Fax: (011) 39 0437 216150
High syndecan-1 expression in breast carcinoma is related to an aggressive phenotype and to poorer prognosis
Article first published online: 13 JUN 2003
Copyright © 2003 American Cancer Society
Volume 98, Issue 3, pages 474–483, 1 August 2003
How to Cite
Barbareschi, M., Maisonneuve, P., Aldovini, D., Cangi, M. G., Pecciarini, L., Angelo Mauri, F., Veronese, S., Caffo, O., Lucenti, A., Palma, P. D., Galligioni, E. and Doglioni, C. (2003), High syndecan-1 expression in breast carcinoma is related to an aggressive phenotype and to poorer prognosis. Cancer, 98: 474–483. doi: 10.1002/cncr.11515
- Issue published online: 18 JUL 2003
- Article first published online: 13 JUN 2003
- Manuscript Accepted: 8 APR 2003
- Manuscript Revised: 29 MAR 2003
- Manuscript Received: 29 JAN 2003
- Associazione Italiana Ricerca Cancro (AIRC)
- Associazione Artigiani e Piccole Impresse
- Lega Italiana per la Lotta ai Tumori Sezione di Trento—Progetto Per L'Oncologia In Trentino
- Italian Ministry of Health, Ricerca Finalizzata 2002–2004
Syndecan-1 is a transmembrane heparan sulphate proteoglycan that is involved in cell–cell adhesion, organization of cell–matrix adhesion, and regulation of growth factor signaling.
Specimens from 254 consecutive breast carcinoma (BC) cases (110 N0, 144 N1/2) with long-term follow-up (median, 95 months) were immunostained for syndecan-1, estrogen receptor (ER), progesterone receptor (PgR), and p53; in 154 cases, c-erbB-2 status was known. Syndecan-1 mRNA and protein expression also were evaluated in 20 breast tissue samples (10 normal and tumor pairs).
Syndecan-1 was expressed at high levels in 106 (42%) BCs; syndecan-1 up-regulation was confirmed by reverse transcriptase–polymerase chain reaction (RT-PCR) studies. High syndecan-1 expression was associated with high histologic grade, large tumor size, high mitotic count, c-erbB-2 overexpression, and ER and PgR negative status. At univariate survival analysis syndecan overexpression was related to poor prognosis (P < 0.01 for both overall survival (OS) and disease-free survival). Bivariate survival analysis showed an additive adverse effect for syndecan-1 and c-erbB-2 overexpression. At multivariate analysis, syndecan-1 overexpression was independently associated with poor OS (hazard ratio [HR], 1.71; 95% confidence interval [CI], 1.08–2.69). High syndecan-1 expression also was of independent prognostic value for OS in the group of 102 ER-negative patients (HR, 2.42; 95% CI, 1.21–4.82). Stratifying patients on the basis of the type of adjuvant therapy given, high syndecan-1 expression was associated with a higher risk of death only in patients treated with the cyclophosphamide-methotrexate-fluorouracil regimen (HR, 1.9; P = 0.09); at multivariate analysis for OS, this association proved to be of independent statistical significance (P = 0.03; HR, 2.15).
Syndecan-1 is expressed at high levels in a significant percentage of breast carcinomas and is related to an aggressive phenotype and poor clinical behavior. Cancer 2003;98:474–83. © 2003 American Cancer Society.
Patients affected by breast carcinoma may have very different clinical outcomes, and every effort should be made to identify the markers that will allow the best prediction of prognosis and response to therapy in an individual patient. Beyond classic clinical and pathologic parameters, there is a growing need to identify molecular markers that are related to the natural history of the disease.
