Correlation of Vascular Endothelial Growth Factor Expression and Microvessel Density in Cervical Intraepithelial Neoplasia

Transcription

1 JM, Blot WJ. Tobacco use and nasopharyngeal carcinoma in a cohort of US veterans. Int J Cancer 1993;55: (37) Yi Z, Ohshima H, Bouvier G, Roy P, Zhong J, Li B, et al. Urinary excretion of nitrosamino acids and nitrate by inhabitants of high- and low-risk areas for nasopharyngeal carcinoma in southern China. Cancer Epidemiol Biomarkers Prev 1993;2: (38) Yu MC, Nichols PW, Zou XN, Estes J, Henderson BE. Induction of malignant nasal cavity tumors in Wistar rats fed Chinese salted fish. Br J Cancer 1989;60: (39) Huang DP, Ho JH, Saw D, Teoh TB. Carcinoma of the nasal and paranasal regions in rats fed Cantonese salted marine fish. IARC Sci Publ 1978;(20): (40) Raucy JL, Kraner JC, Lasker JM. Bioactivation of halogenated hydrocarbons by cytochrome P4502E1. Crit Rev Toxicol 1993;23: (41) Lee CK, Fulp C, Bombick BR, Doolittle DJ. Inhibition of mutagenicity of N-nitrosamines by tobacco smoke and its constituents. Mutat Res 1996;367: (42) Patten CJ, Smith TJ, Tynes RW, Freseu M, Lee T, Yang CS, et al. Evidence for cytochrome P450 2A6 and 3A4 as catalysts for N- nitrosonornicotine alpha hydroxylation in human liver microsomes. Carcinogenesis. In press. (43) Poirier S, Bouvier G, Malaveille C, Ohshima H, Shao YM, Hubert A, et al. Volatile nitrosamine levels and genotoxicity of food samples from high-risk areas for nasopharyngeal carcinoma before and after nitrosation. Int J Cancer 1989;44: (44) Poirier S, Ohshima H, de The G, Hubert A, Bourgade MC, Bartsch H. Volatile nitrosamine levels in common food from Tunisia, south China and Greenland, high-risk areas for nasopharyngeal carcinoma (NPC). Int J Cancer 1987;39: (45) Shao YM, Poirier S, Ohshima H, Malaveille C, Zeng Y, de The G, et al. Epstein-Barr virus activation in Raji cells by extracts of preserved food from high risk areas for nasopharyngeal carcinoma. Carcinogenesis 1988;9: (46) Bouvier G, Hergenhahn M, Polack A, Bornkamm GW, de The G, Bartsch H. Characterization of macromolecular lignins as Epstein-Barr virus inducer in foodstuff associated with nasopharyngeal carcinoma risk. Carcinogenesis 1995;16: (47) Chen W, Weisburger JH, Fiala ES, Spratt TE, Carmella SG, Chen D, et al. Gastric carcinogenesis: 2-chloro-4-methylthiobutanoic acid, a novel mutagen in salted, pickled Sanma hiraki fish, or similarly treated methionine. Chem Res Toxicol 1996;9: Notes Present address: J. R. Idle, Institute of Cancer Research and Molecular Biology, Norwegian University of Science and Technology, Trondheim, Norway. Manuscript received January 13, 1997; revised May 30, 1997; accepted June 13, Correlation of Vascular Endothelial Growth Factor Expression and Microvessel Density in Cervical Intraepithelial Neoplasia Andreas Obermair, Dagmar Bancher-Todesca, Selcuk Bilgi, Alexandra Kaider, Petra Kohlberger, Silvia Müllauer-Ertl, Sepp Leodolter, Gerald Gitsch* Background: Angiogenesis (the formation of new blood vessels) appears to be required for the growth of invasive tumors, but little information exists about its role in the development of preinvasive lesions. We examined microvessel density and expression of vascular endothelial growth factor in specimens of cervical intraepithelial neoplasia (CIN), a preinvasive lesion of the uterine cervix, to determine whether a connection could be established between these parameters of angiogenesis and the grade of dysplasia (i.e., tissue abnormality). Methods: Sections of biopsy specimens from 83 patients with grades I III CIN were examined retrospectively. Microvessels were localized by use of a polyclonal antibody directed against factor VIII-related antigen; vascular endothelial growth factor was detected by means of a monoclonal antibody. Reported P values are two-sided. Results: Highest microvessel densities and highest expression of vascular endothelial growth factor were found in a narrow border region between CIN lesions and the underlying stroma. A significant correlation was observed between high vascular endothelial growth factor expression and high microvessel density (Kendall s = 0.27; 95% confidence interval [CI] = ; P =.018). Mean microvessel density values ± standard deviations for CIN I, CIN II, and CIN III lesions were 19.4 ± 5.8, 21.9 ± 7.0, and 34.1 ± 14.8, respectively (Kendall s = 0.46; 95% CI = ; P<.0001). Corresponding values for vascular endothelial growth factor expression were 8.3 ± 3.5, 8.4 ± 2.0, and 12.2 ± 3.6, respectively (Kendall s = 0.41; 95% CI = ; P<.0001). Conclusions: Our