1263 P-Cadherin Expression in Breast Carcinoma Indicates Poor Survival Alejandro Peralta Soler, M.D., Ph.D.1 Karen A. Knudsen, Ph.D.1 Hernando Salazar, M.D., M.P.H.2 Aaron C. Han, M.D., Ph.D.2 Albert A. Keshgegian, M.D., Ph.D.3 1 The Lankenau Medical Research Center, Wynnewood, Pennsylvania. 2 Department of Pathology, The Reading Hospital and Medical Center, Reading, Pennsylvania. 3 Department of Pathology, The Lankenau Hospital, Wynnewood, Pennsylvania. Presented in part as “Expression of E- and PCadherin in Breast Tumors” at the XXII Congress of the International Academy of Pathology and the 13th World Congress of Academic and Environmental Pathology, Nice, France, October 18 –23, 1998, and as “P-Cadherin Expression in Breast Cancer Indicates Poor Survival” at the 90th annual meeting of the American Association for Cancer Research, Philadelphia, PA, April 10 –14, 1999. Supported by the John S. Sharpe Research Foundation of the Bryn Mawr Hospital. The authors thank Gary D. Harner, Gwendolyn Gilliard, and Elaine Johnston for technical assistance; Dr. Mike Free, Greg Maislin, and Jacqueline Cater for statistical analysis; and the Editorial Office of the Lankenau Medical Research Center for helping with the article preparation. They also thank Drs. Margaret J. Wheelock and Keith R. Johnson of the University of Toledo, Ohio, for providing anti-P-cadherin monoclonal antibody clone 6A9 and anti-g-catenin monoclonal antibody clone 15F11. Address for reprints: Alejandro Peralta Soler, M.D., Ph.D., The Lankenau Medical Research Center, 100 Lancaster Avenue, Wynnewood, PA 19096. Received November 30, 1998; revision received April 28, 1999; accepted April 28, 1999. © 1999 American Cancer Society BACKGROUND. The cadherin family of cell-cell adhesion molecules and their associated proteins, the catenins, are essential to embryonic development and the maintenance of adult tissues. During development, the homotypic interaction of a particular cadherin with an identical cadherin expressed on a neighboring cell results in the sorting of cells to form distinctive tissues. Cadherins are believed to be tumor suppressors, and their altered expression and function have been associated with tumor development. METHODS. The authors examined the expression of P-cadherin, E-cadherin, and N-cadherin, and a-catenin and b-catenin in 183 cases of invasive breast carcinoma by immunohistochemistry on paraffin sections using specific antibodies and a steam-based antigen retrieval method. RESULTS. P-cadherin was positive in 95 cases and negative in 88 cases of breast carcinoma. Positive P-cadherin expression in breast carcinoma showed a strong correlation with poor patient prognosis. Five years after surgery, 90% of the patients with P-cadherin negative tumors were alive in contrast to only 59% of patients with P-cadherin positive tumors. The difference in survival reached statistical significance (P 5 0.0001) as early as 2 years after surgical treatment. Expression of N-cadherin, a-catenin, and b-catenin did not correlate with patient survival. Multivariable statistical analyses of the data showed that expression of P-cadherin was independent of tumor size and lymph node metastases, but correlated inversely with estrogen/progesterone receptor status. In ductal carcinomas, positive P-cadherin expression correlated with a higher histologic grade. In contrast, expression of E-cadherin was low in high grade ductal carcinomas but negative tumors were uncommon. Negative or low E-cadherin expression did not correlate with poor survival. In lobular carcinomas, E-cadherin expression frequently was negative or low, and P-cadherin always was negative. CONCLUSIONS. Expression of P-cadherin in breast carcinoma is associated strongly with poor survival and constitutes an independent prognostic predictor. P-cadherin expression is a better indicator of clinical outcome than alterations in the expression of E-cadherin, N-cadherin, a-catenin, or b-catenin. Cancer 1999;86: 1263–72. © 1999 American Cancer Society. KEYWORDS: cadherins and catenins in breast carcinoma, P-cadherin, poor survival, cap cell carcinoma of the breast. he cadherins are calcium-dependent cell-cell adhesion proteins.1 The best characterized and most widely distributed members of the family are the classical cadherins,2 a group that includes epithelial (E)-, nerve (N)-, and placental (P)-cadherin. Intracellularly, the cadherins interact with several proteins termed catenins, including a-catenin, b-catenin, g-catenin (plakoglobin), and p120ctn, a substrate for the tyrosine kinase Src.3– 6 The catenins link the cadherins to the cytoskeleton7 and mediate signal-transduction mechanisms that con- T 1264 CANCER October 1, 1999 / Volume 86 / Number 7 trol cellular events, including cell polarity and differentiation, cell growth, and cell death. During development, the homotypic binding of a particular cadherin to an identical cadherin expressed on an adjacent cell produces the sorting of cells into distinctive tissues.1 In adult organisms, the cadherin-catenin adhesion system maintains the differentiated state of the tissues and suppresses tumor development.8 –12 Alterations in the cadherin-catenin cell-cell adhesion system are associated with loss of differentiation and tumor formation,13 increased invasiveness, metastasis,14 and unfavorable prognosis for patients with breast carcinoma.15,16 In breast carcinoma, several lines of evidence indicate that reduced expression and function of cadherins