Breast carcinoma results from molecular alterations that are genetically and/or environmentally induced and result in uncontrolled cell proliferation. A promising approach to studying breast carcinoma is to identify subgroups of tumors characterized by specific genetic and molecular markers, which may allow selective therapeutic approaches. One of these markers, for example, is the c-erbB-2 receptor, which is considered of pivotal importance in breast tumorigenesis and is the therapeutic target of the humanized monoclonal antibody trastuzumab. The mechanism of c-erbB-2 activation has not yet been fully elucidated, but it probably involves ligands belonging to the heparin-binding growth factor family. These growth factors require the presence of specific heparan sulfate proteoglycans (HSPGs) as coreceptors. Cell-associated HSPGs include two main families, syndecans and glypicans, which differ significantly in core protein domain structure.1 Syndecans are a family of four transmembrane HSPGs expressed on adherent cells and play important roles in cell proliferation, cell–cell and cell–matrix adhesion, cell motility, and invasiveness.2, 3 Syndecans act as coreceptors for heparin-binding growth factors (HBGFs) and may increase local concentration of these ligands, allowing enhanced receptor activation even at low ligand concentration.4–8
Syndecan-1, also known as CD138, is the most extensively studied member of the syndecan family, and a few studies have examined the role of syndecan-1 in breast oncogenesis. In a mouse model, it has been shown that syndecan-1 is critical for Wnt-1–induced tumorigenesis of the mammary gland.9 A recent study also demonstrated that syndecan-1 is overexpressed in breast carcinoma cell lines and tumors.10 Based on these experimental data and given the fundamental role of syndecan-1 in the cellular response to growth factors, we investigated syndecan-1 expression by immunohistochemistry (IHC) in a consecutive series of 254 breast carcinoma cases with long-term follow-up. In this set of tumors, we evaluated whether syndecan-1 expression was associated with pathologic and clinical parameters and with survival. We also evaluated the syndecan-1 mRNA and protein expression levels using the reverse transcriptase–polymerase chain reaction (RT-PCR) and IHC in 20 frozen samples of paired normal and tumor breast tissues. In the current study, high syndecan-1 expression proved to be a new independent prognostic factor for breast carcinoma: syndecan-1 was up-regulated in a significant percentage of breast tumors and was related to an aggressive phenotype and poor clinical behavior. Our preliminary data also suggest that syndecan-1 up-regulation might be a marker of poor response to cyclophosphamide-methotrexate-fluorouracil (CMF) chemotherapy.
MATERIALS AND METHODS
We investigated 254 patients with breast carcinoma (BC) who underwent surgery between January 1986 and December 1991 at the Santa Chiara Hospital (Trento, Italy) or the Niguarda-Cà Granda Hospital (Milan, Italy). The main clinicopathologic features of these patients are listed in Table 1. Eligibility criteria were as follows: histologic diagnosis of infiltrating BC, axillary lymph node dissection, no distant metastasis (M0), unilateral BC, and no other previous or concomitant primary malignancy. Patients were staged according to the International Union Against Cancer Tumor Node Metastasis (UICC-TNM) Classification. One hundred seventy-two patients received an adjuvant systemic therapy according to the risk of relapse: 80 patients were treated with CMF-based chemotherapy, and 92 patients received adjuvant hormone therapy with tamoxifen; 8 patients were treated with both hormone therapy and chemotherapy. All patients were followed after surgical treatment–once monthly during adjuvant chemotherapy, then every 4 months for the first 3 years, and finally once per year. The median follow-up duration was 86 months (range, 6–178 months) for disease-free survival (DFS) and 95 months (range, 11–178 months) for overall survival (OS).
|Variable||ALL||Syndecan-1 expression||P value|
|Mean age (yrs)||56.1||57.0||54.9|
|Tumor size (T)|
|Mitotic count (mean)||15.6||13.8||18.2||0.02|
Surgical samples were collected shortly after surgical removal, fixed in buffered 10% formalin for 24 to 48 hours at room temperature, and routinely processed. BCs were classified as follows: 207 were infiltrating ductal carcinomas and 47 were of other histotypes, including infiltrating lobular, medullary, tubular, mucinous, and cribriform carcinomas. Tumor grading, according to Elston and Ellis,11 was performed by one pathologist (M.B.). Mitotic counts were performed using a BH-2 microscope (Olympus, Melville, NY) at high magnification (10 fields with a 10/20L eyepiece at ×40), starting the counts in the most mitotically active areas. Twenty samples from a different group of 10 tumors and corresponding normal tissue were collected shortly after surgical removal and were snap-frozen and processed.