results are consistent with the idea that progression of cervical dysplasia is dependent on angiogenesis. [J Natl Cancer Inst 1997;89: ] The role of angiogenesis in tumor growth is well defined for a variety of solid malignant tumors (1). Once a tumor has become established, a prevascular phase of uncertain length, during which the tumor is dormant, is followed by a phase of intense endothelial cell proliferation and subsequent tumor growth (2). The angiogenic response is mediated by diffusible signals, so-called angiogenic factors, which can be secreted by malignant tumor cells (2). Observations in vivo suggest that the inhibition of blood vessel formation results in a reduction in tumor size and a reduction in metastasis (3 5). The vascular endothelial growth factor (VEGF), which is also known as the vascular permeability factor (VPF), is a multifunctional cytokine that acts on endothelial cells as a highly specific mitogen (6,7), binding to specific class III receptor tyrosine kinases (flt-1 and KDR) and opening calcium channels to increase intracellular calcium levels (8). VEGF also stimulates angiogenesis by increasing vascular permeability (7,9). Four VEGF isoforms (polypeptides containing 206, 189, 165, and 121 amino acids) have been isolated, but they apparently express identical biologic activities (10). Although extensive data are available on the role of angiogenesis in the development of a variety of invasive malignant tumors, little information exists concerning preinvasive lesions. Cervical intraepithelial neoplasia (CIN) is a preinvasive lesion of the uterine cervix that is charac- *Affiliations of authors: A. Obermair, D. Bancher-Todesca, P. Kohlberger, S. Leodolter, G. Gitsch (Department of Gynecology and Obstetrics), S. Bilgi, S. Müllauer-Ertl (Department of Pathology, Gynecopathology Unit), A. Kaider (Department of Medical Computer Sciences/Biometrics), University Hospital of Vienna, Austria. Correspondence to: Andreas Obermair, M. D., Department of Gynecology and Obstetrics, University Hospital of Vienna, Währinger Gürtel 18-20, A-1090 Vienna, Austria. andreas. obermair@akh-wien.ac.at See Notes following References. Oxford University Press 1212 REPORTS Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997

2 terized by the presence of mitotic figures above the basal layer, enlarged and irregular nuclei, and an increased nuclear/ cytoplasmic ratio (11). CIN is usually detected by means of an abnormal Pap smear and can be visualized by colposcopy after the application of acetic acid. On colposcopic examination, CIN lesions appear as white plaques. The lesions are graded according to the extent of normal epithelium that is replaced by dysplasia. An additional adverse feature visualized by colposcopy is punctuation or mosaicism on a white background, which represents an area of higher blood vessel count relative to that of normal epithelium (12,13). The aim of this study was to describe the immunostaining patterns of VEGF expression and microvessel density in CIN and to detect a connection, if any, between these parameters of angiogenesis and the grade of dysplasia. Methods Patients and Tissues The tissue of 83 patients with CIN grades I III (see below) who underwent cone biopsy or loop excision at the University Hospital of Vienna from 1990 through 1993 were examined retrospectively. All surgical specimens were embedded in paraffin, and sections were mounted on glass slides and stained with hematoxylin eosin. An experienced pathologist reviewed all slides to confirm the diagnoses and to determine whether a sufficient amount of CIN was present before assessing microvessel density and VEGF expression. CIN lesions were classified histologically as mild (CIN I), moderate (CIN II), or severe dysplasia/carcinoma in situ (CIN III), depending on the grade of nuclear abnormality and the extent of epithelium replaced by dysplasia (11). Immunostaining for factor VIII-related antigen (F8-RA) (to evaluate microvessel density) and for VEGF was performed from the same tissue block for 39 of the 83 cases. Because of the small size of CIN I lesions and for forensic reasons, only one histologic section was available for immunostaining for the remaining 44 cases. Three CIN lesions of different grades on one slide were available for two cases, and two CIN lesions of different grades on one slide were available for 14 cases. These 16 cases were also included in this analysis. Immunohistochemistry Fifty-seven slides with 65 CIN I III lesions were immunostained for F8-RA. The lesions consisted of CIN I, CIN II, and