are associated with tumor development and invasion. E-cadherin is considered a tumor suppressor in the breast, and breast carcinoma frequently exhibits loss of heterozygosity on the long arm of chromosome 16 (16q), which contains the E-cadherin gene (16q22.1). Low E-cadherin expression is found in half of infiltrating ductal carcinomas17–19 as a result of hypermethylation of the E-cadherin promoter region.20 Lobular carcinoma contains a high frequency of E-cadherin mutations, resulting in decreased or absent expression.21,22 Low expression of E-cadherin in breast carcinoma has been associated with dedifferentiation,17 increased invasiveness,16 and high metastatic potential.23 Conversely, transfection of E-cadherin cDNA into invasive breast carcinoma cells suppresses osteolytic bone metastases in mice,24 reduces invasiveness, and induces posttranscriptional up-regulation of syndecan-1, a proteoglycan associated with cell differentiation and cell-matrix anchorage.25 Another member of the cadherin family, Hcadherin, a mesenchymal cadherin lacking the cytoplasmic domain, also has a tumor-suppressing role in breast carcinoma, and its expression is reduced in human breast tumors.26 Loss or alterations in the cadherin-associated proteins, a-catenin13 and p120ctn,27 contribute to additional mechanisms of poor cell adhesion and breast carcinogenesis. Deletions of a-catenin were found in some tumors.13,28,29 Alterations also have been found in b-catenin,30 a mediator in the oncogenic Wnt-1 signaling pathway that has been implicated in breast tumorigenesis in mice.31,32 The expression and role of P-cadherin in breast carcinoma is still understood poorly. In one study, P-cadherin was not detected in a series of patients with ductal carcinoma,19 but, in another study, strong P-cadherin expression was found in some cases of infiltrating ductal carcinoma, and it was associated with reduced E-cadherin expression and advanced histologic grade.33 In the current study, a comprehensive analysis was conducted of the expression of cadherins and catenins in breast carcinoma. P-cadherin was expressed in more than half of the cases of invasive ductal carcinoma studied. More importantly, we found that the strong expression of P-cadherin in breast carcinoma correlates with poor patient outcome. Furthermore, our data indicate that the expression of P-cadherin represents a more accurate predictor of poor patient survival than the altered expression of other cadherins and catenins. MATERIALS AND METHODS Clinical Material Five-micron-thick sections were obtained from formalin fixed, paraffin embedded, archival tissues from invasive breast carcinomas obtained from the Department of Pathology of The Bryn Mawr Hospital, Bryn Mawr, Pennsylvania. The tumor tissues were obtained from a series used in a previous study.34 The samples used in that study included biopsy, lumpectomy, and mastectomy specimens that were large enough to provide material for flow cytometry and for biochemical and immunocytochemical analysis of hormone receptors. Out of the original 300 cases stained by immunohistochemistry, 183 cases were selected for this study based on a minimum of a 5-year patient follow-up after surgery for living patients and breast carcinoma as the cause of death for those patients that died. The surgical procedures were performed between 1989 and 1992. Patient status was classified as died of disease, alive with disease, or alive with no evidence of disease. The original pathology data included tumor size, histologic grade, estrogen and progesterone receptor status, and number of lymph node metastases. Histologic grading was according to the Bloom and Richardson method,35 as modified by Elston and Ellis. Tumors were classified as Grade 1 (well differentiated), Grade 2 (moderately differentiated), and Grade 3 (poorly differentiated). Histopathologic diagnoses included 160 invasive ductal carcinomas, 18 invasive lobular carcinomas, 4 medullary carcinomas, and 1 metaplastic carcinoma. Three cases of normal breast tissues from reduction mammoplasty also were included in the study. Antibodies A mouse monoclonal antibody (MoAb) against P-cadherin was purchased from Transduction Laboratories (Lexington, KY) and was used at a 2 mg/mL. Anti-Pcadherin MoAb clone 6A9 was a gift of Dr. M.J. Wheelock (University of Toledo, OH) and was used as conditioned supernatant fluid. Anti-E-cadherin MoAb clone HECD-1 was purchased from Zymed Laboratory P-Cadherin in Breast Carcinoma/Peralta Soler et al. (San Francisco, CA) and was used at 5 mg/mL. AntiN-cadherin MoAbs 13A9 and 3B9 (Zymed Laboratory) were developed in our laboratories and were used as conditioned supernatant fluids.5 They recognize intracellular domains of human N-cadherin in paraffin sections without cross-reactivity to other cadherins.36 Anti-a-catenin MoAb was purchased from Zymed Laboratory and was used at 1 mg/mL. Anti-b-catenin MoAb 5H10 (Zymed Laboratory) was developed in collaboration with Drs. Margaret J. Wheelock and Keith R. Johnson (University of Toledo) and was used as conditioned supernatant fluid.37 This antibody recognizes b-catenin in routinely processed paraffin sections of tumors from archival tissues. For g-catenin (plakoglobin), we used MoAb 15F11,38 (a gift from Drs. Margaret J. Wheelock and Keith R. Johnson, University of Toledo). The absence of cross-reactivity of the antibodies has been tested previously by Western immunoblotting.4,5,36 –38 All of the antibodies have been used