BCL-2, topoisomerase IIα, microvessel density and prognosis of early advanced breast cancer patients after adjuvant anthracycline-based chemotherapy

Adjuvant systemic therapy of patients with early advanced breast cancer (BC) significantly reduces recurrence risk and overall mortality. Anthracyclines (ATC) are the most commonly used drugs in the adjuvant chemotherapy (CT) of breast cancer. The last meta-analysis of Early Breast Cancer Trialists’ Collaborative Group demonstrated that regimens based on higher dosages of ATC are associated with around 4 % reduction of recurrence rate and overall mortality as compared with cyclophosphamide, methotrexate and fluorouracil scheme (CMF) (EBCTCG et al. 2012). There is a possibility to improve treatment outcome by more differentiated therapy, tailored to the biological characteristics of the tumour cells. For this purpose, it is necessary to undertake translational researches, consisting in the development of biological prognostic and predictive factors for adjuvant chemotherapy based on ATC. Today, there is no reliable prognostic factor which allows discrimination between patients who will benefit from adjuvant ATC treatment and those who will not. In our previous papers, in the group of 167 patients with breast cancers in clinical stage T1–T2, N1–N2, M0, treated with ATC in adjuvant setting, we have shown prognostic potential of topoisomerase IIα (TOPOIIα) expression (Biesaga et al. 2011) and tumour microvessel density (MVD) (Biesaga et al. 2012). However, chemotherapeutic agents such as ATC, irrespective of their intracellular target, act through induction of apoptosis in cancer cells. Therefore, alterations in the regulatory mechanism of apoptosis are responsible not only for cancer progression, but also for different response to treatment.

Antiapoptotic protein B cell lymphoma 2 (BCL-2) is one of the most extensively studied apoptosis regulators in translational studies, but its role in ATC resistance/sensitivity is unclear. There are a few reports (see Table 1), in which prognostic potential of BCL-2 expression was investigated in the breast cancer patient groups consisting of subgroups of women treated with ATC in adjuvant setting, although in all these studies, expect one (Abdel-Fatah et al. 2010) not analysed separately. In these studies, contradictory results were obtained. Moreover, in two meta-analyses, performed to investigate BCL-2 prognostic potential, opposite results were reported. Callagy et al. (2006) have shown positive prognostic significance of BCL-2 overexpression, whereas Yang et al. (2013) have found significant relation between BCL-2 negativity and good prognosis. These contradictions may be the effect of at least four causes: (1) other than antiapoptotic function of BCL-2 [the influence on cell cycle progression (Zinkel et al. 2006; Pietenpol et al. 1994)], (2) regulation of BCL-2 expression through oestrogen pathway (Perillo et al. 2000), (3) lack of homogeneity in patient group studied (according to clinical stage and adjuvant treatment type) and (4) differences in IHC procedure and cut-off points used to identified BCL-2 positivity and negativity (Table 1). Thus, there is still a need to evaluate the prognostic role of BCL-2 in BC patients.

The aim of the present study was to evaluate the prognostic significance of BCL-2 expression in the group of 172 women with breast cancer in clinical stage T1–T2, N1–N2, M0, treated with ATC in adjuvant settings. For this purpose, we assessed the correlation between this parameter and patient disease-free survival (DFS). Because in the present patient group we reanalysed follow-up data, we decided to assess once again the influence of all clinical (patients’ age tumour size, lymph node metastases, grade) and biological features (expression of steroid hormone receptors, cytokeratin 5/6 (CK5/6), HER2, TOPOIIα, Ki-67, P53 and MVD), which were previously studied (Biesaga et al. 2011, 2012) on DFS of breast cancer patients.

