%) DR5 expression n (%) TRAIL expression n (%) n Staining intensity 0 1 (0.3) 1 (0.3) 6 (1.9) 1 25 (7.9) 43 (12.6) 48 (15.3) (60.1) 221 (64.6) 199 (63.4) (31.7) 77 (22.5) 61 (19.4) in disease free and overall survival according to expression of DR4, DR5 and TRAIL were analysed using Kaplan-Meier survival analysis using log-rank testing and Cox proportional hazard analysis. Cox proportional hazard analyses were performed to estimate the effect of high DR4, DR5 and TRAIL expression on disease free and overall survival with adjustment for the following potential confounders: treatment arm, age and sex. Since we have previously shown that p53 mutation status is predictive of disease-free survival, we repeated all Cox analyses with additional adjustment for p53 mutation status, for which data were obtained from the original report 5. Disease free survival was defined as the period from the date of randomisation to the date of any documented relapse or the date of last contact without any documented relapse. Overall survival was defined as the time period from the date of randomisation to the date of documented death or to the date of last contact and documented alive. To analyse whether DR4, DR5 and TRAIL expression affected time to recurrence within those with relapse, we analysed the time till tumour recurrence, defined as the period from the date of randomisation to the date of documented tumour recurrence within those subjects with tumour recurrence, using Kaplan-Meier. Cox proportional hazard analysis was performed in order to estimate (Odds Ratio (OR), 95 % confidence interval (CI)) the effect of DR4, DR5 and TRAIL expression on time till tumour recurrence within those subjects with tumour recurrence, with adjustment for the following potential confounders: treatment arm, age, sex and p53 mutation status. P values < 0.05 were considered significant. SPSS for Windows 12.0 (SPSS Inc., Chicago, IL) was used for all statistical analysis. Results Expression of DR4, DR5 and TRAIL in stage III colon cancer Immunohistochemistry for TRAIL, DR4 and DR5 was performed on all 376 tumours. Interpretable data were obtained for TRAIL in 314, for DR4 in 316 and for DR5 in 342 tumours. DR4, DR5 and TRAIL staining was cytoplasmic and no apparent membranous staining was detected. Most tumours showed high expression of DR4, DR5 and TRAIL (Table 1, figure 1). Strong expression (staining intensity 2 and 3) of TRAIL receptors DR4 and DR5 was observed in 91.8 % and 87.1 % of the tumours, respectively. When coupling 95

97 Chapter 6 Figure 1. DR4, DR5 and TRAIL expression in tissue microarrays of colon cancer tumors. Examples of immunohistochemical staining results for DR4 (A and B, low and high expression respectively), DR5 (C and D, low and high expression respectively) and TRAIL (E and F, low and high expression respectively). See appendix for colour pictures. interpretable staining intensity data of DR4 and DR5 expression, we found that 8 of 291 tumours showed low expression of both receptors; in all other cases expression of at least one receptor was high. TRAIL was high in 82.8 % of the tumours. There was no difference in DR4 and DR5 expression between tumours with high TRAIL expression versus those with low TRAIL expression. DR4, DR5 or TRAIL expression and disease free survival Using Kaplan-Meier survival analysis, high DR4 expression was associated with a worse disease free survival compared to low DR4 expression with borderline significance (p = 0.06). We found no such association with disease free survival for DR5 and TRAIL expression. In Cox proportional hazard analyses high DR4 was associated with worse disease free survival (OR = 2.19 ( ); p = 0.03), whereas DR5 and TRAIL expression were not. 96

98 TRAIL and its receptors and colon cancer survival A 1,0 P= Cumulative survival 0,8 0,6 0,4 0,2 DR4-HIGH DR4-LOW 0, Time to recurrence (months) B 1,0 P= Cumulative survival 0,8 0,6 0,4 DR5-HIGH 0,2 DR5-LOW 0, Time to recurrence (months) C 1,0 P= Cumulative survival 0,8 0,6 0,4 TRAIL-HIGH 0,2 TRAIL-LOW 0, Time to recurrence (months) Figure 2. Time to recurrence in patients with recurrent disease in adjuvant-treated stage III colon cancer patients according to DR4 (A), DR5 (B) and to TRAIL (C) expression. DR4, DR5 or TRAIL expression and overall survival Kaplan-Meier survival analysis revealed a trend towards worse overall survival for individuals with high DR4 expression compared to those with low DR4 expression (p=0.06). In contrast, there was a trend towards better overall survival associated with high expression of DR5 (p = 0.07). Cox proportional hazard analyses showed that high DR4 expression was associated with a worse survival (OR (95 % CI) = 2.22 ( ); p = 0.04) while high DR5 expression was borderline significant for a better survival (OR = 0.65 ( ); p = 0.07). High or low TRAIL expression was not associated with overall survival. 97

99 Chapter 6 DR4, DR5 or TRAIL expression and time to recurrence From the cohort of 376 patients, 168 (44.7 %) experienced recurrent disease during followup. Kaplan-Meier survival curves in these 168 patients indicated a longer time to recurrence for high DR5 (p = 0.03) or high TRAIL (p = 0.007) expression compared to low DR5 or TRAIL expression levels (figures 2B and 2C). DR4 expression was not associated with time to recurrence (figure 2A). The distribution of time to recurrence according to high versus low DR4 expression did not allow for Cox proportional hazard analysis. Cox proportional hazard analyses confirmed that those patients with high DR5 and TRAIL expression were less likely to have early recurrence (OR = 0.60 ( ); p = 0.04 and OR = 0.54 ( ); p = 0.02 respectively). DR4, DR5 and TRAIL expression in relation to p53 mutation status Given that DR4 and DR5 expression are reported to be at least in part p53 dependent 24,25, expression patterns of DR4 and DR5 were compared between tumours with p53 mutations and those with wild-type p53 (table 2). No significant differences were found between the two groups. We also found no differences between tumours with p53 mutations and those with wild-type p53 and expression patterns of TRAIL. We repeated all Cox proportional hazard analyses in order to estimate the effect of DR4, DR5 and TRAIL expression on disease free survival, overall survival and time to tumour recurrence within those subjects with tumour recurrence, with additional adjustment for p53 mutation status. The above described effect estimates did not change substantially following adjustment for p53 mutation status. Discussion In the present study, the prognostic value of TRAIL and its death receptors DR4 and DR5 was studied in colon cancers of patients who received adjuvant chemotherapy. With the tissue microarray technique, 376 tumour samples of stage III colon cancer patients previously participating in a randomised adjuvant chemotherapy study were analysed. The majority of the tumours showed high expression of DR4, DR5 and TRAIL. High DR4 expression was associated with worse overall and disease free survival. There was a trend towards better disease free survival with high DR5 expression as compared to low DR5 expression. Furthermore, high expression of TRAIL and high expression of DR5 were associated with a longer time to recurrence of the disease. Previous immunohistochemical studies have demonstrated that the pro-apoptotic TRAIL-receptors DR4 and DR5 are expressed in normal colon mucosa as well as colorectal adenomas and carcinomas 17,18. In contrast to our results, Sträter et al. demonstrated that high DR4 expression was associated with better disease-free survival rates in a heterogeneous group of 128 patients with colon cancer, including both stage II and III tumours 18. In their study 18.6 % of the patients received adjuvant chemotherapy. The strengths of our study are the high number of patients, the inclusion of only stage III colon cancer patients and the fact that all patients were treated with 5-fluorouracil based chemotherapy within a prospective clinical trial. This means that results obtained give a level II evidence for the 98

100 TRAIL and its receptors and colon cancer survival Table 2. Patterns of DR4 and DR5 expression in tumors with wild-type p53 versus tumors with p53 mutations. wild-type p53 (%) mutant p53 (%) p DR4 Staining intensity 0/ ns DR5 Staining intensity 0/ ns ns: not significant utility of the tumour marker studied 26. In our study high DR4 expression was associated with a shorter disease free and overall survival. This indicates that DR4 positive stage III colon cancers may have a growth advantage. Spierings et al speculated that a number of other signalling pathways, such as the ERK1/2 pathway, are activated by death receptors and that these might have a dominant protecting effect over the apoptotic signal from death receptors, stimulating cell proliferation 27. Increasingly it is thought that as an endpoint for treatment effects disease free survival with a follow up of at least 3 years is preferable over data obtained with overall survival 28. This makes also our findings for DR5 and TRAIL of interest. High DR5 expression was associated with longer disease free survival with borderline significance and significantly associated with a longer time to recurrence. It has been demonstrated that in human colon cancer cell lines, in which the expression of DR5 was decreased using inducible RNA interference, silencing of DR5 in vivo led to accelerated growth of bioluminescent tumour xenografts. This study also showed that silencing of DR5 conferred resistance of the tumour cells to the chemotherapeutic agent 5-fluorouracil 19. Tumour cells with DR5 expression may be more susceptible to TRAIL-induced apoptosis than DR4 positive tumour cells. Truneh et al have demonstrated that at physiological conditions TRAIL binds with a higher affinity to DR5 than to DR4 29. Additionally, in order to induce apoptosis, DR4 and DR5 have distinct cross-linking requirements. DR4 equally responds to cross-linked (e.g. membrane bound) TRAIL and non-cross-linked (soluble) rhtrail, whereas DR5 signals only in response to cross-linked rhtrail 30. Several studies have demonstrated that DR4 and DR5 can contribute differently to TRAIL-induced apoptosis 31,32. Furthermore, DR4 could be less expressed or non-functional in colon cancers due to mutations. Alterations or mutations in or near the ligand-binding domain of DR4 have the potential to affect TRAIL binding to DR4 and are found in several cancers 33,34. 99

101 Chapter 6 Nevertheless, DR4 mutations in colorectal cancer have not been studied so far and no somatic DR5 mutations were found in 41 colorectal cancers 35. Interestingly high TRAIL expression was associated with a longer time to recurrence of the disease. Previously, it has been demonstrated that high TRAIL mrna expression levels in ovarian cancer were associated with longer survival (n = 120) 36. In accordance with this, another study reported that reduced TRAIL protein levels in tumours of ovarian cancer patients were associated with a poor prognosis 37. There are several indications that TRAIL expression is involved in immune surveillance 38,39. In agreement with this, TRAIL expression could contribute to the immune escape of tumour cells 40. In that case, it is expected that high expression of TRAIL would be associated with an unfavourable disease. However, elevated levels of cell-associated and secreted TRAIL can also contribute to apoptosis induction in cancer cells in autocrine or paracrine manners In our study, tumours with high expression of TRAIL may have higher levels of autocrine or paracrine TRAIL-induced apoptosis compared to tumours with low TRAIL expression levels. All patients received adjuvant chemotherapy. Thus in tumour cells with high TRAIL expression the intrinsic (drug-induced) as well as the extrinsic (TRAIL interaction with DR5) pathway may have been activated, enhancing the levels of apoptosis in tumour cells and thus increase the time to recurrence. A few other studies have investigated the prognostic value of TRAIL, DR4 and DR5 expression. DR4 and DR5 expression in acute myelogenous leukemia patients (n = 29) showed no prognostic significance 44. Evaluation of DR4, DR5 and TRAIL expression in non-small cell lung cancer patients revealed that positive DR5 staining was associated with increased risk of death 27. In colon cancer cells rhtrail induces apoptosis in vitro and in xenograft models 45,46, whereas normal human colon epithelium is resistant to TRAIL-induced apoptosis 47. In addition, anti-cancer drugs are known to increase DR4 and DR5 membrane expression on the cell surface and to enhance TRAIL-mediated apoptosis in colon cancer cells 45, Furthermore, the natural phytoalexin resveratrol was reported to enhance death receptor membrane expression, induce their redistribution into lipid rafts and to facilitate death ligand induced apoptosis 51. Combinations of these agents can enhance the anti-tumour effect of rhtrail. In vitro, the combination of 5-fluorouracil and rhtrail has an additive antitumour effect in colon cancer cell lines 52. Administration of rhtrail showed no toxicity in animal models 8,13. Several clinical phase I-II studies with rhtrail and agonistic anti-dr4 and anti-dr5 antibodies are ongoing 53,54. The fact that the majority of the tumours showed high expression levels of DR4 and DR5, independent of p53 mutations, and the finding that high TRAIL expression was associated with longer time to recurrence, supports the further development of rhtrail or TRAIL receptor agonistic antibodies as anti-cancer agents for stage III colon cancer patients. In conclusion, high DR4 expression is associated with worse overall and disease free survival and both DR5 and TRAIL expression are associated with a longer time to recurrence in adjuvant-treated stage III colon cancer patients. The fact that the majority of the tumours showed high DR4 and DR5 expression indicates that rhtrail or agonistic antibodies may be interesting molecules for use in adjuvant treatment of stage III colon cancer patients. 100

102 TRAIL and its receptors and colon cancer survival References 1. Andre T, Boni C, Mounedji-Boudiaf L, et al. Oxaliplatin, fluorouracil, and leucovorin asadjuvant treatment for colon cancer. N Engl J Med 2004;350: Lowe SW, Ruley HE, Jacks T, Housman DE. p53-dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell 1993;74: Lowe SW, Bodis S, McClatchey A, et al. p53 status and the efficacy of cancer therapy invivo. Science 1994;266: Ravi R, Jain AJ, Schulick RD, et al. Elimination of hepatic metastases of colon cancer cells via p53- independent cross-talk between irinotecan and Apo2 ligand/trail. Cancer Res2004;64: Westra JL, Schaapveld M, Hollema H, et al. Determination of TP53 mutation is more relevant than microsatellite instability status for the prediction of disease-free survival inadjuvant-treated stage III colon cancer patients. J Clin Oncol 2005;23: Pitti RM, Marsters SA, Ruppert S, Donahue CJ, Moore A, Ashkenazi A. Induction of apoptosis by Apo- 2 ligand, a new member of the tumor necrosis factor cytokine family. J Biol Chem 1996;271: Wiley SR, Schooley K, Smolak PJ, et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity. 1995;3: Ashkenazi A, Pai RC, Fong S, et al. Safety and antitumor activity of recombinant soluble Apo2 ligand. J Clin Invest 1999;104: Kelley SK, Harris LA, Xie D, et al. Preclinical studies to predict the disposition of Apo2L/tumor necrosis factor-related apoptosis-inducing ligand in humans: characterization of in vivo efficacy, pharmacokinetics, and safety. J Pharmacol Exp Ther 2001;299: Lawrence D, Shahrokh Z, Marsters S, et al. Differential hepatocyte toxicity of recombinant Apo2L/ TRAIL versions. Nat Med 2001;7: Mitsiades CS, Treon SP, Mitsiades N, et al. TRAIL/Apo2L ligand selectively induces apoptosis and overcomes drug resistance in multiple myeloma: therapeutic applications. Blood 2001;98: Nagane M, Pan G, Weddle JJ, Dixit VM, Cavenee WK, Huang HJ. Increased death receptor 5 expression by chemotherapeutic agents in human gliomas causes synergistic cytotoxicity with tumor necrosis factor-related apoptosis-inducing ligand in vitro and in vivo. Cancer Res 2000;60: Walczak H, Miller RE, Ariail K, et al. Tumoricidal activity of tumor necrosis factor related apoptosis inducing ligand in vivo. Nat Med 1999;5: Held J, Schulze-Osthoff K. Potential and caveats of TRAIL in cancer therapy. Drug ResistUpdat 2001;4: Hengartner MO. The biochemistry of apoptosis. Nature 2000;407: Reed JC. Mechanisms of apoptosis. Am J Pathol 2000;157: Koornstra JJ, Kleibeuker JH, van Geelen CM, et al. Expression of TRAIL (TNF-related apoptosisinducing ligand) and its receptors in normal colonic mucosa, adenomas, and carcinomas. J Pathol 2003;200: Sträter J, Hinz U, Walczak H, et al. Expression of TRAIL and TRAIL receptors in colon carcinoma: TRAIL-R1 is an independent prognostic parameter. Clin Cancer Res 2002;8: Wang S, El-Deiry WS. Inducible silencing of KILLER/DR5 in vivo promotes bioluminescent colon tumor xenograft growth and confers resistance to chemotherapeutic agent 5-fluorouracil. Cancer Res 2004;64:

103 Chapter Bleeker WA, Mulder NH, Hermans J, Otter R, Plukker JT. The addition of low-dose leucovorin to the combination of 5-fluorouracil- levamisole does not improve survival in the adjuvant treatment of Dukes C colon cancer. IKN Colon Trial Group. Ann Oncol 2000;11: Hamilton SR, Aaltonen LA (ed). WHO classification of tumours: pathology and genetics of tumours of digestive system. Geneva, Switzerland, WHO, Westra JL, Hollema H, Schaapveld M, et al. Predictive value of thymidylate synthase and dihydropyrimidine dehydrogenase protein expression on survival in adjuvantly treated stage III colon cancer patients. Ann Oncol 2005;16: Hoos A, Cordon-Cardo C. Tissue microarray profiling of cancer specimens and cell lines opportunities and limitations. Lab Invest 2001;81: Takimoto R, El-Deiry WS. Wild-type p53 transactivates the KILLER/DR5 gene through an intronic sequence-specific DNA-binding site. Oncogene 2000;19: Liu X, Yue P, Khuri FR, Sun SY. p53 upregulates death receptor 4 expression through an intronic p53 binding site. Cancer Res 2004;64: Hayes DF, Bast RC, Desch CE, et al. American Society of Clinical Oncology Tumor Markers Expert Panel. Tumor marker utility grading system: a framework to evaluate clinical utility of tumor markers. J Natl Cancer Inst 1996;88: Spierings DC, de Vries EG, Timens W, Groen HJ, Boezen HM, de Jong S. Expression of TRAIL and TRAIL death receptors in stage III non-small cell lung cancer tumors. Clin Cancer Res 2003;9: Grem J. Adjuvant therapy for colon cancer: a historical perspective. J Natl Cancer Inst 2004;96: Truneh A, Sharma S, Silverman C, et al. Temperature-sensitive differential affinity of TRAIL for its receptors. DR5 is the highest affinity receptor. J Biol Chem 2000;275: Wajant H, Moosmayer D, Wuest T, et al. Differential activation of TRAIL-R1 and -2 by soluble and membrane TRAIL allows selective surface antigen-directed activation of TRAIL-2 by a soluble TRAIL derivative. Oncogene 2001;20: Kelley RF, Totpal K, Lindstrom SH, et al. Receptor-selective mutants of Apo2L/TRAIL reveal a greater contribution of DR5 than DR4 to apoptosis signaling. J Biol Chem 2005;280: Leverkus M, Sprick MR, Wachter T, et al. TRAIL-induced apoptosis and gene induction in HaCaT keratinocytes: differential contribution of TRAIL receptors 1 and 2. J Invest Dermatol 2003;121: Dechant MJ, Fellenberg J, Scheuerpflug CG, Ewerbeck V, Debatin KM. Mutation analysis of the apoptotic death-receptors and the adaptors TRADD and FADD/MORT-1 in osteosarcoma tumor samples and osteosarcoma cell lines. Int J Cancer 2004;109: Fisher MJ, Virmani AK, Wu L, et al. Nucleotide substitution in the ectodomain of trail receptor DR4 is associated with lung cancer and head and neck cancer. Clin Cancer Res 2001;7: Arai T, Akiyama Y, Okabe S, Saito K, Iwai T, Yuasa Y. Genomic organization and mutation analyses of the DR5/TRAIL receptor 2 gene in colorectal carcinomas. Cancer Lett 1998;133: Lancaster JM, Sayer R, Blanchette C, et al. High expression of tumor necrosis factor-related apoptosisinducing ligand is associated with favorable ovarian cancer survival. Clin Cancer Res 2003;9: Arts HJ, de Jong S, Hollema H, ten Hoor K, van der Zee AG, de Vries EG. Chemotherapy induces death receptor 5 in epithelial ovarian carcinoma. Gynecol Oncol 2004;92: Corazza N, Brumatti G, Schaer C, Cima I, Wasem C, Brunner T. TRAIL and immunity: more than a 102

