Versican but not decorin accumulation is related to malignancy in mammographically detected high density and malignant-appearing microcalcifications in non-palpable breast carcinomas

This is a retrospective, single-center study conducted at the University Hospital of Patras, Greece. The conduct of the study was approved by the institutional review board of Patras Medical School and written informed consent was obtained from the patients. Between 1989 and 2002, 310 patients with non-palpable suspicious breast lesions detected during screening mammography were evaluated. The median age of the patients was 55 years, (range, 35-74 years) and none of the patients had previously been screened or shown clinical signs of breast disease.

The standard craniocaudal and lateral views were carried out in all patients. Mammographic findings requiring further exploration with breast biopsy were considered the following: (1) microcalcifications; (2) mass with or without microcalcifications; (3) architectural distortion with or without microcalcifications; and (4) asymmetric density with the greater diameter < 1 cm with or without microcalcifications [10]. Based on published observations, we evaluated MAMCs according to their shape (e.g. pleomorphic, irregular, fragmented, casting), density (highly variable), distribution (e.g. clustered) and size (highly variable) [36].

All patients with suspicious and highly suggestive of malignancy non-palpable breast lesions classified as BI-RADS 4 and BI-RADS 5 underwent pre-operative mammographically guided breast lesion localization with a Kopan breast localization needle (19G, 9 cm length) [37].

Histological examination of the specimens showed 83/310 carcinomas. Immunohistochemical analysis was carried out in 62 out of 83 carcinomas, on the basis of tissue availability, 9 normal breast tissues, 14 breast tissues with high MD and 14 breast tissues with MAMCs without pathologic findings. MAMCs was the prominent finding in 48 cases and 14 cases were characterized by increased MD without microcalcifications. Tissue sections were obtained from the files of the Department of Pathology in the University Hospital of Patras. A panel of antibodies was employed for versican (2-B-1 Seikagaku), decorin (6-B-6 Seikagaku), ERα (6F11 Novocasta), PR (1A6 Novocasta), c-erbB2 (CB11 Biogenex). Avidin-biotin-peroxidase complex method (Dako Co., Copenhagen, Denmark) and microwave antigen retrieval was carried out. Tissue samples were fixed in 10% buffered formalin and embedded in paraffin. Serial 5- μm sections were taken and deparaffinized with xylene and dehydrated with 98% ethanol. The sections were treated with 1 U/ml chondroitinase ABC for 15 min at 37°C to detach glycosaminoglycan side chains from the protein core for versican and decorin staining. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide for 5 min at room temperature. Non-specific protein binding of the antibodies was blocked by incubation with 3% normal swine serum in PBS for 20 min at room temperature. Slides were incubated with the antibodies diluted in PBS containing 1% normal swine serum for 1 h at room temperature. The resulting antigen- antibody complexes were visualized by 30 min incubation at room temperature, using appropriate biotinylated secondary antibodies diluted 1:200 and the avidin-biotin-peroxidase technique (Dako Co., Copenhagen, Denmark). The staining was developed with 3, 3-diaminobenzidine/hydrogen peroxide for 5 min at room temperature and slides were counterstained with haematoxylin. A positive tissue control and a negative reagent control (without primary antibody) were run in parallel. The level of decorin and versican expression was assessed by semiquantitative scoring devised by Alowami et al., [38] which includes (i) the overall percentage of the tissue section stained positive (0-100%), and (ii) the signal intensity (4-point scale) for each proteoglycan. The scoring used for the 4-point scale was: 1, negative or very weak staining; 2, weak positive; 3, moderate positive; 4, strong positive. The semiquantitative immunohistochemical (IHC) score for the expression level for each proteoglycan was provided by the multiplication of the percentage (0-100) of the tissue section staining positive by the factor (1-4) corresponding to the staining intensity of the tissue section. Positivity scoring for ERα, PR and c-erbB2 was carried out as described previously [12]. Three independent researchers randomly evaluated the specimens using this method.

Data were analysed using GraphPad Prism (Version 3.0; GraphPad Software Inc., San Diego, CA, USA). Statistical analyses were performed using the Mann-Whitney U-test, Kruskal-Wallis Anova by ranks and linear regression analysis (Pearson). All tests were two-tailed and statistical significance was set at P < 0.05.

