Markers of subtypes in inflammatory breast cancer studied by immunohistochemistry: Prominent expression of P-cadherin

We collected a series of 91 Tunisian T4d tumors (TNM, UICC) treated between 1994 and 1998 at the Salah Azaiz Institute (ISA, Tunis, Tunisia). Paraffin-embedded specimens were collected prior chemotherapy. The presence of dermal lymphatic emboli was not mandatory. All patients received neo-adjuvant chemotherapy with anthracyclin-based (FEC 100 or FAC 50) regimen. Forty-three patients were metastatic at diagnosis and did not have complementary mastectomy. External beam radiation was done for all patients.

This series was compared with a consecutive series of 547 non-IBC cases deposited in tissue microarrays and previously used in a study of IBC cases that established IBC immunohistochemical profile [22] and for other purposes [23]. The 547 non-IBC cases were selected from cases included in the database of the Paoli-Calmettes Institute (IPC, Marseille, France) treated between 1990 and 1999. They were defined as T1, T2, T3 tumors (TNM, UICC), collected prior to adjuvant chemotherapy and embedded in paraffin.

The clinical records were reviewed to determine the following patient characteristics: age, metastatic status, therapeutic regimen and survival. Pathological slides were reviewed by two pathologists (JJ, KM) according to the European guidelines [24]. Criteria evaluated were histological type, Elston-and-Ellis grade and, whenever possible, peritumoral vascular invasion.

Two tissue microarrays (TMA) were used, one for the IBCs (ISA set) and the other one for the non-IBCs (IPC set). TMAs were prepared as described, with slight modifications [22‐25]. Briefly, for each tumor, three representative tumor areas were carefully selected from a hematoxylin-eosin-safran stained section of a donor block. Core cylinders with a diameter of 0.6 mm each were punched from each of these areas and deposited into a recipient paraffin block using a specific arraying device (Alphelys, Plaisir, France). In addition to tumor tissues, the recipient blocks also received normal breast tissues and cell lines. Five-μm sections of the resulting microarray blocks were made and used for IHC analysis after transfer to glass slides.

Immunohistochemistry (IHC) was performed on 5-μm sections. The characteristics of the antibodies used and pre-treatment conditions are listed in Table 1. A good concordance of IHC results has been reported between standard full tissue sections and TMA [21]. TMA data were evaluated by the mean score of two cores biopsies minimum for each case. The slides were dewaxed, pre-treated according to the supplier's recommendations (Table 1). This was followed by the use of a streptavidin/biotin kit (Dako, Trappes, France). Diaminobenzidine, (DAB) or 3-amino-9-ethylcarbazole (AEC) was used as chromogen. Sections were counterstained with hematoxylin and coverslipped. Slides were evaluated under a light microscope by two independent observers on the Spot Browser device (Alphelys). Immunoreactivities were classified by estimating the percentage (P) of tumor cells showing characteristic staining (from 0%, undetectable level, to 100%, homogeneous staining) and by estimating the intensity (I) of staining 1, weak staining; 2, moderate staining; 3, strong staining). Results were scored by multiplying the percentage of positive cells by the intensity, i.e. by the so-called quick-score (Q). Internal positive controls such as epidermis or benign breast lobules were used. Hormone receptors (ER, PR) were positive when at least 1% of tumor cell nuclei were stained. For ERBB2, the Dako scale was used; staining was considered as positive when limited to a membrane staining of more than 10% of tumor cells and scored as 1+, 2+ or 3+ according to intensity and partial/complete staining. Protein overexpression was considered for 2+ and 3+. Cytokeratin 5/6, cytokeratin 14 and EGFR were scored positive if any (weak or strong) membranous invasive carcinoma cell staining was observed. Caveolin 1 (CAV1) and Caveolin 2 (CAV2) status of tumor cells was evaluated by light microscopy as either positive (any tumor cell with IHC staining) or negative. For CAV1 negative cases, endothelial cells and interstitial fibroblasts were used as internal positive controls.

Data were summarized by frequencies and percentages for categorical variables. Furthermore, for continuous variables the means, the median and range were computed. To investigate the association between categorical variables, univariate statistical analysis were done, using Pearson's Chi-2 test or Fisher's exact test for small sample size and using non-parametric Mann and Whitney test for continuous variables [26]. Multivariate analysis was done using logistic regression model with backward stepwise selection procedure to evaluate the effect of interactions between the different variables. All statistical tests were two-sided at the 5% level of significance. All the statistical analyses were done using R.2.4.0 statistical software[27].

A total of 91 IBC and 547 non-IBC informative cases from women patients were analyzed. Their characteristics are listed in Table 2. Age was different (p < 0.001) for the two series: for IBCs, age ranged from 22 to 76 years (mean 43.47 years) whereas for non-IBCs, age ranged from 25 to 94 years (mean 59.43 years).

Among the 91 IBCs, 86 were invasive ductal carcinomas (IDC) and 5 were invasive micropapillary carcinomas. The distribution of histological types in non-IBCs was representative of unbiased populations. IBCs were more frequently of grade II and III (65.93%, 30.77%) than non-IBCs (p < 0.001) (Table 2).

IHC was done on IBCs and non-IBCs using three TMA slides for each marker. We compared the expression of the markers only for the infiltrating ductal cases, i.e. 86 IBC cases and 386 non-IBC cases. We observed differences between the two series for: ER negativity (53.85% IBCs, 26.26 % non-IBCs; p < 0.001), PR negativity (52.86% IBCs, 35.39% non-IBCs; p < 0.01), ERBB2 positivity (33.33% IBCs, 14.49% non-IBCs; p < 0.001), CK14 positivity (19.15% IBCs, 5.45% non-IBCs; p < 0.01), P-cadherin positivity (75.93% IBCs, 48.16% non-IBCs; p < 0.001) (Figure 1) and proliferative rate expressed by a Ki67 index (MIB1) higher than 20% (41.1% IBCs, 12.98% non-IBCs; p < 0.001) (Table 3).

We classified tumors in terms of subtypes based on ER/PR and ERBB2 IHC expression [12, 28]. The analysis was restricted to the infiltrating ductal cases (86 IBC and 386 non-IBC), and allowed a classification in three groups (Table 4). ER-positive and PR-positive tumors were classified as luminal. ERBB2-positive tumors were classified as ERBB2-overexpressing. ER-negative and ERBB2-negative (double negative) were considered as basal (ER-/ERBB2-). Using this protein-based definition of subtypes, 33.8% of IBCs vs 15.93 % of non-IBCs (p < 0.001) were basal, 33.33 % of IBCs vs 14.49% of non-IBCs (p < 0.001) were ERBB2-overexpressing, and 29.33% of IBCs vs 58.93% of non-IBCs (p < 0.001) were luminal.