104 TRAIL and its receptors and colon cancer survival license to kill tumor cells. Cell Death Differ 2004;11 Suppl 2:S122-S Lee HO, Herndon JM, Barreiro R, Griffith TS, Ferguson TA. TRAIL: a mechanism of tumor surveillance in an immune privileged site. J Immunol 2002;169: Strebel A, Bachmann F, Wernli M, Erb P. Tumor necrosis factor-related, apoptosis inducing ligand supports growth of mouse mastocytoma tumors by killing tumor-infiltrating macrophages. Int J Cancer 2002;100: Abadie A, Wietzerbin J. Involvement of TNF-related apoptosis-inducing ligand (TRAIL) induction in interferon gamma-mediated apoptosis in Ewing tumor cells. Ann NYAcad Sci 2003;1010: Ludwig AT, Moore JM, Luo Y, et al. Tumor necrosis factor-related apoptosis-inducing ligand: a novel mechanism for Bacillus Calmette-Guerin-induced antitumor activity. Cancer Res 2004;64: Papageorgiou A, Lashinger L, Millikan R, et al. Role of tumor necrosis factor-related apoptosis-inducing ligand in interferon-induced apoptosis in human bladder cancer cells. Cancer Res 2004;64: Min YJ, Lee JH, Choi SJ, et al. Prognostic significance of Fas (CD95) and TRAIL receptors (DR4/DR5) expression in acute myelogenous leukemia. Leuk Res 2004;28: Naka T, Sugamura K, Hylander BL, Widmer MB, Rustum YM, Repasky EA. Effects of tumor necrosis factor-related apoptosis-inducing ligand alone and in combination with chemotherapeutic agents on patients colon tumors grown in SCID mice. Cancer Res 2002;62: Van Geelen CM, de Vries EG, Le TK, van Weeghel RP, de Jong S. Differential modulation of the TRAIL receptors and the CD95 receptor in colon carcinoma cell lines. Br J Cancer 2003;89: Sträter J, Walczak H, Pukrop T, et al. TRAIL and its receptors in the colonic epithelium: a putative role in the defense of viral infections. Gastroenterology 2002;122: He Q, Huang Y, Sheikh MS. Proteasome inhibitor MG132 upregulates death receptor 5 and cooperates with Apo2L/TRAIL to induce apoptosis in Bax-proficient and -deficient cells. Oncogene 2004;23: Kim YH, Park JW, Lee JY, Kwon TK. Sodium butyrate sensitizes TRAIL-mediated apoptosis by induction of transcription from the DR5 gene promoter through Sp1 sites in colon cancer cells. Carcinogenesis 2004;25: LeBlanc H, Lawrence D, Varfolomeev E, et al. Tumor-cell resistance to death receptor-induced apoptosis through mutational inactivation of the proapoptotic Bcl-2 homolog Bax. Nat Med 2002;8: Delmas D, Rebe C, Micheau O, et al. Redistribution of CD95, DR4 and DR5 in rafts accounts for the synergistic toxicity of resveratrol and death receptor ligands in colon carcinoma cells. Oncogene 2004;23: Shimoyama S, Mochizuki Y, Kusada O, Kaminishi M. Supra-additive antitumor activity of 5FU with tumor necrosis factor-related apoptosis-inducing ligand on gastric and colon cancers in vitro. Int J Oncol 2002;21: Pukac L, Kanakaraj, Humphreys R, et al. HGS-ETR1, a fully human TRAIL-receptor 1 monoclonal antibody, induces cell death in multiple tumour types in vitro and in vivo. Br J Cancer 2005;92: Motoki K, Mori E, Matsumoto A, et al. Enhanced apoptosis and tumor regression induced by a direct agonist antibody to tumor necrosis factor-related apoptosis-inducing ligand receptor 2. Clin Cancer Res 2005;11:

105

106 Chapter 7 Non-steroidal anti-inflammatory drugs and molecular carcinogenesis of colorectal carcinomas. G. Huls 1, J.J. Koornstra 2, J.H. Kleibeuker 2. Department of 1 Hematology and 2 Gastroenterology and Hepatology, University of Groningen Medical Centre, the Netherlands. Lancet 2003; 362:

107 Chapter 7 Abstract Colorectal cancer is the second most common cause of cancer-related mortality in the western world. The high incidence and mortality make effective prevention an important public-health and economic issue. Nonsteroidal anti-inflammatory drugs (NSAIDs) can inhibit colorectal carcinogenesis and are among the few agents known to be chemopreventive. Randomised trials have shown that sulindac and celecoxib suppress the development of adenomatous polyps and cause regression of existing polyps in patients with familial adenomatous polyposis (FAP), who have a high risk for developing colorectal cancer. The mechanisms by which NSAIDs inhibit neoplastic growth are not fully known. Two randomised placebo-controlled trials have recently shown a chemopreventive effect of aspirin in populations other than those with FAP (Robert Sandler and colleagues, N Engl J Med 2003; 348: ; John Baron and colleagues, N Engl J Med 2003; 348: ). In the Sandler study 635 patients with colorectal cancer were randomised to receive 325 mg aspirin or placebo daily. After a follow-up of around 31 months, the mean number of adenomas was lower in the aspirin group than in the placebo group, corresponding to a relative risk of any recurrent adenoma in the aspirin group of In the Baron study 1121 patients with colorectal adenomas were assigned to receive 81 or 325 mg aspirin or placebo daily. Follow-up colonoscopy, 32 months after the index endoscopy, showed an incidence of one or more adenomas of 38 % in the 81 mg aspirin group, 45 % in the 325 mg aspirin group, and 47 % in the placebo group. Together, these studies indicate a moderate chemopreventive effect of aspirin in populations with an intermediate risk of developing colorectal cancer. The anticancer properties of NSAIDs have been demonstrated in vitro as well as in vivo (animal studies, epidemiological reports, and intervention studies). Several mechanisms through which NSAIDs alter colonic carcinogenesis have been elucidated, including the induction of apoptosis in neoplastic cells, via mechanisms dependent and independent of cyclo-oxygenase. Some studies have suggested an important role for the cell-cycle regulating protein p21 in mediating the chemopreventive effect of sulindac. A decrease in p21 expression may be one of the main oncogenic events in the development of colorectal cancer. Thus p21 could be the molecular link in the chemopreventive effects of NSAIDs. Introduction Colorectal cancer (CRC) is the second most common cause of cancer-related deaths in the western world. Current strategies to reduce mortality from this disease focus on early detection of colorectal cancer or its precursor lesion, the adenoma, e.g. by endoscopic screening. The use of drugs to prevent the development of colorectal adenomas or carcinomas is gaining interest. Chemopreventive measures might be important, especially in patients who have an increased risk of developing colorectal neoplasia. Indeed, epidemiological studies have shown a consistent % reduction in the risk of developing colorectal neoplasia associated with the use of non-steroidal anti-inflammatory drugs (NSAIDs), despite differences in study design 1,2. Several randomised trials have shown a decrease in the number and size of adenomas in patients with familial adenomatous polyposis (FAP) 106

108 Review: NSAIDs and colorectal carcinogenesis Table. Overview of randomised placebo-controlled trials on NSAIDs and colorectal adenomas. Study population Patients NSAID Result Source High-risk FAP 10 sulindac regression of adenomas in all patients Labayle, FAP 22 sulindac reduction in number and size Giardiello, FAP 14 sulindac regression of adenomas in 71 % of cases Nugent, FAP 77 celecoxib reduction in number and size Steinbach, FAP prephenotype 42 sulindac no effect on adenoma formation Giardiello, Intermediate risk Current adenomas 44 sulindac no decrease in number and size Ladenheim, Previous CRC 635 aspirin 35 % reduction of recurrent adenomas Sandler, Previous adenomas 1121 aspirin 12 % reduction of recurrent adenomas Baron, who received the NSAIDs celecoxib or sulindac (table). Recently two randomised placebocontrolled trials showed that aspirin reduced the risk of colorectal adenomas in populations with an intermediate risk of developing adenomas (table) 9,10. However, before deciding whether aspirin should be recommended for secondary chemoprevention in persons with a history of colorectal cancer, the clinical importance of these two studies requires further thought. Despite better knowledge of molecular colorectal carcinogenesis and the mechanisms by which NSAIDs act as chemopreventive agents, the mechanisms by which NSAIDs intervene with colorectal carcinogenesis remains a matter of debate. Colorectal carcinogenesis and Wnt signaling Colorectal tumours have provided an excellent model for studying the genetic alterations involved in the development of human neoplasms, as tumours at various stages of development, from very small adenomas to large metastatic carcinomas, can be obtained for molecular genetic analysis 11,12. Correlation of morphological and genetic data in colorectal neoplasia led to a genetic model of colorectal carcinogenesis proposed by Vogelstein and colleagues 12. According to this model, carcinogenesis proceeds through the accumulation of a series of genetic mutations involving several tumour-suppressor genes (APC, TP53, and tumour-suppressor genes located at chromosome 18q) and oncogenes (k-ras), as well as epigenetic changes (methylation). The accumulation of these changes is associated with gradually increasing size, disorganisation, and malignancy of colorectal tumours. Most colorectal tumours have somatic mutations in the APC gene. Germline mutations in this gene cause the inherited cancer-predisposition syndrome FAP. Among the remaining colorectal tumours with wild-type APC, most have mutations in β-catenin 13. Much evidence suggests that the formation of benign adenomas in the intestine is initiated by mutational events occurring in either APC or the gene for β-catenin 13. APC and β-catenin both operate 107

109 Chapter 7 in the same signalling cascade, the Wnt-signalling pathway 13. Loss of functional APC or mutations in the gene for β-catenin lead to the nuclear accumulation of β-catenin, which binds and activates the transcription factor TCF4 14,15. Consequently, activated TCF4 activates a genetic programme that is presumed to be responsible for early adenoma formation. A gene-disruption experiment in mice provided some insight into the nature of the genetic programme 16. Mice deficient in TCF4 develop normally, but die shortly after birth due to the absence of cycling epithelial precursor cells in the developing crypts of the small intestine. This insight was extended by a study, which demonstrated that active Wntsignalling imposes a crypt-progenitor phenotype on colorectal cells 17. In this model, activated Wnt-signalling decreases p21 concentrations, which consequently prevents the cells from entering G1 arrest or differentiation, thereby allowing cells to proliferate. These findings led to the proposal that β-catenin/tcf4 activity acts as a switch controlling proliferation versus differentiation in the intestinal epithelium. This mechanism is illustrated by the clonal amplification of crypt-progenitor cells at the intestinal surface epithelium when β- catenin/tcf4 is activated, giving rise to aberrant crypt foci, the earliest precursor lesion of colorectal cancer 18. In addition to p21 as an important mediator of the differentiation/proliferation switch by β-catenin/tcf4 activity, there are many other pathways important for colorectal carcinogenesis. The hundreds of genes regulated by β-catenin/tcf4 activity could also affect many cellular processes (e.g. tissue invasion, apoptosis, and angiogenesis). In this issue of The Lancet, P Kim and colleagues show that the antiapoptotic protein survivin is a target of β-catenin/tcf4 activity 19. The expression of survivin, following β-catenin/tcf4 activity, could provide the adenoma cell with a mechanism to resist apoptosis. Two forms of genetic instability In colorectal cancer, two forms of genetic instability have been described: microsatellite instability and chromosomal instability. Microsatellite instability is caused by defects in mismatch-repair machinery, which results in a mutator phenotype at the nucleotide level and a consequent instability of repetitive sequences or microsatellites. 20 Microsatellite instability is the characteristic molecular defect in the cancer syndrome hereditary nonpolyposis colorectal cancer (HNPCC). Chromosomal instability is the hallmark of most colorectal cancers. Mutations in APC (not β-catenin) are involved in chromosomal instability, and are likely to elicit such instability with the synergistic co-operation of other somatically acquired mutations 21,22. Thus inactivation of APC provides the intestinal cell with two essential requirements for neoplastic development. First, activation of the β-catenin/tcf4 complex, resulting from inactivation of APC, gives a selective advantage to allow for the initial clonal expansion. Second, inactivation of APC facilitates chromosomal instability to allow for multiple hits in other genes that are responsible for tumour progression and malignant transformation. 108

110 Review: NSAIDs and colorectal carcinogenesis Mechanisms of chemoprevention by NSAIDs How do NSAIDs mediate their antineoplastic activity? NSAIDs bind and inhibit the cyclooxygenase (COX) enzymes, COX-1 and COX-2, which catalyse the conversion of arachidonic acid to prostaglandins. COX-1 is expressed constitutively and is required for physiological processes such as maintenance of gastrointestinal mucosa and platelet aggregation, whereas COX-2 is induced by cytokines, growth factors, and mitogens. NSAIDs vary in their abilities to inhibit COX-1 and COX-2 1. Tumour inhibition by NSAIDs may be mediated by distinct cellular processes. These processes involve the ability of NSAIDs to restore apoptosis, induce cell-cycle arrest, and inhibit angiogenesis 1,2. One of the main ways by which NSAIDs exert their effects is modulation of apoptosis, although there is considerable debate about how these effects are mediated 2. Because COX-2 expression is increased in up to 90% of sporadic colon carcinomas and 40% of adenomas, but not in normal colonic mucosa, NSAIDs were presumed to mediate apoptosis via COX-2 inhibition. The relevance of COX-2 for adenoma formation was genetically demonstrated by a reduced number of adenomas in APC min mice, the mouse model for FAP, with an additional targeted deletion of COX However, compounds that do not inhibit COX-2, such as sulindac sulphone, also induce apoptosis in vitro and inhibit colorectal carcinogenesis in animal models 2. In addition, 81 mg aspirin, which has virtually no COX-2 inhibitory effects, had a chemopreventive effect in individuals at increased risk for developing colorectal cancer 10. COX-independent mechanisms are also suggested by the finding that some NSAIDs inhibit proliferation and induce cell death in cells that do not express COX-1 and COX Other mechanisms of apoptosis induction have been described 1,2. NSAIDs, colorectal carcinogenesis, and p21 There may be other mechanisms than the induction of apoptosis by which NSAIDs mediate their anticancer effects. Because 85 % of sporadic colorectal carcinomas have activated Wnt-signalling, which is presumed to be important in early adenoma formation, it might be expected that NSAIDs somehow affect either Wnt- signalling itself or its target genes. But microarray data did not demonstrate COX-2 to be a target gene of β-catenin/ TCF 4,17 However, it has been shown by microarray data of rectal biopsy specimens and colonic cells in culture that sulindac induced expression of p In vitro sulindac-mediated induction of p21 expression in colon carcinoma cell-lines is associated with both cell-cycle arrest and apoptosis 26. Further, homozygous inactivation of p21 in APC min mice eliminated the ability of sulindac to reduce the number of small intestinal tumours in these mice 25. So, sulindac could mediate its effect on intestinal adenoma formation by modifying p21 expression (figure). Confirmation of these results with other NSAIDs has to be awaited. A study addressing the chemopreventive effect of NSAIDs in patients with hereditary nonpolyposis colorectal cancer would be of great value, as a subset of these tumours develops independently of Wnt signalling. Such studies are ongoing. In conclusion, the role of β-catenin/tcf4 activity as a master switch that controls proliferation versus differentiation in the intestinal epithelium, by controlling the expression of p21, extends the possible mechanisms by which NSAIDs could mediate their anticancer effect. 109

111 Chapter 7 Figure. Model for the role of p21 in the chemopreventive effects of NSAIDs. Intestinal Stem Cell Active complex of Tcf / β-catenin e.g. proliferative region of colon cryp t or adenoma Low p21 Proliferating crypt progenitors No complex of Tcf / β-catenin e.g. differentiated region of colon crypt High + p21 NSAID sulindac Differentiated intestinal cell + APOPTOSIS The β-catenin/tcf4 complex is the master switch that controls proliferation versus differentiation by regulating p21 expression. Cells undergoing mutation in either APC or β-catenin become independent of the physiological signals controlling β-catenin/tcf4 activity. This mechanism is currently believed to be responsible for early adenoma formation. Sulindac can induce expression of p21 resulting in differentiation or G1 arrest. This may be one of the explanations for the chemopreventive effects of sulindac in addition to its effect on apoptosis. Acknowledgements We thank Dr. E.G.E. de Vries for helpful discussions and comments. 110

112 Review: NSAIDs and colorectal carcinogenesis References 1. Thun MJ, Henley SJ, Patrono C. Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues. J Natl Cancer Inst 2002;94: Chan TA. Nonsteroidal anti-inflammatory drugs, apoptosis, and colon-cancer chemoprevention. Lancet Oncol 2002;3: Labayle D, Fischer D, Vielh P, et al. Sulindac causes regression of rectal polyps in familial adenomatous polyposis. Gastroenterology 1991;101: Giardiello FM, Hamilton SR, Krush AJ, et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 1993;328: Nugent KP, Farmer KC, Spigelman AD, Williams CB, Phillips RK. Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis. Br J Surg 1993;80: Steinbach G, Lynch PM, Phillips RK, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000;342: Giardiello FM, Yang VW, Hylind LM, et al. Primary chemoprevention of familial adenomatous polyposis with sulindac. N Engl J Med 2002;346: Ladenheim J, Garcia G, Titzer D, et al. Effect of sulindac on sporadic colonic polyps. Gastroenterology 1995;108: Sandler RS, Halabi S, Baron JA, et al. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 2003;348: Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003;348: Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. Cell 1996;87: Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61: Bienz M, Clevers H. Linking colorectal cancer to Wnt signaling. Cell 2000;103: Korinek V, Barker N, Morin PJ, et al. Constitutive transcriptional activation by a beta-catenin-tcf complex in APC-/- colon carcinoma. Science 1997;275: Morin PJ, Sparks AB, Korinek V, et al. Activation of beta-catenin-tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 1997;275: Korinek V, Barker N, Moerer P, et al. Depletion of epithelial stem-cell compartments in the small intestine of mice lacking Tcf-4. Nat Genet 1998;19: Van de Wetering M, Sancho E, Verweij C, et al. The beta-catenin/tcf-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 2002;111: Shih IM, Wang TL, Traverso G, et al. Top-down morphogenesis of colorectal tumors. Proc Natl Acad Sci USA 2001;98:

