| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Human Cancer Biology |
Authors' Affiliations: Departments of 1 Oncology and 2 Pathology and 3 Statistical Service, Tel Aviv Sourasky Medical Center; 4 Department of Cell Research and Immunology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel
Requests for reprints: Adit Ben-Baruch, Department of Cell Research and Immunology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel. Phone: 972-3-6407933; Fax: 972-3-6422046; E-mail: aditbb{at}tauex.tau.ac.il.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: The expression of CCL5, alone and in conjunction with estrogen receptor (ER)-
, ER-ß, progesterone receptor (PR), and HER-2/neu (ErbB2), was determined in breast tumor cells by immunohistochemistry. The study included 142 breast cancer patients, including individuals in whom disease has progressed.
Results: Using Cox proportional hazard models, univariate analysis suggested that, in stage I breast cancer patients, CCL5 was not a significant predictor of disease progression. In contrast, in stage II patients, the expression of CCL5 (CCL5+), the absence of ER- (ER-–), and the lack of PR expression (PR–) increased significantly the risk for disease progression (P = 0.0045, 0.0041, and 0.0107, respectively). The prognostic strength of CCL5, as well as of ER-–, improved by combining them together (CCL5+/ER-–: P = 0.0001), being highly evident in the stage IIA subgroup [CCL5+/ER-– (P = 0.0003); ER-– (P = 0.0315)]. In the stage II group as a whole, the combinations of CCL5–/ER-+ and CCL5–/PR+ were highly correlated with an improved prognosis. Multivariate analysis indicated that, in stage II patients, ER- and CCL5 were independent predictors of disease progression.
Conclusions: CCL5 could be considered as a biomarker for disease progression in stage II breast cancer patients, with the CCL5+/ER-– combination providing improved prediction of disease progression, primarily in the stage IIA subgroup.
The aim of the present study was to determine the prognostic strength of the CCL5 chemokine alone and to analyze its ability to strengthen the prognostic value of other molecules, which have already been characterized as markers for progression in breast cancer. CCL5 was selected in view of findings, suggesting that it is a potential contributor to breast malignancy. CCL5 belongs to the family of chemokines, primarily identified as potent inducers of leukocyte motility (10, 11). This chemokine was first identified in our published research to be expressed almost exclusively in breast tumor cells and was only minimally detected in adjacent normal breast epithelial cells (12). Moreover, CCL5 was found to be highly expressed in breast cancer patients compared with healthy individuals and was significantly associated with advanced disease course (12). These findings are suggestive of a role for CCL5 in breast malignancy and were supported by the study of Niwa et al. (13), whose results indicated that elevated CCL5 serum levels are correlated with advanced breast cancer. Later clinical studies on CCL5 supported a role for the chemokine in breast cancer, and animal studies indicated that CCL5 has a causative role in breast malignancy (14–17). Finally, various in vitro and in vivo analyses provided evidence for a variety of tumor-supporting functions that may be exerted by the chemokine in this disease, such as acting on the tumor microenvironment and on the tumor cells themselves to promote tumor-associated functions (15–26).
The other proteins investigated in this research included conventional markers in this disease [i.e., estrogen receptor (ER)-, progesterone receptor (PR), and ErbB2; refs. 6–9]. Whereas ER- and PR are considered as markers for good prognosis (1–7), the expression of ErbB2, documented in 15% to 30% of invasive breast cancers, has been associated with poor prognosis (6–9). We also analyzed the prognostic value of CCL5 together with ER-ß, whose role in breast cancer has not yet been fully elucidated (1–3).
Using Cox proportional hazard models, univariate and multivariate analyses were done to assess the applicability of the different proteins, alone or together, as biomarkers for breast cancer progression. By immunohistochemistry, we have determined the expression of the different proteins in breast tumor cells, in a total of 142 breast carcinoma patients, and correlated their expression with the clinical status of the patients. The study included patients who remained disease-free in the course of a follow-up that was carried out for at least 5 years after diagnosis, whereas others relapsed with local tumor or distant metastasis or died of breast cancer. To improve the prognostic strength of our investigations, all the analyses were done separately on stage I patients and on stage II patients.
The results of the present research suggest that, in patients diagnosed in stage I of disease, CCL5 does not have a significant value in predicting disease progression. In contrast, in stage II patients, CCL5 expression by breast tumor cells may be considered as a potential marker for prediction of disease progression, and its combined analysis with the lack of ER- improves the prognostic value of each of these two proteins, especially in patients diagnosed at stage IIA of disease. Moreover, our findings suggest that a favorable outcome for stage II patients may be predicted in individuals who are negative for CCL5 expression, however, positive for ER- or PR.
