In total, 168 breast cancer tissue specimens were collected for this study from 2005 to 2010 from the Department of Pathology, Beijing Tiantan Hospital Affiliated Capital Medical University (Beijing, China). Of these specimens, 82 were obtained from the left breast and 86 from the right. The mean age of the participants was 54.5 years (range, 27–82 years). Tumors were divided into three groups according to size: 56 patients had a tumor <2 cm in diameter, 91 patients had tumors 2–5 cm in diameter and 21 patients had tumors >5 cm in diameter. Histologically, 126 cases involved invasive ductal carcinoma, not otherwise specified (IDC, NOS), whereas 42 cases were classified according to other categories: 15 cases of invasive lobular carcinomas (ILC), six tubular carcinomas, three medullary carcinomas, eight mucinous carcinomas, three invasive papillary carcinomas, one apocrine carcinoma and six neuroendocrine tumors were observed. According to The Diagnosed and Treated Standard of Breast Carcinoma (2011 version) (8), 46 cases were classified as grade 1 (well differentiated), 66 as grade 2 (moderately differentiated), and 56 as grade 3 (poorly differentiated). In terms of clinical tumor, node, metastasis (TNM) stage, 127 cases were stage I/II and 47 were stage III/IV. There were 43 cases of local and distant lymph node metastasis, and 115 without metastasis. In total, 44 samples were frozen and stored at −70ºC for DNA sequencing. In addition, 72 breast intraductal proliferative lesion and 40 normal breast tissue samples were also obtained.

All specimens were fixed in 10% neutral formalin, embedded in paraffin and cut into 4-μm sections for immunohistochemical staining. The EnVision™ two-step method was used (DAKO, Hamburg, Germany), as well as the following antibodies: antibodies against β-catenin, c-myc and cyclin D1, were purchased from Cell Signaling Technology (Danvers, MA, USA). Prior to staining, sections were pretreated with microwaves for 18 min in a 0.01 M citrate buffer (pH 6.0) for antigen retrieval. 3,3′-Diaminobenzidine was used as a chromogen and 0.01 M phosphate-buffered saline (pH 7.4) was substituted for primary antibodies as a negative control.

Under normal conditions, positive staining of β-catenin was located within the cytomembrane. According to the method described by Maruyama et al(9), the evaluation of β-catenin staining is based on its expression in the cytomembrane, cytoplasm, and cell nucleus. Normal expression was observed in >70% of the cell expressed in the cytomembrane; reduced expression was observed in the remainder. More than 10% cells expressed in the cytoplasm or cell nucleus, which was regarded as ectopic expression. Reduced expression and ectopic expression were regarded as abnormal expression.

Positive axin expression was observed in the cell cytoplasm. The evaluation of axin staining was based on the staining intensity and staining cell area, according to the method described by Nakajima et al(10). Staining intensity scores: 0, no expression; 1, mildly positive; 2, moderately positive; and 3, markedly positive. Staining cell area scores: 0, <10% of cells stained; 1, 11–25% stained; 2, 26–50% stained; 3, >50% stained. A sum of scores >2 was considered to be positive, whereas a score of <2 was regarded as negative.

A positive expression of cyclin D1 was observed in the cell nucleus, and c-myc was in the cell nucleus and/or cytoplasm. The slides showed either negative (−), positive (+), or overexpression, depending on the count of positive cells; <10% was regarded as negative, 10–50% was regarded as positive and >50% was regarded as overexpression.

DNA was extracted using the phenol-chloroform method for polymerase chain reaction (PCR) amplification with the upstream primer (5′-GCTGATTTGATGGAGTTGGA-3′) and the downstream primer (5′-GCTACTTGTTCTTGAGTGAA-3′). PCR conditions included denaturation at 94ºC for 3 min, amplification for 35 cycles for 40 sec at 94ºC, annealing at 54ºC for 40 sec, and extension at 72ºC for 60 sec. Finally, the samples underwent extension at 72ºC for 10 min, followed by cooling at 12ºC. The size of the PCR products was 199 bp. Synthesis of PCR primers, and purification and direct sequencing of PCR products, was performed by the Beijing Genomics Institute (Beijing, China).