previously in immunohistochemistry on formalin fixed, paraffin embedded human tissues.5,39 1265 TABLE 1 Semiquantitative Evaluation of P-Cadherin Expression and Patient Statusa P-cadherin Total ANED (%) AWD (%) DOD (%) Negative Total positive Level 1 Level 2 Level 3 Level 4 88 95 16 38 21 20 76 (86.5) 47 (49.5) 11 (69) 26 (68) 7 (33) 3 (15) 4 (4.5) 9 (9.5) 2 (12) 0 (0) 3 (14.5) 4 (20) 8 (9) 39 (41) 3 (19) 12 (32) 11 (52.5) 13 (65) ANED: alive with no evidence of disease; AWD: alive with disease; DOD: died of disease. a P-cadherin positive tumors were separated into 4 levels based on the percentage of positive cells; Level 4, 75–100% positive cells; Level 3, 50–75% positive cells; Level 2, 20–50% positive cells; Level 1, 10–20% positive cells. Patient status was classified as (died of disease), alive with disease, and alive with no evidence of disease. Note a 86.5% disease free survival rate in patients with P-cadherin negative tumors 5 years after surgery compared with a 15% disease free survival rate in patients with high levels of P-cadherin expression (Level 4). (plasma membrane, cytoplasm, nucleus) and staining intensity were recorded for tumor and nontumor tissues. Immunohistochemistry Immunohistochemistry was performed in most cases in an automated immunostainer (Biotek Techmate 500; Ventana Medical Systems, Tuczon, AZ) using an avidin-biotin system, following the manufacturer’s instructions (Vector Laboratories, Burlingame, CA). A heat-induced antigen retrieval method40 was applied. Deparaffinized 5-mm-thick sections were placed in prewarmed 0.1 M citrate buffer, pH 6.0 (Dako Corp., Carpinteria, CA) for 20 minutes in a steamer (Black and Decker, Shelton, CT), as described previously.39,41 Primary antibodies were incubated for 1 hour at room temperature. When immunohistochemistry was performed manually, the primary antibodies were incubated overnight at 4°C in a humid chamber. Sections were counterstained with hematoxylin. Data Recording The identities of patients were kept confidential, and the samples were coded in accordance with the guidelines of the Institutional Review Board of the Bryn Mawr Hospital for the use of human tissues for research. The follow-up data were provided by the Tumor Registry of the Bryn Mawr Hospital. A semiquantitative analysis of the immunohistochemistry was performed to determine the approximate percentage of cells expressing each cadherin and catenin. Cadherin and catenin positive tumors were classified into 4 levels based on the percentage of positive cells: Level 1, 10 –20% positive cells; Level 2, 20 –50% positive cells; Level 3, 50 –75% positive cells; and Level 4, 75–100% positive cells. Also, the patterns of cellular distribution Statistical Methods Relative mortality rates comparing patients with and without P-cadherin expression were estimated by incidence density ratios and 95% confidence intervals. Potential confounding or effect-modifying variables (including tumor grade, tumor size, lymph node involvement, and estrogen/progesterone receptor status) were investigated by using Cox proportional hazards regression models.42 Two patients (1.1%) were missing estrogen/progesterone receptor status and were excluded from multivariable analyses. P-cadherin positive and P-cadherin negative groups were compared with regard to each factor by using chisquare statistics and generalized Fisher exact tests or two-group t-tests, as appropriate.43 Multivariable analyses were performed by constructing adjusted incidence density ratios to compare patients with positive P cadherin expression with patients with negative P-cadherin expression while controlling for each covariate factor one at a time. Multivariable models were estimated to assess the impact of P-cadherin expression on survival after adjusting for multiple covariates. After assessing associations among covariate factors, a final multivariable model was constructed that included P-cadherin expression, tumor grade category, and lymph node status categories. RESULTS Both anti-P-cadherin MoAbs used in this study (the MoAb from Transduction Laboratories and MoAb 1266 CANCER October 1, 1999 / Volume 86 / Number 7 FIGURE 1. Kaplan–Meier survival curve of patients with placental (P)-cadherin positive (solid squares) and P-cadherin negative (open squares) tumors. The difference in survival reached statistical significance 2 years after surgical treatment (P 5 0.0001). 6A9) produced similar staining patterns. Western blot analysis showed that the anti-P- and anti-E-cadherin MoAbs used in this study did not have cross-reactivity. Both anti-P-cadherin MoAbs recognize a 118-kilodalton (kDa) protein, and the anti-E-cadherin MoAb (Zymed Laboratory) recognizes a 120-kDa protein in cells induced to express either P-cadherin or E-cadherin (data not shown). Anti-N-cadherin MoAbs, 13A9 and 3B9, have been shown previously to recognize a 135-kDa protein corresponding to N-cadherin.38 Both anti-N-cadherin MoAbs used here produced similar staining patterns. Anti-g-catenin (plakoglobin) MoAb was used only in a limited number of cases, and the data were not included in the correlative analysis with patient outcome. P-cadherin positive cells were found in 52% of the cases, and a semiquantitative analysis of the expression was recorded as the percentage of positive tumor cells (see Data Recording, above). When semiquantitative evaluation of P-cadherin expression was correlated with the clinical status of the patients (Table 1), the results