113 Chapter Kim PJ, Plescia J, Clevers H, Fearon ER, Altieri DC. Survivin and molecular pathogenesis of colorectal cancer. Lancet 2003;362: Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 1993;363: Fodde R, Kuipers J, Rosenberg C, et al. Mutations in the APC tumour suppressor gene cause chromosomal instability. Nat Cell Biol 2001;3: Kaplan KB, Burds AA, Swedlow JR, Bekir SS, Sorger PK, Nathke IS. A role for the adenomatous polyposis coli protein in chromosome segregation. Nat Cell Biol 2001;3: Oshima M, Dinchuk JE, Kargman SL, et al. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87: Hanif R, Pittas A, Feng Y, et al. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin- independent pathway. Biochem Pharmacol 1996;52: Yang W, Velcich A, Mariadason J, et al. p21(waf1/cip1) is an important determinant of intestinal cell response to sulindac in vitro and in vivo. Cancer Res 2001;61: Goldberg Y, Nassif II, Pittas A, et al. The anti-proliferative effect of sulindac and sulindac sulfide on HT-29 colon cancer cells: alterations in tumor suppressor and cell cycle- regulatory proteins. Oncogene 1996;12:

114

115

116 Chapter 8 Sulindac inhibits β-catenin expression in normal appearing colon of hereditary nonpolyposis colorectal cancer and familial adenomatous polyposis patients. J.J. Koornstra 1, F.E.M. Rijcken 1, C.N.A.M. Oldenhuis 1, N. Zwart 2, T. van der Sluis 3, H. Hollema 3, E.G.E. de Vries 2, J.J. Keller 4, G.J.A. Offerhaus 4, F.M. Giardiello 5, J.H. Kleibeuker 1. University of Groningen Medical Centre, Departments of 1 Gastroenterology and Hepatology; 2 Medical Oncology; 3 Pathology, the Netherlands; 4 Academic Medical Center, Department of Pathology, Amsterdam, the Netherlands; 5 Johns Hopkins Medical Institutions, Department of Medicine, Baltimore, USA. Cancer Epidemiology Biomarkers & Prevention 2005;14:

117 Chapter 8 Abstract Background: Sulindac reduces colorectal cancer risk in genetically susceptible humans and animals. The molecular mechanisms underlying these effects are incompletely understood. Many studies suggest an important role for induction of apoptosis involving the mitochondrial pathway and the death receptor (DR) pathway. Alternatively, mechanisms involving the APC-β-catenin-Wnt pathway have been suggested, possibly mediated by p21. The effects of sulindac were determined on apoptosis and expression of DR4 and DR5, β- catenin and p21 in normal appearing colorectal epithelium. Methods: Biopsies were obtained before and after sulindac treatment during two chemoprevention studies. Patients (n = 18) with hereditary non-polyposis colorectal cancer (HNPCC) received 150 mg sulindac twice daily for 4 weeks in a placebo-controlled cross-over design. Patients (n = 6) with familial adenomatous polyposis (FAP) received 150 mg sulindac twice daily for 6 months. Apoptosis was assessed by M30 immunoreactivity and expression patterns of DR4, DR5, β-catenin and p21 were studied by immunohistochemistry. Results: In HNPCC patients, apoptotic indices were similar following placebo and sulindac. Also in FAP patients, apoptotic indices were not different after sulindac compared to pre-treatment values. Expression of DR4 and DR5 was observed in all samples with no consistent differences between placebo/baseline and sulindac. Intensity of membranous β-catenin staining was lower in HNPCC samples following sulindac compared to placebo (p < 0.001). Similar results were obtained in FAP samples (p < 0.01). p21 expression before and after sulindac treatment was similar, in both patient groups. Conclusion: Sulindac inhibits β-catenin expression in normal colorectal epithelium from HNPCC and FAP patients without affecting apoptotic indices and DR4, DR5 and p21 expression. Introduction Colorectal cancer (CRC) is the second leading cause of cancer death in the western world. Familial adenomatous polyposis (FAP) and hereditary non-polyposis colorectal cancer (HNPCC) are well-defined conditions predisposing to CRC 1. Several studies have established the chemopreventive effects of nonsteroidal anti-inflammatory drugs (NSAIDs) as sulindac 2-4 and celecoxib 5 in patients with FAP, while studies in HNPCC patients are ongoing 6. The precise mechanisms by which NSAIDs mediate their effects are incompletely understood, but likely involve induction of apoptosis 7-9. Apoptosis is controlled via two major pathways, one originating at the cell membrane, the extrinsic pathway, and one involving the mitochondria, the intrinsic pathway 10. Apoptotic pathways originating at the cell membrane involve death receptors like FAS, TNF-R1, DR3, DR4 and DR5, which are activated upon binding to their respective ligands 10. Recent reports reveal that sulindac mediates apoptosis through the mitochondrial pathway in colon cancer cells, involving caspase 9 and BAX 11. Sulindac-induced apoptosis has also been shown to involve DR5 as sulindac induced up-regulation of DR5 mrna and protein levels, but not of DR4 in vitro 12,

118 Effects of sulindac on normal colon in HNPCC and FAP The initial event in the neoplastic transformation of normal colon epithelium is assumed to be the activation of the Wnt-signalling pathway, caused by mutations in the APC or the β-catenin gene 14. This leads to cytoplasmic and subsequent nuclear accumulation of β-catenin. In the nucleus, β-catenin binds and activates the transcription factor TCF4. Finally, activated TCF4 activates a program presumed to be responsible for early adenoma formation 14. Several in vitro studies suggest that sulindac mediates its antineoplastic effect by inhibition of the Wnt-pathway This is supported by studies in APC min mice 18 and in adenomas of FAP patients 15. Recent reports reveal that sulindac affects Wnt-signalling by modifying expression of p21, a cyclin-dependent kinase inhibitor 19. To provide further insight into mechanisms involved in the chemopreventive action of sulindac, we studied the effects on apoptosis and expression of DR4, DR5, β-catenin and p21 in normal epithelium of FAP and HNPCC patients. Materials and Methods Patient selection. Recently, a chemoprevention biomarker study was performed in proven or probable HNPCC patients at the University of Groningen Medical Centre 20. Proven patients were carriers of a mutation in one of the mismatch repair genes (hmlh1, hmsh2, hmsh6). Probable HNPCC patients had a family history meeting the revised Amsterdam criteria 21 and a medical history of a HNPCC associated cancer, a colorectal adenoma at an early age (<40 years) or an adenoma with advanced neoplastic characteristics. Individuals with prior colorectal surgery were enrolled when the estimated length of the remaining colon exceeded 50 % of the original length. In this randomised double-blind placebo-controlled cross-over study, patients were assigned to receive sulindac 150 mg orally twice daily or an identically appearing placebo for 4 weeks. Both were produced by the Pharmacy Department of the University Medical Centre. After a wash out period of 4 weeks, patients crossed over to the alternative treatment for another 4 weeks. Full colonoscopy was performed at 4 and 12 weeks. Biopsies were taken of macroscopically normal mucosa with a standard biopsy forceps at four locations: ascending, transverse and sigmoid colon and rectum. Samples were formalin-fixed, embedded in paraffin and coded to disguise the subjects treatment assignment. The local medical ethical committee approved the study. Reasons for exclusion from the study were: use of an NSAID in the three months before the study, pregnancy or a history of peptic ulcer disease or gastrointestinal bleeding. For the present study, sufficient residual material was available from 18 patients (12 men, 6 women; mean age 44.6 years). Nine of these patients were proven carriers of a mutation in the mismatch repair gene hmsh2 (n = 6) or hmlh1 (n = 3). Samples from FAP patients were obtained from a study in which FAP patients had been treated with sulindac 150 mg twice daily during 9 months, as described previously 2,22. From 6 patients, tissue sections were available from normal appearing rectal mucosa before and after 6 months of treatment. These 6 patients had adenomas at baseline and showed regression of adenomas after treatment with sulindac. 117

119 Chapter 8 Immunohistochemistry for apoptosis, DR4, DR5, β-catenin and p21 For immunohistochemistry, 3 µm thick sections were cut from paraffin blocks and deparaffinised in xylene. Apoptosis was determined using the murine monoclonal antibody M30 (Boehringer Mannheim, Mannheim, Germany) directed against cleaved cytokeratin-18 that is expressed during early apoptosis 23. Staining procedures for M30, DR4 and DR5 were performed as previously described 23,24. For β-catenin staining (1:1000; clone 14, Transduction Laboratories, Lexington, KY, USA), antigen retrieval was carried out by microwave treatment for 8 min at 700 W in 0.01 M citrate buffer (ph 6.0). For p21 staining (1:50, clone WAF1, Oncogene Research), antigen retrieval was performed by heating slides three times for 5 min at 115 C with 5 min cooling in between in maleate buffer in a preheated autoclave (Presto deluxe, Presto, Eclaire, WI, USA). After blocking of endogenous peroxidase with 0.3 % hydrogen peroxide for 30 min and incubation with avidin and biotin blocking solutions (Vector Laboratories, Burlingame, CA, USA), primary antibodies were applied for 1 h at room temperature. After washing with phosphatebuffered saline, slides were incubated with appropriate secondary and tertiary antibodies. Slides were counterstained with haematoxylin. As negative controls, slides were stained in absence of the primary antibody. As positive controls, sections of normal human liver (DR4, DR5) and colorectal cancer (β-catenin, p21, M30) were included. For each antibody, slides were stained in one batch. Evaluation of staining Slides were evaluated independently by light microscopy by two investigators in a coded fashion. For M30 and p21, positive cells were expressed as percentage of the total number of cells counted (apoptotic and p21 index respectively). Only complete longitudinal crypts and at least 500 cells were counted. Intensity of DR4, DR5 and β-catenin staining was graded semi-quantitatively using a scale from 1 to 3 (1: weak staining; 2: moderate staining; 3: intense staining). For β-catenin, staining was separately recorded as membranous, cytoplasmic or nuclear. To assess changes in staining intensity as a consequence of treatment, intensities were compared in paired slides and scored as increased, decreased or unchanged. When the observers scores differed, cases were re-evaluated using a multi-headed microscope and the final grade was reached by consensus. Statistics For statistical assessment of changes in apoptotic indices, p21 indices and cumulative DR4, DR5 and β-catenin expression scores following sulindac treatment versus placebo, the Wilcoxon rank sum test for paired samples was used. Changes in distribution of staining intensities of DR4, DR5 and β-catenin were assessed using chi-square tests. To determine differences between various colonic regions in HNPCC patients, Mann-Whitney tests for continuous variables and chi-square tests for discontinuous variables were conducted. Differences between proven and probable HNPCC patients were assessed using Mann- Whitney test for continuous variables and chi-square tests for discontinuous variables. Reported p-values are two-tailed and significance was assumed if p < SPSS for Windows software (SPSS Inc, Chicago, IL) was utilised for all statistical analyses. 118

120 Effects of sulindac on normal colon in HNPCC and FAP Table 1. Mean apoptotic and p21 indices in sample pairs from normal colon mucosa following sulindac compared to placebo (HNPCC) and baseline values (FAP). AI * (%) p21 * (%) HNPCC n Placebo Sulindac n Placebo Sulindac Cumulative ± ± ± ± 0.9 Ascendens ± ± ± ± 2.4 Transverse ± ± ± ± 1.2 Sigmoid ± ± ± ± 1.2 Rectum ± ± ± ± 2.0 FAP Baseline Sulindac Baseline Sulindac Rectum ± ± ± ± 12.4 AI: apoptotic index; * expressed as mean ± SEM Results To assess changes in apoptosis, DR4, DR5, β-catenin and p21 expression following placebo and sulindac, samples were analysed pair-wise, comparing staining results in biopsies from the same patient in the same colonic region. The analysis of sample pairs was hampered by limited availability of material. In case one of a pair of samples contained insufficient material to allow evaluation, it meant that these samples were not evaluated. Not all biopsies obtained in HNPCC patients were of sufficient quality, limiting the number of sample pairs analysed to 55 for apoptotic indices, 64 for DR4, 67 for DR5, 48 for β-catenin and 63 for p21 expression. The number of sample pairs analysed from FAP patients was 6 for all staining procedures. Apoptosis When comparing cumulative AI s between placebo and sulindac treatment (in HNPCC) and between pre- and post treatment with sulindac (in FAP), no statistically significant differences were observed (Table 1, left panel). Given the predilection for the proximal colon in the development of colorectal neoplasia in HNPCC, AI s were compared in different colonic regions in HNPCC patients. For each region, AI s were not significantly different following sulindac compared to placebo although in biopsies from the proximal colon there was a trend towards lower apoptotic indices following sulindac. DR4, DR5 and β-catenin expression following placebo (HNPCC) and at baseline (FAP) In all patient samples, cytoplasmic staining of DR4 and DR5 was observed. For DR4, the immunoreactivity of epithelial cells increased gradually from the crypt base to the luminal 119

121 Chapter 8 Table 2. Distribution of DR4, DR5 and β-catenin staining intensities in sample pairs from normal colon mucosa following sulindac, compared to placebo (HNPCC) and baseline values (FAP). HNPCC FAP Staining score * Placebo Sulindac Baseline Sulindac n/n tested (%) n/n tested DR4 1 5/64 (8%) 6/64 (9%) 0/6 1/6 2 31/64 (48%) 32/64 (50%) 5/6 4/6 3 28/64 (44%) 26/64 (41%) 1/6 1/6 DR5 1 10/67 (15%) 9/67 (14%) 4/6 3/6 2 36/67 (54%) 39/67 (58%) 2/6 3/6 3 21/67 (31%) 19/67 (28%) 0/6 0/6 β-catenin 1 4/48 (8%) 9/48 (19%) 0/6 2/6 2 24/48 (50%) 33/48 (69%) 4/6 3/6 3 20/48 (42%) 6/48 (12%) 2/6 1/6 * Assessed as described in the Materials and Methods section number of samples varied in patient groups as a consequence of limited availability of slides distribution of staining intensities of β-catenin in HNPCC following placebo vs sulindac, p < distribution of staining intensities of β-catenin in FAP at baseline vs sulindac, p < 0.01 surface. DR5 immunoreactivity was seen along the entire crypt axis. β-catenin expression was membranous in all investigated samples, i.e. no cases of cytoplasmic or nuclear staining were seen. In HNPCC patients, DR4, DR5 and β-catenin staining intensities were similar in all four investigated regions of the colon. No differences were seen between proven carriers of MLH1 or MSH2 gene mutations and patients without an established mutation. Also, no differences in expression patterns were observed between MLH1 and MSH2 mutation carriers. Changes in DR4, DR5 and β-catenin expression following sulindac (HNPCC and FAP) Alterations in expression patterns of DR4, DR5 and β-catenin were analysed studying the distribution of staining intensities and changes in absolute and cumulative staining intensity scores. To assess whether changes in staining intensities were consistent in HNPCC patients, cumulative scores were calculated for each patient by adding the respective intensity scores in the samples from different parts of the colon. Cumulative scores were calculated when at least 2 sample pairs per patient were available. Table 2 summarises changes in the distribution of staining intensities of DR4, DR5 and β-catenin expression following sulindac compared to placebo (HNPCC) and baseline (FAP). For DR4 and DR5, staining scores were similarly distributed following sulindac compared to placebo in HNPCC patients and compared to baseline values in FAP patients. For β-catenin, staining scores were distributed differently following sulindac compared to placebo in HNPCC (p < 0.001) and compared to baseline values in FAP samples (p < 0.01, 120

122 Effects of sulindac on normal colon in HNPCC and FAP 10 Cumulative scores 5 0 Placebo Sulindac Figure 1. Cumulative β-catenin intensity scores in paired samples from HNPCC patients (n = 12) following placebo and sulindac. table 2), with lower scores in both patient groups after sulindac. In sample pairs, the intensity scores of DR4 staining were not consistently different following sulindac compared to placebo in HNPCC: higher in 26/64 pairs, lower in 27/64 pairs and unchanged in 11/64 pairs (not significant). For DR5, similar results were obtained: higher in 19/67, lower in 20/67 and unchanged in 28/67 pairs (not significant). The intensity scores of membranous β-catenin staining following sulindac compared to placebo was higher in 7/48, lower in 26/48 and unchanged in 15/48 pairs (p < 0.05). In paired FAP samples, DR4 and DR5 staining intensities were similar between baseline and sulindac in all 6 samples. For membranous β-catenin, staining intensities were lower following sulindac in 3/6 and unchanged in 3/6 FAP pairs (not significant). In cases with lower staining intensity of membranous β-catenin following sulindac, no apparent increase in cytoplasmic or nuclear staining was seen. With respect to cumulative staining intensity scores in HNPCC, scores were similar for DR4 and DR5 following placebo and sulindac (data not shown). However, for β-catenin, cumulative intensity scores were significantly lower following sulindac compared to placebo (p < 0.01, figure 1). p21 Mean percentages of p21 positive cells following placebo and sulindac are shown in table 1 (right panel). After placebo, p21 indices were comparable in different colon regions in HNPCC patients. p21 indices were not significantly different between HNPCC and FAP patients. Following sulindac, p21 indices were similar compared to placebo (HNPCC) and baseline (FAP) values. 121

123 Chapter 8 Discussion The efficacy of chemopreventive agents in the colorectum is routinely assessed by measuring one or more endpoints: biomarker modulation in the at-risk mucosa, adenoma regression, adenoma suppression or adenoma prevention 25. Our biomarker modulation study evaluated changes in apoptosis and expression of DR4, DR5, β-catenin and p21 occurring in normal-appearing mucosa following treatment with sulindac in HNPCC and FAP patients. Although, in general, few conclusions can be drawn from biomarker studies, they provide an opportunity to identify mechanisms of action of chemopreventive agents. In particular, quantitative measurements of apoptosis are considered a sensitive index of the biological effects of NSAIDs 6. Whereas several biomarker modulation studies are available in FAP patients, our placebo-controlled crossover study is one of only a few in HNPCC patients. We found that sulindac did not alter the apoptotic index in normal colorectal mucosa from HNPCC and FAP patients compared to placebo (HNPCC) or baseline (FAP). As anticipated from these null results, no changes were seen in expression of the death receptors DR4 and DR5. However, reduced membranous β-catenin expression patterns were observed following sulindac in both patient groups, suggesting an inhibiting effect of sulindac on the APC-β-catenin-Wnt pathway. Sulindac is one of the most extensively studied NSAIDs in the setting of chemoprevention of CRC 7. An important mechanism behind the chemopreventive effect of sulindac appears to be induction of apoptosis 7. Sulindac is a prodrug that is converted into sulindac sulfide and then sulindac sulfone by colonic bacteria 26. In vitro, both metabolites induce apoptosis in colon cancer cells 7,11,27, including mismatch repair deficient cells 28. In APC Min mice, a mouse model of FAP, sulindac had an anti-tumour effect and was associated with induction of apoptosis 29. Also in a mismatch repair deficient APC Min mouse model, carrying genetic features of both FAP and HNPCC, sulindac inhibited intestinal adenoma development 30. Whether this effect was mediated by induction of apoptosis was not studied. In normal rectal mucosa of FAP patients with adenomas, an increase or change of apoptosis has been observed following sulindac therapy 8,31. We did not find a significant effect on apoptosis in our FAP material, but this may be due to the limited number of cases. Interestingly, in presymptomatic, phenotypically unaffected FAP patients, no changes in apoptosis were seen upon sulindac treatment 32. In accordance with these data, sulindac did not have a preventive effect on the development of adenomas in these phenotypically unaffected patients 33. There is no data on efficacy of chemoprevention in HNPCC patients, although studies are ongoing 6,34. Taken together, the chemopreventive action of sulindac in FAP patients seems to be mediated by induction of apoptosis, but limited to the stage when adenomas have already developed. Recent studies have suggested that β-catenin is a target for the chemopreventive action of NSAIDs 16,35,36. In vitro, NSAIDs including sulindac, prevented nuclear accumulation of β-catenin 16,35. Oncogenic activation of the Wnt-signalling pathway resulting in nuclear translocation of β-catenin is considered critical for the initiation in intestinal epithelial neoplastic transformation 37. A recent report reveals that adenomas from FAP patients showed less nuclear β-catenin staining after sulindac treatment 15. Similar results were obtained in APC Min mice, in normal intestinal mucosa 38 as well as in adenomas 39. Our 122