CIN III in 16, 21, and 28 cases, respectively. The sections were deparaffinized and treated with 1% protease (type XIV; Sigma Chemical Co., St. Louis, MO) for 15 minutes at 37 C, followed by treatment with 3% hydrogen peroxide. The sections were then incubated with a polyclonal antibody directed against F8-RA (factor VIII polyclonal; BioGenex, San Ramon, CA) at room temperature for 1 hour. Avidin biotin complex staining was performed by incubating the sections with an alkaline phosphatase-conjugated biotinylated second antibody (BioGenex) at room temperature for 20 minutes, followed by incubation with Fast Red (Bio- Genex) as the chromogen. Counterstaining was performed with hematoxylin. After scanning each section under low magnification ( 40), the area of clear-cut CIN with the greatest number of distinctly highlighted microvessels ( hot spot ) was selected. Microvessel density was then determined by counting all vessels under high magnification ( 200) with the use of an eyepiece screen with an edge length of 10 mm/100 (Olympus, Vienna, Austria) and an examination area of 0.25 mm 2. Evaluation of immunostaining for F8-RA was confined strictly to the subepithelial region. Forty-eight slides with 57 CIN I III lesions were immunostained for VEGF protein expression. The lesions consisted of CIN I, CIN II, and CIN III in 19, 17, and 21 cases, respectively. Immunostaining for VEGF protein was performed as follows. A monoclonal anti-human VEGF neutralizing antibody (R&D Systems Europe, Oxford, U.K.), used at a dilution of 1:50, was the primary antibody, and sections were incubated with this antibody for 1 hour at room temperature. The sections were then treated with reagents from the StrAviGen High Performance Immunodetection System (BioGenex) for 20 minutes at room temperature; 3-amino-9-ethylcarbozole (AEC; BioGenex) was used as the chromogen. Finally, the sections were washed with distilled water and counterstained with hematoxylin. To determine the number of VEGF-positive cells accurately, the counts were performed on a single field of mm 2 under 400 magnification. Moreover, VEGF-positive cells were quantified strictly in the subepithelial zone of CIN lesions or the adjacent normal epithelium. Negative-control staining reactions were carried out by substituting normal, nonimmune serum for the primary antibodies. As positive controls for F8- RA and VEGF immunostaining, we used sections from a breast cancer and a vulvar cancer, which were previously shown to be strongly positive for F8-RA and VEGF expression (data not shown). One investigator (S. Bilgi) evaluated all sections for microvessel density and VEGF expression. Statistical Methods Individual values for microvessel density and the number of VEGF-positive cells are reported as the mean + standard deviation (SD). To assess the correlation between microvessel density, VEGF expression, and the grade of cervical dysplasia, Kendall s correlation was applied; 95% confidence intervals (CIs) for Kendall s correlation coefficient were calculated by use of the bootstrap method with 2000 resampling runs (14,15). In some cases, the chisquared test was used to analyze the data. In addition, comparison of microvessel density values and VEGF-positive cell counts between the three CIN groups (CIN I, CIN II, and CIN III) was performed by use of the Kruskal Wallis test (16). Reported P values are two-sided. Results Microvessel Density The highest level of anti-f8-ra staining of microvessels was observed in a narrow border region between CIN lesions and the underlying stroma (Fig. 1, A). The subepithelial zone of the adjacent normal epithelium and the deep stroma both exhibited much less immunostaining. High-power light microscopy revealed an area of enhanced microvessel density beneath CIN lesions in comparison with the density found in the adjacent normal epithelium. The mean microvessel density value ± SD was 25.6 ± 5.9 for all CIN lesions, and the mean values + SDs were 19.4 ± 5.8, 21.9 ± 7.0, and 34.1 ± 14.8 for CIN I, CIN II, and CIN III lesions, respectively (Kendall s 0.46; 95% CI ; P<.0001) (Fig. 2). VEGF Expression VEGF expression was found to a great extent within the subepithelial zone of CIN lesions and to a lesser extent beneath the normal epithelium and in the deep stroma (Fig. 1, B). We found VEGF expressed in the cytoplasm of macrophagelike cells. The mean number of VEGFpositive cells ± SD was 9.8 ± 3.6 for all CIN lesions, and the mean numbers + SDs were 8.3 ± 3.5, 8.4 ± 2.0, and 12.2 ± 3.6 for CIN I, CIN II, and CIN III lesions, respectively (Kendall s 0.41; 