Anatomic Pathology / CD105 IN BREAST CARCINOMAS

Transcription

1 Anatomic Pathology / CD105 IN BREAST CARCINOMAS CD105 Expression Is a Marker of High Metastatic Risk and Poor Outcome in Breast Carcinomas Correlations Between Immunohistochemical Analysis and Long-Term Follow-up in a Series of 929 Patients Jean-Philippe Dales, MD, 1 Stephane Garcia, MD, 1 Pascal Bonnier, MD, 2 Florence Duffaud, MD, 3 Lucile Andrac-Meyer, MD, 1 Olivier Ramuz, MD, Marie-Noëlle Lavaut, MD, 1 Claude Allasia, PhD, 1 and Colette Charpin, MD 1 Key Words: CD105; Endoglin; Immunocytochemistry; Frozen tissue; Breast carcinoma; Prognosis DOI: /1KF54L6RB625556W Abstract CD105 (endoglin) is expressed significantly in activated endothelial cells in culture and in tumor microvessels. Quantification of CD105 immunocytochemical expression that may be clinically relevant has not been accurately evaluated. We studied CD105 expression on frozen tissue sections by using immunohistochemical assays in a series of 929 patients and correlated the findings with long-term follow-up (median, 11.3 years). Univariate (Kaplan-Meier) analysis showed that the number of CD105+ microvessels (cutoff, 15 vessels) correlated significantly with poor overall survival among all patients (P =.001). This correlation was less significant in node-negative patients (P =.035). Marked CD105 expression also correlated with a high risk for metastasis among all patients (P =.006) and among node-negative patients (P =.001). Multivariate analysis (Cox model) identified CD105 immunodetection as an independent prognostic indicator. Our results suggest that immunohistochemical expression of CD105 has practical clinical relevance for identifying node-negative patients with a poor prognosis. Moreover, immunodetection of CD105 also may be considered a potential tool for selecting patients who could benefit from specific antiangiogenic therapy, using anti-cd105 conjugates. Angiogenesis is a complex multistep process required for tumor growth and metastasis. It involves endothelial cell migration and proliferation, microvessel differentiation and anastomosis, and extracellular matrix remodeling. 1-3 The quantification of angiogenesis in human solid tumors can be used as a prognostic marker in particular circumstances. 4,5 In invasive breast carcinomas, extensive neovascularization has been reported to be an indicator of a poor prognosis. 6 In addition, the tumor microvasculature may constitute a relevant target for antiangiogenic therapy. 3 Intratumoral microvessel density has been previously studied using panendothelial markers such as CD34, CD31, and von Willebrand factor. 3,7 Anti-CD105 monoclonal antibody was found to bind preferentially to activated endothelial cells in vitro, 8-11 and CD105 was expressed strongly in endothelial cells of tumor blood vessels and not in most normal tissue. 10,11 Moreover, antihuman CD105 immunotoxin inhibited the growth of human breast tumor xenografts in immunodeficient mice. 8,12 Since the expression of CD105 in breast carcinoma reflects neoangiogenesis and endothelial activation as well, CD105 detection may be useful not only in terms of prognostic significance, 13 but also in the potential selection of candidates for specific antiangiogenic therapy with anti-cd105 antibodies. 12 Several immunohistochemical studies in breast cancer using monoclonal antibodies raised against panendothelial antigens including CD34, CD31, and anti von Willebrand factor have been reported, 3-7 but only 1 studied the correlation between the immunohistochemical expression of CD105 and clinical outcome. 13 In that study, frozen tissue sections from 106 breast carcinomas were tested with anti-cd105 and Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W 1