showed that, 5 years after surgery, 41% of the patients with P-cadherin positive tumors died of the disease, 9.5% were alive with disease, and 49.5% were alive with no evidence of disease. In contrast, only 9% of the patients with P-cadherin negative tumors died of the disease, 4.5% were alive with disease, and 86.5% were alive with no evidence of disease. High P-cadherin expression correlated with poor survival: 65% of the patients with tumors expressing a high level of P-cadherin (Level 4) died of the disease within 5 years after surgery compared with 19% among patients with low-expressing (Level 1) tumors TABLE 2 Correlation between P-Cadherin Expression and Estrogen/Progesterone Receptor Status Status ER/PR negativea ER/PR positiveb P-cadherin positivec P-cadherin negative 50 17 43 71 ER: estrogen receptor; PR: progesterone receptor. a These include tumors negative for both receptors and tumors negative for either estrogen receptor or progesterone receptor. b These include tumors positive for both receptors. c Two of the 95 P-cadherin positive patients did not have hormone receptor data. TABLE 3 Correlation between P-Cadherin Expression and Histologic Grade in Invasive Breast Carcinomas Histologic tumor grade P-cadherin Total Grade 1 (%) Grade 2 (%) Grade 3 (%) Negative Positive 88 95 20 (23) 8 (8) 55 (63) 57 (60) 13 (14) 30 (32) (Table 1). The results of a Kaplan–Meier survival curve (Fig. 1) showed that the number of patients with Pcadherin positive tumors that died of the disease was significantly higher than that for patients with P-cadherin negative tumors as early as 2 years after surgical treatment (P 5 0.0001). The comparison between patients with positive and negative P-cadherin-express- P-Cadherin in Breast Carcinoma/Peralta Soler et al. 1267 TABLE 4 Two-Variable Regression Analysis of Mortality Rates Modeling P-Cadherin Expression plus One Covariate Covariate (adjusted one at a time with P-cadherin expression) Covariate Unadjusted mortality ratio Grade 3 vs. 1 2 vs. 1 Tumor size Positive lymph nodes .5 positive vs. negative 1–5 positive vs. negative Negative ER/PR receptor status P value P-cadherin expression IDR (95% CI)a ,0.029 ,0.158 ,0.002 9.5 (1.3–71.9) 4.2 (0.6–31.6) 1.5 (1.2–1.9)b,c ,0.0001 ,0.002 ,0.01 6.9 (2.7–17.9) 4.2 (1.7–10.4) 2.2 (1.2–4.2) P value IDR (95% CI)a ,0.0001 5.5 (2.6–11.7) ,0.0001 5.5 (2.5–12.3) ,0.0001 6.1 (2.8–13.2) ,0.0001 5.0 (2.3–10.8) ,0.0002 4.4 (2.0–9.7) IDR: incidence density ratio; 95% CI: 95% confidence interval. a Incidence density ratio computed by exponentiating estimated slope coefficient from Cox regression. b Incidence density ratio is per 1 standard deviation (16.8 mm) increase in size. c Includes 15 imputed size values. ing tumors was performed for each covariate factor. Table 2 shows the correlation between P-cadherin expression and estrogen/progesterone receptor status. There was a greater proportion of patients in the positive P-cadherin group with tumors with negative estrogen/progesterone receptor status compared with the negative P-cadherin group. Estrogen or progesterone receptor negative status was found in 53% of the P-cadherin positive tumors. In contrast, only 19% of the P-cadherin negative tumors were estrogen/progesterone receptor negative. A Fisher exact test showed that this difference was statistically significant (P , 0.0001). Table 3 shows the frequency of P-cadherin expression in invasive ductal carcinoma divided according to the histologic tumor grade. Grade 3 (poorly differentiated) tumors were more frequently P-cadherin positive than negative. In contrast, Grade 1 (well differentiated) tumors were more frequently P-cadherin negative. A Wilcoxon signed rank test for ranked qualified data with ties and a Fisher exact test showed a significant difference (P 5 0.0006 and P 5 0.002, respectively) between the distribution of P-cadherin positive and P-cadherin negative tumors when grouped according to histologic grade. However, there was no difference in the distribution of P-cadherin positive and P-cadherin negative tumors in Grade 2 (moderately differentiated) tumors. This finding is particularly significant, because 55% of the patients who died of the disease had Grade 2 tumors, and 90% of the tumors from those patients were P-cadherin positive. Furthermore, positive lymph node status was not significantly different between P-cadherin positive (47%) and P cadherin negative (41%) patients with TABLE 5 Multivariable Cox Regression Analysis of Mortality Rates Variable P value IDR (95% CI)a P-cadherin expression Grade 3 vs. 1 2 vs. 1 Positive lymph nodes .5 positive vs. negative 1–5 positive vs. negative ,0.0001 5.4 (2.4–12.2) ,0.036 ,0.161 8.8 (1.2–67.3) 4.2 (0.6–31.7) ,0.0006 ,0.003 5.4 (2.1–14.0) 4.0 (1.6–10.0) IDR: incidence density ratio; 95% CI: 95% confidence interval. a Incidence density ratio was computed by exponentiating estimated slope coefficient from Cox regression. Grade 2 tumors. Also, there was no significant difference in tumor size between patients with P-cadherin positive (25.04 mm 6 14.41 mm) and P-cadherin negative (24.33 mm 6 14.11 mm) Grade 2 tumors. When all of the tumor samples were studied together, regardless of their histologic grade, a t-test analysis of tumor size and a Fisher exact test for the presence of lymph node metastases did not correlate with P-cadherin expression (P 5 0.516 and P 5 0.103, respectively), although the value of this analysis is relative, because the criteria originally used for the