HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by BREWER, C. A. Articles by MCASEY, M. E. Search for Related Content PubMed PubMed Citation Articles by BREWER, C. A. Articles by MCASEY, M. E.

Endoglin Expression as a Measure of Microvessel Density in Cervical Cancer

From the Department of Obstetrics and Gynecology, University of Illinois College of Medicine at Peoria, Peoria, Illinois; the Department of Obstetrics and Gynecology, Southern Illinois University School of Medicine, Springfield, Illinois; and the Department of Pathology, Memorial Medical Center, Springfield, Illinois.

Address reprint requests to: Cheryl A. Brewer, MD University of Illinois College of Medicine Department of Obstetrics and Gynecology Division of Gynecologic Oncology One Illini Drive, Box 1649 Peoria, IL 61656 E-mail: bactd{at}uic.edu

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
Objective: To evaluate endoglin, a membrane protein and member of the transforming growth factor ß-1 receptor complex, as an endothelial marker of angiogenesis in cervical cancer tissues.

Methods: Tumor tissue was collected from 31 surgically treated stage IB nonbulky (under 5 cm) cervical cancer subjects, and samples were fixed in formalin and embedded in paraffin. Endoglin was stained on 5-µm slide sections by the DAKO Catalyzed Signal Amplification method (DAKO Corporation, Carpinteria, CA). Factor VIII was stained by standard immunohistochemistry. Positively stained microvessels were counted in "hot spots" at 200x magnification. Clinical data were correlated with vessel counts by Spearman correlation. Mean differences in counts were tested using paired t tests.

Results: This staining method for endoglin identified significantly more vessels than the factor VIII method (mean 92 ± 45 versus 33 ± 16, P < .001). Endoglin and factor VIII counts correlated significantly with deep stromal invasion (Spearman rho 0.466 and 0.522, respectively, P < .05); however, only endoglin counts correlated significantly with lymph node metastases (rho = .495, P < .01).

Conclusion: Endoglin is stimulated in tumor angiogenesis and might be relatively more specific than commonly used endothelial markers. The endoglin system was more sensitive for staining capillaries in neoplastic cervical tissue, better predicted lymph node metastases, and should be widely applicable for the study of other tumors.

Characteristics of primary solid tumors such as histologic grade, tumor size, and regional lymph node status have been used for a long time as prognostic features to identify patients who need adjuvant therapy. Over several years, biologic and molecular tumor characterization has become an important aspect of cancer therapy. Angiogenesis directly affects growth and metastasis of solid tumors. A detailed characterization of that process is under way.¹ Tumors promote angiogenesis by secreting growth factors that stimulate endothelial migration, proliferation, proteolytic activity, and capillary morphogenesis. Newly formed blood vessels supply tumors with nutrients and oxygen, dispose of metabolic wastes, and generate paracrine stimuli that further promote tumor cell proliferation and invasiveness. Therefore, there is strong biologic plausibility that the number of microvessels, particularly those associated with tumor neoangiogenesis, correlate with clinical tumor behavior.

There is correlative clinical evidence that supports microvessel density as an important predictor of tumor behavior. The intensity of angiogenesis measured by density of microvessels has been associated with poor prognoses in invasive breast,^1–4 cervical,^5–8 ovarian,⁹ and endometrial¹⁰ cancer. Tumor vascularity in breast cancer has been a potential tool for identifying lymph node–negative women who might benefit most from adjuvant chemotherapy.^2,3 Evidence exists that microvessel counts might be one of the most important clinical prognosticators. Whereas some investigators have shown that microvessel density counts might be an independent prognostic factor, others have been unable to reproduce those findings.^11,12 Differences likely are caused by variability in staining and counting microvessels. The groundwork for analysis of microvessels and standardization of many aspects of those techniques recently was suggested in an international consensus paper.¹³ Most commonly, areas of highest vessel density, or hot spots, are identified by an experienced investigator. Various fields containing hot spots are counted, and an average or highest value is used. Different investigators have used varying magnification powers and field sizes for their studies. Different capillary antigens have been targeted in staining procedures, and various modifications in immunohistochemistry have been described.