Immunostaining was performed on paraffin sections of primary tumors. Briefly, 4-micron paraffin sections were treated with the microwave antigen retrieval system, incubated for 1 hour at room temperature with the primary antibodies and processed using the StreptABC technique (Dako Co., Glostrup, Denmark). All samples were stained for syndecan-1, ER, PgR, and p53; a subset of cases also was investigated for c-erbB-2 overexpression. Primary antibodies are shown in Table 2. Negative controls were obtained by omitting primary antibodies. Positive controls were sections of normal human skin; positive internal controls were normal ducts and lobules, stromal elements, and plasma cells. Syndecan-1 immunoreactivity of tumor cells was recorded as membrane associated, cytoplasmic, or both. The percentage of immunoreactive tumor cells was evaluated by two observers (M. B. and D. A.), using a double-headed microscope, scanning the whole sections at medium and high magnification, and evaluating at least 1000 cells. Syndecan-1 staining intensity was scored as follows: 0, no staining; 1+, weak, 2+, moderate, and 3+, strong staining; cases were considered 2+ and 3+ when the staining intensity was higher than the one of normal breast epithelial cells. A final score (syndecan-1 score) was obtained by multiplying the percentage of reacting tumor cells with their staining intensity. The syndecan-1 staining pattern of the stromal component also was evaluated using a semiquantitative score (−/+/++). A group of 30 randomly selected cases form the study series were stained twice and blindly evaluated to ensure reproducibility of staining and of the evaluation system. Cells were considered positive for ER, PgR, and p53 only when distinct nuclear staining was identified, and cases were considered positive when the percentage of reacting cells was higher than 10%, as suggested in the literature.12 c-erbB-2 was evaluated with a semiquantitative score, as follows: 0 = no staining or faint staining in <10% of tumor cells; 1 = moderate to strong membrane staining in > 10% of tumor cells.
|Syndecan-1||BB4||Serotec (Bicester, United Kingdom)|
|Estrogen receptor||ER1D5||Dako (Glostrup, Denmark)|
|Progesterone receptor||1A4||Novocastra (Newcastle, United Kingdom)|
RNA Extraction and RT-PCR
We also examined syndecan-1 mRNA expression in 20 frozen breast tissue samples, representing normal and tumor pairs from an additional 10 different patients. These tumors included 6 cases with positive syndecan-1 expression at the immunohistochemical level, which ranged from 10–90% of reactive cells, and 4 cases that were unreactive. A tissue dismembrator was used to grind frozen tissue samples, and both frozen tissue and cell-line RNA were extracted using a guanidine isothiocyanate protocol (Tri Reagent; Molecular Research Center, Inc., Cincinnati, OH). After DNAse treatment, equal amounts of total RNA (1 μg) were reverse transcribed into cDNA using random primers and MuLV reverse transcriptase (2.5 units/μL) according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). The human 7S rRNA gene was used as an endogenous control. PCR for syndecan-1 and 7S was performed in a 25 μL reaction mixture containing DNA template, PCR buffer (50 mM KCl, 10 mM Tris-HCl, pH 8.3), 1.5 mM MgCl2, 200 mM of each dNTP, 0.4 mM of each primer (syndecan-1: upstream primer, 5′-AGGACGAAGGCAGCTACTCCT-3′; downstream primer, 5′-TTTGGTGGGCTTCTGGTAGG-3′; 7S: upstream primer, 5′-ACCACCAGGTTGCCTAAGGA-3′; downstream primer, 5′-CACGGGAGTTTTGACCTGCT-3′), and 0.75 units of AmpliTaq DNA polymerase (Perkin-Elmer, Norwalk, CT). PCR was performed for 35 cycles consisting of 95 °C for 30 seconds, 60 °C for 30 seconds, and 72 °C for 30 seconds. The resulting amplification products were then analyzed by agarose gel electrophoresis using standard methods, and normal and tumor tissue pairs were compared.