2 Dales et al / CD105 IN BREAST CARCINOMAS anti-cd34 antibodies and showed that marked CD105 expression (and not CD34 expression) correlated with poor survival. 13 However, the authors also pointed out that their findings required further investigation and confirmation in larger cohorts of patients and a follow-up period of longer than 5 years. We studied CD105 expression in frozen tumor sections from a large series (n = 929) of patients with breast carcinoma and correlated the results with patient outcome after long-term follow-up (median, 11.3 years). We found that immunohistochemical expression of CD105 is a marker of high metastatic risk and poor outcome. Materials and Methods Cases The study included 929 patients aged 25 to 81 years (mean ± SD = 56.1 ± 13.3 years) with breast carcinoma who underwent surgery from January1986 to December They did not receive chemotherapy or hormone therapy before surgery. The patients underwent axillary node excision combined with wide local excision with margin clearance or mastectomy in the Department of Oncologic Gynecology, Hospital of the Conception, Marseilles, France. All specimens were examined in the same department of pathology by experienced pathologists (C.C., L.A.M., S.G.). The follow-up periods ranged from 6 to 15 years (median, 11.3 years). The records for 2001 showed that 340 patients (36.6%) experienced relapse, among whom 228 died (median survival, 82 months), and 589 (63.4%) were disease-free. Overall survival was calculated as the period from surgery until date of death. Metastasis-free and recurrence-free survival were calculated as the period from surgery until date of metastasis or recurrence. The mean ± SD tumor size was 20.7 ± 13.8 mm; 23.2% of tumors were 10 mm or smaller, 41.2% were between 10 and 20 mm, 22.5% were more than 20 mm but not more than 30 mm, and 15.1% were larger than 30 mm. Histologic examination of surgical specimens was performed on paraffin-embedded sections stained with hematoxylin, eosin, and safranin O. Tumors were classified as ductal carcinoma (n = 632 [68.0%]), lobular carcinoma (n = 167 [18.0%]), and other types, including mucinous, medullary, papillary, apocrine, and mixed (n = 130 [14.0%]). Tumor grades were as follows: 1, 24.0% (n = 223); 2, 51.0% (n = 474); and 3, 25.0% (n = 232). Tumor grading, initially assessed by using the grading methods of Scarf, Bloom, and Richardson, was reevaluated according to the Elston and Ellis 15 method. A mean ± SD of 14.7 ± 4.3 lymph nodes was found in axillary node excision, and 506 (54.5%) of 929 patients were node-negative. Immunostaining Procedure and Quantification of CD105-Immunostained Microvessels Fresh tissue fragments were sampled by pathologists (C.C, L.A.M., S.G.) immediately after intraoperative diagnosis. The fragment size varied according to the tumor size (average size, 5 mm long, 4 mm wide, 3 mm thick). Fragments were obtained from dense tumor areas that lacked grossly visible adipose tissue. They then were dipped promptly in liquid nitrogen and stored at 80 C in the laboratory tumor library. Immunodetection studies were performed on 5-µm-thick sections (cryostat, Leica CM 3050, Leica Microsystèmes, Rueil Malmaison, France). An immunoperoxidase procedure was performed using monoclonal (8 E11) 16 mouse (1:20 dilution) antihuman CD105 (Novocastra, Tebu Le Perray en Yvelynes, France) and the Ventana Gene II device with Ventana kits (Ventana, Tucson, AZ). The microvessel count was assessed in the most vascular areas (so-called hot spots) in 4 fields (0.4 mm 2 ) using a 20 objective (1.060 µm field diameter) with a Zeiss Axioplan microscope (Carl Zeiss International, Göttingen, Germany). The mean value of the vessel count in the 4 fields was retained as the final value. The CD105-immunostained surface was evaluated for the percentage of the total tumor area with the same 20 objective. Statistical Analysis The Kaplan-Meier method was used to analyze diseasefree and overall survival rates. The difference between curves was evaluated with the Mantel Cox test (or log-rank test) for observations regarding censored survival or events. All computations were done with NCSS 2000 statistical software (NCSS, Kaysville, UT). Both the percentage of CD105- stained tumor surface and the vessel count were stratified and correlated with major events during the disease course (distant metastasis or local recurrence) and with the overall survival rate to define immunohistochemical thresholds of prognostic significance. The optimal CD105 cutoff point of positive surface endowed with prognostic significance was determined after statistical validation. 