selection of the tumors was based on their large size.34 Multivariable analyses of assessment of mortality was performed in patients with P-cadherin positive and P-cadherin negative tumors (Table 4). The unadjusted mortality ratio comparing patients with P-cadherin positive tumors with patients with P-cadherin negative tumors was 5.5 (95% confidence interval, 2.6 –11.7). These results were statistically significant 1268 CANCER October 1, 1999 / Volume 86 / Number 7 FIGURE 2. Immunohistochemical patterns of P-cadherin expression in breast carcinoma detected with anti-P-cadherin monoclonal antibody (MoAb; Transduction Laboratories, Lexington, KY). (A) P-cadherin in the plasma membrane in invasive cells of a ductal carcinoma. (B) Cytoplasmic P-cadherin expression in invasive ductal carcinoma. (C,D) P-cadherin expression mostly in the peripheral cells of invading ductal carcinoma nests. (E) P-cadherin negative invasive ductal carcinoma (small arrow). Note the positive myoepithelial cells (large arrow) in nonneoplastic breast ducts. (F) Lobular carcinoma negative for P-cadherin in both the in situ and invasive cells. Note the P-cadherin positive myoepithelial cells (arrow). (P , 0.0001), indicating that the probability of dying of breast carcinoma is 5.5 times greater in patients with P-cadherin positive tumors compared with patients with P-cadherin negative tumors. The mortality ratio values determined by P-cadherin expression ranged from 4.4 to 6.1 after adjusting for 1 covariate factor at a time. All of the covariates analyzed, including tumor grade, tumor size, positive lymph node status, and negative estrogen/progesterone receptor status, also were statistically significant predictors of patient out- come when controlling for P-cadherin expression. More importantly, there were no statistically significant interactions between any of the covariates and P-cadherin expression. A multiple variable model that included all factors simultaneously resulted in an adjusted mortality ratio for P-cadherin expression of 4.8 (range, 2.0 –11.3) with P 5 0.0003. A model including P-cadherin expression, tumor grade, and lymph node status (Table 5) showed that all three factors that were included did not modify the predictive value of P- P-Cadherin in Breast Carcinoma/Peralta Soler et al. cadherin. These results indicate that expression of P-cadherin in breast carcinoma constitutes an independent predictor of poor patient outcome. Distribution of P-cadherin in tumor cells showed several distinctive tissue patterns (Fig. 2). In invasive, P-cadherin positive ductal carcinomas, P-cadherin was either predominantly at the plasma membrane or in the cytoplasm. P-cadherin-expressing carcinoma cells often were located at the periphery of invasive cell clusters. P-cadherin frequently was expressed by cells penetrating the stroma in early invasive ductal carcinomas (not shown), suggesting an association between P-cadherin expression and invasion. Myoepithelial cells always were positive for P-cadherin, both in nontumoral ducts and in ducts and lobules containing in situ carcinoma. In P-cadherin negative invasive ductal carcinomas, P-cadherin positive mesenchymal spindle cells, resembling myoepithelial cells, were observed occasionally around the tumor cells. Lobular carcinomas were negative for P-cadherin, and negative in situ lobular carcinomas often were surrounded by P-cadherin positive myoepithelial cells. Expression of E-cadherin was positive in 98% of ductal carcinomas, and E-cadherin expression was correlated inversely with tumor grade. High levels of E-cadherin expression (Levels 3 and 4) were observed in 83% of patients with Grade 1 (well differentiated), 77% of patients with Grade 2 (moderately differentiated), and 51% of patients with Grade 3 (poorly differentiated) tumors. Low E-cadherin expression (Levels 1 and 2) was found in 44% of patients with Grade 3 tumors. However, only 5% of Grade 3 tumors were E-cadherin negative. In contrast to ductal carcinomas, 44% of lobular carcinomas were E-cadherin negative, and 39% had very low expression (Level 1). When the semiquantitative analysis of E-cadherin expression was correlated with patient survival, unexpectedly, there was not a significant association between low E-cadherin expression and poor survival. Five years after surgery, 60% of the patients with E-cadherin negative tumors and 68% of the patients with tumors expressing low E-cadherin (Level 1) were alive. Within 5 years after surgery, 20% of the patients with Ecadherin negative tumors and 26% and of the patients with E-cadherin positive tumors died of the disease (Table 6). N-cadherin was found in 48% of the tumors, but its expression was restricted to a small population of cells and was mostly cytoplasmic (Fig. 3). There was no correlation between the expression of N-cadherin and patient survival (not shown). The expression of a-catenin and b-catenin (Fig. 3) was variable. The staining of b-catenin was particularly intense in some ductal carcinoma samples. Low or negative b-catenin 1269 TABLE 6 Semiquantitative Evaluation of E-Cadherin Expression and Patient Status E-cadherina Total ANED (%) AWD (%) DOD (%) Negative Total positive Level 1 Level 2 Level 3 Level 4 10 173 19 30 64 60 6 (60) 117 (68) 13 (68) 16 (53) 38 (59.5) 50 (83.5) 2 (20) 11 (6) 3 (16) 0 (0) 6 (9.5) 2 (3.5) 2 (20) 45 (26) 3 (16) 14 (47) 20 (31) 8 (13) ANED: alive with no evidence of disease; AWD: alive with disease; DOD: died of disease. a See Table 1 