One of three different monoclonal antibodies to endothelial cell antigens is used frequently to view tumor blood vessels. These are anti-CD 31, anti-CD 34, and the most common, factor VIII–related antigen, or von Willebrand’s factor, which are expressed heterogeneously on most endothelial capillaries and termed panendothelial markers. The present study targeted the endoglin molecule for staining tumor blood vessels. Endoglin is an integral membrane glycoprotein and a member of the transforming growth factor-beta 1 receptor complex. The endoglin or anti-CD 105 antibody, although present in all vasculature, binds preferentially to activated endothelial cells that are participating in angiogenesis^14,15; therefore, it is potentially a more specific marker for tumor neovascularization.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
All 31 patients with nonbulky (less than 5 cm by histologic measurements) stage IB cervical cancer who were treated by primary radical hysterectomy at Southern Illinois University Memorial Medical Center from June 1993 to December 1997 were included. The subjects had adjacent tissue sections cut for endoglin staining; however, only 28 had enough material available for additional adjacent sections for factor VIII staining. Institutional review board exempt approval was obtained from the Springfield Committee on Research in Human Subjects. As per Title 45, Part 46 of the Code of Federal Regulations "Protection of Human Subjects," we did not need additional informed consent to use the specimens in this study because only archived material was used.

Vascularization was assessed by immunohistochemical staining for endoglin and factor VIII of adjacent serial sections. Slides were cut at 5 µm from formalin-fixed, paraffin-embedded blocks. Staining was done with monoclonal primary antibodies to factor VIII (Clone F8/86, DAKO Corporation, Carpinteria, CA) and endoglin (Clone SN6h, DAKO Corporation). Liver and heart tissue were used as positive controls, and the primary antibody was replaced with mouse immunoglobulin (Ig)G (DAKO Corporation) for negative controls. Positive and negative controls were run for each batch of slides. Immunohisto-chemical staining for endoglin was completed with a catalyzed signal amplification system (DAKO Corporation), as described. Staining for factor VIII was done with a standard immunohistochemistry protocol, using 3,3'-diaminobenzadine tetrahydrochloride (DAB; Sigma Chemical, St. Louis, MO) as the chromogen.

The DAKO catalyzed signal amplification horseradish peroxidase system incorporates a signal amplification method based on peroxidase-catalyzed deposition of a biotinylated phenolic compound, tyramine, followed by a secondary reaction with streptavidin peroxidase. The specimens are first incubated with 3% hydrogen peroxide for 5 minutes to quench endogenous peroxidase activity, then incubated for 5 minutes with a protein block to suppress nonspecific binding of subsequent reagents. That is followed by incubation with an appropriately characterized and diluted mouse primary antibody or negative control reagent, which is followed by sequential 15-minute incubations with biotinyl tyramide, streptavidin-biotin-peroxidase complex, amplification reagent, and streptavidin-peroxidase. Staining is completed by 5-minute incubation with 3,3'-diaminobenzadine tetrahydrochloride (Sigma Chemical), which results in a brown-colored precipitate at the antigen site. All slides were counterstained with Mayer’s hematoxylin (Newcomer Supply, Middleton, WI).

For microvessel counting, slides were scanned at low magnification (10x/40x) to identify hot spots or areas of densest vascularity. An experienced pathologist selected the hot spots in each tumor section. The pathologist and the principal investigator reviewed the slides, which were presented in nonspecific order for counting. Both individuals were masked to primary antibody and patient identity. They simultaneously counted vessels, and each reported his results to a third investigator. A 200x magnification field of 0.25 mm² including and surrounding the hot spot was counted in each case. Any single cell or spot that stained specifically was counted as a vessel. As in previous reports, a visible lumen was not required.² The highest number of vessels counted by each investigator was recorded. The average of those numbers was used in the statistical analysis. Clinical data including tumor size, tumor grade, depth of invasion into the cervical stroma, and lymph node metastases were correlated with vessel counts. Correlation with clinical data was by Spearman correlation. Mean differences in vessel counts were tested with the use of paired t tests. All statistical analysis was done with SPSS (Statistical Package for Social Sciences; SPSS Inc., Chicago, IL).