In the present study, we have shown, to the best of our knowledge, for the first time, 100 % DFS for patients having tumours characterized by low topoisomerase IIα expression and high microvessel density and BCL-2 positivity (all assessed by immunohistochemistry). The results suggest that these three parameters may constitute an input into the new prognostication model specially dedicated to adjuvant anthracycline-based chemotherapy of breast cancer patients. This suggestion is in agreement with other results showing that BCL-2 expression can improve prognostication of BC patients treated with different CT regimens. Abdel-Fatah et al. (2010), in the group of 245 patients with invasive BC treated with adjuvant ATC-based chemotherapy, found a good prognosis for patients with active or partially inactive P53 pathway (assessed by expression of BCL-2, p21, p27, hormone receptors, P53, Ki-67, EGFR, CK 5/6 and HER2 gene amplification status). In turn, Hwang et al. (2012) assessed prognostic influence of BCL-2 expression in the group of 7,230 primary BC patients in clinical stage 1, 2 or 3, of whom 4,532 received different systemic therapy. They have shown that the risk groups of cancer patients, identified by St. Gallen model or Nottingham Prognostic Index model or TNM staging, could be subdivided by addition of the BCL-2 assessment in all three models. Callagy et al. (2006) have also found that BCL-2 was the only marker from a panel of 13 proteins (HER2, c-Myc, cyclin E, CK5/6, CK17, CK8/18, ER, Ki-67, MCM-2, p27, P53 and PgR) that improved prognostication in the group of 930 breast cancer patients, some of whom were treated with adjuvant CT.

The positive prognostic significance of low TOPOIIα expression and high MVD is in concordance with results obtained by some authors (TOPOIIα: Fritz et al. 2005; Depowski et al. 2000; MVD: Gunel et al. 2002; Protopapa et al. 1993). The positive effect of low TOPOIIα may be explained by hypothesis suggesting that low enzyme expression is related to less aggressive tumour phenotype, characterized by low histological grade, hormone receptors positivity, CK5/6 negativity and low proliferation rate, what was shown by us (Biesaga et al. 2011) and other authors (Fritz et al. 2005; Durbecq et al. 2004; Depowski et al. 2000). In turn, the positive influence of high MVD on ATC treatment outcome may be related to enhancement of production of reactive oxygen species under ATC or increased drug access to tumour cells in well-vascularized tumours. All these suggestions were carefully discussed in our earlier publications (Biesaga et al. 2011, 2012); therefore, in the present paper, we would like to focus on prognostic potential of BCL-2 expression.

We have shown that BCL-2 expression is independent prognostic factor for adjuvant chemotherapy with ATC, what is in agreement with results of a few other reports in which analysis of subgroups of patients treated with ATC in adjuvant settings was provided (see Table 1). Some of these data were summarized in meta-analysis performed by Callagy et al. (2008), who included seventeen papers published since 2006, reporting data of almost 6,000 patients with BC, regardless of the treatment type. This meta-analysis supports that BCL-2 overexpression indicates good prognosis. This effect was independent of lymph node status, tumour size or grade as well as a range of other biological variables. However, another meta-analysis from 2013, in which 23 studies involving almost 2,500 BC patients, showed that negative BCL-2 expression was associated with pathological complete response in women treated with ATC-based neoadjuvant chemotherapy (Yang et al. 2013). We mentioned about the possible causes of these contrary results in Introduction section.

The reasons of positive prognostic potential of BCL-2 overexpression are not fully understood. It is well known that this protein has oncogenic function related to inhibition of apoptosis and hence tumour progression and metastatic potential (Short and Johnstone 2012). However, in some types of tumour, BCL-2 seems to play both oncogenic and suppressive role. It is postulated that the dominance of one of these functions over another may depend on the cell type and physiology and that suppressive effect dominates in solid epithelial tumours including breast cancer (Zinkel et al. 2006). BCL-2 suppressive function may be a consequence of its non-apoptotic functions, particularly its ability to control cell cycle machinery, what was supported by cell line studies in which relation between BCL-2 expression and G0 prolongation and delay of G1-S transition was shown (Zinkel et al. 2006; Pietenpol et al. 1994).