Three hundred and ten patients with suspicious and highly suggestive of malignancy mammographically found non-palpable breast lesions, classified as BI-RADS 4 and BI-RADS 5, who underwent mammographically guided needle excision biopsy were evaluated. Eighty-three out of 310 (26.8%) non-palpable breast lesions proved histologically to be carcinomas. Histological subtypes were 46/83 ductal invasive carcinomas (55.5%), 31/83 ductal in situ carcinomas (37.3%) and 6/83 lobular invasive carcinomas (7.2%). MAMCs as an isolated finding or in combination with a mass, density or distortion were detected in 65 out of 83 (78.3%) patients with non-palpable breast carcinoma (Table 1). Immunohistochemistry was performed in 62/83 carcinomas on the basis of tissue availability, 9 normal breast tissues, 14 breast tissues with high MD and 14 breast tissues with MAMCs without pathologic findings. The median age of the patients was 58 years (range, 41-74 years). Non-palpable breast carcinomas were divided in two groups according to their mammographic findings, one which is characterized by the presence of increased MD, mass and/or distortion (14/62) (Figure 1A) and the other characterized by the presence of MAMCs as an isolated finding or in combination with a mass, density or distortion (48/62) (Figure 1B-D). The expression of proteoglycans was evaluated in breast carcinomas and compared to normal breast tissues with identical mammographic findings.

In normal breast, staining for decorin was observed in the interstitial connective tissue surrounding the glands (Figure 2A), whereas negligible deposits of versican were also identified in stroma surrounding normal glands (Figure 2B). Interestingly, there was a marked stromal decorin and versican accumulation surrounding intraductal epithelial proliferations and in situ tumor components (Figure 2C and 2D). This was more diffusely distributed in the case of decorin, whereas versican staining was limited to the immediate proximity of the basement membrane. Prominent immunostaining for versican and decorin was observed in stroma associated with malignant areas of sectioned breast tissue. In the majority of patients with breast cancer, increased staining for versican and decorin was found, in the peritumoral and intratumoral stroma. In the central areas of malignant tumors, the staining was generally strong for both decorin and versican. The staining intensity for both proteoglycans was strong but varied considerably at the periphery of the same tumors. The expression of decorin and versican differed greatly in carcinomas with increased MD (Figure 3A and 3B) and carcinomas with MAMCs (Figure 3C and 3D). The expression of decorin was markedly elevated in breast carcinomas and normal tissues that are characterized by increased MD and MAMCs (Figure 3A, C, E and 4A). The overall expression of decorin was significantly higher in carcinomas and normal tissues with MAMCs (median ± SE: 91.5 ± 13.6 for carcinomas and 96.0 ± 18.9 for normal tissues) compared to tissues with increased MD (median ± SE: 46.5 ± 17.4 for carcinomas and 43.0 ± 13.7 for normal tissues) and normal tissue (median ± SE: 20.5 ± 2.5) (Figure 4A). No statistically significant differences were observed between sub-groups of carcinoma and normal tissues characterized by the presence of MD and MAMCs, indicating that decorin accumulation in the stroma may not correlate specifically with malignant transformation (Figure 4A). In contrast, the expression of versican was significantly higher only in carcinomas showing increased MD and MAMCs compared to normal tissues with identical mammographic findings (Figure 3B, D, F and 4B). The overall expression of versican was significantly higher in carcinomas with MAMCs (median ± SE: 42.0 ± 9.1 for carcinomas and 14.0 ± 5.8 for normal tissues) compared to tissues with increased MD (median ± SE: 22.5 ± 10.1 for carcinomas and 11.0 ± 4.4 for normal tissues) and normal tissue (median ± SE: 10.0 ± 2.0) (P < 0.01) (Figure 4B).

Higher expression of decorin in the tumor stroma was found to associate with in situ breast cancer only in patients with MAMCs (P = 0.01) (Figure 5A-D). On the other hand the elevated expression of versican in the tumor stroma was strongly correlated with higher tumor grade and invasiveness in patients with MD and MAMCs (Figure 6A-D). The accumulation of decorin was significantly related to higher expression of ERα (P = 0.02) and PR (P < 0.01) in tumor cells only in patients with MAMCs (Figure 7A-D). Similarly, the expression of versican is significantly associated with higher expression of ERα (P < 0.01) and PR (P < 0.01) in tumor cells only in patients with MAMCs (Figure 7E-H). c-erbB2 expression was found in 1/14 carcinoma with MD and 13/48 of patients with MAMCs. The expression of both proteoglycans was not related to the presence of c-erbB2 in tumor cells in patients with MD and MAMCs (figure not shown).