	Results

Top Abstract Materials and Methods Results Discussion References

 
All 31 cervical tumor specimens stained well for endoglin. Among the 28 with sufficient material for comparative staining, three had respectively poor staining for factor VIII and were excluded. Therefore, tissue from 25 subjects had adjacent slide sets of endoglin and factor VIII staining, which were used for comparative analysis. The racial distribution was 20 white, four black, and one Asian subject. The mean and median ages were 37 years and 36 years, respectively. The age range was from 25 to 60 years. Mean gravity and parity were 2.6 and 2.3, respectively. All tumors were either squamous cell carcinoma (n = 18), adenocarcinoma (n = 6), or adenosquamous carcinoma (n = 1). Ten tumors had positive regional lymph node metastasis, whereas 15 tumors did not.

In all instances, endoglin identified more blood vessels than factor VIII. Endoglin and factor VIII counts for each subject and regional lymph node status are shown in Table 1. For the 25 subjects, the mean number of vessels identified in each tumor by the endoglin catalyzed signal amplification assay was significantly greater (m = 92 ± 45 standard deviation [SD]) than the mean number identified by staining for factor VIII (m = 33 ± 16 [SD], P < .001). Figures 1 and 2 are examples of endoglin and factor VIII staining in a cervical adenocarcinoma. Our technique relied on an experienced pathologist to select the hot spots, which were readily identified. Interobserver variability for counts within the selected field was less than 20% in all subjects and less than 10% in all but four subjects. Intraobserver variability was less than 10% in all cases.

View larger version (109K): [in this window] [in a new window]   Figure 1. Low magnification of cervical adenocarcinoma with factor VIII immunohistochemical staining.

View larger version (102K): [in this window] [in a new window]   Figure 2. Low magnification of same tumor with endoglin immunohistochemical staining.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
The staining procedure for endoglin, using a catalyzed signal amplification system (DAKO Corporation), elucidated more tumor vessels and better predicted lymph node metastases than factor VIII immunohistochemistry. Amplified endoglin staining potentially will enhance the reliability of microvessel density counts as a prognostic biomarker. Endoglin is a receptor for the TGF-ß 1 molecule and is upregulated in tumor neoangiogenesis. The endoglin antibody binds preferentially to activated endothelial cells that are participating in angiogenesis, such as in tumors, healing wounds, psoriasis, embryonic tissue, and stroke tissues. The endoglin antigen, as a marker for neoangiogenesis, has certain advantages over the common used panendothelial markers (CD 31, CD 34, factor VIII). Several studies^16–18 were published recently that examined endoglin expression in tumors, by immunohistochemistry. Two of those studies used fresh or snap-frozen tissue specimens because endoglin was unreactive in paraffin-embedded tissue.^13,14 One study used a histogold amplification detection system (Zymed Laboratories, South San Francisco, CA) to examine semiquantitatively endoglin expression in paraffin-embedded breast tissue.¹⁵

Antibodies against the panendothelial markers generally react well with endothelial cells of larger vessels, but their expression is sometimes diminished or lost in tumor microvessels. The monoclonal antibody to anti-factor VIII stains large vessels with high sensitivity and capillaries with variable and focal staining.¹⁶ Our experience with factor VIII staining was similar to that of previous reports. Factor VIII is also nonspecific because it is present on megakaryocytes, platelets, and some lymphatic endothelial cells.¹⁰ Some tumors, as in our study, do not stain at all for factor VIII.^4,11 Anti-CD 31 is reported to stain large and small vessels with equal intensity in normal and tumor tissue. However, some nonspecific staining is also seen as in plasma cells and other inflammatory cells.¹⁶ The reliability of anti-CD 31 antibody staining has been inconsistent between laboratories.⁵ The addition of a microwave antigen retrieval step has reduced, but not eliminated, that inconsistency.¹⁸

In our study, endoglin staining was consistent and satisfactory for staining microvessels in all specimens. Endoglin stained small vessels with high sensitivity but stained large vessels with some variable and focal results. In one comparative study between endoglin (CD 105) and CD 31, blood vessels in or around the tumor tissues stained intensely for endoglin. Those same vessels stained weakly or did not stain at all for CD 31.¹⁶ For normal tissues in that study, endoglin stained only a portion (20%) of blood vessels highlighted by CD 31,¹⁶ which suggests that endoglin is detected preferentially in vessels undergoing neoangiogenesis. Endoglin expression has been reported in proerythroblasts, activated monocytes, and lymphocytes in childhood leukemia but is absent from circulating monocytes.¹⁸ Experience to date with endoglin in solid tumors suggests that staining is very selective for the blood vessel endothelium. Endoglin reacts specifically with angiogenic endothelial cells, so it should result in reduced false-positive staining compared with other commonly used antigens.