Overall, our investigation calls for multicenter studies that will further determine and possibly validate the prognostic value of CCL5, alone and together with ER-, for better identification of breast cancer patients who are at risk for progression and for assessment of its therapeutic implications in this group of patients.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
Immunohistochemistry. Tissues from malignant breast lesions were formalin fixed and embedded in paraffin. All the tissues were obtained from the Oncology Department, Tel Aviv Sourasky Medical Center (Tel Aviv, Israel) with the approval of the institutional Helsinki Committee. Serial sections (5-µm thick) were prepared from the blocks and processed as described previously (12, 18).
The slides were deparaffinized in alcohol and treated with hyaluronidase and 3% H2O2. Nonspecific binding was blocked by normal goat serum, and staining was done overnight at 4°C by antibodies having well-determined specificities, as indicated by published articles, using immunohistochemistry and Western blotting (see below). For detection of CCL5, monoclonal mouse IgG2b recognizing human CCL5 by Western blotting (data not shown) was used (50 µg/mL; PeproTech 500-M75, Rocky Hill, NJ). The binding specificity of the antibodies against CCL5 was verified in our laboratory compared with an isotype-matched antibody and in published studies (12). In addition, the following antibodies were used: monoclonal mouse IgG1 anti-ER- (1D5; 10 µg/mL; Zymed, South San Francisco, CA), polyclonal goat anti-ER-ß (SC-6820; 20 µg/mL; Santa Cruz Biotechnology, Santa Cruz, CA) (for antibodies against ER- and ER-ß: refs. 27–30), monoclonal mouse IgG1 anti-PR (1A6 antibody; 1:40 dilution; according to the manufacturer's instructions; Novocastra Laboratories, Newcastle, United Kingdom; refs. 28–30), and monoclonal mouse IgG1 anti-ErbB2 (3B5; 0.5-1 µg/mL; Oncogene, Boston, MA; refs. 31, 32). The antibodies against ER-, PR, and ErbB2 were all IgG1 and therefore served as internal controls for each other's specificity: The ER- and PR specimens showed nuclear staining but often with different patterns, and the ErbB2 specimens showed membrane and/or cytoplasmic staining. The staining of all biopsies was negatively controlled by omission of the antibodies and by buffer substitution. Antigen retrieval by microwave was used for staining by antibodies against ER-, ER-ß, PR, and ErbB2; however, it was not used for staining by antibodies to CCL5.
The sections were washed thoroughly in PBS and stained by biotinylated anti-broad spectrum second antibody (Histostain kit, Zymed) according to the manufacturer's instructions. They were then stained with horseradish peroxidase–streptavidin (Histostain kit), counterstained, and processed as described previously (18). The staining pattern of the tested proteins in tumor cells was evaluated on the whole section area in a blind manner by an expert pathologist and by two researchers and is summarized in Table 2 .
|
2 or Fisher exact tests as applicable. Survival analysis according to the Kaplan-Meier method was carried out to estimate the disease-free time function according to protein expression. The SAS for Windows version 9.1 was used for all statistical analyses.
|
|
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
, ER-ß, PR, and ErbB2. Figure 1
displays representative examples of the staining patterns of each of these proteins in the examined breast tumor cells. In line with the previously described patterns of their expression, CCL5 staining was mainly cytoplasmatic, whereas ER-, ER-ß, and PR were primarily nuclear. ErbB2 staining in the tumor cells ranged between a membranous staining (graded 3+) and weak cytoplasmic expression (graded 1+).
|
As ErbB2 was found to be the sole protein that had the potential to predict progression in stage I patients, we also examined the distribution and statistical significance of the CCL5+/ErbB2+ combination by univariate analysis and found that it had no significant prognostic value for disease progression for this group (Table 3).
Protein expression in stage II breast cancer patients. When compared with the group of stage I breast cancer patients, a different pattern of protein expression was noted in patients diagnosed at stage II of disease (Tables 2 and 3). Three of the tested proteins, CCL5, ER-, and PR yielded significant values for prediction of progression using univariate analysis in Cox proportional hazard models. As shown in Table 3, CCL5 expression highly increased the risk for disease progression (P = 0.0045), a finding that was also shown by Kaplan-Meier plots, which showed the distribution of disease-free patients according to the presence or absence of CCL5 expression (Fig. 2A
).