95% CI ; P<.0001) (Fig. 2). Correlation Between Microvessel Density and VEGF Expression The area of highest VEGF protein expression and anti-f8-ra staining of microvessels was found to be the border region between the dysplastic epithelium and the underlying normal stroma. We observed a significant correlation between high VEGF expression and high microvessel density (Kendall s 0.27; 95% CI ; P.018). An overall Kruskal Wallis test showed that the difference in microvessel density across the CIN I, CIN II, and CIN III lesions studied was statistically significant (P<.0001). Subgroup analysis also revealed statistically significant differences between the CIN I and CIN III lesions and between the CIN II and CIN III lesions (Kruskal Wallis test; P<.0001 and P.001, respectively), but the differ- Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997 REPORTS 1213

3 ence between the CIN I and CIN II lesions was not statistically significant (Kruskal Wallis test; P.067). Similarly, an overall Kruskal Wallis test showed a statistically significant difference in VEGF-positive cell counts across the CIN I, CIN II, and CIN III lesions (P.0004). Subgroup analysis revealed statistically significant differences in VEGF-positive cell counts between the CIN I and CIN III lesions and between the CIN II and CIN III lesions (Kruskal Wallis test; P.0007 and P.0015), whereas the difference between the CIN I and CIN II lesions was not statistically significant (Kruskal Wallis test; P.397). Evaluation of anti-f8-ra staining of microvessels and of VEGF-positive cells revealed the presence of intralesional vessels stained for F8-RA in none of the CIN I and CIN II lesions and in one of 26 cases of CIN III (chi-squared test 1.464; P Fig. 1. Specimens of cervical intraepithelial neoplasia (CIN) immunostained for factor VIII-related antigen (F8-RA) (A) and vascular endothelial growth factor (VEGF) (B). The area of highest F8-RA expression (red staining that indicates the location of microvessels) and of VEGF protein expression (reddish-brown staining) was shown to be the border region between the dysplastic epithelium (uppermostcell layers) and the normal stroma (lower layers). Both lesions were graded histologically as CIN III. Magnification was 200 for both (A) and (B)..691); VEGF expression within the CIN lesions was noted in three (16%) of 19, in four (23%) of 17, and in five (24%) of 21 of CIN I, CIN II, and CIN III lesions, respectively (chi-squared test 1.388; P.728). Discussion To our knowledge, our data demonstrate for the first time that the expression Fig. 2. Microvessel density and vascular endothelial growth factor (VEGF) expression in normal epithelium and in cervical intraepithelial neoplasia (CIN) grades I III. Values are given as the mean ± the standard deviation (see Methods section for more details). An overall Kruskal Wallis test demonstrated that the difference in microvessel density across the CIN I, CIN II, and CIN III lesions was statistically significant (P<.0001). Subgroup analysis also revealed statistically significant differences between the CIN I and CIN III lesions and between the CIN II and CIN III lesions (Kruskal Wallis test; P<.0001 and P.001, respectively), but the difference between the CIN I and CIN II lesions was not statistically significant (Kruskal Wallis test; P.067). Similarly, an overall Kruskal Wallis test showed a significant difference in VEGFpositive cell counts across the CINI, CIN II, and CIN III lesions (P.0004). Subgroup analysis revealed that the differences in VEGF-positive cell counts between the CIN I and CIN III lesions and between the CIN II and CIN III lesions were statistically significant (Kruskal Wallis test; P.0007 and P.0015, respectively), whereas the difference between the CIN I and CIN II lesions was not (Kruskal Wallis test; P.397). Dark-shaded bars indicate neighboring normal epithelium; light-shaded bars indicate dysplastic epithelium REPORTS Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997