The detection method we used also was superior in sensitivity to standard immunohistochemistry. An amplification method was needed because endoglin cannot be detected readily by standard immunohistochemistry in paraffin-embedded tissue sections. However, factor VIII is detectable by standard immunohistochemistry, and an amplification technique would have increased nonspecific staining. Comparisons to standard labeled streptavidin-biotin methods have shown that the DAKO catalyzed signal amplification horseradish peroxidase system (DAKO Corporation) is approximately 50 times greater in sensitivity, which detects extremely small quantities of endoglin. Not only is this technique more reliable and more sensitive, it is potentially more specific for identifying activated and leaky capillaries believed to be important in tumor metastases, an idea supported by the ability of the endoglin technique to predict lymph node metastases in this small series of early stage cervical cancers.

Microvessel density has been an independent prognostic factor in cervical cancer and other malignancies. Vessel counts might help identify high-risk patients for whom adjuvant or combined modality treatments are appropriate. As biomolecular advances are made in antiangiogenic therapy, an angiogenic panel that includes specific microvessel density information for individual patients might prove useful. It has been predicted that one day physicians might be as concerned about this angiogenic information as they are today about clinical staging and histologic parameters.¹⁹

One important aspect of this technique, which proved to be very sensitive, is that it can be done easily on formalin-fixed, paraffin-embedded tissues. We believe this study reinforces other data that suggest that microvessel density counts are likely to provide clinically useful information that would be applicable not only to cervical cancer, but other solid tumors. For microvessel density counting to become widely acceptable as a clinical technique, practical and reproducible staining and counting methods are necessary. Confirmation of clinical correlation and prognostic use in multiple well-controlled studies will be necessary.

	Footnotes

 
Supported by departmental funds, Southern Illinois School of Medicine and University of Illinois College of Medicine at Peoria.

PII S0029-7844(00)00864-4

Received November 5, 1999. Received in revised form February 15, 2000. Accepted March 2, 2000.

	References

Top Abstract Materials and Methods Results Discussion References

 
1. Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis: Correlation in invasive breast carcinoma. N Engl J Med 1991;324:1–8.[Abstract]

2. Weidner N, Folkman J, Pazza F, Bevilacqua P, Allred EN, Moore DH, et al. Tumor angiogenesis: A new significant and independent prognostic indicator in early stage breast carcinoma. J Natl Cancer Inst 1992;84:1875–87.[Abstract/Free Full Text]

3. Martin L, Holcombe C, Green B, Leinster SJ, Winstanley J. Is a histological section representative of whole tumor vascularity in breast cancer? Br J Cancer 1997;76:40–3.[Medline]

4. Guidi AJ, Fischer L, Harris JR, Schnitt SJ. Microvessel density and distribution in ductal carcinoma in situ of the breast. J Natl Cancer Inst 1994;86:614–9.[Abstract/Free Full Text]

5. Smith-McCune KK, Weidner N. Demonstration and characterization of the angiogenic properties of cervical dysplasia. Cancer Res 1994;54:800–4.[Abstract/Free Full Text]

6. Obermair A, Wanner C, Bilgi S, Speiser P, Kaider A, Reinthaller A, et al. Tumor angiogenesis in stage IB cervical cancer: Correlation of microvessel density with survival. Am J Obstet Gynecol 1998;178: 314–9.[Medline]

7. Abulafia O, Triest WE, Sherer DM. Angiogenesis in malignancies of the female genital tract. Gynecol Oncol 1999;72:220–31.[Medline]

8. Bremer GL, Tiebosch AT, van der Putten HW, Schouten HJ, de Haans J, Arends JW. Tumor angiogenesis: An independent prognostic parameter in cervical cancer. Am J Obstet Gynecol 1996;174: 126–31.[Medline]