The Fisher exact test was used to examine the association between different measures and syndecan-1 expression. For OS, the endpoint of the analysis was death from any cause; subjects still alive at the end of the study or known to be alive at some time and lost to follow-up were censored. Disease-free interval was calculated from the date of surgery to the date of progression or the date of last follow-up. Survival estimates were calculated using the Kaplan–Meier method and compared using the log rank test. The Cox proportional hazards regression model was used to evaluate the simultaneous effect of explanatory variables on survival time. All models were fitted on the whole series of patients with inclusion of dummy variables to represent patients with missing values. All analyses were carried-out using SAS statistical software (SAS Institute, Inc., Cary, NC). All P values were based on two-sided testing; P values below 0.05 were considered statistically significant.
In normal breast, syndecan-1 immunoreactivity was observed in most ducts and in a small percentage of lobules: myoepithelial cells showed strong staining, whereas luminal cells showed a much more heterogeneous reactivity, always along the basolateral cell membrane, with minor degrees of cytoplasmic staining. Normal stromal tissue was unreactive (Fig. 1A).
In breast carcinomas, both the epithelial and stromal components showed a heterogeneous staining pattern. Invasive carcinomas showed three different patterns of reactivity: 1) diffuse; 2) focal, with single or small clusters of reactive tumor cells; and 3) negative staining (Fig. 1B–D). Syndecan-1 immunoreactivity was seen in 180 (71%) cases with membranous (58%), membranous and cytoplasmic (37%), and cytoplasmic-only (5%) staining patterns. The percentage of immunoreactive cells ranged from 1% to 90%, with a median value of 10; the intensity of staining was as follows: 92 cases had an intensity of 1+, 76 had an intensity of 2+, and 12 had an intensity of 3+; the scores ranged from 0 to 270, with a median value of 10. For statistical analysis, cases were considered as exhibiting high syndecan-1 expression when the percentage of reacting cells or the score was higher than the median value. Results for the percentage and score values were almost the same, and we will show in detail only the results for the percentages. High syndecan-1 expression (cases with more than 10% immunoreactive cells) was seen in 106 (42%) cases. Our immunohistochemical data were supported by the mRNA analysis. Syndecan mRNA was expressed at higher levels in tumors compared with corresponding normal tissue (Fig. 2). Moreover, high syndecan-1 mRNA levels were correlated with high protein expression detected by immunohistochemistry in epithelial tumor cells.
The stromal compartment of invasive tumors showed heterogeneous and inconstant syndecan-1 reactivity: 37% cases showed no stromal reactivity, 31% showed moderate staining, and 32% showed strong staining. Staining was seen both in stromal cells and in collagen bands. Dense desmoplastic stroma showed the most intense staining reactivity, whereas loose connective tissue with prominent lymphocytic infiltrate showed the least intense staining.
Clinicopathologic Correlates of Syndecan-1 Expression
High syndecan-1 expression in tumor cells was significantly associated with higher grade and mitotic count, c-erbB-2 overexpression, and a negative steroid receptor status; there was also a trend for association with larger tumor size (Table 1). No relation was seen between the pattern of syndecan-1 immunostaining in tumor cells (membranous versus cytoplasmic) and any clinicopathologic parameter. No correlation was seen between stromal syndecan-1 expression and any clinicopathologic parameter, even when the pattern of stromal expression was stratified on the basis of the level of expression seen in tumor cells.