for definition of levels of cadherin expression and patient status. expression in ductal carcinoma samples was very uncommon. In contrast, low expression was frequent in lobular carcinoma samples (not shown). Semiquantitative evaluation of a-catenin and b-catenin expression was difficult, and we did not find a correlation between catenin staining and patient survival (not shown). DISCUSSION The results of this study show that the expression of P-cadherin in breast carcinoma is a valuable indicator of poor patient survival. P-cadherin expression is a better indicator of poor survival than the altered expression of other cadherins and catenins, including a decrease in E-cadherin expression. In previous studies, assessment of the down-regulation or altered subcellular distribution of E-cadherin, a-catenin, and b-catenin as prognostic markers of breast carcinoma has proven problematic. Although E-cadherin frequently is mutated in lobular carcinomas of the breast, lack of expression is rare. The value of assessing reduced E-cadherin expression as a predictor of shorter survival for patients with breast carcinoma remains controversial. Although decreased E-cadherin in breast carcinoma cells has been associated with dedifferentiation,17 invasiveness,16 and metastases,23 low E-cadherin expression did not correlate with patient prognosis.27 In another study, the down-regulation of E-cadherin alone failed to correlate with the metastatic capacity of breast carcinomas, and the concomitant evaluation of catenins was necessary for predicting the development of metastases.29 In contrast, a recent study correlated reduced E-cadherin expression with shorter overall patient survival, although the study was limited to a population of lymph node negative patients.44 These opposite findings may reflect the difficulty in evaluating decreased protein expression in tumor cells, which requires optimal tissue pro- 1270 CANCER October 1, 1999 / Volume 86 / Number 7 FIGURE 3. Immunohistochemical expression of epithelial (E)-cadherin (Zymed Laboratory, Inc., San Francisco, CA MoAb), b-catenin (5H10 MoAb), and nerve (N)-cadherin (13A9 MoAb) in patients with breast carcinoma. (A) Cell membrane and cytoplasmic distribution of E-cadherin in ductal carcinoma. (B) shows strong b-catenin expression in an invasive ductal carcinoma. (C) Low E-cadherin expression in lobular carcinoma cells (small arrow) growing under the epithelial layer of a duct. Note the high level of E-cadherin in the ductal nonneoplastic epithelial cells (large arrow). (D) Focal cytoplasmic expression of N-cadherin in an invasive ductal carcinoma. cessing and computer-assisted analysis of the immunohistochemical staining. Alternatively, they may indicate the current poor understanding of the dynamics in the regulation of E-cadherin expression during the different phases of tumor invasion and metastasis. The expression of N-cadherin in breast carcinoma has not been evaluated previously. In this study, N-cadherin expression failed to correlate with poor patient survival. The role of P-cadherin in breast carcinoma is not known. However, the phenotypical characteristics of cells expressing P-cadherin during breast develop- ment provide clues that might help to explain the aggressive behavior of P-cadherin expressing tumors. During the development of the mouse mammary gland, P-cadherin and E-cadherin are expressed differentially in the breast epithelium. P-cadherin is expressed by the cap cells, a layer of growing cells located at the tip of the end buds, and E-cadherin is expressed by the differentiated epithelial cells facing the lumen of ducts.45 The spatially selective expression of E-cadherin and P-cadherin is required for both mammary tissue integrity and normal DNA synthesis.45 The P-cadherin expressing cap cells are consid- P-Cadherin in Breast Carcinoma/Peralta Soler et al. ered to be a stem cell population in the breast.46,47 They have a higher proliferation rate and lower frequency of apoptosis than the luminal epithelial cells48 as a result of paracrine regulation by growth factors, including transforming growth factor-a and epidermal growth factor. These growth factors stimulate the advance of ductal structures into the surrounding stroma.49 Cap cells differentiate into myoepithelial cells, contributing to the morphogenesis of the branching mammary gland.46 Evidence that P-cadherin plays a critical role in maintaining the growth of the ductal structures of the breast is supported further by the mammary phenotype of the P-cadherin gene knockout mouse model. P-cadherin-deficient mice show abnormal breast development. The mammary glands of virgin females do not have a ductal morphology, and, instead, they resemble those of pregnant mice.50 The data from the current study that show expression of P-cadherin in a subset of patients with aggressive breast carcinomas suggest a histogenetic origin in cap cells or acquisition of a tumor phenotype with characteristics similar to cap cells. Additional data support this postulate. Cap cells do not express estrogen receptors,51 a common characteristic of tumors expressing P-cadherin. P-cadherin expressing tumors are highly aggressive, and the paracrine growth factors that stimulate the growth of cap cells also are associated with the up-regulation of matrix metalloproteinases52 and increased invasiveness in breast carcinomas.53 In addition, the presence of prostate specific antigen (PSA), which has been found by polymerase chain reaction in '30% of breast carcinomas,54 indicates favorable patient prognosis55 and is expressed in a mutually exclusive manner with P-cadherin in prostate and prostate carcinomas.56 We propose that it is appropriate to reevaluate the current classification of breast tumors in light of more recent information on tumor progeny and prognostic significance of the different phenotypes. REFERENCES 1. 