|
and PR: the absence of ER- expression significantly increased the risk for disease progression (P = 0.0041) as did the absence of PR expression (P = 0.0107; Table 3). The contribution of the lack of hormone receptor expression to progression of the disease in our study is in line with the current and accepted views on ER- and PR in breast cancer (1–7). With regard to ErbB2, its incidence of expression was in line with published studies denoting its overexpression in 15% to 30% of breast cancer patients (22 of the 89 patients were positive to ErbB2; Table 2; refs. 6–9). Nevertheless, in contrast to other studies associating ErbB2 amplification with poor outcome in patients with nodal involvement (6–9), this protein did not have a significant predictive value for disease progression in this group of patients in our study (Table 3). In addition to ErbB2, ER-ß also did not have prognostic significance in this group of patients. Taken together, then, these results indicate that the presence of CCL5 and the absence of ER- or PR are predictive markers for disease progression in stage II breast cancer patients.
Furthermore, univariate analysis of the CCL5+/ER-– combination as well as of the CCL5+/PR– combination provided significant statistical values for disease progression (P = 0.0001 and 0.0253, respectively; Table 3) as was also indicated by the Kaplan-Meier plots shown in Figs. 2B and 2C. Importantly, combining CCL5+ with ER-– improved the prognostic value not only of CCL5 but also of ER- in predicting disease progression: whereas the statistical significance of CCL5+ alone or of ER-– alone was P = 0.0045 and 0.0041, respectively, the combination of CCL5+/ER-– improved the statistical value for prediction of breast cancer progression to P = 0.0001 (see also the Kaplan-Meier plot in Fig. 2B). In addition, although ErbB2 was not predictive of disease progression in our stage II patients, the combination of CCL5+/ErbB2+ showed an elevated risk for progression in that group (P = 0.0140; Fig. 2D; Table 3).
In view of the improved predicting value of the CCL5+/ER-– combination in stage II breast cancer patients compared with CCL5+ or ER-– alone, further investigation of this observation was done. To this end, the group of stage II patients was divided into two subgroups, stage IIA and stage IIB, according to tumor size and lymph node involvement (for details, see Materials and Methods). Of the two subgroups, that of stage IIA has yielded a similar pattern to the stage II group as a whole, providing evidence to the major advantage of the CCL5+/ER-– combination (P = 0.0003) over ER-– alone (P = 0.0315) in predicting disease progression (Table 3). These results indicate that the CCL5+/ER-– combination has a highly significant value in this respect in the stage IIA breast cancer patients.
To analyze the possibility that the prognostic protective value of the presence of ER- and PR will take effect when there was no CCL5 expression, we determined the statistical significance of specific combinations, in which the absence of CCL5 expression was combined with presence of ER- and PR. First, this analysis was done for the group of stage II patients as a whole. As shown in Table 3 and Fig. 2E, the combination CCL5–/ER-+ was predictive for improved prognosis (P = 0.0029) as was the combination of CCL5–/PR+ (P = 0.0005; Fig. 2F; Table 3). More specifically, these combinations were predictive for a favorable outcome in the subgroup of stage IIA patients [CCL5–/ER-+ combination (P = 0.0488); CCL5–/PR+ combination (P = 0.0092)].
To conclude, these results indicate that positive expression of CCL5 and the absence of ER- and PR are predictors of progression and that the combination of CCL5+/ER-– is advantageous for predicting progression primarily in stage IIA patients. In contrast, a favorable outcome was indicated by the absence of CCL5, combined with the presence of ER- or with PR in breast tumor cells.
Dependence between variables. We did stepwise multivariate analyses using Cox proportional hazard models to determine the dependence between the different proteins that were found to be of significance for disease progression. The results indicated that none of the proteins had an independent prognostic significance in stage I patients, whereas the absence of ER- and the expression of CCL5 had an independent significance in predicting disease progression for stage II patients (P = 0.049 and 0.005, respectively). The other studied proteins, ER-ß, PR, and ErbB2, did not have any independent predictive value for disease progression at this stage of disease.
Coexpression of the tested proteins. In addition to the above analyses, we looked at the possibility that several of the proteins that were investigated in this study are coexpressed in the different stages of disease. The 2 and Fisher exact tests (as applicable) found coexpression between CCL5 and ER-ß (P = 0.027), between ER- and PR (P < 0.001), and also between PR and ErbB2 in stage I patients. A similar analysis of the coexpression of the different proteins in stage II patients revealed coexpression between ER- and PR (P < 0.001) and between ER-ß and ErbB2 (P = 0.01). These results provide evidence to complex associations between the expression patterns of the proteins analyzed in the present study.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Our focus on CCL5 was based on previous findings, suggesting that this chemokine contributes to breast malignancy (see above). Taken together with our earlier demonstration of CCL5 being highly expressed in advanced stages of disease (12), we now hypothesized that CCL5 may be a powerful biomarker for breast cancer progression and determined its prognostic significance in prediction of breast cancer progression.