Univariate survival analysis in the whole series of cases showed that high syndecan-1 expression (i.e., syndecan expression in more than 10% of tumor cells) was related to poor DFS and OS (Table 3, Fig. 3A). The percentage of syndecan-1–expressing cells was also evaluated also as a continuous variable and proved to be associated with poor OS and DFS (P < 0.01 and P = 0.02, respectively). The other prognostically relevant parameters were tumor size, nodal status, mitotic count, grade, ER status, and p53 and c-erbB-2 overexpression. Stratifying patients on the basis of nodal status, tumor size and grade, and hormone receptor status, high syndecan-1 expression was related to poor DFS and OS in node-positive patients and in high grade and in ER-negative tumors (Fig. 3B–D). In a subset of 154 cases with known c-erbB-2 status, bivariate survival analysis showed that there was an additive adverse effect when both markers were overexpressed (Fig. 4).
|No. of patients||Overall survival||Disease-free survival|
|5-year||10-year||Log rank test||5-year||10-year||Log rank test|
|Tumor size (T)|
|4||3||0.50||—||< 0.01||—||—||< 0.01|
|1–2 positive nodes||67||0.75||0.67||0.69||0.51|
|≥ 3 positive nodes||77||0.63||0.45||< 0.01||0.48||0.35||< 0.01|
|≥ 12||123||0.69||0.61||< 0.01||0.63||0.52||0.03|
|3||133||0.65||0.56||< 0.01||0.58||0.47||< 0.01|
|Pos (> 10%)||155||0.87||0.80||< 0.01||0.70||0.55||< 0.01|
|Pos (> 10%)||130||0.85||0.70||0.07||0.75||0.59||0.09|
|Pos (> 15%)||60||0.60||0.53||< 0.01||0.57||0.46||0.02|
|Pos||35||0.54||0.44||< 0.01||0.46||0.36||< 0.01|
|Pos (> 10%)||106||0.65||0.56||< 0.01||0.59||0.48||< 0.01|
|Pos (> 10%)||129||0.82||0.72||0.88||0.73||0.56||0.40|
Multivariate analysis was performed using three different models: the first model included syndecan-1 expression along with all other available clinicopathologic parameters, the second model included the biologic parameters (ER, PgR, p53, and c-erbB-2), and the third model was restricted to the ER-negative subgroup of patients. In the first model, syndecan-1 proved to be of independent prognostic value for OS (HR, 1.71; 95% confidence interval [CI], 1.08–2.69) along with tumor size, nodal status, and grade (Table 4). In the second model, syndecan-1 did not show any independent prognostic value: this loss of significance was due to the association between syndecan-1 and ER status. The third model, which considered only ER-negative cases, showed that high syndecan-1 expression was of independent prognostic value for OS (Table 4).
|Model 1a (n = 254)||Model 2b (n = 254)||Model 3c (n = 102)|
|OS HR (95% CI)||DFS HR (95% CI)||OS HR (95% CI)||DFS HR (95% CI)||HR (95% CI)||DFS HR (95% CI)|
|Syndecan-1 (> 10% vs. 0–10%)||1.71 (1.08–2.69)d||1.36 (0.92–2.01)||1.29 (0.80–2.06)||1.10 (0.74–1.66)||2.42 (1.21–4.82)d||1.72 (0.91–3.22)|
|Age (≥ 50 yrs vs. < 50 yrs)||0.81 (0.52–1.28)||0.88 (0.59–1.30)d||1.05 (0.64–1.71)||1.02 (0.66–1.56)||1.92 (1.02–3.64)d||1.52 (0.84–2.75)|
|Size (T2-4 vs. T1)||1.75 (1.02–3.02)d||1.66 (1.06–2.60)d||1.54 (0.86–2.76)||1.54 (0.96–2.46)||1.26 (0.59–2.72)||1.16 (0.56–2.41)|
|LN status (1–2 positive nodes vs. 0 positive nodes)||2.15 (1.13–4.07)d||1.84 (1.09–3.12)d||2.44 (1.24–4.82)d||1.96 (1.12–3.43)d||4.27 (1.65–11.1)d||3.62 (1.54–8.50)d|
|LN status (≥ 3 positive nodes vs. 0 positive nodes)||4.00 (2.25–7.10)d||3.52 (2.19–5.66)d||4.94 (2.73–8.93)d||4.19 (2.57–6.85)d||8.57 (3.72–19.7)d||8.84 (4.04–19.4)d|
|Grade (3 vs. 1–2)||2.44 (1.44–4.12)d||1.76 (1.16–2.68)d||2.60 (1.47–4.60)d||1.75 (1.11–2.77)d||1.61 (0.79–3.29)||1.51 (0.77–2.97)|
|ER (> 10% vs. ≤ 10%)||—||—||0.45 (0.26–0.77)d||0.68 (0.43–1.08)d||—||—|
|PgR (> 10% vs. ≤ 10%)||—||—||1.28 (0.75–2.18)||0.99 (0.63–1.54)d||1.32 (0.65–2.70)||0.90 (0.46–1.78)d|
|p53 (> 15% vs. ≤ 15%)||—||—||1.36 (0.84–2.20)||1.18 (0.76–1.82)||1.36 (0.74–2.51)||1.52 (0.84–2.75)|