4 of VEGF protein, a highly selective mitogen for endothelial cells, is significantly correlated with microvessel density in CIN. Angiogenic parameters, such as microvessel density and the expression of the highly angiogenic peptide VEGF, increase with the grade of cervical dysplasia, whereas microvessel density and VEGF expression in the adjacent normal epithelium remain nearly constant at a significantly lower level. Experimental models demonstrate that preinvasive malignant cells are dormant until they induce angiogenesis. Evidence of angiogenic properties has been found in certain dysplastic lesions, such as preneoplastic bladder epithelium, carcinoma in situ of the breast, and hyperplasia in pancreatic islets. In bladder biopsy specimens, increased numbers of capillaries were observed focally in about one half of the cases of flat dysplasia (17). In mouse mammary papillary hyperplasia, angiogenic properties appeared much earlier than any morphologic or clinical signs of carcinoma (18). In breast tissue, the frequency of neoplastic transformation was observed to be higher for those hyperplastic lesions with a high frequency of angiogenic response (19). In pancreatic islets, hyperplasia per se did not obligate angiogenesis; however, it was shown that angiogenic activity first appeared in a subset of hyperplastic islets before the onset of a tumor (20). All of these studies demonstrated that angiogenic activity preceded the development of invasive malignant neoplasia. By means of F8-RA immunostaining in CIN lesions, Smith-McCune et al. (21) highlighted vascular endothelial cells and demonstrated that a region of neovascularization develops along the basement membrane subtending the dysplastic epithelium. In that study, angiogenesis was closely correlated with the grade of cervical dysplasia. Additional vascular stalks in the upper half of the epithelium were found in CIN II and CIN III lesions but not in CIN I lesions. Neither the degree of inflammation nor the type of human papillomavirus (HPV) infection influenced the extent of angiogenesis. Guidi et al. (22) reported VEGF messenger RNA expression and microvessel density to be significantly increased in patients with invasive carcinoma of the cervical epithelium and in high-grade intraepithelial lesions compared with what was observed in low-grade lesions and benign epithelium. In contrast, Soini et al. (23) did not find a significant association between the number of blood vessels and the degree of dysplasia, but they did find a higher, although not statistically significant, microvessel density in lesions containing high-risk HPV DNA than in lesions with no or low-risk HPV DNA. By determining the number of proliferating cell nuclear antigen (PCNA)-positive vascular endothelial cells, Soini et al. observed moderate or strong PCNA staining significantly more often in CIN II and III lesions than in condylomas and CIN I lesions. No correlation was found between HPV DNA and PCNA positivity. Recently, Abulafia et al. (24) reported that microvessel counts did not differ between microinvasive squamous cell cancer (n 14) and in situ squamous cell carcinoma (n 18) of the uterine cervix. In all of the studies mentioned above (21 24) as well as in our investigation, immunostaining of the vascular endothelium was carried out by use of an antibody directed against F8-RA. This approach is considered to be the gold standard for highlighting vascular endothelial cells (1). In our study, mean values are given as the mean ± SD of the number of microvessels in a 0.25-mm 2 field. The mean microvessel density in CIN I lesions was 19.4 ± 5.8 per 0.25-mm 2 field, whereas it was 34.1 ± 14.8 per 0.25-mm 2 field in CIN III lesions. By assessing microvessel density in cervical biopsy specimens at 400 magnification, Guidi et al. (21) reported significantly lower microvessel counts in normal benign epithelium and in low-grade intraepithelial lesions (38.8 ± 14.4 and 44.5 ± 13.0, respectively) than in high-grade intraepithelial lesions and invasive carcinomas (67.0 ± 29.1 and 62.1 ± 11.1, respectively). Soini et al. (23), using a 100- m 2 field for microvessel counting, recorded density values of 0.24 per 100- m 2 field and 0.36 per 100- m 2 field in CIN 0-I and CIN II III lesions, respectively. In the study by Abulafia et al. (24), a magnification of 400 yielded median microvessel density values of 17 per highpower field (HPF) for benign control specimens, 19 per HPF for in situ carcinomas, and 34.5 per HPF for microinvasive carcinomas. In contrast, Smith- McCune et al. (21) did not give absolute mean blood vessel counts but provided an angiogenic ratio, which consists of the vessel count of a lesion divided by that of the normal epithelium from the same patient. The angiogenic ratio in CIN I and CIN II lesions was approximately 1.4, and it was higher than 2.0 in CIN III lesions. Since mean microvessel density values depend on the size of the examined area and other methodologic considerations, the values given above for the different studies cannot be compared directly. However, the angiogenic ratio observed by Smith-McCune et al. (21) was similar to our