Statistical analysis was performed using the Pearson χ² test. The likelihood ratio and Spearman rank correlation coefficient analysis was performed by SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

In the 40 cases with normal breast tissue, the epithelial cells revealed an equally strong membranous expression of β-catenin protein at the cell-cell boundaries. Axin expression was positive in 87.5% (35/40) of cases, but the expression of cyclin D1 and c-myc were negative in all cases. In the 72 cases of breast intraductal proliferative lesions, the abnormal rate of β-catenin expression was 44.44% (32/72), and the expression of axin was positive in 68.06% (49/72) of cases. The overexpression of cyclin D1 and c-myc were observed in 33.33% (24/72) and 27.78% (20/72) of cases, respectively. In 168 cases of breast carcinoma, positive axin staining was observed in the cytoplasm in 47.62% (80/168) of cases, amongst which one sample exhibited nuclear axin expression and 11 samples showed membrane and cytoplasmic axin expression (Fig. 1). The abnormal expression rate of β-catenin was 70.83% (119/168). The majority of β-catenin expression was observed in the cytoplasm, whereas membranous staining was reduced or disappeared (Fig. 2). The nuclear expression of β-catenin was observed in 27.38% (46/168) of cases. The positive expression of cyclin D1 was observed in the cell nucleus (Fig. 3) and c-myc was observed in the cell nucleus and/or cytoplasm (Fig. 4). The overexpression of cyclin D1 and c-myc was observed in 52.98% (89/168) and 50.00% (84/168) of cases, respectively.

The abnormal expression rate of β-catenin was higher in breast carcinoma compared with breast intraductal proliferative lesions (119/168, 70.83% vs. 32/72, 44.44%; P<0.01). Positive axin expression was lower in breast carcinoma compared with breast intraductal proliferative lesions (80/168, 47.62% vs. 49/72, 68.06%; P=0.004), and also compared with normal breast tissues (80/168, 47.62% vs. 35/40, 87.50%;P<0.01). The overexpression of cyclin D1 and c-myc was higher in breast carcinoma compared with breast intraductal proliferative lesions (89/168, 52.98% vs. 24/72, 33.33%; P<0.01 and 84/168, 50.00% vs. 20/72, 27.78%;P<0.01), respectively.

Axin expression was inversely correlated with tumor size, histological grade, clinical TNM stage and lymph node metastasis, but not with patient age or left/right location of the breast carcinoma (Table I). The rate of expression in tumors with a diameter of >5 cm was lower than those 2–5 cm in diameter (5/21, 23.81% vs. 44/91, 48.35%; P=0.041) and those <2 cm in diameter (5/21, 23.81% vs. 31/56, 55.36%; P=0.038), but no difference in axin expression was observed between tumors with a diameter of <2 cm and those between 2–5 cm in diameter (31/56, 55.36% vs. 44/91, 48.35%; P=0.409). The difference in expression was also observed between tumors with histologic grade I and grade II (37/46, 80.43% vs. 30/66, 45.45%; P<0.01), grade II and grade III (30/66, 45.45% vs. 13/56, 23.21%; P<0.01), and grade I and grade III (37/46, 80.43% vs. 13/56, 23.21%; P<0.01). A higher expression rate of axin expression was observed in patients with clinical TNM stage I/II of the disease compared with those with stage III/IV of the disease (69/127, 54.33% vs. 11/41, 26.83%; P<0.01). The positive expression rate of axin was 24.53% (13/53) in tumors with lymph node metastasis, in contrast to 58.26% (67/115) in tumors without lymph node metastasis (P=0.002) (Table I).