Positive prognostic significance of BCL-2 overexpression may be explained by its influence on BC cell chemosensitivity. Del Bufalo et al. (2002) have shown that human BC cell lines overexpressing BCL-2 displayed an increased sensitivity to drugs known to induce multidrug resistance phenomenon (MDR) such as doxorubicin, vincristine, vinblastine and actinomycin, whereas they show increased resistance to drugs not related to the MDR (cisplatin and bischloroethylnitrosourea). The enhanced sensitivity to doxorubicin was associated with an increased drug accumulation and its decreased efflux as well as with reduction of the adenosine triphosphate level and kinase C activity, both of which participate in the regulation of MDR. In earlier studies, the above-referred authors have demonstrated that BCL-2 overexpression was related to inhibition of mitochondrial metabolism and reduction of oxygen uptake and ¹⁴CO₂ production (Biroccio et al. 1999).

The positive prognostic power of BCL-2 overexpression may be also related to its correlation with other favourable prognostic markers such as lower tumour grade, oestrogen and progesterone positivity, cytokeratin negativity, lack of HER2 overexpression, which was shown in our study and is in agreement with other authors (van Slooten et al. 1996; Mottolese et al. 2000; Yang et al. 2003; Lee et al. 2007; Kim et al. 2010). We have also found the higher percentage of BCL-2-positive tumours among luminal A and B cancer immunophenotypes as compared with HER2⁺ and TN immunophenotypes. Therefore, BCL-2 expression may also reflect lower degree of tumour differentiation.

In the presented analysis, based on BCL-2 expression, we could stratify patients with triple negative tumours into two subgroups with different prognosis. TN breast cancer patients with BCL-2 overexpression had significantly higher DFS than those with BCL-2 negativity. These results should be, however, interpreted with care, because of the small number of patients in each subgroup (7 and 13, respectively). Nonetheless, presented results are partly in agreement with those obtained by other authors. Abdel-Fatah et al. (2013) in the large cohort of 635 patients with early advanced TN breast cancers have found that BCL-2 negativity was associated with approximately twice the risk of death and recurrence in patients who did not receive CT or were treated with CMF. In turn, they have shown longer survivals in BCL-2-negative TN BC patients exposed to anthracyclines than in those who did not receive chemotherapy or were treated with CMF. Relatively high DFS in BCL-2-positive TN tumours obtained by us could also confirm results of earlier reports, showing that a subset of TN breast cancer patients responded to standard chemotherapy and had survival rate similar to women with other breast cancer immunophenotypes (Liedtke et al. 2008). However, Tawfik et al. (2012), contrary to us, have demonstrated association between BCL-2 positivity and poorer survival only among non-TN breast cancer patients, whereas in the group of TN breast cancer patients, BCL-2 status had no influence on treatment outcome. These discrepancies may be explained by heterogeneity of TN subtype, which represents distinct molecular subgroups that differentially respond to chemotherapy. Lehmann et al. (2011), analysing gene expression profiles, have identified six distinct molecular subtypes of TNBC including two basal like, an immunomodulatory, a mesenchymal, a mesenchymal stem-like and a luminal androgen receptor subtype. In the next report of this team, in the group of 130 patients with TN BC who received neoadjuvant chemotherapy containing sequential taxane and ATC chemotherapy, significant relation between TN BC subtypes and pathological complete response status was found (Masuda et al. 2013). Thus, in the light of our and some other author’s results, we postulate that BCL-2 expression may be one of the surrogate markers allowing distinguishing TN BC subtypes with different sensitivity to ATC; however, this supposition requires confirmation in further studies.

In conclusion, in the present study, we have shown, to the best of our knowledge for the first time, that by combination of three parameters: BCL-2 positivity, low topoisomerase IIα expression and high microvessel density, it is possible to identify subgroups of BC patients with very good prognosis after adjuvant anthracycline-based chemotherapy. All these three parameters were assessed by immunohistochemistry, which is an inexpensive and routine method, widely used in pathology and radiobiology departments. Therefore, assessment of these proteins is easy for implementation in daily clinical practise. However, the presented results should be confirmed in larger, prospective clinical trials.