2. 3. 4. 5. 6. Takeichi M. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 1991;51:1451–5. Grunwald GB. The structural and functional analysis of cadherin calcium-dependent cell adhesion molecules. Curr Opin Cell Biol 1993;5:797– 805. Tsukita S, Tsukita S, Nagafuchi A, Yonemura S. Molecular linkage between cadherins and actin filaments in cell-cell adherens junctions. Curr Opin Cell Biol 1992;4:834 –9. Knudsen KA, Wheelock MJ. Plakoglobin, or an 83-kD homologue distinct from b-catenin, interacts with E-cadherin and N-cadherin. J Cell Biol 1992;118:671–9. Knudsen KA, Peralta Soler A, Johnson KR, Wheelock MJ. Interaction of a-actinin with the cadherin/catenin cell-cell adhesion complex via a-catenin. J Cell Biol 1995;130:67–77. Reynolds AB, Daniel J, McCrea PD, Wheelock MJ, Wu J, 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 1271 Zhang Z. Identification of a new catenin: the tyrosine kinase substrate p120cas associates with E-cadherin complexes. Mol Cell Biol 1994;14:8333– 42. Wheelock MJ, Knudsen KA. Cadherins and associated proteins. In Vivo 1991;505–13. Vleminckx KL, Vakaet L, Marell MM, Fiers W, van Roy FM. Genetic manipulation of E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 1991; 66:107–19. Behrens J, Vakaet L, Friis R, Winterhager E, van Roy FM, Mareel MM, et al. Loss of epithelial differentiation and gain of invasiveness correlates with tyrosine phosphorylation of the E-cadherin/b-catenin complex in cells transformed with a temperature-sensitive v-SRC gene. J Cell Biol 1993;120: 757– 66. Hedrick L, Cho KR, Vogelstein B. Cell adhesion molecules as tumour suppressors. Trends Cell Biol 1993;3:36 –9. Peifer M. Cancer, catenins, and cuticle pattern: a complex connection. Science 1993;262:1667– 8. Takeichi M. Cadherins in cancer: implications for invasion and metastasis. Curr Opin Cell Biol 1993;5:806 –11. Rimm DL, Sinard JH, Morrow JS. Reduced a-catenin and E-cadherin expression in breast cancer. Lab Invest 1995;72: 506 –12. Oka H, Shiozaki H, Kobayashi K, Inoue M, Tahara H, Kobayashi T, et al. Expression of E-cadherin cell adhesion molecules in human breast cancer tissues and its relationship to metastasis. Cancer Res 1993;53:1696 –701. Sommers CL, Thompson EW, Torri JA, Kemler R, Gelmann EP, Byers W. Cell adhesion molecule unomorulin expression in human breast cancer cell lines: relationship to morphology and invasive capacities. Cell Growth Diff 1991;2:365–72. Siitonen SM, Kononen JT, Helin HJ, Rantala IS, Holli KA, Isola JJ. Reduced E-cadherin expression is associated with invasiveness and unfavorable prognosis in breast cancer. Am J Clin Pathol 1996;105:394 – 402. Gamallo C, Palacios J, Suarez A, Pizarro A, Quintanilla M, Cano A. Correlation of E-cadherin expression with differentiation grade and histological type in breast carcinoma. Am J Pathol 1993;142:987–93. Moll R, Mitze M, Frixen UH, Birchmeier W. Differential loss of E-cadherin expression in infiltrating ductal and lobular breast carcinomas. Am J Pathol 1993;143:1731– 42. Rasbridge SA, Gillett CE, Sampson SA, Walsh FS, Millis RR. Epithelial (E-) and placental (P-) cadherin cell adhesion molecule expression in breast carcinoma. J Pathol 1993;169: 245–50. Graff JR, Herman JG, Lapidus RG, Chopra H, Xu R, Jarrard DF, et al. E-cadherin expression is silenced by DNA hypermethylation in human breast and prostate carcinomas. Cancer Res 1995;55:5195–9. Berx G, Cleton-Jansen A-M, Nollet F, de Leeuw W, van de Vijver M, Cornelisse C, et al. E-cadherin is a tumour/invasion suppressor gene mutated in human lobular breast cancers. EMBO J 1995;14:6107–15. De Leeuw WJF, Berx G, Vos CBJ, Peterse JL, van de Vijver MJ, Litvinov S, et al. Simultaneous loss of E-cadherin and catenins in invasive lobular breast cancer and lobular carcinoma in situ. J Pathol 1997;183:404 –11. Hunt NCA, Douglas-Jones AG, Jasani B, Morgan JM, Pignatelli M. Loss of E-cadherin expression associated with lymph node metastases in small breast carcinomas. Virchows Arch 1997;430:285–9. 1272 CANCER October 1, 1999 / Volume 86 / Number 7 24. Mbalaviele G, Dunstan CR, Sasaki A, Williams PJ, Mundy GR, Yoneda T. E-cadherin expression in human breast cancer cells suppresses the development of osteolytic bone metastases in an experimental metastasis model. Cancer Res 1996;56:4063–70. 25. Leppa S, Vlemincks K, Van Roy F, Jalkanen M. Syndecan-1 expression in mammary epithelial tumor cells is E-cadherin-dependent. J Cell Sci 1996;109:1393– 403. 26. Lee SW. H-cadherin, a novel cadherin with growth inhibitory functions and diminished expression in human breast cancer. Nature Med 1996;2:776 – 82. 27. Dillon DA, D’Aquila T, Reynolds AB, Fearon ER, Rimm DL. The expression of p120ctn protein in breast cancer is independent of alpha- and beta-catenin and E-cadherin. Am J Pathol 1998;152:75– 82. 28. Shimoyama Y, Nagafuchi A, Fujita S, Gotoh M, Takeichi M, Tsukita S, et al. Cadherin dysfunction in a human cancer cell line: possible involvement of loss of a-catenin expression in reduced cell-cell adhesiveness. Cancer Res 1992;52:5770 – 4. 