124 Effects of sulindac on normal colon in HNPCC and FAP results in normal colon mucosa, with a reduction in membranous expression of β-catenin following sulindac, are consistent with these data. Whether this phenomenon is limited to subjects with a predisposition for colorectal adenoma development or also applies to the general population is unknown. Finally, we assessed whether changes in β-catenin expression were associated with alterations in p21 expression. Recent data indicated that active Wnt-signalling decreases p21 concentrations preventing cells from entering G1 arrest or differentiation, thereby allowing cells to proliferate 40. In a previous study of three patients treated with sulindac, p21 expression increased in two compared to pre-treatment values in rectal biopsy specimens 41. Our results in a larger patient group do not confirm these data. Although we recently postulated that sulindac could mediate its effect on intestinal adenoma formation by modifying p21 expression 19, the present study does not support such a mechanism. In summary, in normal colorectal mucosa from HNPCC and FAP patients, sulindac had an inhibiting effect on β-catenin expression without affecting apoptotic indices, DR4, DR5 and p21 expression. Our data provide further support for inhibition of the Wnt-signalling as a contributing mechanism of chemoprevention by sulindac. Whether this effect is universal or limited to patients genetically predisposed to colorectal cancer is unclear. 123

125 Chapter 8 References 1. Jass JR, Whitehall VL, Young J, et al. Emerging concepts in colorectal neoplasia. Gastroenterology 2002;123: Giardiello FM, Hamilton SR, Krush AJ, et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 1993;328: Nugent KP, Farmer KC, Spigelman AD,et al. Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis. Br J Surg 1993;80: Winde G, Schmid KW, Brandt B, et al. Clinical and genomic influence of sulindac on rectal mucosa in familial adenomatous polyposis. Dis Colon Rectum 1997;40: Steinbach G, Lynch PM, Phillips RK, et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000;342: Thun MJ, Henley SJ, Patrono C. Nonsteroidal anti-inflammatory drugs as anticancer agents: mechanistic, pharmacologic, and clinical issues. J Natl Cancer Inst 2002;94: Chan TA. Nonsteroidal anti-inflammatory drugs, apoptosis, and colon-cancer chemoprevention. Lancet Oncol 2002;3: Pasricha PJ, Bedi A, O Connor K, et al. The effects of sulindac on colorectal proliferation and apoptosis in familial adenomatous polyposis. Gastroenterology 1995;109: Rigas B, Shiff SJ. Nonsteroidal anti-inflammatory drugs and the induction of apoptosis in colon cells: evidence for PHS-dependent and PHS-independent mechanisms. Apoptosis 1999;4: Ashkenazi A, Dixit VM. Apoptosis control by death and decoy receptors. Curr Opin Cell Biol 1999;11: Zhang L, Yu J, Park BH, et al. Role of BAX in the apoptotic response to anticancer agents. Science 2000;290: He Q, Luo X, Huang Y, et al. Apo2L/TRAIL differentially modulates the apoptotic effects of sulindac and a COX-2 selective non-steroidal anti-inflammatory agent in Bax-deficient cells. Oncogene 2002;21: Huang Y, He Q, Hillman MJ, et al. Sulindac sulfide-induced apoptosis involves death receptor 5 and the caspase 8-dependent pathway in human colon and prostate cancer cells. Cancer Res 2001;61: Bienz M, Clevers H. Linking colorectal cancer to Wnt signaling. Cell 2000;103: Boon EM, Keller JJ, Wormhoudt TA, et al. Sulindac targets nuclear beta-catenin accumulation and Wnt signalling in adenomas of patients with familial adenomatous polyposis and in human colorectal cancer cell lines. Br J Cancer 2004;90: Gardner SH, Hawcroft G, Hull MA. Effect of nonsteroidal anti-inflammatory drugs on beta-catenin protein levels and catenin-related transcription in human colorectal cancer cells. Br J Cancer 2004;91: Rice PL, Kelloff J, Sullivan H, et al. Sulindac metabolites induce caspase- and proteasome-dependent degradation of beta-catenin protein in human colon cancer cells. Mol Cancer Ther 2003;2: Orner GA, Dashwood WM, Blum CA, et al. Suppression of tumorigenesis in the Apc(min) mouse: down-regulation of beta-catenin signaling by a combination of tea plus sulindac. Carcinogenesis 2003;24:

126 Effects of sulindac on normal colon in HNPCC and FAP 19. Huls G, Koornstra JJ, Kleibeuker JH. Non-steroidal anti-inflammatory drugs and molecular carcinogenesis of colorectal carcinomas. Lancet 2003;362: Rijcken FE, Hollema H, van der Sluis T, Boersma-van Ek W, Kleibeuker JH. Sulindac increases epithelial cell proliferative activity in the proximal colon of HNPCC patients. Eur J Gastroenterol Hepatol 2005;17: A Vasen HF, Watson P, Mecklin JP, et al. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative group on HNPCC. Gastroenterology 1999;116: Giardiello FM, Offerhaus JA, Tersmette AC, et al. Sulindac induced regression of colorectal adenomas in familial adenomatous polyposis: evaluation of predictive factors. Gut 1996;38: Koornstra JJ, Rijcken FE, De Jong S, et al. Assessment of apoptosis by M30 immunoreactivity and the correlation with morphological criteria in normal colorectal mucosa, adenomas and carcinomas. Histopathology 2004;44: Koornstra JJ, Kleibeuker JH, van Geelen CMM, et al. Expression of TRAIL (TNF-related apoptosisinducing ligand) and its receptors in normal colonic mucosa, adenomas, and carcinomas. J Pathol 2003;200: Hawk E, Lubet R, Limburg P. Chemoprevention in hereditary colorectal cancer syndromes. Cancer 1999;86(11 Suppl): Rao CV, Rivenson A, Simi B, et al. Chemoprevention of colon carcinogenesis by sulindac, a nonsteroidal anti-inflammatory agent. Cancer Res 1995;55: Shiff SJ, Qiao L, Tsai LL, et al. Sulindac sulfide, an aspirin-like compound, inhibits proliferation, causes cell cycle quies2cence, and induces apoptosis in HT-9 colon adenocarcinoma cells. J Clin Invest 1995;96: Ruschoff J, Wallinger S, Dietmaier W, et al. Aspirin suppresses the mutator phenotype associated with hereditary nonpolyposis colorectal cancer by genetic selection. Proc Natl Acad Sci U S A 1998;95: Mahmoud NN, Boolbol SK, Dannenberg AJ, et al. The sulfide metabolite of sulindac prevents tumors and restores enterocyte apoptosis in a murine model of familial adenomatous polyposis. Carcinogenesis 1998;19: Lal G, Ash C, Hay K, et al. Suppression of intestinal polyps in Msh2-deficient and non-msh2-deficient multiple intestinal neoplasia mice by a specific cyclooxygenase-2 inhibitor and by a dual cyclooxygenase- 1/2 inhibitor. Cancer Res 2001;61: Keller JJ, Offerhaus GJ, Polak M, et al. Rectal epithelial apoptosis in familial adenomatous polyposis patients treated with sulindac. Gut 1999;45: Keller JJ, Offerhaus GJ, Hylind LM, et al. Rectal epithelial apoptosis does not predict response to sulindac treatment or polyp development in presymptomatic familial adenomatous polyposis patients. Cancer Epidemiol Biomarkers Prev 2002;11: Giardiello FM, Yang VW, Hylind LM, et al. Primary chemoprevention of familial adenomatous polyposis with sulindac. N Engl J Med 2002;346: Annie Yu HJ, Lin KM, Ota DM, et al. Hereditary nonpolyposis colorectal cancer: preventive management. Cancer Treat Rev 2003;29: Hawcroft G, D Amico M, Albanese C, et al. Indomethacin induces differential expression of betacatenin, gamma-catenin and T-cell factor target genes in human colorectal cancer cells. Carcinogenesis 2002;23:

127 Chapter Smith ML, Hawcroft G, Hull MA. The effect of non-steroidal anti-inflammatory drugs on human colorectal cancer cells: evidence of different mechanisms of action. Eur J Cancer 2000;36: Giles RH, van Es JH, Clevers H. Caught up in a Wnt storm: Wnt signaling in cancer. Biochim Biophys Acta 2003;1653: Mahmoud NN, Boolbol SK, Bilinski RT, et al. APC gene mutation is associated with a dominantnegative effect upon intestinal cell migration. Cancer Res 1997;57: McEntee MF, Chiu CH, Whelan J. Relationship of beta-catenin and Bcl-2expression to sulindac-induced regression of intestinal tumors in Min mice. Carcinogenesis 1999;20: Van de Wetering M, Sancho E, Verweij C, et al. The beta-catenin/tcf-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 2002;111: Yang W, Velcich A, Mariadason J, et al. p21(waf1/cip1) is an important determinant of intestinal cell response to sulindac in vitro and in vivo. Cancer Res 2001;61:

128

129

130 Chapter 9 Summary, conclusions and future perspectives

131 Chapter 9 Summary, conclusions and future perspectives Colorectal cancer (CRC) is the second cause of cancer related deaths in the western world. Although improvements have been made in surgical techniques and the development of new chemotherapeutic modalities, 5-year survival rates are still around 60 % 1, leaving ample room for improvement. Nevertheless, progress is being made in various fields of investigation. Before summarising the studies performed in this thesis, some recent developments in colorectal cancer research are discussed below, which may help to put things in perspective. One area of research is dedicated to improve early detection of colorectal cancer or its precursor lesion, the adenoma, for example by screening the general population and surveillance of high-risk populations. The latter include subjects with the CRC predisposition syndromes Familial Adenomatous Polyposis (FAP) and Hereditary Non- Polyposis Colorectal Cancer (HNPCC), subjects with a family history of CRC, patients with previous adenomas or CRC or patients with longstanding inflammatory bowel disease. Efforts are made to develop strategies to better identify subjects at risk for HNPCC. An early diagnosis of patients with HNPCC is particularly important since periodic colonoscopic screening has been shown to improve mortality in HNPCC patients. With respect to current screening procedures, several options are available, including faecal occult blood testing, flexible sigmoidoscopy and colonoscopy, each of which has its advantages and disadvantages. Potential future screening modalities include virtual colonoscopy, serum proteomics and the detection of molecular abnormalities in stool samples 1. A related area of study is further improvement in endoscopic detection techniques of early neoplasia, such as high-magnification chromoscopic colonoscopy. Next, the potential of chemopreventive measures is increasingly being recognised, again especially in high-risk populations. Chemoprevention is the use of pharmacological or natural agents to prevent the development of cancer or precancerous lesions. Although there is evidence that certain non-steroidal anti-inflammatory drugs (NSAIDs), calcium and ursodeoxycholic acid reduce the risk of adenoma or carcinoma development, the effects are incomplete and there is a need for regimens with better efficacy. A promising strategy lies in combining different chemopreventive agents. Such an approach in the animal model of FAP, using the combination of a NSAID with an epidermal growth factor receptor (EGFR) kinase inhibitor, showed encouraging results 2. The efficacy of similar combination regimens, combining NSAIDs with other anticancer drugs, needs further exploration. Further, efforts are made to increase the efficacy of chemotherapy. Chemotherapy in CRC is used in the adjuvant setting in patients with stage III colon tumours and subject of study in stage II/III rectal cancer. In the palliative setting, chemotherapy is given in patients with stage IV colorectal cancer. The benefit of adjuvant chemotherapy in stage II colon cancer patients is controversial, but there appears to be a high-risk subpopulation that may benefit. Attempts are being made to identify genetic or protein markers in stage II colon cancer patients that predict recurrence 3 and in stage II/III patients to predict sensitivity to 130

132 Summary chemotherapy 4. It is expected that newly developed genomic and proteomic approaches may provide more sensitive markers or indicators of response than current biochemical, immunohistochemical or morphological methods. Built on an improved understanding of the molecular biology of the disease, advancement in the treatment of CRC is also being made by the identification of several molecular targets in colon tumours, potentially amenable for directed specific therapies. It can be expected that such a strategy results in direct interference with the cellular pathway involved, with less toxicity compared to the toxicity associated with the less specific effect of chemotherapy. Recently, the first of such compounds have been introduced into clinical practice, in particular the management of advanced CRC, namely antibodies directed against EGFR and the vascular endothelial growth factor respectively 5,6. Increasing knowledge of the molecular events characterising colorectal carcinogenesis will allow the development of novel therapies aiming at specific molecular targets. It is known that the development of CRC is characterised by a sequence of events during which normal colonic epithelium gradually transforms to carcinoma, in most cases via the development of colorectal adenomas, the so-called adenoma-carcinoma sequence. The adenoma-carcinoma sequence is driven by an accumulation of molecular (epi)genetic alterations causing progressive disorders in cell growth, differentiation and apoptosis. Apoptosis, or programmed cell death, plays an important role in the development and maintenance of tissue homeostasis. Multicellular organisms also use apoptosis to eliminate potentially dangerous cells, such as genetically damaged cells, including tumour cells. Abnormalities in apoptotic function have been identified as contributing events in the pathogenesis of colorectal cancer. Furthermore, resistance to apoptosis induction by chemotherapeutic drugs or radiotherapy frequently hampers its efficacy in the treatment of established tumours. During apoptosis, a complex death program is initiated that ultimately leads to the fragmentation of the cell. The death program can be induced by two major apoptosis signalling pathways namely the extrinsic and the intrinsic pathway. The extrinsic pathway is initiated by triggering cell death receptors on the cell surface, leading to activation of the intracellular apoptotic machinery. The intrinsic pathway of apoptosis is initiated via the mitochondria by cellular stress, for example following DNA damage caused by chemotherapeutic drugs and radiation. Mitochondria induce apoptosis by releasing cytochrome-c into the cytosol, which causes the assemby of a multiprotein caspase-activating complex, known as the apoptosome. Among the regulators of the intrinsic pathway are Bcl- 2 protein family members and inhibitors of apoptosis proteins (IAPs). The elucidation of the molecular mechanisms regulating these processes is of primary interest as this may guide specific targeted therapies. Tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is a transmembrane protein belonging to the TNF superfamily. TRAIL triggers the extrinsic apoptotic pathway by binding to its membrane-bound death receptors DR4 and DR5, which transmit an apoptotic signal via their intracellular death domains. TRAIL can also bind to decoy receptors DcR1 and DcR2 that are unable to initiate an apoptotic signal. Although 131

133 Chapter 9 the physiological role of TRAIL is not exacty known, there are data suggesting a role in immunomodulation. In vitro, TRAIL induces apoptosis in a wide variety of cancer cell lines but not in normal cells. Preclinical data in mice and nonhuman primates have shown that recombinant TRAIL induces tumour regression without serious systemic toxicity. Therefore, TRAIL is considered a promising new anti-cancer agent. The aims of this thesis were to clarify the changes occurring in the degree of apoptosis during the gradual development of colorectal cancer. Furthermore, the potential of TRAIL as a therapeutic agent was investigated in colorectal adenomas and carcinomas. Tumours were studied from patients with a sporadic background as well as from FAP and HNPCC patients. Finally, mechanisms behind the chemopreventive action of the NSAID sulindac on colorectal epithelial cells were studied, as recent data suggested a mechanism involving TRAIL mediated apoptosis. Following a brief introduction and outline of this thesis in Chapter 1, current knowledge of mechanisms of apoptosis and changes in the degree of apoptosis in the development of colorectal cancer was reviewed in Chapter 2. Although there is a multitude of studies available on apoptosis in colon tissue at different stages of colorectal cancer development, results from these studies are contradictory. Considerable controversy exists as to whether the frequency of apoptosis increases or decreases during the adenoma-carcinoma sequence. Therefore, a systematic review of all available studies, which addressed changes in frequency, and distribution of apoptosis in colon epithelium was performed. Studies were included if either normal or neoplastic colonic epithelium, or both, was examined. A PUBMED search revealed 53 articles. Several aspects from these articles were studied such as methods used to determine apoptotic cell death, frequency of apoptosis and locoregional distribution of apoptotic cells in normal mucosa, adenomas and carcinomas, correlations between the degree of apoptosis and proliferation and the prognostic significance of the degree of apoptosis in colorectal cancer. There was a wide variety in the methods used to determine apoptotic cells in the reviewed studies. Most studies had used terminal deoxyribonucleotidyl transferase-mediated nick end labelling (TUNEL, n = 29). Other methods used were in situ end labelling (ISEL, n = 7), morphology by light microscopy (n = 7), TUNEL or ISEL in conjunction with morphology by light microscopy (n = 9), ISEL in conjunction with morphology by electron microscopy (n = 1), morphology by light microscopy in conjunction with electron microscopy (n = 2) and the recently developed method of detection of cleaved cytokeratin 18 by immunohistochemistry (n = 2). Caspase-cleaved cytokeratin 18 is an early marker of apoptosis that can be recognised in epithelial cells using an antibody (M30) directed against this neo-epitope. Percentages of apoptotic cells (apoptotic indices) reported with these methods varied widely. This variation is probably partly due to a major limitation of the TUNEL and ISEL methods, i.e. false positive marking of necrotic cells in addition to apoptotic cells. Others have already seriously questioned the applicability of the TUNEL method to evaluate apoptotic cell death in the gastrointestinal tract. Although recently new methods of identifying apoptotic cells with higher specificity have been developed, such as M30 immunoreactivity, few studies have used these techniques. 132