findings for CIN I and CIN II lesions (approximately 1.4), and they also found a significantly higher ratio for CIN III lesions (approximately 2.0). Overall, it is remarkable that, with the exception of the study by Abulafia et al. (24), the difference between mild and strong cervical dysplasia measured in terms of microvessel density is similar in all studies, despite the different methods used (21 23). We did not perform within-block or within-person correlation analyses, but we did test whether within-block values affected mean values for microvessel density and number of VEGF-positive cells. The mean values ± SDs for each CIN group were similar when cases with two or three different CIN lesions of different grades within one slide were either included or excluded. VEGF is the most specific mitogen known for endothelial cells, and it is considered, therefore, to play a key role in tumor angiogenesis (25). Toi et al. (26) reported 29 (26.6%) of 109 breast cancer tissues to be rich in VPF/VEGF. Takahashi et al. (26) reported not only VEGF but also the VEGF-receptor KDR to be significantly associated with microvessel density in colon cancer. In a study on the prognostic effect of microvessel density and VEGF expression in patients with vulvar cancer (28), we found overall survival to be substantially influenced by the level of angiogenic activity. In all of these studies, VEGF expression was closely associated with high microvessel density and poor prognosis. Thus, our data are in accord with those of previous studies, i.e., we found a significant association between VEGF expression and the grade of dysplasia as well as between VEGF expression and microvessel density. Our data confirm the hypothesis that VEGF is an important mediator of angiogenesis in the dysplastic process. In contrast with Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997 REPORTS 1215

5 previously reported intense immunostaining reactions for VEGF in the cytoplasm of tumor cells (29,30), we found VEGF to be highly expressed by macrophage-like cells and less so by tumor cells and vascular endothelial cells. The implications of these findings remain unclear and will be the subject of further investigation. In clinical trials, spontaneous regression has been observed frequently in cases of mild or moderate CIN. Spontaneous regression rates for histologically proven CIN range from 50% to 60% for CIN I and from 27% to 53% for CIN II, but they drop to 13% 31% in cases of CIN III (31 33). Progression to invasive carcinoma has been reported to be directly correlated with the severity of CIN that is detected in the initial biopsy specimen, with progression rates varying from 14% to 40% for CIN I and from 71% to 73% for CIN III (31,32). Our data on the angiogenic properties of CIN, therefore, are consistent with the hypothesis that CIN I and CIN II lesions are nonaggressive precursors, since they express significantly less angiogenic activity than CIN III lesions, which may be the actual clinically relevant precursors of invasive cervical cancer, as indicated by an intense expression of VEGF protein and a highly dense network of microvessels below the dysplastic epithelium. Our data give indirect evidence that the growth of cervical dysplasia is dependent on angiogenesis and that VEGF may play an important role in the switch to the angiogenic phenotype in CIN. This switch has been reported to be essential for the development of invasive cervical and other cancers (34). Since the growth of CIN seems to be dependent on angiogenesis, severe dysplasia of the uterine cervix might represent a target for antiangiogenic therapy strategies in the future. References (1) Weidner N. Intratumoral microvessel density as a prognostic factor in cancer. Am J Pathol 1995;147:9 19. (2) Folkman J. Tumor angiogenesis. In: Mendelsohn J, Howley PM, Israel MA, Liotta LA, editors. The molecular basis of cancer. Philadelphia: Saunders, 1995: (3) Melnyk O, Shuman MA, Kim KJ. Vascular endothelial growth factor promotes tumor dissemination by a mechanism distinct from its effect on primary tumor growth. Cancer Res 1996;56: (4) Teicher BA, Sotomayor EA, Huang ZD. Antiangiogenic agents potentiate cytotoxic cancer therapies against primary and metastatic disease. Cancer Res 1992;52: (5) Brown PD, Giavazzi R. Matrix metalloproteinase inhibition: a review of anti-tumour activity. Ann Oncol 1995;6: (6) Leung DW, Cachianes G, Kuang W, Goeddel DV, Ferrara N. Vascular endothelial growth factor is a secreted angiogenic mitogen. Science 1989;246: (7) Senger DR, Connolly DT, Van de Water