|c-erbB-2 (positive vs. negative)||—||—||1.96 (1.08–3.57)d||2.38 (1.37–4.12)d||2.31 (1.09–4.93)d||3.23 (1.52–6.83)d|
|c-erbB-2 (unknown vs. negative)||—||—||1.18 (0.69–2.03)||1.57 (1.00–2.46)d||1.90 (0.88–4.07)||2.67 (1.29–5.55)d|
Relation between Syndecan-1 and Adjuvant Treatment
Patients were stratified on the basis of the type of adjuvant therapy given. After adjustment for lymph node and ER status and grade, high syndecan-1 expression was found to be associated with a trend toward a higher risk of death only in the group of 80 patients treated with chemotherapy (HR, 1.896; P = 0.09) and not in patients treated with tamoxifen or in the group of patients who did not receive any adjuvant therapy (Fig. 5). Among women who received adjuvant chemotherapy, a restricted model (without ER status, which was not significant) showed that high syndecan-1 expression was significantly associated with poorer OS (HR, 2.15; P = 0.03) along with nodal status (HR, 6.94; P = 0.06) and grade (HR, 2.04; P = 0.04).
The current study shows that syndecan-1 is highly expressed, both at the immunohistochemical and mRNA levels, in a relevant percentage of breast carcinomas. High syndecan-1 expression in the current series of cases is associated with an aggressive phenotype characterized by larger tumor size, higher tumor grade, higher mitotic count, negative steroid status, and overexpression of c-erbB-2 and p53. High syndecan-1 expression is a newly discovered independent marker of poor prognosis. The current data also suggest that syndecan-1 could be used to identify a group of patients who were less responsive to CMF adjuvant chemotherapy, but further studies are needed to confirm this point.
In keeping with previous data on cell lines, tumors, and animal models,9, 10, 13 our results suggest an important role for syndecan-1 in the biologic natural history of human breast carcinoma. Syndecan-1 is in fact implicated in several essential physiologic cell functions, such as control of cell proliferation, differentiation, adhesion, and migration.2 One of the best known biologic functions of syndecan-1 is related to its interaction with HBGFs.2, 14 Syndecan-1 can interact, for example, with fibroblast growth factors (FGFs), which are known angiogenic and mitogenic growth factors (GFs) for breast carcinoma cells,15 binding to FGFs and to their receptors in a ternary signaling complex.3 Syndecan-1 also can function as a potent FGF-2 activator through physiologic shedding and degradation of its extracellular domain by enzymes, such as heparanase.16
The data indicating that syndecan-1 expression is high in breast tumors compared with normal tissue and that syndecan-1 is related to a more aggressive phenotype are at variance with the study of Stanley et al.13 These authors, by using the same antisyndecan antibody as we did, suggested that syndecan-1 levels are indeed reduced in breast carcinoma and could not identify any correlation with any pathobiologic parameter. However, the limited number of cases in their series (n = 20), probably did not allow them to determine the true frequency of syndecan-1 expression and its clinicopathologic correlations.
In the current study, high syndecan-1 expression was of independent prognostic value in the entire series. After the series was stratified on the basis of other known pathologic and biologic factors (e.g., nodal status, grade, and steroid receptor status), syndecan-1 was found to have prognostic value in the subgroups identified by other unfavorable prognostic factors (node positivity, high tumor grade, and ER negativity). These findings suggest that high syndecan-1 expression may contribute to the identification of a patient population at particularly high risk of relapse and death.