134 Summary Concerning the degree of apoptosis in normal mucosa, reported apoptotic indices varied between 0.11 and 11 %. It was noted that in most studies sections of normal colonic mucosa were in fact retrieved from macroscopic normal tissue adjacent to adenomas or carcinomas and one could argue whether this represents truly normal epithelium. One study revealed that apoptosis in normal mucosa obtained from resection margins indeed differed from genuinely normal mucosa, which warrants caution when interpreting studies on apoptosis in normal colonic mucosa. When studying the transition from normal colonic mucosa to adenomas, most studies indicated a shift towards higher apoptotic indices. The real controversy centres on the transition from adenoma to carcinoma. Most studies reported that the frequency of apoptotic cell death in carcinomas was higher than in adenomas. However, in other studies, no differences or even opposite results were found. Concerning the distribution of apoptotic cells during the adenoma-carcinoma sequence, there was general agreement that in adenomas apoptosis predominates at the base of the crypt, with proliferative activity prevailing at the luminal surface. This is a complete reversal of the pattern seen in normal epithelium. With respect to the correlation between rates of apoptosis and proliferation in adenomas and carcinomas, again conflicting results were obtained. Some studies indicated a correlation between the two, whereas others did not. Regarding the prognostic value of apoptotic indices it was found that most studies suggest that colorectal tumours with high apoptotic indices were associated with better prognosis and survival. Given the limitations mentioned earlier of the techniques most often used for the detection of apoptotic cells, TUNEL and ISEL, a recently developed technique to investigate apoptosis in colorectal tissue was studied in Chapter 3. The method used, assessing M30 immunoreactivity, is based on recognition in epithelial cells of caspase-cleaved cytokeratin 18 which is an early and specific marker of apoptosis. The technique was validated against the gold standard of identification of apoptotic cells. i.e. by morphological criteria. Following the observation from Chapter 2 that most studies on apoptosis in normal colorectal epithelium were in fact performed in macroscopic normal tissue adjacent to adenomas or carcinomas, it was investigated whether apoptotic indices differ between normal mucosa obtained from resection margins and genuinely normal mucosa. Paraffin sections of normal colonic mucosa (n = 30), normal mucosa obtained from resection margins from carcinomas (n=30), colorectal adenomas (n = 84) and carcinomas (n = 40) were studied. Apoptosis of epithelial cells was assessed by M30 immunoreactivity as well as by morphological criteria and expressed as a proportion of the total number of cells counted (apoptotic index). Mean apoptotic indices assessed by M30 immunoreactivity were 0.18 ± 0.04 % in normal mucosa, 0.42 ± 0.04 % in adenomas and 1.97 ± 0.24 % in carcinomas. Apoptotic indices determined by morphological criteria were 0.23 ± 0.03 %, 0.62 ± 0.06 % and 1.78 ± 0.19 %, respectively. Apoptotic counts were higher in normal mucosa obtained from resection margins than in genuinely normal mucosa (p < ). There was a good correlation between apoptotic indices obtained by M30 immunoreactivity and morphological criteria (r = 0.71, p < 0.01). These findings support the use of the M30 method in the study of apoptosis in colorectal tissues. 133

135 Chapter 9 To investigate the potential use of TRAIL or TRAIL receptor agonists as therapeutic agents in colorectal neoplasms, the expression of the four membrane-bound TRAIL receptors DR4, DR5, DcR1 and DcR2 was studied in colon tissue sections of different stages of colon cancer development, i.e. normal colon, adenomas and carcinomas. The expression and localisation of TRAIL and its receptors were investigated by immunohistochemistry in normal mucosa (n = 10), adenomas (n = 19) and carcinomas (n = 21). Tissues were obtained from patients with a sporadic background of their disease. In addition, correlations between expression of TRAIL and its receptors and the degree of apoptosis (assessed by M30 immunoreactivity) and histopathological characteristics were explored. The results are described in Chapter 4. In normal mucosal epithelial cells, TRAIL and TRAIL receptor expression was observed. In adenomas and carcinomas, expression of all four receptors was seen. However, TRAIL expression was lost in a subset of colorectal tumours in comparison to adjacent normal mucosa. This occurred more frequently in carcinomas than in adenomas (p < 0.05). Intensity of DR4 and DR5 staining was stronger in neoplastic cells compared to adjacent normal epithelial cells, and was accompanied by a higher degree of apoptosis. No differences were found between neoplastic and normal cells regarding DcR1 and DcR2 expression. No correlations were found between TRAIL or TRAIL receptor expression and histopathological characteristics. The stronger expression of DR4 and DR5 in neoplastic cells as opposed to normal cells holds promise for the future use of TRAIL or TRAIL receptor agonists as therapeutic agents in colon neoplasms. Given these encouraging results in patients with sporadic disease, the investigations were expanded to tissues obtained from patients with FAP and HNPCC, described in Chapter 5. Expression of DR4 and DR5 was studied using immunohistochemistry in colorectal adenomas and carcinomas from patients with sporadic disease (n = 74 and 56 respectively), FAP (n = 41 and 4 respectively) and HNPCC (n = 50 and 21 respectively). In addition, sporadic tumours with microsatellite instability (MSI) were studied, to examine the role of BAX gene mutations. BAX was reported to play a role in sensitivity to TRAIL-mediated apoptosis in vitro and colorectal tumours with MSI are prone for mutations in the BAX gene. The presence of BAX mutations in MSI positive tumours could potentially limit the use of TRAIL or TRAIL receptor agonists. Carcinomas with high frequency MSI (MSI-H, n = 42, of which 27 sporadic and 15 HNPCC-associated) were analysed for apoptotic cell death, assessed by M30 immunoreactivity, and BAX mutations. Immunohistochemistry revealed that most adenomas from all three patient groups expressed DR4 and DR5. Most carcinomas expressed DR4, except for six cases, all with mucinous histology. All carcinomas, including those with mucinous histology, showed DR5 expression. BAX mutations were found in 6 out of 42 MSI-H cancers, with similar apoptotic indices and expression of DR4, DR5 and TRAIL in BAX mutant and BAX wild-type cases. These data do not support the presence of BAX inactivation as a critical mechanism in vivo to evade apoptosis in MSI-H tumours. Taken together, these results indicate that selectively targeting TRAIL death receptors in sporadic and hereditary settings has potential as prevention or treatment strategy. 134

136 Summary It is important to identify prognostic factors that can help to develop new patient-tailored treatment strategies for CRC patients. Resistance of tumour cells to induction of apoptosis is a limiting factor in the efficacy of chemotherapeutic treatment of colon cancer. In vitro, chemotherapy and recombinant human (rh) TRAIL were shown to act synergistically on induction of apoptosis in colon cancer cells. The prognostic significance of expression of TRAIL, DR4 and DR5 on overall and disease free survival and disease recurrence was studied and described in Chapter 6. Tissue microarrays were constructed containing tumour tissue of 376 stage III colon cancer patients who had participated in an adjuvant chemotherapy study in the past 7. These arrays were analysed by immunohistochemistry for TRAIL, DR4 and DR5 expression. Median follow-up was 43 months. Kaplan-Meier survival analysis and Cox proportional hazard analysis with adjustment for treatment arm, age and gender was performed. Most colon cancers showed high expression of TRAIL (82.8 % of cases), DR4 (91.7 %) and DR5 (87.1 %). High DR4 expression was associated with worse overall survival {odds ratio (OR; 95 % CI) = 2.22 ( ); p = 0.04} and worse disease free survival {OR = 2.19 ( ); p = 0.03}, compared to low DR4 expression. High DR5 expression was associated with better survival with borderline significance {OR = 0.65 ( ); p = 0.07}. Within the group of patients with recurrent disease, the time to recurrence was longer for those with high DR5 expression (p = 0.03) or high TRAIL expression (p = 0.007) compared to those with low DR5 or TRAIL expression levels. The finding that most tumours showed high DR4 and DR5 expression indicates that rhtrail may potentially be used as an adjunct to adjuvant chemotherapeutic treatment of stage III colon cancer patients. Recently, two randomised placebo-controlled trials were published showing a chemopreventive effect of aspirin in patients with previous colorectal adenomas or carcinomas 8,9. These studies were discussed in Chapter 7, and mechanisms behind the chemopreventive action against colorectal cancer development of NSAIDs were briefly reviewed. Both studies indicated a moderate chemopreventive effect of aspirin in these populations. The anticancer properties of NSAIDs have been demonstrated in vitro as well as in vivo (animal studies, epidemiological reports, and intervention studies). Several mechanisms through which NSAIDs alter colorectal carcinogenesis have been elucidated, including the induction of apoptosis in neoplastic cells, via mechanisms dependent and independent of cyclo-oxygenase. The exact mechanisms involved in apoptosis induction are unknown but recent evidence suggests that sulindac-induced apoptosis involves DR5. There is also evidence suggesting that NSAIDs work by modulating the Wnt-pathway. The formation of adenomas in the intestine is thought to be initiated by mutational events occurring in either APC or the gene for β-catenin, both operating in the Wnt-signalling pathway. In normal cells, β-catenin is membrane bound, with cytoplasmic β-catenin being degraded by the APC/axin/GSK3β-complex. Loss of functional APC or mutations in the β-catenin gene lead to the nuclear accumulation of β-catenin, which binds and activates the transcription factor T-cell factor 4 (TCF4). Consequently, activated TCF4 activates a genetic programme that is presumed to be responsible for early adenoma formation. The importance of the Wnt-pathway in which β-catenin/tcf4 activity acts as a switch controlling 135

137 Chapter 9 proliferation versus differentiation in the intestinal epithelium seems critical in the early development of colorectal neoplasia. It has been shown that activated Wnt-signalling decreases concentrations of the cell cycle regulating protein p21, thereby allowing cells to proliferate. Recent evidence suggests that part of the chemopreventive effect of sulindac is mediated via p21. For example, microarray data of rectal biopsy specimens and colonic cells in culture showed that sulindac induced expression of p21. Furthermore, homozygous inactivation of p21 in APC min mice (the mouse model of FAP) eliminated the ability of sulindac to reduce the number of small intestinal tumours in these mice. So, p21 could be the link between the Wnt pathway and the chemopreventive effect of sulindac on intestinal adenoma formation. These data served as the background for studying the effects of sulindac on apoptosis and expression of DR4 and DR5, β-catenin and p21 in normal appearing colon mucosa. The results are described in Chapter 8. Biopsies obtained before and after sulindac treatment during two chemoprevention studies 10,11 were collected. Patients (n = 18) with HNPCC had received 150 mg sulindac twice daily for 4 weeks in a placebo-controlled cross-over design. Patients (n = 6) with FAP had received 150 mg sulindac twice daily for 6 months. Apoptosis was assessed by M30 immunoreactivity and expression patterns of DR4, DR5, β-catenin and p21 were studied by immunohistochemistry. In HNPCC patients, apoptotic indices were similar following placebo and sulindac. Also in FAP patients, apoptotic indices were not different after sulindac compared to pretreatment values. Expression of DR4 and DR5 was observed in all samples with no consistent differences following sulindac versus placebo/baseline. Intensity of membranous β-catenin staining was lower in HNPCC samples following sulindac compared to placebo (p < 0.001). Similar results were obtained in FAP samples (p < 0.01). p21 expression before and after sulindac treatment was similar, in both patient groups. In conclusion, sulindac inhibits β-catenin expression in normal colorectal epithelium from HNPCC and FAP patients without affecting apoptotic indices and DR4, DR5 and p21 expression. An inhibiting effect of sulindac on β-catenin expression has also been observed in APC Min mice as well as in human adenomas. Whether this phenomenon is important in the chemopreventive effects of sulindac remains to be investigated. Conclusions and future perspectives Several genes that regulate apoptosis are inappropriately expressed or mutated in colorectal cancer. Although the idea of decreased apoptosis during colorectal cancer development is conceptually attractive, the available evidence from in vivo material suggests that the proportion of apoptotic cells increases gradually during the adenoma-carcinoma sequence. In vitro, tumour progression is usually associated with resistance to apoptosis induction. Although these data seem contradictory at first sight, and potentially confusing, they are not conflicting. The contradiction resolves when one realises that in vivo studies describe spontaneous apoptosis whereas in vitro studies describe induced apoptosis, whether by chemotherapy or other agents. Considering the limitations of the current methods of in 136

138 Summary situ end labelling of DNA fragments, used in the vast majority of in vivo studies, improved diagnostic criteria for the identification of apoptotic cells are needed and can be expected in the near future. The results from this thesis using the method of M30 immunoreactivity support its use in the study of apoptosis in colorectal tissues. Successful nonsurgical elimination of cancer cells ultimately involves apoptosis. Essentially all cytotoxic drugs work by induction of apoptosis, usually not only of malignant cells but also of normal cells, manifested by toxicity. Over the last years, knowledge has increased of mediators of apoptosis in colon cells, which are dysregulated in tumour cells, such as Bcl-2 family members and IAPs 12. These suppressors of apoptosis have become targets for drug development aimed at restoring apoptosis sensitivity of tumour cells. Alternatively, one can imagine strategies aimed at direct activation of agonists of apoptosis, via the extrinsic pathway. Unlike its family members TNF and FasL, TRAIL is not associated with induction of sepsis and hepatotoxicity respectively. Based on preclinical toxicity and activity data and the results described in this thesis showing broad expression of the pro-apoptotic TRAIL receptors in colorectal neoplasms, recombinant human (rh) TRAIL is considered of great interest for clinical use. Agonistic TRAIL receptor antibodies, directed towards the death receptors DR4 and DR5 have shown to exert the same effects as rhtrail 13,14. The first phase- I trials using rhtrail and agonistic TRAIL receptor antibodies are underway and results from these studies will indicate the value of these agents as anticancer therapeutics. Furthermore, combination regimens including rhtrail should be explored. Findings from several reports indicate that the combination of rhtrail with certain chemotherapeutic drugs, NSAIDs, proteasome inhibitors and interferon-γ all induce apoptosis in an additive or synergistic fashion 15. Another intriguing approach is the development of bi-specific antibodies. Bi-specific antibodies combine immune cell activation with tumour cell recognition, as a result of which pre-defined effector cells kill tumour cells. Recently, a technique that combines EGFR-signalling inhibition with target cell-restricted apoptosis induction using a TRAIL fusion protein with engineered specificity for EGFR showed promising results 16. Analogously, a TRAIL fusion protein with specificity for the carcinoma-associated antigen EGP2 has been developed 17. Such combination strategies potentially reduce the dose of TRAIL required for anti-tumour activity and should be further explored. Apart from the possible future application of TRAIL in colorectal cancer, there may also be a potential role in treatment or prevention of adenomas. Although adenomas are generally removed during endoscopy, there are situations in which this cannot be accomplished. For example, large, flat sessile adenomas sometimes cannot be removed for technical reasons. Another example is the vast number of adenomas that is usually encountered in FAP patients, which precludes endoscopic removal. Recent data show that TRAIL also has apoptosis-inducing effects on colon adenoma cells 18. This has been confirmed in a study in which adenoma cell lines and ex vivo cultures of human adenomas were examined 19. The finding in Chapters 4 and 5 of this thesis of expression of TRAIL receptors DR4 and DR5 in colorectal adenomas supports further investigation of rhtrail or other TRAIL receptor agonistic agents in the treatment of adenomas. 137

139 Chapter 9 In conclusion, new treatment options for CRC have been developed based on insights into mechanisms involved in the regulation of apoptosis. The potential clinical application of rhtrail or other TRAIL receptor agonistic agents in the treatment of colorectal adenomas and carcinomas, not only in sporadic cases but even more so in the setting of hereditary predisposition syndromes, should be further explored. Results from the first phase-i studies with rhtrail and TRAIL receptor agonistic antibodies are eagerly awaited and should guide future therapeutic strategies. The key to further improvement of prevention and therapeutic strategies for colorectal neoplasms is probably not to search for the Holy Grail: one ideal anticancer agent. Rather, it is the hunt for defining the optimal combination of agents, targeting different intracellular pathways that will ultimately result in killing neoplastic cells. 138

140 Summary References 1. Weitz J, Koch M, Debus J, Höhler T, Galle PR, Büchler MW. Colorectal cancer. Lancet 2005;365: Torrance CJ, Jackson PE, Montgomery E, et al. Combinatorial chemoprevention of intestinal neoplasia. Nat Med 2000;6: Wang Y, Jatkoe T, Zhang Y, et al. Gene expression profiles and molecular markers to predict recurrence of Dukes B colon cancer. J Clin Oncol 2004;22: Ribic CM, Sargent DJ, Moore MJ, et al. Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 2003;349: Cunningham D, Humblet Y, Siena S, et al. Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 2004;351: Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350: Bleeker WA, Mulder NH, Hermans J, Otter R, Plukker JT. The addition of low-dose leucovorin to the combination of 5-fluorouracil-levamisole does not improve survival in the adjuvant treatment of Dukes C colon cancer. IKN Colon Trial Group. Ann Oncol 2000;11: Sandler RS, Halabi S, Baron JA, et al. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 2003;348: Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003; 348: Rijcken FE, Hollema H, van der Sluis T, Boersma-van Ek W, Kleibeuker JH. Sulindac increases epithelial cell proliferative activity in the proximal colon of HNPCC patients. Eur J Gastroenterol Hepatol 2005;17: A Giardiello FM, Hamilton SR, Krush AJ et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 1993;328: Reed JC. Apoptosis-targeted therapies of cancer. Cancer Cell 2003;3: Pukac L, Kanakaraj, Humphreys R, et al. HGS-ETR1, a fully human TRAIL-receptor 1 monoclonal antibody, induces cell death in multiple tumour types in vitro and in vivo. Br J Cancer 2005;92: Motoki K, Mori E, Matsumoto A, et al. Enhanced apoptosis and tumor regression induced by a direct agonist antibody to tumor necrosis factor-related apoptosis-inducing ligand receptor 2. Clin Cancer Res 2005;11: Van Geelen CM, de Vries EG, de Jong S. Lessons from TRAIL-resistance mechanisms in colorectal cancer cells: paving the road to patient-tailored therapy. Drug Resist Updat 2004;7: Bremer E, Samplonius DF, van Genne L, et al. Simultaneous inhibition of epidermal growth factor receptor (EGFR) signaling and enhanced activation of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptor-mediated apoptosis induction by an scfv:strail fusion protein with specificity for human EGFR. J Biol Chem 2005;280: Bremer E, Kuijlen J, Samplonius D, Walczak H, de Leij L, Helfrich W. Target cell-restricted and enhanced apoptosis induction by a scfv:strail fusion protein with specificity for the pancarcinomaassociated antigen EGP2. Int J Cancer 2004;109: Hague A, Hicks DJ, Hasan F, Smartt H, Cohen GM, Paraskeva C, MacFarlane M. Increased sensitivity to TRAIL-induced apoptosis occurs during the adenoma to carcinoma transition of colorectal carcinogenesis. Br J Cancer 2005;92: Jalving M, Koornstra JJ, de Jong S, et al. TRAIL induces apoptosis in colorectal adenoma cell lines and ex vivo adenoma cultures. Gastroenterology 2005;128(Suppl2):A

