Expression of bcl-2 protein predicts efficacy of adjuvant treatments in operable node-positive breast cancer.
Gasparini G., Barbareschi M., Doglioni C., Palma PD., Mauri FA., Boracchi P., Bevilacqua P., Caffo O., Morelli L., Verderio P.
The proto-oncogene bcl-2 encodes a protein that inhibits apoptosis, a common mechanism of cell death caused by hormone and chemotherapy. We have analyzed bcl-2 protein expression by immuno-cytochemistry in primary node-positive breast cancers in two groups of patients (for a total of 180 cases). One group received adjuvant hormone therapy, the other chemotherapy (cyclophosphamide, methotrexate, and fluorouracil), and both groups were followed for a median time of 63 months. We compared our findings with conventional clinicopathological indicators [menopausal status, number of axillary nodes, histological grade, tumor size and type, estrogen receptor (ER), and progesterone receptor] and with p53 protein expression. bcl-2 protein was present in 65% of the carcinomas (117/180) and it was significantly associated with ER and progesterone receptor and inversely associated with p53 in both the groups of patients treated with adjuvant chemotherapy and tamoxifen. In patients treated either with adjuvant chemotherapy or tamoxifen, relapse-free survival at 5 years was significantly better among patients with bcl-2-positive tumors than in those with bcl-2 negative ones (P = 0.05 and 0.02, respectively). As far as overall survival is concerned, patients with bcl-2-positive tumors had a significantly better outcome in the group treated with adjuvant chemotherapy (P = 0.03). Multivariate analyses were performed for the two treatment groups. In the group treated with tamoxifen, lack of expression of ER and of bcl-2 was the only significant and independent predictor for poor relapse-free survival (P < 0.01). A number of nodes above 3 was the only significant and independent predictor for poor overall survival (P < 0.01). In the cyclophosphamide-methotrexate-fluorouracil-treated group, bcl-2 absence was significant for poor overall survival (P = 0.02) as well as a number of nodes above 3 (P = 0.04) and a tumor size above 2 cm (P = 0.05). For poor relapse-free survival only a number of nodes above 3 (P < 0.01) and progesterone negativity (P = 0.02) were significant and independent predictors of a higher probability of relapse. Thus, in contrast to in vitro data on drug resistance, bcl-2 expression was associated with better outcomes in patients treated with hormone and chemotherapy. Overall, these results suggest that expression of bcl-2 protein and the number of metastatic lymph nodes are independent features predictive of clinical outcome in patients with node-positive breast cancer, irrespective of the type of adjuvant treatment. The determination of bcl-2 protein may prove to be a useful tool to distinguish patients for whom conventional forms of adjuvant therapy are beneficial from those with bcl-2 negative and ER-negative tumors for whom novel therapeutic strategies are needed.