L, Feder J, Dvorak HF. Purification and NH2- terminal amino acid sequence of guinea pig tumor-secreted vascular permeability factor. Cancer Res 1990;50: (8) Brock TA, Dvorak HF, Senger DR. Tumorsecreted vascular permeability factor increases cytosolic Ca2+ and von Willebrand factor release in human endothelial cells. Am J Pathol 1991;138: (9) Ferrara N, Heinsohn H, Walder CE, Bunting S, Thomas GR. The regulation of blood vessel growth by vascular endothelial growth factor. Ann N Y Acad Sci 1995;752: (10) Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 1995;146: (11) Ferenczy A, Winkler B. Cervical intraepithelial neoplasia and condyloma. In: Kurman RJ, editor. stein s pathology of the female genital tract. Hamburg: Springer, (12) Hinselmann H. Der Nachweis der aktiven Ausgestaltung der Gefasse beim jungen Portiokarzinom als neues differentialdiagnostisches Hilfsmittel. Zentralbl Gynakol 1940;64:1810. (13) Stafl A, Mattingly RF. Angiogenesis of cervical neoplasia. Am J Obstet Gynecol 1975;121: (14) Efron B, Tibshirani R. An introduction to the bootstrap. New York: Chapman & Hall, (15) Kendall MG. Rank correlation methods. 4 th ed. London: Charles Griffin, (16) Kruskal WH, Wallis WA. Use of ranks in onecriterion variance analysis. J Am Stat Assoc 1952;47: (17) Deen S, Ball RY. Basement membrane and extracellular interstitial matrix components in bladder neoplasia evidence of angiogenesis. Histopathology 1994;25: (18) Brem S, Gullino PM, Medina D. Angiogenesis: a marker for neoplastic transformation of mammary papillary hyperplasia. Science 1977; 195: (19) Brem SS, Jensen HM, Gullino PM. Angiogenesis as a marker of preneoplastic lesions of the human breast. Cancer 1978;41: (20) Folkman J, Watson K, Ingber D, Hanahan D. Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature 1989; 339: (21) Smith-McCune KK, Weidner N. Demonstration and characterization of the angiogenic properties of cervical dysplasia. Cancer Res 1994;54: (22) Guidi AJ, Abu-Jawdeh G, Berse B, Jackman RW, Tognazzi K, Dvorak HF, et al. Vascular permeability factor (vascular endothelial growth factor) expression and angiogenesis in cervical neoplasia. J Natl Cancer Inst 1995;87: (23) Soini Y, Pollanen R, Kemppainen S, Paakko P, Lehto VP. The association of vascular proliferation with HPV status and epithelial PCNA positivity in cervical intraepithelial lesions. APMIS 1996;104: (24) Abulafia O, Triest WE, Sherer DM. Angiogenesis in squamous cell carcinoma in situ and microinvasive carcinoma of the uterine cervix. Obstet Gynecol 1996;88: (25) Zhang HT, Craft P, Scott PA, Ziche M, Weich HA, Harris AL, et al. Enhancement of tumor growth and vascular density by transfection of vascular endothelial cell growth factor into MCF-7 human breast carcinoma cells. J Natl Cancer Inst 1995;87: (26) Toi M, Hoshina S, Takayanagi T, Tominaga T. Association of vascular endothelial growth factor expression with tumor angiogenesis and with early relapse in primary breast cancer. Jpn J Cancer Res 1994;85: (27) Takahashi Y, Kitadai Y, Bucana CD, Cleary KR, Ellis LM. Expression of vascular endothelial growth factor and its receptor, KDR, correlates with vascularity metastasis, and proliferation of human colon cancer. Cancer Res 1995;55: (28) Obermair A, Kohlberger P, Bancher-Todesca D, Tempfer C, Sliutz G, Reinthaller A, et al. Influence of microvessel density and vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) expression on prognosis in vulvar cancer. Gynecol Oncol 1996;63: (29) Dvorak HF, Sioussat TM, Brown LF, Berse B, Nagy JA, Sotrel A, et al. Distribution of vascular permeability factor (vascular endothelial growth factor) in tumors: concentration in tumor blood vessels. J Exp Med 1991;174: (30) Qu-Hong, Nagy JA, Senger DR, Dvorak HF, Dvorak AM. Ultrastructural localization of vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) to the albuminal plasma membrane and vesiculovacuolar organelles of tumor microvascular endothelium. J Histochem Cytochem 1995;43: (31) Kataja V, Syrjanen S, Mantyjarvi R, Yliskoski M, Saarikoski S, Syrjanen K. Prognostic factors in cervical human papillomavirus infections. Sex Transm Dis 1992;19: (32) Hellberg D, Nilsson S, Gad A, Hongxiu J, Fuju C, Syrjanen S, et al. Behavior of cervical intraepithelial neoplasia (CIN) associated with various human papillomavirus (HPV) types [published erratum appears in Arch Gynecol Obstet 1993;253:116]. Arch Gynecol Obstet 1993;252: (33) Meyskens FL Jr, Surwit E, Moon TE, Childers JM, Davis JR, Dorr RT, et al. Enhancement of regression of cervical intraepithelial neoplasia II (moderate dysplasia) with topically applied all-trans-retinoic acid: a randomized trial. J Natl Cancer Inst 1994;86: (34) Holmgren L, O Reilly MS, Folkman J. Dormancy of micrometastases: balanced prolifera REPORTS Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997