141

142 Chapter 10 Samenvatting, conclusies en toekomstperspectieven

143 Chapter 10 Samenvatting, conclusies en toekomstperspectieven. Dikkedarmkanker of colorectaal carcinoom (CRC) is de tweede meest voorkomende oorzaak van sterfte aan kanker in de westerse wereld. Hoewel de afgelopen jaren vooruitgang is geboekt met betrekking tot chirurgische technieken en de ontwikkeling van nieuwe chemotherapeutische behandelingen is de 5-jaars overleving nog steeds rond de 60 % 1. Alvorens de studies die in het kader van dit proefschrift zijn verricht samen te vatten zullen enkele nieuwe ontwikkelingen binnen het onderzoek naar oorzaken, diagnostiek en behandeling van CRC worden besproken. Veel onderzoek houdt zich bezig met het verbeteren van vroege opsporing van CRC of het voorstadium ervan, de adenomateuze poliep, bijvoorbeeld door screening van de bevolking en door periodiek endoscopisch onderzoek van mensen met een verhoogd risico op de ontwikkeling van CRC. Hiertoe behoren personen met een bepaalde genetische predispositie zoals familiaire adenomateuze polyposis (FAP) en hereditair non-polyposis colorectaal carcinoom (HNPCC), personen met een belaste familieanamnese voor CRC, patiënten met eerdere adenomen of CRC of patiënten met langer bestaand inflammatoir darmlijden. De laatste tijd is veel onderzoek gedaan ter verbetering van het identificeren van personen die behoren tot HNPCC families aangezien periodieke endoscopische controles bij deze populatie de overlevingscijfers blijken te verbeteren. Op dit moment zijn verschillende screeningstechnieken beschikbaar, te weten het testen van feces op occult bloed, sigmoïdoscopie en colonoscopie, waarbij elk onderzoek zijn voor- en nadelen kent. Screeningstechnieken zullen mogelijk in de nabije toekomst worden uitgebreid met methoden als CT- of MR-colografie (zogenaamde virtuele colonoscopie), detectie van eiwitten in serum ( proteomics ) en de opsporing van moleculair genetische afwijkingen in ontlasting van patiënten met CRC 1. Daarnaast is men bezig om endoscopische technieken te verbeteren, met name om maligne of premaligne weefsel beter te herkennen, zoals door het gebruik van bepaalde kleurstoffen of endoscopen met de mogelijkheid om beelden sterk te vergroten. Daarnaast wordt het potentieel belang van chemopreventieve maatregelen toenemend erkend, met name voor populaties met een hoog risico op CRC. Chemopreventie is het gebruik van geneesmiddelen of natuurlijk voorkomende stoffen die de ontwikkeling van CRC of adenomateuze poliepen remmen. Hoewel van een aantal middelen een beschermend effect is aangetoond, zoals bepaalde non-steroidale anti-inflammatoire drugs (NSAIDs), calcium en ursodeoxycholzuur, is het effect niet compleet en is er behoefte aan meer effectieve middelen. Een veelbelovende aanpak bestaat uit het combineren van verschillende chemopreventieve middelen. Een dergelijke strategie in het diermodel voor FAP, waarbij de combinatie van een NSAID met een remmer van het zogenaamde epidermal growth factor receptor (EGFR) kinase werd gegeven, leverde veelbelovende resultaten op 2. De werkzaamheid van dergelijke combinaties verdient zeker verder onderzocht te worden. Een andere gebied van ontwikkeling betreft het verbeteren van de effectiviteit van 142

144 Samenvatting chemotherapie. Chemotherapie bij CRC wordt gebruikt in de adjuvante setting bij patiënten met stadium III colontumoren en zijn onderwerp van studie bij stadium II/ III rectumtumoren. In de palliatieve setting wordt chemotherapie gegeven bij patiënten met stadium IV CRC. Of adjuvante chemotherapie ook van waarde is bij patiënten met stadium II colonkanker is controversieel, maar er lijkt een zekere subpopulatie te bestaan die er voordeel van kan hebben. Pogingen worden ondernomen om genetische of eiwitmerkers te identificeren in patiënten met stadium II colonkanker die terugkeer van de ziekte voorspellen 3 en in patiënten met stadium II/III CRC om de gevoeligheid voor chemotherapie te voorspellen 4. De toegenomen kennis van de moleculaire biologie van CRC heeft ook geleid tot het herkennen van moleculaire aangrijpingspunten in tumoren van CRC patiënten, die specifiek geattaqueerd kunnen worden. De verwachting is dat een dergelijke aanpak resulteert in een directe verstoring van het groeiproces van kankercellen, waarbij gezonde cellen relatief ongemoeid worden gelaten, in tegenstelling tot de situatie bij de huidige cytostatica. Kortgeleden zijn de eerste van dergelijke middelen geïntroduceerd in de klinische praktijk. Het betrof in dit geval antilichamen gericht tegen EGFR en VEGF (vascular endothelial growth factor) 5,6. Over het algemeen geldt dat een toenemend kennisniveau van de moleculaire achtergrond van CRC nieuwe aangrijpingspunten zal opleveren voor de behandeling ervan. De ontwikkeling van CRC wordt gekenmerkt door een aantal opeenvolgende gebeurtenissen waarbij normaal colonepitheel geleidelijk verandert naar kanker, in de meeste gevallen via het tussenstadium van een adenomateuze poliep. Dit wordt de zogenaamde adenoom-carcinoom sequentie genoemd. Deze sequentie wordt gedreven door een opeenstapeling van moleculaire veranderingen met als gevolg progressieve verstoringen in normale celgroei, celdifferentiatie en apoptose. Apoptose, ook wel bekend als geprogrammeerde celdood, speelt een belangrijke rol in de ontwikkeling van weefsels en in het onderhouden van een zeker evenwicht in die weefsels. Apoptose is ook een mechanisme om potentieel bedreigende cellen op te ruimen, bijvoorbeeld cellen die genetische schade hebben opgelopen, zoals tumorcellen. Verstoringen in de normale functie van apoptotische mechanismen zijn een bijdragende factor in de ontwikkeling van CRC. Daarnaast is het zo dat resistentie tegen apoptose nogal eens de effectiviteit van chemotherapie of radiotherapie belemmert. Tijdens apoptose wordt een complex programma geïnitieerd dat uiteindelijk leidt tot de dood van de cel. Dit programma kan grofweg op twee verschillende manieren worden aangeschakeld namelijk via de extrinsieke route en de intrinsieke route. De extrinsieke route wordt geactiveerd door het aanzetten van zogenaamde death receptoren op het celoppervlak, leidend tot activatie van de intracellulaire apoptotische machinerie. De intrinsieke route wordt geactiveerd via mitochondrieën door cellulaire stress, bijvoorbeeld na DNA schade veroorzaakt door chemotherapie of radiotherapie. De intrinsieke route wordt onder andere gereguleerd door eiwitten van de Bcl-2 familie en eiwitten genaamd inhibitors of apoptosis (IAPs). De verdere ontrafeling van de moleculaire mechanismen die een rol spelen bij de regulatie van deze processen is van belang omdat kennis hierover ongetwijfeld aangrijpingspunten zal bieden voor gerichte therapieën. 143

145 Chapter 10 Tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) is een eiwit dat behoort tot de TNF superfamilie. TRAIL initieert de extrinsieke apoptotische route na binding aan de membraangebonden receptoren DR4 en DR5. TRAIL kan ook binden aan de zogenaamde decoy receptoren DcR1 en DcR2 die geen apoptotisch signaal doorgeven. De fysiologische betekenis van TRAIL is nog niet goed bekend. Er zijn gegevens die een rol in immuunmodulatie suggereren. In vitro induceert TRAIL apoptose in kankercellijnen van diverse origine maar niet in normale cellen. Dierexperimenteel onderzoek in muizen en apen gaf aan dat recombinant TRAIL sterke tumorreducerende eigenschappen heeft en niet gepaard gaat met veel systemische toxiciteit. Om deze redenen wordt TRAIL beschouwd als een veelbelovend nieuw middel in de strijd tegen kanker. Het doel van dit proefschrift was om meer duidelijkheid te verkrijgen omtrent de veranderingen die optreden met betrekking tot apoptose in de loop van de adenoomcarcinoom sequentie. Daarnaast werd de potentiële toepassing van TRAIL onderzocht bij colorectale adenomen en carcinomen. Hiervoor werden colorectale tumoren onderzocht van patiënten met een sporadische (niet erfelijke) ziekteachtergrond en van FAP en HNPCC patiënten. Tenslotte werden mechanismen achter de chemopreventieve werking van het NSAID sulindac op colorectale epitheliale cellen bestudeerd, aangezien recente gegevens hierbij een rol voor TRAIL gemedieerde apoptose suggereren. Na een korte inleiding in Hoofdstuk 1 wordt in Hoofdstuk 2 een overzicht gegeven van de huidige kennis van de mechanismen van apoptose en veranderingen die optreden in de ontwikkeling van colorectaal carcinoom. Hoewel er tientallen onderzoeken beschikbaar zijn waarin apoptose in colonslijmvlies wordt bestudeerd in verschillende stadia van de ontwikkeling van CRC zijn de resultaten hiervan tegenstrijdig. Zo bestaat er een aanzienlijke controverse over de vraag of de mate van apoptose toe- of afneemt in de loop van de adenoom-carcinoom sequentie. Dit was reden om een systematisch onderzoek te doen van de beschikbare studies. Hierbij werden alle onderzoeken die veranderingen in de mate en distributie van apoptose in normaal of neoplastisch colonslijmvlies hadden bestudeerd onder de loep genomen. Een zoekactie met behulp van PUBMED leverde 53 relevante artikelen op. Bij nauwkeurige beschouwing van deze 53 studies viel allereerst een grote variatie op in de technieken die gebruikt waren om apoptotische cellen aan te tonen. In de meeste studies was de TUNEL (terminal deoxyribonucleotidyl transferase-mediated nick end labelling) techniek gebruikt (n = 29). Andere gebruikte technieken waren ISEL (in situ end labelling, n = 7), morfologische analyse met lichtmicroscopie (n = 7), TUNEL of ISEL samen met lichtmicroscopische morfologie (n = 9), ISEL samen met morfologie vastgesteld met electronenmicroscopie (n = 1), morfologie vastgesteld met licht microscopie samen met electronenmicroscopie (n = 2) en de recent ontwikkelde techniek van het aantonen van gekliefd cytokeratine 18 met behulp van immuunhistochemie (n = 2). Door caspase gekliefd cytokeratine 18 is een vroeg kenmerk van apoptose dat aangetoond kan worden in epitheliale cellen met gebruik van een antilichaam (M30). Zoals verwacht mag worden met al deze verschillende technieken was er een grote variatie in het gerapporteerde percentage 144

146 Samenvatting apoptotische cellen (apoptotische indices). Deze variatie is waarschijnlijk grotendeels terug te voeren op de voornaamste beperking van de TUNEL en ISEL techniek: het foutpositief aanmerken van necrotische cellen naast apoptotische cellen. Anderen hebben reeds hun vraagtekens gezet bij de toepasbaarheid van TUNEL in het aantonen van apoptotische cellen in de tractus digestivus. Hoewel de laatste tijd nieuwe technieken zijn ontwikkeld met een hogere specificiteit dan TUNEL voor apoptotische cellen, zoals de methode van M30 immunoreactiviteit, hebben nog weinig studies van deze technieken gebruik gemaakt. Met betrekking tot de mate van apoptose in normaal colonslijmvlies varieerden gerapporteerde apoptotische indices tussen 0,11 en 11 %. Daarnaast viel op dat in de meeste studies het bestudeerde normale colonslijmvlies in feite was verkregen van macroscopisch normaal ogend slijmvlies in resectievlakken van tumoren waarbij men zich kan afvragen in hoeverre dit daadwerkelijk als normaal colonslijmvlies kan worden beschouwd. In een studie bleek dat de mate van apoptose inderdaad verschillend was in normale mucosa van sneevlakken ten opzichte van echte normale mucosa. Dit dient men zich te realiseren bij de interpretatie van studies waarin apoptose in normaal colon is bestudeerd. De overgang van normaal colonepitheel naar het adenoomstadium gaat in de meeste studies gepaard met een toename in mate van apoptose. Over wat er gebeurt bij de overgang van adenoom naar carcinoom bestaat veel discussie. De meeste onderzoeken geven aan dat de mate van apoptose in carcinomen hoger is dan in adenomen. Er zijn echter ook studies waarin er geen verschil was, of zelfs tegenovergestelde resultaten werden gezien. Als het gaat om de distributie van apoptotische cellen in de loop van de adenoom-carcinoom sequentie komt uit vrijwel alle studies naar voren dat in adenomen de meeste apoptose gezien wordt aan de basis van crypten, met de meeste prolifererende cellen aan de luminale zijde van crypte. Dit patroon is volstrekt tegenovergesteld aan dat wat gezien wordt in normaal epitheel. Wanneer de correlatie werd bekeken tussen de mate van apoptose en de mate van proliferatie in adenomen en carcinomen werden opnieuw tegenstrijdige resultaten verkregen. Waar sommige studies een correlatie tussen beide vonden, wordt dit in andere bestreden. Met betrekking tot de prognostische waarde van apoptotische indices werd in de meeste studies gevonden dat colorectale tumoren met een hoge mate van apoptose geassocieerd waren met een betere prognose. Gegeven de beperkingen van de juist beschreven technieken TUNEL and ISEL om apoptotische cellen aan te tonen werd in Hoofdstuk 3 de recent ontwikkelde techniek van M30 immunoreactiviteit (zie eerder) nader bestudeerd. De techniek werd gevalideerd tegen de methode die over het algemeen als de gouden standaard wordt beschouwd: de identificatie van apoptotische cellen door herkennen van bepaalde morfologische criteria. Naar aanleiding van de observatie in Hoofdstuk 2 dat de meeste studies over apoptose in normaal colorectaal epitheel in feite waren uitgevoerd in macroscopisch normaal ogend weefsel in sneevlakken werd onderzocht of de mate van apoptose verschillend was tussen normale mucosa van sneevlakken en werkelijk normale mucosa. Paraffine materiaal van normale colonmucosa (n = 30), normale mucosa van sneevlakken (n = 30), colorectale adenomen (n = 84) en carcinomen (n = 40) werd bestudeerd. Apoptotische cellen werden aangetoond met behulp van M30 immunoreactiviteit en door het gebruik van morfologische 145

147 Chapter 10 criteria en de mate van apoptose werd uitgedrukt als percentage van het totaal aantal getelde cellen (apoptotische index). De gemiddelde apoptotische indices bepaald volgens de M30 methode waren 0,18 ± 0,04 % in normale mucosa, 0,42 ± 0,04 % in adenomen en 1,97 ± 0,24 % in carcinomen. Apoptotische indices vastgesteld met morfologische criteria waren respectievelijk 0,23 ± 0,03 %, 0,62 ± 0,06 % en 1,78 ± 0,19 %. De mate van apoptose was hoger in normale mucosa afkomstig van sneevlakken vergeleken met werkelijk normale mucosa (p < 0,0001). Er was een goede correlatie tussen apoptotische indices verkregen met de M30 techniek en met morfologische criteria (r = 0,71, p < 0,01). Deze bevindingen ondersteunen het gebruik van de techniek van M30 immunoreactiviteit bij het bestuderen van apoptose in colorectaal epitheel. Om de potentiële toepassing van TRAIL of TRAIL receptor agonistische stoffen als therapeutische middelen bij colorectale tumoren te bestuderen werd de expressie van de vier membraangebonden TRAIL receptoren DR4, DR5, DcR1 en DcR2 in de darm onderzocht. De expressie en lokalisatie van TRAIL en TRAIL receptoren werd geanalyseerd met behulp van immuunhistochemie in normale mucosa (n = 10), adenomen (n = 19) en carcinomen (n = 21). Het materiaal was afkomstig van patiënten met een sporadische achtergrond van hun aandoening. Daarnaast werden mogelijke correlaties bekeken tussen expressie van TRAIL en zijn receptoren en de mate van apoptose (bepaald met M30 immunoreactiviteit) en histopathologische kenmerken. De resultaten zijn beschreven in Hoofdstuk 4. Zowel in normale mucosa als in adenomen en carcinomen werd expressie van TRAIL en TRAIL receptoren gezien. Het viel op dat TRAIL expressie verloren was gegaan in een aantal colorectale tumoren in vergelijking met aangrenzend normaal epitheel. Dit was vaker het geval in carcinomen dan in adenomen (p < 0,05). De intensiteit van DR4 en DR5 aankleuring was sterker in neoplastische cellen dan in aangrenzende normale epitheliale cellen, en ging samen met een hogere mate van apoptose. Er werden geen verschillen gevonden tussen normale cellen en neoplastische cellen met betrekking tot DcR1 en DcR2 expressie. Er waren geen correlaties tussen TRAIL of TRAIL receptor expressie en histopathologische kenmerken. De sterkere expressie van DR4 en DR5 in neoplastische cellen in vergelijking met normale cellen suggereert dat neoplastische cellen gevoeliger zijn voor TRAIL en TRAIL receptor agonistische stoffen dan normale cellen, hetgeen een bemoedigende bevinding is in het licht van een toekomstige klinische toepassing. Naar aanleiding van de resultaten in patiënten met een sporadische achtergrond van hun colorectale tumor werd het onderzoek uitgebreid met materiaal van patiënten met FAP en HNPCC en dit is beschreven in Hoofdstuk 5. De expressie van DR4 en DR5 werd bestudeerd met behulp van immuunhistochemie in colorectale adenomen en carcinomen van patiënten met een sporadische achtergrond (n = 74 respectievelijk 56), FAP (n = 41 en 4) en HNPCC (n = 50 en 21). Verder werden sporadische tumoren met zogenaamde microsatelliet instabiliteit (MSI) bestudeerd om de rol van mutaties in het BAX gen te onderzoeken. Er zijn aanwijzingen dat BAX een rol speelt bij de gevoeligheid voor TRAIL- 146

148 Samenvatting gemedieerde apoptose in vitro. Colorectale tumoren met MSI zijn kwetsbaar voor mutaties in het BAX gen. De aanwezigheid van BAX mutaties in MSI positieve tumoren zou derhalve de potentiële toepassing van TRAIL of TRAIL receptor agonisten kunnen beperken. Carcinomen (n = 42, waarvan 27 sporadische en 15 HNPCC-geassocieerd) met een hoge mate van MSI (MSI-H) werden geanalyseerd op de aanwezigheid van BAX mutaties en de mate van apoptose. De meeste adenomen van alle drie patiëntengroepen toonden DR4 en DR5 expressie. De meeste carcinomen hadden DR4 expressie, met uitzondering van zes tumoren, alle met histologisch een mucineus karakter. Alle carcinomen, inclusief die met mucineuze histologie, toonden DR5 expressie. BAX mutaties werden gevonden in 6 van de 42 MSI-H tumoren, waarbij de apoptotische indices en expressie van DR4, DR5 en TRAIL vergelijkbaar waren in tumoren met en zonder BAX mutaties. Deze gegevens suggereren niet direct dat BAX inactivatie in tumoren met MSI in vivo een belangrijk mechanisme is om apoptose te omzeilen. Al met al geven de resultaten aan dat het selectief aangrijpen van TRAIL receptoren DR4 en DR5 in sporadische en erfelijke colorectale tumoren in potentie mogelijk is en verder uitgezocht dient te worden. De identificatie van prognostische factoren voor patiënten met colorectaal carcinoom stelt ons wellicht in staat behandelingen hierop af te stemmen. Resistentie van tumorcellen tegen apoptose inductie is een beperkende factor in de effectiviteit van chemotherapie van coloncarcinoom. In vitro studies hebben laten zien dat chemotherapie en recombinant humaan (rh)trail synergistisch apoptose induceren in colonkankercellen. De prognostische waarde van de expressie van TRAIL, DR4 en DR5 op overleving, ziektevrije overleving en tijd-tot-recidief werden bestudeerd in Hoofdstuk 6. Van tumoren van een cohort van 376 patiënten met stadium III colon carcinoom, die in het verleden aan een twee-armige studie met adjuvante chemotherapie hadden deelgenomen, werden zogenaamde weefsel microarrays geconstrueerd 7. Deze arrays werden geanalyseerd met behulp van immuunhistochemie op TRAIL, DR4 en DR5 expressie. De mediane follow-up van het cohort was 43 maanden. Kaplan-Meier analyses en Cox proportional hazard analyses werden verricht waarbij werd gecorrigeerd voor behandelarm, leeftijd en geslacht. De meeste carcinomen toonden sterke expressie van TRAIL (82,8 %), DR4 (91,7 %) en DR5 (87,1 %). Sterke DR4 expressie was geassocieerd met een slechtere overleving {odds ratio (OR; 95% CI) = 2,22 (1,03-4,81); p = 0,04} en een slechtere ziekte-vrije overleving {OR = 2,19 (1,06-4,53); p = 0,03}, vergeleken met zwakke DR4 expressie. Er was een trend dat sterke DR5 expressie geassocieerd was met een betere overleving {OR = 0,65 (0,41-1,03); p = 0,07}. Binnen de patiëntengroep met een recidief was het zo dat de tijd tot vaststelling van het recidief langer was voor diegenen met sterke DR5 expressie (p = 0,03) of sterke TRAIL expressie (p = 0,007) vergeleken met hen met een zwakke DR5 of TRAIL expressie. De bevinding dat de meeste tumoren sterke DR4 en DR5 expressie tonen geeft aan dat rhtrail in potentie gebruikt kan worden in de adjuvante behandeling van patiënten met een stadium III coloncarcinoom. 147