9. Brustmann H, Riss P, Naude S. The relevance of angiogenesis in benign and malignant epithelial tumors of the ovary: A quantitative histological study. Gynecol Oncol 1997;67:20–6.[Medline]

10. Abulafia O, Treist WE, Sherer DM, Hansen CC, Ghezzi F. Angiogenesis in endometrial hyperplasia and stage I endometrial carcinoma. Obstet Gynecol 1995;86:479–85.[Abstract]

11. Axelsson A, Clark W. Tumor angiogenesis as a prognostic assay for invasive ductal breast carcinoma. J Natl Cancer Inst 1995;87: 997–1008.[Abstract/Free Full Text]

12. Abulafia O, Triest WE, Sherer DM. Angiogenesis in squamous cell carcinoma in situ and microinvasive carcinoma of the uterine cervix. Obstet Gynecol 1996;88:927–32.[Abstract]

13. Vermeulen PB, Gasparini G, Fox SB, Toi M, Martin L, McCulloch P, et al. Quantification of angiogenesis in solid human tumors: An international consensus on the methodology and criteria of evaluation. Eur J Cancer 1996;32A:2474–84.

14. Westphal JR, Willems HW, Schalkwijk CJM, Ruiter DJ, DeWaal RMW. Characteristics and possible function of endoglin. A TGF-ß binding protein. Behring Inst Mitt 1993;92:15–22.

15. Seon BK, Matsuno F, Haruta Y, Kondo M, Bareas M. Long-lasting complete inhibition of human solid tumors in SCID mice by targeting endothelial cells of tumor vasculature with antihuman endoglin immunotoxin. Clin Cancer Res 1997;3:1031–44.[Abstract]

16. Wang JM, Kumar S, Pye D, Haboubi N, Al-Nakib L. Breast carcinoma: Comparative study of tumor vasculature using two endothelial cell markers. J Natl Cancer Inst 1994;86:386–8.[Free Full Text]

17. Kumar S, Ghellal A, Li C, Byrne G, Haboubi N, Wang JM, et al. Breast carcinoma: Vascular density determined using CD 105 antibody correlates with tumor prognosis. Cancer Res Feb 1999;59: 856–61.

18. Bodey B, Bodey B Jr, Seigel SE, Kaiser HE. Over-expression of endoglin (CD 105): A marker of breast carcinoma-induced neovascularization. Anticancer Res 1998;18:3621–8.[Medline]

19. Hayes DF. Angiogenesis and breast cancer. Hematol Oncol Clin North Am 1994;8:51–71.[Medline]

This article has been cited by other articles:

				G. Bellone, D. Solerio, L. Chiusa, G. Brondino, A. Carbone, A. Prati, T. Scirelli, M. Camandona, G. Palestro, and M. Dei Poli Transforming Growth Factor-{beta} Binding Receptor Endoglin (CD105) Expression in Esophageal Cancer and in Adjacent Nontumorous Esophagus as Prognostic Predictor of Recurrence Ann. Surg. Oncol., November 1, 2007; 14(11): 3232 - 3242. [Abstract] [Full Text] [PDF]

				C. Liu, C. Dickinson, J. Shobe, F. Donate, W. Ruf, and T. Edgington A hybrid fibronectin motif protein as an integrin targeting selective tumor vascular thrombogen Mol. Cancer Ther., July 1, 2004; 3(7): 793 - 801. [Abstract] [Full Text] [PDF]

				S. E. DUFF, C. LI, J. M. GARLAND, and S. KUMAR CD105 is important for angiogenesis: evidence and potential applications FASEB J, June 1, 2003; 17(9): 984 - 992. [Abstract] [Full Text] [PDF]

				E. Fosslien Molecular Pathology of Cyclooxygenase-2 in Cancer-induced Angiogenesis Ann. Clin. Lab. Sci., October 1, 2001; 31(4): 325 - 348. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by BREWER, C. A. Articles by MCASEY, M. E. Search for Related Content PubMed PubMed Citation Articles by BREWER, C. A. Articles by MCASEY, M. E.