In the current series, high syndecan-1 expression possessed independent prognostic value, particularly in the estrogen-independent tumors. In particular, we showed that cases with the ER-negative phenotype and concurrent high expression of syndecan-1 had a poorer outcome when compared with the ER-negative cases with low syndecan-1 levels. This finding could be of possible clinical interest, and the predictive value of the combined steroid receptor/syndecan-1 phenotype should be investigated, as has been suggested for other markers.12, 17–19 These clinical data can be interpreted in the light of the known biologic role of syndecan-1. In fact, it could be hypothesized that in ER-negative tumors that have lost the ability to respond to the estrogen-dependent proliferative pathway, high syndecan-1 expression may confer a particularly important growth advantage by enhancing the response to other growth factors. In fact, syndecan-1, like all HSPGs, are essential for HBGF/HBGF receptor signaling, as they 1) increase the local concentration of GFs; 2) protect GFs from proteolytic cleavage (e.g., transforming growth factor-β); 3) participate in the internalization of ligands; 4) may have distinct binding sites for GFs; and 5) may interact directly with GFs and their receptors in ternary complexes (e.g., FGF-2/FGFR-1).4 The overall effect of syndecan-1 therefore could be to enhance receptor activation at low ligand concentrations.4 Recent data support the idea that in breast carcinomas, syndecan-1 is one of the most important cell-surface HSPGs involved in FGF-2/FGFR-1 complex formation.20
We also analyzed subgroups of patients who received homogeneous adjuvant treatments: high syndecan-1 expression was prognostically relevant only in CMF-treated patients, for whom it was independently associated with poor outcome, and not in patients treated with tamoxifen or those left untreated. These data, although based on a retrospective analysis of a relatively small patient population, suggest that syndecan-1 up-regulation could be useful to identify a group of patients less responsive to CMF adjuvant therapy. The correlation between high syndecan-1 expression and CMF therapy is similar to the one reported for c-erbB-2 overexpression (see Piccart et al.21). In the current series, syndecan-1 and c-erbB-2 were statistically associated and showed some similar associations with other parameters (such as ER status and p53), but their overexpression only partially overlapped. In addition, syndecan-1 and c-erbB-2 were independent prognostic factors, and bivariate survival analysis of the expression of both markers showed an additive adverse effect if both were overexpressed. Because both syndecan-1 and c-erbB-2 are involved in cellular signaling, it is tempting to hypothesize a possible biologic interaction between them. This observation deserves further study, which could disclose interacting pathways of receptor activation, as suggested by observations in colon carcinom cell lines in which c-erbB-2 is up-regulated by syndecans.22 In fact, the HBGFs that could interact with syndecan-1 are epidermal growth factor, amphiregulin, and heregulin, which also are candidate ligands for c-erbB-2. This scenario could expand the opportunities for therapeutic targeting and suggests the possibility of modulating syndecan-1 activity, in addition to counteracting the activity of c-erbB-2 with humanized anti-HER2 antibodies.23
In conclusion, the current study demonstrates that syndecan-1 is expressed at high levels in a significant proportion of breast carcinomas and that this up-regulation is associated with tumor aggressiveness and poor prognosis. The concurrent evaluation of ER and c-erbB-2 status and syndecan-1 expression allows us to split the ER and c-erbB-2 categories into subgroups with vastly different outcomes. Further studies are needed to confirm whether syndecan-1 can be a member of the family of markers that will enhance our ability to customize prognosis and/or therapy according to the risk profile of the individual patient.
The authors thank Mr. Enzo Meggiolaro, Ms. Lucia Tamis, and Dr. Silvia Funes for excellent technical assistance.
- 13Syndecan-1 expression is induced in the stroma of infiltrating breast carcinoma. Am J Clin Pathol. 1999; 112: 371–383., , , .