29. Bukholm, IK, Nesland JM, Karesen R, Jacobsen U, BorresenDale A-L. E-cadherin and a-, b-, and g-catenin protein expression in relation to metastasis in human breast carcinoma. J Pathol 1998;185:262– 6. 30. Oyama T, Kanai Y, Ochiai A, Akimoto S, Oda T, Yanagihara K, et al. A truncated b-catenin disrupts the interaction between E-cadherin and a-catenin: a cause of loss of intercellular adhesiveness in human cancer cell lines. Cancer Res 1994;54:6282–7. 31. Hinck L, Nathke IS, Papkoff J, Nelson WJ. Beta-catenin: a common target for the regulation of cell adhesion by Wnt-1 and Src signaling pathways. Trends Biochem 1994;19:538 – 42. 32. Tsukamoto AS, Grosschedl R, Guzman RC, Parslow T, Varmus HE. Expression of the int-1 gene in transgenic mice is associated with mammary gland hyperplasia and adenocarcinomas in male and female mice. Cell 1988;55:619 –25. 33. Palacios J, Benito N, Pizarro A, Suarez A, Espada J, Cano A, et al. Anomalous expression of P-cadherin in breast carcinoma. Am J Pathol 1995;146:605–12. 34. Keshgegian AA. ErB-2 oncoprotein overexpression in breast carcinoma: inverse correlation with biochemically- and immunohistochemically-determined hormone receptors. Breast Cancer Res Treat 1995;35:201–10. 35. Bloom HJG, Richardson WW. Histological grading and prognosis in breast cancer: a study of 1409 cases of which 359 have been followed for 15 years. Br J Cancer 1957;11:359 –77. 36. Peralta Soler A, Knudsen KA, Jaurand M-C, Johnson KR, Wheelock MJ, Klein-Szanto AJP, et al. The differential expression of N-cadherin and E-cadherin distinguishes pleural mesotheliomas from lung adenocarcinomas. Hum Pathol 1995;26:1363–9. 37. Johnson KR, Lewis JE, Li D, Wahl J, Peralta Soler A, Knudsen KA, et al. P- and E- cadherin are in separate complexes in cells expressing both cadherins. Exp Cell Res 1993;207:252– 60. 38. Sacco PA, McGranahan TM, Wheelock MJ, Johnson KR. Identification of plakoglobin domains required for association with N-cadherin and a-catenin. J Biol Chem 1995;270: 20201– 6. 39. Peralta Soler A, Knudsen KA, Tecson-Miguel A, McBrearty FX, Han AC, Salazar H. The expression of E-cadherin and N-cadherin in surface epithelial-stromal tumors of the ovary distinguishes mucinous from serous and endometrioid tumors. Hum Pathol 1997;28:734 –9. 40. Pasha T, Montone KT, Tomaszewski JE. Nuclear antigen retrieval utilizing steam heat. Mod Pathol 1995;8:167. 41. Han AC, Peralta Soler A, Knudsen KA, Wheelock MJ, Johnson KR, Salazar H. Differential expression of N-cadherin in pleural mesotheliomas and E-cadherin in lung adenocarcinomas in formalin-fixed, paraffin-embedded tissues. Hum Pathol 1997;28:641–5. 42. Cox DR. Regression models with life-tables. J R Stat Soc 1972;66:188 –90. 43. Mehta CR, Patel NR. A network algorithm for performing Fisher’s exact test in r 3 c contingency tables. J Am Stat Soc 1983;78:472– 4. 44. Charpin C, Garcia S, Bonnier P, Martini F, Andrac L, Choux R, et al. Reduced E-cadherin immunohistochemical expression in node-negative breast carcinomas correlates with 10-year survival. Am J Pathol 1998;109:431– 8. 45. Daniel CW, Strickland P, Friedmann Y. Expression and functional role of E- and P-cadherins in mouse mammary ductal morphogenesis and growth. Dev Biol 1995;169:511–9. 46. Williams JM, Daniel CW. Mammary ductal elongation: differentiation of myoepithelium and basal lamina during branching morphogenesis. Dev Biol 1983;97:274 –90. 47. Rudland PS. Histochemical organization and cellular composition of ductal buds in developing human breast: evidence of cytochemical intermediates between epithelial and myoepithelial cells. J Histochem Cytochem 1991;39:1471– 84. 48. Humphreys RC, Krajewska M, Krnacik S, Jaeger R, Weiher H, Krajewski S, et al. Apoptosis in the terminal endbud of the murine mammary gland: a mechanism of duct morphogenesis. Development 122:4013–22. 49. Snedeker SM, Brown CF, DiAugustine RP. Expression and functional properties of transforming growth factor a and epidermal growth factor during mouse mammary gland ductal morphogenesis. Proc Natl Acad Sci USA 1991;88:276 – 80. 50. Radice GL, Ferreira-Cornwell MC, Robinson SD, Rayburn H, Chodosh LA, Takeichi M, et al. Precocious mammary gland development in P-cadherin-deficient mice. J Cell Biol 1997; 139:1025–32. 51. Sapino A, Macri L, Gugliotta P, Pacchioni D, Liu YJ, Medina D, et al. Immunophenotypic properties and estrogen dependency of budding cell structures in the developing mouse mammary gland. Differentiation 1993;55:13– 8. 52. Kondapaka SB, Fridman R, Reddy KB. Epidermal growth factor and amphiregulin up-regulate matrix metalloproteinase-9 (MMP-9) in human breast cancer cells. Int J Cancer 1997;70:722– 6. 53. Castellani R, Visscher DW, Wykes S, Sarkar FH, Crissman JD. Interaction of transforming growth factor-alpha and epidermal growth factor receptor in breast carcinoma. An immunohistologic study. Cancer 1994;73:344 –9. 54. Zarghami N, Grass L, Diamandis EP. Steroid hormone regulation of prostate-specific antigen gene expression in breast cancer. Br J Cancer 1997;75:579 – 88. 55. Yu H, Giai M, Diamandis EP, Katsaros D, Sutherland DJA, Levesque MA, et al. Prostate specific antigen is a new favorable prognostic indicator for women with breast cancer. Cancer Res 1995;55:2104 –10. 56. Peralta Soler A, Harner GD, Knudsen KA, McBrearty FX, Grujic E, Salazar H, et al. The expression of P-cadherin identifies PSA negative cells in epithelial tissues of male sexual accessory organs and in prostatic carcinomas: implications for prostate cancer biology. Am J Pathol 1997;151: 471– 8.