149 Chapter 10 Recent werden twee gerandomiseerde placebo-gecontroleerde onderzoeken gepubliceerd waarin een chemopreventief effect van aspirine werd aangetoond in patiënten met eerder een colorectaal adenoom of carcinoom 8,9. Deze studies worden besproken in Hoofdstuk 7, waarbij een kort overzicht wordt gegeven van de mechanismen achter de chemopreventieve werking van NSAIDs. Beide studies tonen een bescheiden beschermend effect van aspirine in de genoemde patiëntenpopulaties op de ontwikkeling van adenomen. De anti-neoplasmatische eigenschappen van NSAIDs zijn zowel in vitro bewezen als in vivo (dierexperimenteel onderzoek, epidemiologische studies en interventiestudies). Verschillende mechanismen achter de beschermende invloed van NSAIDs op de colorectale carcinogenese zijn in de afgelopen jaren beschreven. Centraal hierbij lijkt inductie van apoptose in neoplastische cellen, via cyclo-oxygenase- afhankelijke en onafhankelijke mechanismen. De precieze mechanismen van apoptose inductie zijn onbekend, maar er zijn aanwijzingen dat dit bij het NSAID sulindac via DR5 wordt gemedieerd. Daarnaast zijn er aanwijzingen dat NSAIDs de Wnt route moduleren. De vorming van adenomen wordt over het algemeen toegeschreven aan mutaties in hetzij het APC gen of het gen coderend voor β-catenine. In normale cellen is β-catenine gebonden aan het celmembraan, waarbij cytoplasmatisch β-catenine wordt afgebroken door het zogenaamde APC/axine/GSK3β-complex. Verlies van functioneel APC of bepaalde mutaties in het β- catenine gen leiden tot stapeling van β-catenine in de celkern, wat leidt tot activatie van de transcriptie factor T-cell factor 4 (TCF4). Als gevolg hiervan activeert TCF4 een genetisch programma dat verantwoordelijk wordt gehouden voor de vorming van adenomen. De Wnt-route, waarbij β-catenine/tcf4 activiteit fungeert als een schakel tussen proliferatie en differentiatie in het darmepitheel speelt waarschijnlijk een kritieke rol in de vroege ontwikkeling van colorectale tumoren. Anderen hebben aangetoond dat geactiveerde Wnt-signalering de concentratie vermindert van het eiwit p21, een eiwit dat de celcyclus beïnvloedt. Via dit mechanisme kunnen cellen ongeremd prolifereren. Recent zijn er aanwijzingen gekomen dat het chemopreventieve effect van het NSAID sulindac deels gemedieerd wordt door p21. Daarnaast is aangetoond dat het remmend effect van sulindac op de ontwikkeling van darmtumoren teniet werd gedaan door homozygote inactivatie van p21 in APC min muizen, het muizenmodel van FAP. p21 is mogelijk de link tussen de Wnt route en het chemopreventieve effect van sulindac op de ontwikkeling van adenomen. Deze informatie vormde de achtergrond voor een studie waarin de effecten van sulindac werden bestudeerd op de mate van apoptose en de expressie van DR4, DR5, β-catenine en p21 in normaal ogend slijmvlies. De resultaten zijn beschreven in Hoofdstuk 8. Biopten voor en na behandeling met sulindac, afgenomen in het kader van twee chemopreventie studies 10,11, werden verzameld. Het betrof patiënten met HNPCC (n = 18) die 2 dd 150 mg sulindac gedurende 4 weken hadden gekregen in een placebo-gecontroleerd cross-over design 10. Uit een andere studie kwam materiaal van patiënten met FAP (n = 6) die behandeld waren met 2 dd 150 mg sulindac gedurende 6 maanden 11. Met behulp van immuunhistochemie werd de mate van apoptose vastgesteld (M30 immunoreactiviteit) en expressiepatronen van DR4, DR5, β-catenine en p

150 Samenvatting In HNPCC patiënten waren de apoptotische indices na sulindac en placebo vergelijkbaar. Ook in FAP patiënten was er geen verschil in mate van apoptose na sulindac vergeleken met de waarden vooraf. Expressie van DR4 en DR5 werd gezien in alle onderzochte coupes waarbij geen consistente veranderingen werden gezien na sulindac in vergelijking met placebo. De intensiteit van membraneuze β-catenine aankleuring was lager in HNPCC materiaal na sulindac dan na placebo (p < 0,001). Vergelijkbare resultaten werden verkregen in het FAP materiaal (p < 0,01). p21 expressie voor en na sulindac was vergelijkbaar in beide patiëntengroepen. Samengevat had sulindac een remmend effect op β-catenine expressie in normaal colorectaal epitheel van HNPCC en FAP patiënten zonder zichtbare veranderingen in mate van apoptose en DR4, DR5 en p21 expressie. Een remmend effect van sulindac op β-catenine expressie is ook geobserveerd in APC Min muizen en in adenomen. Of dit fenomeen van belang is in het chemopreventieve effect van sulindac kan met de huidige gegevens vooralsnog niet worden beantwoord. Conclusies en toekomstperspectieven Meerdere genen die betrokken zijn bij de regulatie van apoptose komen abnormaal tot expressie of zijn gemuteerd bij CRC. Hoewel het idee van geleidelijk verminderde apoptose in de ontwikkeling van CRC conceptueel aantrekkelijk is, duidt het beschikbare bewijs uit in vivo materiaal er op dat het percentage apoptotische cellen in de loop van de adenoomcarcinoom sequentie juist geleidelijk toeneemt. In vitro is tumor progressie vaak geassocieerd met resistentie tegen apoptose inductie. Hoewel deze gegevens elkaar op het eerste gezicht lijken tegen te spreken, doen ze dat in feite niet. De schijnbare tegenstrijdigheid vervalt wanneer in aanmerking wordt genomen dat bij in vivo studies spontane apoptose wordt bestudeerd terwijl in vitro studies gaan over geïnduceerde apoptose. Gegeven de beperkingen van de methoden van apoptose detectie, die in de meeste studies zijn gebruikt, zijn betere testen gewenst en waarschijnlijk in de nabije toekomst beschikbaar. De resultaten in dit proefschrift waarbij in een aantal studies de methode van M30 immunoreactiviteit is gebruikt ondersteunen het nut van deze techniek bij het bestuderen van apoptose in colorectaal epitheel. Succesvolle, niet-chirurgische, eliminatie van kankercellen wordt uiteindelijk bepaald door inductie van apoptose. Vrijwel alle bekende cytostatica werken door inductie van apoptose, vaak niet alleen van kwaadaardige cellen maar ook van normale cellen, zich uitend in toxiciteit. De laatste jaren is de kennis over de mediatoren van apoptose in coloncellen enorm toegenomen. Voorbeelden van mediatoren, waarvan bekend is dat deze verstoord zijn in tumorcellen, zijn Bcl-2 familieleden en IAPs 12. Deze onderdrukkers van apoptose zijn aangrijpingspunten geworden voor de ontwikkeling van medicamenten die erop gericht zijn om de gevoeligheid van tumorcellen voor apoptose te herstellen. Een alternatieve benadering wordt gevormd door strategieën gericht op directe activering van agonisten van apoptose via de extrinsieke route. TRAIL is, in tegenstelling tot familiegenoten TNF-α en FasL, niet geassocieerd met het sepsis syndroom respectievelijk levertoxiciteit. Afgaande op de effectiviteits- en toxiciteitsgegevens uit preklinisch onderzoek en de informatie in dit proefschrift over de uitgebreide expressie van de pro-apoptotische TRAIL receptoren 149

151 Chapter 10 in colorectale tumoren, kan rhtrail in deze setting als veelbelovend worden beschouwd. Agonistische TRAIL receptor antilichamen, gericht tegen DR4 en DR5, blijken dezelfde effecten als rhtrail te hebben 13,14. De eerste fase-i studies met rhtrail en agonistische TRAIL receptor antilichamen zijn onderweg en de resultaten van deze studies zullen waarschijnlijk een indicatie geven van de waarde van deze middelen in de behandeling van CRC. Daarnaast dient de waarde van combinatie regimes, waarbij rhtrail gecombineerd wordt met een andere stof, verder te worden uitgezocht. Meerdere onderzoeken hebben aangetoond dat door rhtrail te combineren met bepaalde chemotherapeutische middelen, NSAIDs, proteosoomremmers of interferon-γ een additief of synergistisch apoptotisch effect kan worden verkregen 15. Een andere veelbelovende strategie vormt de ontwikkeling van zogenaamde bi-specifieke antilichamen. Deze antilichamen combineren herkenning van de tumorcel met activering van het immuunsysteem, met als gevolg dat bepaalde effectorcellen tumorcellen opruimen. Er zijn inmiddels fusie eiwitten geconstrueerd waarbij TRAIL gecombineerd wordt met het aangrijpen van de EGFR-receptor en het met carcinomen geassocieerd antigeen EGP2, met veelbelovende resultaten in vitro 16,17. Dergelijke combinatiestrategieën verminderen waarschijnlijk de dosis TRAIL die nodig is voor een bepaald anti-tumor effect, en daarmee de kans op toxiciteit. Afgezien van een toekomstige toepassing van TRAIL in de behandeling van CRC, is er mogelijk ook een rol voor TRAIL weggelegd in de therapie of preventie van colorectale adenomen. Hoewel adenomen over het algemeen verwijderd worden bij endoscopisch onderzoek zijn er situaties waarin dit niet mogelijk is, bijvoorbeeld in het geval van vlakke sessiele adenomen. Ook komt het voor dat het aantal aanwezige adenomen te groot is om verwijderd te kunnen worden, zoals bij FAP patiënten. Recente gegevens laten zien dat TRAIL ook apoptose induceert in colonadenoomcellen 18. Deze waarneming werd bevestigd in een studie waarbij adenoomcellijnen en ex vivo kweken van humane adenomen werden onderzocht 19. De bevindingen uit hoofdstukken 4 en 5 van dit proefschrift over de aanwezige expressie van TRAIL receptoren DR4 en DR5 in colorectale adenomen ondersteunen het verder exploreren van rhtrail of andere TRAIL receptor agonistische middelen voor de behandeling van colorectale adenomen. Concluderend kan gezegd worden dat de laatste jaren nieuwe behandelingsmogelijkheden zijn ontwikkeld voor CRC, gebaseerd op toenemende kennis over de mechanismen die betrokken zijn bij de regulatie van apoptose. De potentiële klinische toepassing van rhtrail of andere TRAIL receptor agonistische middelen bij de behandeling van colorectale adenomen en carcinomen verdient nader onderzocht te worden. De resultaten van de eerste fase-i studies met rhtrail en TRAIL receptor agonistische antilichamen zullen toekomstige behandelschema s moeten leiden. De sleutel voor verdere verbetering van preventie en behandeling van colorectale tumoren ligt waarschijnlijk níet in het zoeken naar de Heilige Graal: het ene ideale middel tegen kanker. Meer waarschijnlijk zal daadwerkelijke vooruitgang voortkomen uit het speuren naar een optimale combinatie van middelen die, door op verschillende essentiële routes in de cel aan te grijpen, resulteert in de onvoorwaardelijke dood van de kankercel. 150

152 Samenvatting Referenties 1. Weitz J, Koch M, Debus J, Höhler T, Galle PR, Büchler MW. Colorectal cancer. Lancet 2005;365: Torrance CJ, Jackson PE, Montgomery E, et al. Combinatorial chemoprevention of intestinal neoplasia. Nat Med 2000;6: Wang Y, Jatkoe T, Zhang Y, et al. Gene expression profiles and molecular markers to predict recurrence of Dukes B colon cancer. J Clin Oncol 2004;22: Ribic CM, Sargent DJ, Moore MJ, et al. Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 2003;349: Cunningham D, Humblet Y, Siena S, et al. Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 2004;351: Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350: Bleeker WA, Mulder NH, Hermans J, Otter R, Plukker JT. The addition of low-dose leucovorin to the combination of 5-fluorouracil-levamisole does not improve survival in the adjuvant treatment of Dukes C colon cancer. IKN Colon Trial Group. Ann Oncol 2000;11: Sandler RS, Halabi S, Baron JA, et al. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 2003;348: Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 2003; 348: Rijcken FE, Hollema H, van der Sluis T, Boersma-van Ek W, Kleibeuker JH. Sulindac increases epithelial cell proliferative activity in the proximal colon of HNPCC patients. Eur J Gastroenterol Hepatol 2005;17: A Giardiello FM, Hamilton SR, Krush AJ et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 1993;328: Reed JC. Apoptosis-targeted therapies of cancer. Cancer Cell 2003;3: Pukac L, Kanakaraj, Humphreys R, et al. HGS-ETR1, a fully human TRAIL-receptor 1 monoclonal antibody, induces cell death in multiple tumour types in vitro and in vivo. Br J Cancer 2005;92: Motoki K, Mori E, Matsumoto A, et al. Enhanced apoptosis and tumor regression induced by a direct agonst antibody to tumor necrosis factor-related apoptosis-inducing ligand receptor 2. Clin Cancer Res 2005;11: Van Geelen CM, de Vries EG, de Jong S. Lessons from TRAIL-resistance mechanisms in colorectal cancer cells: paving the road to patient-tailored therapy. Drug Resist Updat 2004;7: Bremer E, Samplonius DF, van Genne L, et al. Simultaneous inhibition of epidermal growth factor receptor (EGFR) signaling and enhanced activation of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptor-mediated apoptosis induction by an scfv:strail fusion protein with specificity for human EGFR. J Biol Chem 2005;280: Bremer E, Kuijlen J, Samplonius D, Walczak H, de Leij L, Helfrich W. Target cell-restricted and enhanced apoptosis induction by a scfv:strail fusion protein with specificity for the pancarcinomaassociated antigen EGP2. Int J Cancer 2004;109: Hague A, Hicks DJ, Hasan F, Smartt H, Cohen GM, Paraskeva C, MacFarlane M. Increased sensitivity to TRAIL-induced apoptosis occurs during the adenoma to carcinoma transition of colorectal carcinogenesis. Br J Cancer 2005;92: Jalving M, Koornstra JJ, de Jong S, et al. TRAIL induces apoptosis in colorectal adenoma cell lines and ex vivo adenoma cultures. Gastroenterology 2005;128(Suppl2):A

153

154 List of publications Ulvik A, Koornstra JJ, Røkke O, Halvorsen JF, Haug K, Øgreid D. Determination of the activated proto-oncogene (Ki-ras) in feces. A new laboratory analysis for early diagnosis of colorectal cancer. Tidsskr Nor Laegeforen 1995;115: Koornstra JJ, Ulvik A, Røkke O, Halvorsen JF, Haug K, Øgreid D. Ras oncogendetectie in faeces van patiënten met colorectumtumoren. Ned Tijdschr Geneeskd 1995;139: Koornstra JJ, Ottervanger JP, Fehmers MCO, Stricker BHCh. Klinisch manifeste leverbeschadiging tijdens gebruik van simvastatine. Ned Tijdschr Geneeskd 1996;140: Koornstra JJ, Olffen GH van, Noort G van. Retractile mesenteritis: to treat or not to treat. Hepatogastroenterology 1997;44: Koornstra JJ, Veenendaal D, Bruyn GAW, de Graaf H. A case of septic arthritis caused by Fusobacterium nucleatum. Br J Rheum 1998;37:1249. Schot BW, Koornstra JJ, Maas AH, Coenen JL. Maligne of niet maligne pericarditis?. Ned Tijdschr Geneeskd 1999;143:643. Koornstra JJ, Loualidi A, de Vries CJ. Ticlopidine-induced thrombocytopenia. Neth J Med 1999;54: Koornstra JJ, Spoelstra P, de Graaf H, Queré M. Icterus toegeschreven aan gebruik van ticlopidine. Cardiologie 1999;6: Koornstra JJ, Kerstens MN, Hoving J, Visscher KJ, Schade JH, Gort HBW, Leemhuis MP. Clinical and biochemical changes following 131I therapy for hyperthyroidism in patients not pretreated with antithyroid drugs. Neth J Med 1999;55: Postma AA, Koornstra JJ, Fagel WJ. Fibromuscular dysplasia in a 63-year-old woman. Neth J Med 2000;56: Koornstra JJ, Klaver NS, ter Maaten JC, Limburg AJ, van der Jagt EJ, van der Werf TS. Neostigmine bij pseudo-obstructie van het colon (Ogylvie syndroom). Ned Tijdschr Geneeskd 2001;145: Koornstra JJ, van de Loosdrecht AA, van Imhoff GW. Lactic acidosis: pathophysisiology, diagnosis and treatment. Neth J Med 2001;59: Koornstra JJ, Kleibeuker JH. Colorectaal carcinoom: opsporing verzocht. Analyse 2001;56: van Haelst PL, Koornstra JJ, van der Werf TS. Een kortademige man met koorts en keelpijn: Syndroom van Lemièrre. Ned Tijdschr Geneeskd 2002;146:

155 Koornstra JJ, de Jong S, Hollema H, de Vries EGE, Kleibeuker JH. Changes in apoptosis during the development of colorectal cancer: a systematic review of the literature. Crit Rev Oncol Hematol 2003;45: Scheenstra R, Rijcken FE, Koornstra JJ, Hollema H, Fodde R, Menko FH, Sijmons RH, Bijleveld CM, Kleibeuker JH. Rapidly progressive adenomatous polyposis in a patient with germline mutations in both the APC and MLH1 genes: the worst of two worlds. Gut 2003;52: Koornstra JJ, Kleibeuker JH, van Geelen CM, Rijcken FE, Hollema H, de Vries EGE, de Jong S. Expression of TRAIL (TNF-related apoptosis-inducing ligand) and its receptors in normal colonic mucosa, adenomas, and carcinomas. J Pathol 2003;200: Huls G, Koornstra JJ, Kleibeuker JH. Non-steroidal anti-inflammatory drugs and molecular carcinogenesis of colorectal carcinomas. Lancet 2003;362: Jalving M, Koornstra JJ, Boersma-van Ek W, de Jong S, Karrenbeld A, Hollema H, de Vries EG, Kleibeuker JH. Dysplasia in fundic gland polyps is associated with nuclear beta-catenin expression and relatively high cell turnover rates. Scand J Gastroenterol 2003;38: Jalving M, Koornstra JJ, Gotz JM, van der Waaij LA, de Jong S, Zwart N, Karrenbeld A, Kleibeuker JH. High-grade dysplasia in sporadic fundic gland polyps: a case report and review of the literature. Eur J Gastroenterol Hepatol 2003;15: Koornstra JJ, van Haelst PL. Een man die per ongeluk formaline dronk: Caustisch letsel van slokdarm en maag. Ned Tijdschr Geneeskd 2003;147:2219. Houwerzijl EJ, Blom NR, van der Want JJ, Esselink MT, Koornstra JJ, Smit JW, Louwes H, Vellenga E, de Wolf JT.Ultrastructural study shows morphologic features of apoptosis and para-apoptosis in megakaryocytes from patients with idiopathic thrombocytopenic purpura. Blood 2004;103: Koornstra JJ, Rijcken FE, De Jong S, Hollema H, de Vries EG, Kleibeuker JH. Assessment of apoptosis by M30 immunoreactivity and the correlation with morphological criteria in normal colorectal mucosa, adenomas and carcinomas. Histopathology 2004;44:9-17. Spierings DC, de Vries EG, Vellenga E, van den Heuvel FA, Koornstra JJ, Wesseling J, Hollema H, de Jong S. Tissue distribution of the death ligand TRAIL and its receptors. J Histochem Cytochem 2004;52: Harms MP, Koornstra JJ. Een jonge man met een gezwollen en rode arm: Diepveneuze trombose van de arm. Ned Tijdschr Geneeskd 2004;148:937. Jalving M, Koornstra JJ, de Jong S, de Vries EGE, Kleibeuker JH. The potential of combinational regimen with non-steroidal anti-inflammatory drugs in the chemoprevention of colorectal cancer. Aliment Pharmacol Ther 2005;21: Koornstra JJ, Jalving M, Rijcken FE, Westra J, Zwart N, Hollema H, de Vries EG, Hofstra 154

156 RWM, Plukker JTM, de Jong S, Kleibeuker JH. Expression of TRAIL receptors DR4 and DR5 in sporadic and hereditary colorectal tumours: potential targets for apoptosis induction. Eur J Cancer 2005;41: Thijs WJ, Koornstra JJ, Karrenbeld A, Limburg AJ. Eosinofiele oesofagitis bij jonge mannen met voedselpassageklachten. Ned Tijdschr Geneeskd 2005;149: JJ Koornstra, FE Rijcken, CN Oldenhuis, N Zwart, T van der Sluis, H Hollema, EG de Vries, JJ Keller, JA Offerhaus, FM Giardiello, JH Kleibeuker. Sulindac inhibits beta-catenin expression in normal-appearing colon of hereditary nonpolyposis colorectal cancer and familial adenomatous polyposis patients. Cancer Epidemiol Biomarkers Prev 2005;14: JJ Koornstra, JH Kleibeuker. Statinen ter preventie van colorectaal carcinoom vooralsnog niet aanbevolen. Ned Tijdschr Geneeskd 2005;149:

157

158 Curriculum vitae De auteur werd geboren op 25 juli 1969 in Alkmaar. In 1987 werd het eindexamen VWO afgelegd aan de Rijksscholengemeenschap in Heerenveen. In datzelfde jaar werd begonnen met de studie Geneeskunde aan de Rijskuniversiteit Groningen. Na coschappen in Medisch Spectrum Twente in Enschede en een wetenschappelijke stage in Noorwegen werd in 1995 het artsexamen behaald. Na een jaar assistent-geneeskundige te zijn geweest bij de afdeling Interne Geneeskunde in Enschede begon de opleiding tot internist, van 1996 tot 1999 in Medisch Centrum Leeuwarden en van 1999 tot 2003 in het Academisch Ziekenhuis Groningen. De opleiding werd onderbroken in 2000 om, met ondersteuning van het KWF Kankerbestrijding, een jaar lang fulltime wetenschappelijk onderzoek te kunnen doen. Sinds 2003 is de auteur verbonden aan de afdeling Maag-, Darm- en Leverziekten van het AZG, thans UMCG, aanvankelijk als internist, fellow hepatologie, vanaf februari 2004 als maag-darm-leverarts in opleiding. 157

159

160 Dankwoord Het dankwoord: vaak als laatste geschreven maar als eerste gelezen. Nu is zo n proefschrift ook het eindresultaat van het werk van velen hoewel er maar één naam op de voorkant staat. Velen hebben aan dit proefschrift een bijdrage geleverd, sommigen als co-auteur, anderen als commentator of stimulator en weer anderen als klankbord. Het gebeurt niet vaak dat men in de gelegenheid is om de mensen te bedanken die een rol hebben gespeeld bij het bereiken van een mijlpaal, als men dit proefschrift althans als zodanig kan beschouwen. Ik zal derhalve dan ook geen enkele poging doen me in dit opzicht te beperken. Ten eerste wil ik mijn promotoren danken. Prof. dr. J.H. Kleibeuker, beste Jan, in 1999 legde je de basis voor dit proefschrift. Ik solliciteerde destijds bij jou en Prof. dr. P.L.M. Jansen, voor een opleidingsplaats Maag-, Darm- en Leverziekten. Rond die tijd zocht ik een onderwerp van studie en kwam je met het voorstel om te gaan kijken naar TRAIL, apoptose en colorectaal carcinoom. Alleen met het laatste begrip was ik enigszins vertrouwd. Desondanks verliep het sollicitatiegesprek gunstig. Dank voor het vertrouwen dat je destijds in me hebt gehad. De afgelopen jaren heb ik geweldig veel van je geleerd, niet alleen als clinicus en endoscopist, maar ook als wetenschapper. Je uiterlijke rust bij de begeleiding van wetenschappelijk onderzoek en het schijnbare gemak waarmee je het overzicht houdt is door voorgangers al vele malen geroemd. Ik hoop me de komende jaren onder jouw vleugels verder te kunnen ontwikkelen, om uiteindelijk ooit zelfstandig uit te vliegen. Prof. dr. E.G.E. de Vries, beste Liesbeth, je bent een ware promotor. Van Dale s Groot Woordenboek zegt: pro mo tor: 1 hoogleraar die promovendi begeleidt; 2 persoon die zorgt voor het bijeenbrengen van kapitaal en de voorbereidende werkzaamheden bij de oprichting van een bedrijf enz.; 3 [sport] persoon die een bokser opleidt en introduceert; 4 [scheik.] stof die de katalytische werking van een andere verhoogt; 5 voorstander. Je bent veel meer dan 1 alleen, bij 2 denk ik aan je onvermoeibare inzet bij het schrijven van subsidies en 3 gaat misschien wel over de harde strijd die soms gevoerd moet worden om tegenstanders met wetenschappelijke argumenten te overtuigen. 4, dat gaat helemaal op, je fungeert niet alleen voor mij maar voor velen als een soort überkatalysator. Je hebt het bewonderenswaardige vermogen mensen onophoudelijk te prikkelen, te stimuleren en aan te zetten tot sneller schrijven, hoewel dat op het moment zelf niet altijd als zodanig wordt ervaren. Over snelheid gesproken, er is volgens mij niemand die sneller stukjes corrigeert dan jij. Dank voor je begeleiding en je kritische noten. Prof. dr. H. Hollema, beste Harry, bij jou was het altijd prettig uitblazen. Terwijl ik nerveus met mijn plankje met coupes met de nieuwste kleuringen klaarstond, benieuwd wat je ervan dacht, wilde jij het altijd eerst even over andere dingen hebben. Niet over de wetenschap, maar over het leven. Ik had altijd een gevoel van onthaasting als ik later je kamer weer uit kwam. Je hebt me veel geleerd, over het leven, over de histologie van poliepen en carcinomen, over hoe men een apoptotische cel herkent en over de methodologie bij het opzetten van nieuwe immunohistochemische kleuringen. Dank daarvoor. 159

161 Dr. S. de Jong, copromotor, beste Steven, jij hebt een belangrijke rol gespeeld in het vervolmaken van dit boekje. Maar wat was het in het begin lastig om je te volgen. Van Vestdijk zei men dat hij sneller schreef dan God kan lezen. Van jou zou je kunnen zeggen dat jij sneller denkt dan je kan praten, en dat laatste is al zelden bij te houden. Ik ben je dank verschuldigd voor het me vertrouwd maken met basale laboratoriumtechnieken en je begeleiding van het onderzoek door de jaren heen. Ik weet hoe je veel en met een zekere kritische grondhouding met dokters samenwerkt en ik hoop dat we onze vruchtbare samenwerking in de toekomst kunnen blijven continueren. De leden van de leescommissie, Prof. dr. Peter.L.M. Jansen, Prof. dr. G.J. Offerhaus en Prof. dr. T. Wiggers wil ik danken voor hun snelle beoordeling van dit proefschrift. Beste Peter, dank voor het vertrouwen dat je destijds in me had bij de sollicitatie. Mijn paranimfen, Bart Schot en Sjoerd Hovenga, ben ik dankbaar voor hun vriendschap en het feit dat zij me willen bijstaan bij de verdediging. Hopelijk staan jullie binnenkort weer in hetzelfde theater, maar dan in het midden. Een zekere kiem voor dit boekje ligt in Noorwegen. In 1994 ben ik, als co-assistent, voor het eerst min of meer besmet geraakt met het onderzoeksvirus, tijdens een wetenschappelijke stage in Bergen. Het onderwerp van studie destijds, diagnostiek van colorectale tumoren door fecesanalyse, heeft nog niets aan actualiteit ingeboet. Mijn begeleider destijds, Prof. dr. Dagfinn Øgreid, ben ik hiervoor zeer dankbaar. Kjære Dagfinn, jeg vil gjærne få takke deg for den fantastiske tilliten du ga meg da du høsten 1994 lot meg gå igang med undersøkelsen ved Senter for Klinisk Molekylærmedisin ved Haukeland sykehus i Bergen. Du var den første som viste meg den fascinerende verden som vitenskap og forskning er. Det er med vemod jeg tenker tilbake på vår tid sammen i laboratoriet og hjemme hos deg, våre måltid på Pasta Sentral, Genefecprosjektet og den første presentasjonen på kongressen i Førde. For alt dette: Tusen ganger takk. Mijn opleiders in Enschede (Medisch Spectrum Twente, opleiders Dr. Jordans, respectievelijk Prof. dr. Richel), Leeuwarden (Medisch Centrum Leeuwarden, opleiders Dr. Leemhuis, respectievelijk Dr. Halma) en Groningen (Academisch Ziekenhuis Groningen, opleider Prof. dr. Gans) wil ik danken, met name voor hun medewerking bij het wijzigen van mijn opleidingsschema waardoor ik in Groningen terechtkwam, waar dit oorspronkelijk niet de bedoeling zou zijn. Stafleden uit de genoemde ziekenhuizen dank ik voor hun begeleiding tijdens mijn opleiding. Met name Dr. Piet Spoelstra, MDL-arts in Leeuwarden, ben ik dankbaar voor zijn rol als intermediair bij het leggen van het eerste contact tussen mij en Jan Kleibeuker. De huidige MDL stafleden in het UMCG (Bram Limburg, Hendrik van Dullemen, Frans Peters, Gerard Dijkstra, Aad van den Berg en Els Haagsma) ben ik erkentelijk voor hulp bij het verzamelen van patiëntenmateriaal, niet alleen voor studies waarvan de weerslag in dit boekje te vinden is maar ook voor artikelen die nog op stapel staan, en voor hun begeleiding bij en investering in mijn opleiding tot MDL-arts. Mijn kamergenoten en collega fellows Willem Thijs, Rinse Weersma, Pascal Dekkers, 160

162 Louktje Wormmeester, Antoine Flierman en Robert Verdonk dank ik voor de gezellige uren die we samen doorbrachten. Nogmaals excuses voor al die keren dat jullie de prachtigste verhalen aan het vertellen waren en ik voornamelijk bezig was met het boekje. Willem en Rinse, de Playstation staat altijd klaar. Alle verpleegkundigen en het secretariaat van het endoscopiecentrum ben ik dankbaar voor hun onmisbare hulp bij het verzamelen van patiëntenmateriaal en de fysieke en geestelijke ondersteuning bij het scopiëren. Mijn co-auteurs dank ik voor hun bijdragen in het opzetten en uitwerken van de verschillende artikelen. Met name Fleur Rijcken en Hilde Jalving leverden een onschatbare bijdrage. Lieve Fleur, je generositeit om je data met me te delen is geweldig. Dankzij jou, en je beschikking over veel archiefmateriaal van de pathologie, had ik een vliegende start met dit onderzoek. Daarnaast was het altijd gezellig aan weerszijden van de microscoop, ondanks het vooruitzicht dat er ontelbaar veel cellen geteld moesten worden. Lieve Hilde, ik prijs de dag, het was in 2000, dat je als tweedejaars geneeskunde student, via Steven en Liesbeth, met mijn onderzoek mee kwam kijken waarbij ik een van je begeleiders was. Achter dat verlegen meisje met dat grappige accent bleek een briljante en ambitieuze onderzoekster schuil te gaan. Het was een genoegen om je destijds te begeleiden maar ook om te zien dat je inmiddels je eigen onderzoeksweg gegaan bent. KWF Kankerbestrijding wil ik danken voor het subsidiëren van het onderzoeksjaar voor arts-assistenten waarmee ik in 2000 de basis voor dit boekje kon leggen. Mede dankzij de bemoedigende resultaten die dit onderzoeksjaar opleverden werd een vervolgsubsidie verkregen. De Maag Lever Darm Stichting dank ik zeer voor deze vervolgsubsidie. De analisten die hebben bijgedragen aan dit boekje zijn Wendy Dam, Fiona van den Heuvel, Tineke van der Sluis, Wietske Boersma-van Ek, en vooral Nynke Zwart. Nynke, dank voor al je inspanningen en het feit dat je altijd bereid was weer een nieuw antilichaam te proberen. Mijn collega internisten in opleiding en collega onderzoekers dank ik voor hun geestelijke ondersteuning tijdens de afgelopen jaren. Jentsje Popma (1921) wil ik danken voor zijn toestemming om een van zijn schilderijen af te mogen beelden op het omslag. Naast een zekere betrokkenheid bij de aandoening die centraal staat in dit boekje, delen we een belangrijke liefde voor het (Friese en Groningse) landschap. Dan, vrienden en familie, dank voor jullie belangstelling voor de wetenschap. At en Hendrik, jullie interesse was altijd groot ( Is er nog nieuws? ), ondanks het feit dat er niet altijd nieuws was. Wetenschap gaat langzaam. Mijn zus en zwager, dank voor jullie belangstelling en de inspirerende invloed van jullie kinderen. Mijn schoonouders dank ik voor jullie interesse. Mijn vader en mijn moeder ben ik zeer dankbaar voor de keuzes die jullie voor me 161

163 maakten in het verleden en jullie belangstelling tijdens de loop van het onderzoek. Dank voor jullie onvoorwaardelijke liefde, zorg en toewijding. Tenslotte, Petra, er zijn geen woorden die recht doen aan de rol die je speelt in mijn leven. Ik heb je lief. En is het nu klaar? Dit boekje misschien wel, maar wetenschap is nooit klaar. En het bleek de afgelopen jaren tamelijk spannend om ermee bezig te zijn. Eerst is er de spanning van het bedenken van nieuwe onderzoeken en experimenten (en of anderen niet hetzelfde al eerder hebben bedacht), spanning naar resultaten van proeven, spanning na het schrijven van een eerste versie van een manuscript naar commentaren van co-auteurs, en, na de tiende versie, spanning naar het commentaar van reviewers, nadat een stuk is weggestuurd, spanning of de repliek op het commentaar wel goed is en of het stuk geaccepteerd wordt en tenslotte hoe het artikel er uiteindelijk uit komt te zien. Zoveel spanning, het is eigenlijk zonde om ermee te stoppen Jan Jacob Koornstra 162

164

165

166 Appendix

167 Appendix chapter 3 Figure 2. Examples of M30 immunoreactivity (A-B; magnification x 100), Ki-67 staining (C-D; magnification x 100) and HE-staining (E; magnification x 400) in normal mucosa. (A) Genuinely normal colon mucosa, M30 labelling is predominantly limited to cells at the mucosal surface (arrow). (B) Normal mucosa obtained from a resection margin, M30 labelling is seen along the crypt axis (arrows). (C) Normal mucosa and (D) normal mucosa obtained from resection margins show Ki-67 positive cells limited to the proliferation zone in both cases with comparable proliferative indices. (E) High power view of normal mucosa obtained from a resection margin, showing the presence of apoptotic cells (arrows) in the crypt base. 166

168 Appendix chapter 3 Figure 3. Examples of M30 immunostaining in an adenoma (A-B) and a carcinoma (C-D). (A) M30 positive cells are seen along the entire crypt (magnification x 100). (B) Detail of a crypt containing numerous M30 positive cells with strong cytoplasmic staining (magnification x 400). (C) M30 positive cells are heterogeneously distributed throughout the carcinoma (magnification x 100). (D) High power view of the carcinoma shown in C (magnification x 400). 167

169 Appendix chapter 4 A B C D E Figure 2. Representative examples of expression of TRAIL and TRAIL receptors in adenomas. A: TRAIL; B: DR4; C: DR5; D: DcR1; E: DcR2. Epithelial cells express cytoplasmic immunoreactivity for DR4, DR5, DcR1, DcR2 and TRAIL. The intensity of staining for DR4 and DR5 is generally stronger in adenoma cells as compared to normal cells. Magnification 100x.

170 Appendix chapter 4 A B C D E Figure 3. Representative examples of expression of TRAIL receptors and TRAIL in adenocarcinomas. A: TRAIL; B: DR4; C: DR5; D: DcR1; E: DcR2. Tumour cells express cytoplasmic immunoreactivity for DR4, DR5, DcR1 and DcR2. Loss of TRAIL expression is seen in tumour cells compared to adjacent normal cells. The intensity of staining for DR4 and DR5 is stronger in carcinoma cells as compared to normal cells. Magnification 100x. 169

171 Appendix chapter 5 A B C Figure 1. DR4 and DR5 expression in mucinous adenocarcinoma. Hematoxylin-eosin (A), DR4 (B) and DR5 staining (C) showing absence of DR4 expression with intact DR5 expression. 170

172 Appendix chapter 6 A B C D E F Figure 1. DR4, DR5 and TRAIL expression in tissue microarrays of colon cancer tumors. Examples of immunohistochemical staining results for DR4 (A and B, low and high expression respectively), DR5 (C and D, low and high expression respectively) and TRAIL (E and F, low and high expression respectively) 171