17 The effect of multiple factors on survival was tested with a Cox multivariate proportional hazards model (NCSS 2000). The assumptions of proportional hazards were evaluated by examining data on the log cumulative hazard that were stratified by the histoprognostic factors used in the model (tumor size, histologic grade and type, nodal status) and by examining residual data vs survival time. All P values were based on 2-sided testing. 375 Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W

3 Anatomic Pathology / ORIGINAL ARTICLE Results CD105 Distribution in Tissue Sections CD105 positivity was observed in the endothelial cell membrane and cytoplasm within microvessels Image 1 and Image 2. The CD105 immunostaining was distributed regularly along the small vessels as thin, linear deposits. CD105 immunostaining was more suitable for vessel counting than was CD31 immunostaining (not shown). Image 1 Immunoperoxidase staining of frozen tissue sections in an invasive breast ductal carcinoma. The antibody underlines the fine capillary network of microvasculature in breast cancer tissue (CD105 monoclonal antibody, (130). Percent No. of CD105+ Vessels Figure 1 Distribution of CD105+ immunostaining (number of microvessels stained) in frozen sections from 929 patients with breast carcinoma. The distribution of microvessel counts among the 929 tumors is shown in Figure 1 (mean, 13.9; median, 12.1) and the CD105-immunostained surface in Figure 2 (mean ± SD = 1% ± 3.7%; median, 11.1%). Univariate (Kaplan-Meier and Log-Rank) Analysis and CD105 Prognostic Significance The percentage of CD105-immunostained tumor cell surface (cutoff point, 12%) correlated with overall survival (P =.03) Figure 3. Tumors with a large CD105+ surface Image 2 Immunoperoxidase staining of cryostat tissue sections at a higher magnification of Image 1 (CD105 monoclonal antibody, (480). Percent CD105 Immunostained Surface (%) Figure 2 Distribution of CD105+ immunostaining in frozen sections from 929 patients with breast carcinoma. Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W 3

4 Dales et al / CD105 IN BREAST CARCINOMAS Cumulative Survival CD105 <12% CD105 12% Figure 3 Kaplan-Meier survivorship plot of overall survival for 929 patients with breast carcinoma. A greater risk of death was found for patients with tumors in which 12% or more of the surface was CD105+ (P =.03). were associated with poorer survival than those that exhibited a small CD105+ surface. The validation of the optimal cutoff point 17 for CD105-immunostained surface is shown in Figure 4. However, when patients were divided into nodepositive (n = 423) and node-negative (n = 506) subsets Table 1, CD105 immunoexpression did not retain prognostic significance. Among the total population, a CD105-immunostained surface larger than 12% correlated with early and widespread metastasis (P =.004); the same was not true in node-negative patients (data not shown). The number of microvessels (cutoff, 15 microvessels) correlated with overall survival (P =.001) Figure 5. Tumors with a high number of CD105+ microvessels were associated with a poorer survival than those that exhibited a small number. The validation of the cutoff point for the number of CD105+ microvessels is shown in Figure 4. Table 1 Overall Survival of Patients Correlates With CD105- Immunostained Surface on Frozen Sections * CD105 Immunostained Surface <12 12 All patients (n = 929) Dead 139/626 (22.2) 89/303 (29.4) Alive 487/626 (77.8) 214/303 (70.6) Node-negative patients (n = 506) Dead 62/340 (18.2) 44/166 (26.5) Alive 278/340 (81.8) 119/166 (71.5) * Data are given as number/subtotal (percentage). The median follow-up was 11.3 years. P Value Log Rank All patients, number of vessels Node-negative patients, number of vessels All patients, positive surface staining No. of CD105+ Vessels Figure 4 P value curve (log-rank test) showing optimal cutoff points for CD105 immunostaining for overall survival according to Altman et al. 17 When patients