6 Notes tion and apoptosis in the presence of angiogenesis suppression. Nat Med 1995;1: Supported by a grant from the Medizinisch- Wissenschaftlicher Fonds des Burgermeisters der Bundeshauptstadt Wien. We thank Helga Dammerer and Gerhard Breitenecker for technical advice in performing immunohistochemistry. Manuscript received August 5, 1996; revised June 10, 1997; accepted June 20, Doxorubicin Disaccharide Analogue: Apoptosis-Related Improvement of Efficacy In Vivo Federico Arcamone, Fabio Animati, Marco Berettoni, Mario Bigioni, Giovanni Capranico, Anna Maria Casazza, Claudia Caserini, Amalia Cipollone, Michelandrea De Cesare, Maurizio Franciotti, Paolo Lombardi, Andrea Madami, Stefano Manzini, Edith Monteagudo, Donatella Polizzi, Graziella Pratesi, Sabina C. Righetti, Carmela Salvatore, Rosanna Supino, Franco Zunino* Background: Although doxorubicin remains one of the most effective agents for the treatment of solid tumors, there is an intensive effort to synthesize doxorubicin analogues (compounds with similar chemical structures) that may have improved antitumor properties. We have synthesized a novel doxorubicin disaccharide analogue (MEN 10755) and have characterized some of its relevant biochemical, biologic, and pharmacologic properties. Methods: The antitumor activity of this compound (MEN 10755) was studied in a panel of human tumor xenografts, including xenografts of A2780 ovarian tumor cells, MX-1 breast carcinoma cells, and POVD small-cell lung cancer cells. MEN was compared with doxorubicin according to the optimal dose and schedule for each drug. The drug s cytotoxic effects, induction of DNA damage, and intracellular accumulation were studied in A2780 cells. DNA cleavage mediated by the enzyme topoisomerase II was investigated in vitro by incubating fragments of simian virus 40 DNA with the purified enzyme at various drug concentrations and analyzing the DNA cleavage-intensity patterns. Drug-induced apoptosis (programmed cell death) in tumors was determined with the use of MX-1 and POVD tumor-bearing athymic Swiss nude mice. Results: MEN was more effective than doxorubicin as a topoisomerase II poison and stimulated DNA fragmentation at lower intracellular concentrations. In addition, MEN exhibited striking antitumor activity in the treatment of human tumor xenografts, including those of the doxorubicin-resistant breast carcinoma cell line MX-1. Conclusions: The high antitumor activity of MEN in human tumor xenografts, including doxorubicin-resistant xenografts, and its unique pharmacologic and biologic properties make this disaccharide analogue a promising candidate for clinical evaluation. [J Natl Cancer Inst 1997;89: ] Doxorubicin still remains one of the most effective antitumor agents in clinical use and has the widest spectrum of antitumor activity (1). The observation concerning the minor molecular change differentiating doxorubicin from daunorubicin (the first discovered anthracycline clinically useful against malignant hematologic diseases) and the clinical success of doxorubicin in the treatment of solid tumors have stimulated an intensive effort in analogue synthesis with the aim of improving antitumor properties. Despite the formidable effort, expressed as time and expense devoted to the systematic modification of the parent compound(s), a truly better doxorubicin has not been found (2). Unfortunately, in cancer drug discovery and development, analogue synthesis often does not yield anything better than the original lead compounds. In early attempts to identify anthracyclines more effective than doxorubicin, both empiric and rational approaches have been combined. Although chemical *Affiliations of authors: F. Arcamone, F. Animati, M. Berettoni, M. Bigioni, A. Cipollone, M. Franciotti, P. Lombardi, A. Madami, S. Manzini, E. Monteagudo, C. Salvatore, Menarini Ricerche, Rome, Italy; G. Capranico, C. Caserini, M. De Cesare, D. Polizzi, G. Pratesi, S. C. Righetti, R. Supino, F. Zunino, Istituto Nazionale Tumori, Milan, Italy; A. M. Casazza, The Bristol-Myers Squibb Research Institute, Wallingford, CT. Correspondence to: Graziella Pratesi, Ph.D., Istituto Nazionale Tumori, Via Venezian 1, Milan, Italy. See Notes following References. Oxford University Press Journal of the National Cancer Institute, Vol. 89, No. 16, August 20, 1997 REPORTS 1217