The abnormal expression of β-catenin and the overexpression of cyclin D1 was correlated with tumor size, histological grade, clinical TNM stage and lymph node metastasis, but not with patient age or the location (left/right) of the breast carcinoma (Table I). The abnormal expression of β-catenin in tumors >5 cm in diameter was higher than in those 2–5 cm in diameter (20/21, 95.24% vs. 67/91, 73.63%; P=0.032) and in those <2 cm in diameter (20/21, 95.24% vs. 32/56, 57.14%; P=0.001). It was also higher in tumors 2–5 cm in diameter compared with those <2 cm in diameter (67/91, 73.63% vs. 32/56, 57.14%; P=0.038). The overexpression of cyclin D1 in tumors <2 cm was less than in those between 2–5 cm in diameter (21/56, 37.50% vs. 52/91, 57.14%; P=0.021) and those tumors >5 cm in diameter (21/56, 37.50% vs. 16/21, 76.19%; P=0.002). However, no difference was observed in the overexpression of cyclin D1 between tumors >5 cm and those 2–5 cm in diameter (16/21, 76.19% vs. 52/91, 57.14%; P=0.107). There was a significant difference in the expression of β-catenin and cyclin D1, respectively, between tumors with a histological grade of grade II and grade III (40/66, 60.61% vs. 49/56, 87.50%; P=0.001; 30/66, 45.45% vs. 39/56, 69.64%; P=0.007), and between those with a grade I and grade III (30/46, 65.22% vs. 49/56, 87.50%; P=0.007; 20/46, 43.48% vs. 39/56, 69.64%; P=0.008). No difference was observed between grade I and grade II tumors (30/46, 65.22% vs. 40/66, 60.61%; P=0.620 and 20/46, 43.48% vs. 30/66, 45.45%; P=0.836), respectively. Significant differences in the abnormal expression of β-catenin (82/127, 64.57% vs. 37/41, 90.24%; P=0.002) and the overexpression of cyclin D1 (60/127, 47.24% vs. 29/41, 70.73%; P=0.009) were observed between stage I/II and III/IV of breast carcinoma. Higher rates of abnormal β-catenin expression (44/53, 83.02% vs. 75/115, 65.22%; P=0.018) and the rate of overexpression of cyclin D1 (39/53, 73.58% vs. 50/115, 43.48%) were observed in the primary tumors of patients with lymph node metastasis compared with those who did not.

The overexpression of c-myc does not correlate with age, left/right side location of the breast carcinoma, and clinical TNM stage, but does correlate with tumor size, histologic grade, and lymph node metastasis (Table I). The rate of overexpression of c-myc in tumors <2 cm in diameter was lower than those with tumors 2–5 cm in diameter (18/56, 32.14% vs. 54/91, 59.34%; P=0.001) and those >5 cm in diameter (18/56, 32.14% vs. 12/21, 57.14%; P=0.045), but no difference in expression was observed between tumors >5 cm in diameter and those between 2–5 cm in diameter (12/21, 57.14% vs. 54/91, 59.34%; P=0.854). Significant differences in the expression of c-myc were observed between tumors with a histological grade I and grade II (14/46, 30.43% vs. 33/66, 50.00%; P=0.039) and between those with grade I and grade III (14/46, 30.43% vs. 37/56, 66.07%; P<0.01), but no difference was observed between histological grade II and grade III tumors (33/66, 50.00% vs. 37/56, 66.07%, P=0.074). The overexpression of c-myc in primary tumors was higher in patients with lymph node metastasis than in those without metastasis (34/53, 64.15% vs. 50/115, 43.48%, P=0.013).

The abnormal expression of β-catenin was observed in 74/119 (62.18%) cases with cyclin D1-overexpressing tumors and in 65/119 (54.62%) cases with c-myc overexpressing tumors. There was a positive and significant association between the abnormal expression of β-catenin and the overexpression of cyclin D1 in breast cancer (r=0.288, P<0.01), but the abnormal expression was not associated with the overexpression of c-myc in breast cancer (P=0.062) (Table II).

An increased rate of axin expression was observed in tumors with a nuclear expression of β-catenin compared with those without β-catenin nuclear expression (8/46, 17.39% vs. 72/122, 59.02%; P<0.01), but the expression of axin was similar in tumors with abnormal (reduced) expression of β-catenin to those with normal β-catenin expression (57/119, 47.90% vs. 23/49, 46.94%; P=0.910) (Table III).

The PCR products of β-catenin displayed bands at ~200 bp (Fig. 5). However, mutations in the β-catenin gene were not detected in 44 cases of breast carcinoma.