were divided into node-positive and nodenegative subsets Table 2, the number of CD105+ microvessels (cutoff, 15 microvessels) retained prognostic significance (P =.035) in the node-negative subgroup Figure 6. The number of CD105+ vessels (15 or more) also correlated with a higher risk of metastasis (P =.0002) Figure 7 and relapse (metastasis and local recurrence; P =.0006) Figure 8. The same correlation was observed in the subgroup of node-negative patients (P =.001) Figure 9 and Figure 10. Multivariate (Cox Model) and Prognostic Significance In multivariate analysis, the count of CD105+ vessels (not the percentage of CD105+ surface) proved to be a prognostic indicator that exhibited a predictive value stronger than tumor size, grade, and histologic type, even in nodenegative patients Table 3 and Table 4. Table 2 Overall Survival of Patients Correlates With Microvessel Density Evaluated by CD105-Immunostained Microvessels on Frozen Sections * No. of CD105+ Microvessels <15 15 All patients (n = 929) Dead 137/631 (21.7) 98/298 (32.9) Alive 494/631 (78.3) 200/298 (67.1) Node-negative patients (n = 506) Dead 64/352 (18.2) 42/154 (27.3) Alive 288/352 (81.8) 112/154 (72.7) * Data are given as number/subtotal (percentage). The median follow-up was 11.3 years. 377 Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W

5 Anatomic Pathology / ORIGINAL ARTICLE Cumulative Survival Cumulative Survival Figure 5 Kaplan-Meier survivorship plot of overall survival for 929 patients with breast carcinoma. A greater risk of death was found for patients with tumors in which 15 or more vessels were CD105+ (P =.001). Discussion In many tumor types, including breast carcinoma, marked features of angiogenesis have correlated with a high rate of metastasis. 1,3,5,16-20 Intratumoral microvessels can be identified by using an immunocytochemical method to stain endothelial cells with so-called panendothelial cell markers such as anti-cd34, anti-cd31, and anti von Willebrand factor. However endothelial cells from tumor and normal tissue are heterogeneous in terms of reactivity with these vascular markers, which may not stain all tumor blood vessels to the same degree. 5-7 In this respect, antibodies to CD34 and von Willebrand factor used as panendothelial markers generally react Cumulative Metastasis Figure 6 Kaplan-Meier survivorship plot of overall survival for node-negative patients with breast carcinoma. A greater risk of death was found for patients with tumors in which 15 or more vessels were CD105+ (P =.035). strongly with endothelial cells in large blood vessels, but their expression is weak or absent in microvessels from most normal tissue and from many tumoral tissues. 21 Moreover, anti-cd34 and anti-cd31 may not be entirely specific for endothelial cells. 1,6,7,21 This might explain why anti-cd34 immunohistochemical expression failed to correlate with survival in the study by Kumar et al 13 or why anti-cd31 immunohistochemical expression proved to be of limited prognostic significance in node-negative patients. 22 For these reasons, antipanendothelial antibodies may not be ideal reagents for visualizing tumor blood vessels. Therefore the use of antibodies specifically reacting with endothelial cells of angiogenic tissue and not with most normal tissue has the Cumulative Disease Figure 7 Kaplan-Meier univariate analysis showing a higher risk of metastasis in 929 patients with tumors in which 15 or more vessels were CD105+ (P =.0002). Figure 8 Kaplan-Meier univariate analysis showing a higher risk of metastasis, recurrence, or both in patients with tumors in which 15 or more vessels were CD105+ (P =.0006). Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W 5