To this end, we have characterized CCL5 when solely analyzed and have determined its ability to strengthen the prognostic value of other biomarkers as well. Our current findings showed that CCL5 may be used as a prognostic marker for disease progression in stage II breast cancer patients and that its combination with the lack of ER- expression improves the prognostic value of each of the two proteins. Of importance, our analysis indicates that the CCL5+/ER-– combination has a major advantage over ER-– alone in the subgroup of stage IIA patients, suggesting that these patients can considerably benefit from the use of the CCL5+/ER-– combination for prediction of disease progression.
Our findings also indicated that the CCL5–/ER-+ and CCL5–/PR+ combinations may be predictive of a favorable outcome, suggesting that the protective properties of the hormone receptors become evident when CCL5 is absent. These results imply that, when the chemokine is expressed by breast tumor cells, it actually dominates the potential protective role of ER- and PR. Subsequently, it is possible that, in CCL5+ patients, endocrine therapy may be ineffective. Only when CCL5 is absent, the protective effects of ER- and PR may come into effect; therefore, the CCL5–/ER-+ and CCL5–/PR+ combinations are indicative of favorable prognosis.
Overall, CCL5, either alone or together with the lack of ER-, emerged as a potential novel biomarker that could be used for prediction of progression in stage II breast cancer patients (primarily in patients diagnosed in stage IIA). It is important to note the differences between patients diagnosed in stages I and II of disease, as indicated by the fact that the markers that were associated with disease progression were different for the two groups of patients. Accordingly, CCL5 was significantly associated with progression in stage II but not stage I patients. This observation is in line with the conclusions of our earlier study that has also shown a more significant relevance of CCL5 to stage II as indicated by the fact that CCL5 was more prevalent in stage II patients than in stage I patients (12).
We would like to note that, in contrast to our published study, the current investigation is advantageous as it includes an additional variable, being the criteria of disease progression. This approach allows us to better analyze the prognostic value of CCL5 in breast cancer and to determine its ability to predict progression in this disease.
Accordingly, differences were observed between our two studies in the incidence of CCL5 expression, as a sole factor, in stages I versus II of disease. In the present study, the incidence or CCL5 expression was not markedly different between patients diagnosed at stage I or II of disease (43% in both stages), whereas in our previous investigation, the incidence of CCL5 expression was significantly higher in stage II patients than in stage I patients (83% versus 55%, respectively). Because in our published study (12) we did not include the criteria of progression of disease, it is difficult to assess the incidence of progression in stage I versus stage II patients in that study. Therefore, the possibility exists that the group of stage II patients included relatively more patients who have progressed compared with the stage I group, giving rise to elevated incidence of CCL5 expression in stage II of disease compared with stage I. In contrast, in the current study, the two groups of patients, at stages I and II, had a similar incidence of individuals whose disease has progressed. This may explain the observation that, in the current study, no marked differences have been observed in the incidence of CCL5 expression between the two stages of breast carcinoma.
To conclude, the findings presented herein show the prognostic value of CCL5 in predicting the progression of stage II breast cancer patients to local relapse, metastasis, or death. Our observations further support the importance of CCL5 as a prognostic factor when combined with ER- and propose their joint roles as potential prognostic markers for disease progression in this group of patients, more specifically in patients belonging to the stage IIA subgroup.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: N. Barak and S. Shina contributed equally to this work.
Received 1/12/06; revised 5/ 4/06; accepted 5/11/06.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
and ß in human breast cancer and its clinical application. Endocr Relat Cancer 2003;10:193–202.[Abstract] and a role for estrogen receptors in asymmetric division and differentiation of estrogen-dependent cells. Reprod Biol Endocrinol 2003;1:36.[CrossRef][Medline] protein in normal and abnormal term placentae and stimulation of trophoblast differentiation by estradiol. Reprod Biol Endocrinol 2003;1:13.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  G. Sauer, N. Schneiderhan-Marra, C. Kazmaier, K. Hutzel, K. Koretz, R. Muche, R. Kreienberg, T. Joos, and H. Deissler Prediction of Nodal Involvement in Breast Cancer Based on Multiparametric Protein Analyses from Preoperative Core Needle Biopsies of the Primary Lesion Clin. Cancer Res., June 1, 2008; 14(11): 3345 - 3353. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