6 Dales et al / CD105 IN BREAST CARCINOMAS Cumulative Metastasis (Node-Negative) Figure 9 Kaplan-Meier univariate analysis showing a higher risk of metastasis in node-negative patients with tumors in which 15 or more vessels were CD105+ (P =.001). Table 3 Cox Proportional Hazards Regression Model for Overall Survival in All Patients and Node-Negative Patients Variable CD105+ surface, all patients Tumor grade Histologic type Histologic size CD105, No. of positive vessels, all patients Tumor grade Histologic type Histologic size CD 105, No. of vessels, node-negative patients Tumor grade Histologic type Histologic size Table 4 Cox Proportional Hazards Regression Model for Metastasis and Metastasis and Recurrence in All Patients and Node-Negative Patients Metastasis CD 105, number of vessels (all patients) Histologic grade Tumor size Histologic type CD105, No. of vessels, node-negative patients Histologic grade Tumor size Histologic type Metastasis and recurrence (disease) CD105, No. of vessels, all patients Histologic grade Histologic type Tumor size P P P Value Log Rank Metastasis (all patients) Metastasis (node-negative patients) Disease (all patients) No. of CD105+ Vessels Figure 10 P value curve (log-rank test) showing optimal cutoff points for CD105 immunostaining for risk of metastasis and metastasis and/or recurrence. 17 advantage that normal blood vessels entrapped within a tumor are not be stained, thereby reducing the probability of false-positive staining. Endoglin (CD105) is a cell membrane glycoprotein overexpressed on highly proliferation-activated endothelial cells in culture, and CD105 is the receptor for the transforming growth factor beta superfamily. 10,11 Immunocytochemical studies using anti-cd105 antibodies have shown that anti- CD105 reacts with neovessels in neoplastic tissues. 8-11,23 However, only 1 study 13 dealing with the prognostic significance of CD105 expression in breast carcinomas has been reported in the literature. This study included relatively few patients (n = 106) with a short follow-up. In the present study, we found in a substantially larger (n = 929) series of breast carcinomas with a substantially longer follow-up that marked expression of CD105 in tumor microvasculature correlated with a poor prognosis in node-negative patients. Moreover, multivariate analysis revealed anti-cd105 expression in microvessels as an independent prognostic indicator. Our findings not only confirm the preliminary results of Kumar et al, 13 but also demonstrate the potential practical interest of CD105 immunodetection for identifying nodenegative patients with a likelihood of shorter survival and a higher risk of metastasis. In this respect, the count of CD105+ microvessels proved to be more predictive than the percentage of CD105+ tumor surface. Since angiogenesis is essential for tumor growth and metastasis, tumor microvasculature is an appropriate target for specific therapy. 3 Although the response to first-generation antiangiogenic therapy in trials did not show much tumor regression, in xenograft studies, substantial tumor regression was observed (see review by Fox et al 3 ). In particular, in preformed solid tumors induced by inoculation of 379 Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W

7 Anatomic Pathology / ORIGINAL ARTICLE immunodeficient mice with MCF-7 human breast cancer cells, a long-lasting complete regression of tumors was observed after administration of anti-cd105 monoclonal antibody conjugates. 12 This finding strongly supports the beneficial potential of these conjugates in clinical application and, more specifically, after radiotherapy. 24 To effectively target tumor vasculature and to avoid side effects in clinical trials, it is imperative that reagents be developed to selectively destroy tumor-associated vasculature without severely damaging normal tissues. Therefore, an appropriate proliferation-associated antigene on endothelial cells could be an ideal target for antiangiogenic therapy with negligible side effects. In this respect, our immunohistochemical procedure on frozen sections enabling identification of tumors with strong CD105 expression may be relevant for selecting patients who could benefit from antiangiogenic therapy with CD105 conjugates, similar to the way in which c-erbb-2 immunodetection is used to select patients for specific therapy with trastuzumab. Our study showed that CD105 immunoexpression in breast carcinoma predicts a poor outcome, especially in node-negative patients. Ultimately, CD105 immunohistochemical evaluation in this particular group may be useful for patient monitoring. From the Departments of 1 Pathology, Nord Hospital, 2 Oncologic Gynecology, Hospital of the Conception, and 3 Medical Oncology, Timone Hospital, CHU, Marseilles, France. Supported by the GEFLUC. Address reprint requests to Dr Charpin: Faculté de Médecine Secteur Nord, Service d Anatomie et de Cytologie Pathologiques, Bd Pierre Dramard, Marseille Cedex 20 France. References 1. Folkman J. Clinical applications of research on angiogenesis. N Engl J Med. 1995;333: Ellis LM, Fidler IJ. Angiogenesis and metastasis. Eur J Cancer. 1996;14: Fox SB, Gasparini G, Harris AL. Angiogenesis: pathological, prognostic, and growth-factor pathways and their link to trial design and anticancer drugs. Lancet Oncol. 2001;2: Toi M, Taniguchi T, Yamamoto Y, et al. Clinical significance of the determination of angiogenic factors. Eur J Cancer. 1996;32A: Vermeulen PB, Gasparini G, Fox SB, et al. Quantification of angiogenesis in solid human tumours: an international consensus on the methodology and criteria of evaluation. Eur J Cancer. 1996;32A: Gasparini G. Clinical significance of the determination of angiogenesis in human breast cancer: update of the biological background and overview of the Vicenza studies. Eur J Cancer. 1996;32A: Gasparini G. Angiogenesis research up to 1996: a commentary on the state of art and suggestions for future studies. Eur J Cancer. 1996;14: Seon BK, Matsuno F, Haruta Y, et al. 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: Wang JM, Kumar S, Pye D, et al. A monoclonal antibody detects heterogeneity in vascular endothelium of tumours and normal tissues. Int J Cancer. 1993;54: Kumar P, Wang JM, Bernabeu C. CD105 and angiogenesis. J Pathol. 1996;178: Fonsatti E, Del Vecchio L, Altomonte M, et al. Endoglin: an accessory component of the TGF-beta-binding receptorcomplex with diagnostic, prognostic, and bioimmunotherapeutic potential in human malignancies. J Cell Physiol. 2001;188: Matsuno F, Haruta Y, Kondo M, et al. Induction of lasting complete regression of preformed distinct solid tumors by targeting the tumor vasculature using two new anti-endoglin monoclonal antibodies. Clin Cancer Res. 1999;5: Kumar S, Ghellal A, Li C, et al. Breast carcinoma: vascular density determined using CD105 antibody correlates with tumor prognosis. Cancer Res. 1999;59: Bloom HG, Richardson WW. Histological grading and prognosis in breast cancer. Br J Cancer. 1957;11: Elston CN, Ellis IO. Pathological prognostic factors in breast cancer, I: the value of the histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology. 1991;19: Lastres P, Bellon T, Cabanas C, et al. Regulated expression on human macrophages of endoglin, an Arg-Gly-Asp-containing surface antigen. Eur J Immunol. 1992;22: Altman DG, Lausen B, Sauerberi W, et al. Dangers of using optimal cutpoints in the evaluation of prognostic factors. J Natl Cancer Inst. 1994;86: Weidner N, Folkman J, Pozza F, et al. Tumor angiogenesis: a new significant and independent prognostic indicator in early-stage breast carcinoma. J Natl Cancer Inst. 1992;84: Martin L, Green B, Renshaw C, et al. Examining the technique of angiogenesis assessment in invasive breast cancer. Br J Cancer. 1997;76: Gasparini G, Pozza F, Harris AL. Evaluating the potential usefulness of new prognostic and predictive indicators in node-negative breast cancer patients. J Natl Cancer Inst. 1993;85: Wang JM, Kumar S, Pye D, et al. Breast carcinoma: comparative study of tumor vasculature using two endothelial cell markers. J Natl Cancer Inst. 1994;86: Charpin C, Garcia S, Bouvier C, et al. CD31/PECAM automated and quantitative immunocytochemical assays in breast carcinomas: correlation with patient follow-up. Am J Clin Pathol. 1997;107: Bodey B, Bodey B Jr, Siegel S, et al. Over-expression of endoglin (CD105): a marker of breast carcinoma induced neo-vascularization. Anticancer Res. 1998;18: Wang JM, Kumar S, Van Agthoven A, et al. Irradiation induces up-regulation of E9 protein (CD105) in human vascular endothelial cells. Int J Cancer. 1995;62: Am J Clin Pathol 2003;119: DOI: /1KF54L6RB625556W 7