Polymorphisms of BCL2 and BAX Genes Associate with Outcomes in Advanced Non-small cell lung cancer Patients treated with platinum-based Chemotherapy

Single-nucleotide polymorphisms (SNP) of the gene belonging to the BCL2 family are thought to play a role in chemotherapy resistance. This study investigated the association of BCL2-938C>A(rs2279115) and BAX-248G>A(rs4645878) promoter region SNPs and the clinical responses and outcomes of 235 non-small cell lung cancer (NSCLC) patients treated with platinum-based chemotherapy. The data suggested that BAX-248GA and GA+AA genotype was associated with poor response [odds ratio (OR) 1.943, p = 0.039; OR 1.867, p = 0.038, respectively] to chemotherapy, and BCL2-938CA, CA+AA and BAX-248GA, AA and GA+AA were associated with poor progression-free survival (PFS) [hazard ratio (HR) 1.514, p = 0.004; HR 1.456, p = 0.009; HR 1.449, p = 0.013; HR 2.006, p = 0.010; HR 1.506, p = 0.003, respectively] and BCL2-938CA, AA and CA+AA and BAX-248GA, AA and GA+AA were associated with poor overall survival (OS) (HR 2.006, p < 0.001; HR 2.322, p < 0.001; HR 2.096, p < 0.001; HR 1.632, p = 0.001; HR 2.014, p = 0.010; HR 1.506, p < 0.001, respectively). Furthermore, combination of these two polymorphisms showed patients with 2–4 variant alleles of these two genes associated with poor PFS and OS (HR 1.637, p = 0.001; HR 2.365, p < 0.001). The data from the current study provide evidence that BCL2-938C>A and BAX-248G>A polymorphisms may be useful in predicting clinical outcomes of patients with advanced inoperable NSCLC to platinum-based chemotherapy.


Material and Methods
Study Population, Response Assessment, Toxicity Evaluation and Survival Calculation. In this study, we enrolled a total of 235 inoperable NSCLC (stage III/IV) patients treated with at least two cycles of first line platinum-based chemotherapy between July 2007 and July 2012 from Daping Hospital, The Third Military Medical University (Chongqing, China). The study was approved by the ethics committee of the Daping Hospital and also carried out according to the protocols approved by the ethics committee. Only patients who understood the purpose of the study and signed the informed consent were included in the study. All patients had routine blood, hepatic and renal function tests, and an electrocardiogram. Patients had not received previous chemotherapy or radiotherapy and also had no other malignancies in the 5 years preceding this study. Patients were assessed for their Eastern Cooperative Oncology Group performance status (ECOG PS) and all participants were graded as level 0 ~ 2 before chemotherapy. Patients were treated with 75 mg/m 2 cisplatin on day 1 plus 135 mg/m 2 Taxol, 75 mg/m 2 docetaxel, or 1000 mg/m 2 gemcitabine on day 1 and day 8. The therapy cycles were repeated every 3-4 weeks.
Standard Response Evaluation Criteria in Solid Tumors (RECIST 1.0) were used to evaluate the treatment response, and the response was assessed by comparison of the baseline MRI or CT images with the follow-up images after every two cycles of chemotherapy. Patients were categorized as responders (complete response and partial response; CR or PR) or nonresponders (stable disease and progressive disease; SD or PD).
Chemotherapy-related toxicities were recorded for each treatment cycle, including leukocytopenia, anemia, thrombocytopenia, nausea, vomiting, diarrhea, neuropathy, weakness, hypersensitivity reaction, and renal toxic effects. Grade 3/4 toxicity (defined by the National Cancer Institute common toxicity criteria version 3.0) was assessed twice a week during chemotherapy.Progression-free survival (PFS) was calculated from the start of treatment to documentation of the first date of disease progression (death was considered a progression event in patients who died before disease progression). Overall survival (OS) was calculated from the start of treatment to death. Patients without documented death or objective progression at the time of the final analysis were censored at the time of their last objective tumor assessment or at the date they were last known to be alive. The survival data in this study were censored on July 23, 2014. Ten (4.3%) cases were censored. Genotyping. A 5 mL whole-blood sample was obtained from each patient before chemotherapy.
Genotyping of BCL2-938C> A SNPs and BAX-248G> A was conducted by using the PCR-restriction fragment length polymorphism (PCR-RFLP) method. Each PCR amplification was performed in a 25 μ l reaction mixture containing 12.5 pmol of each primer, 2 μ l genomic DNA, 0.5 U of TaqMan SNP Genotyping Assay Mix (40x), 0.30 mM of dNTPs, 1.5 mM total MgCl 2 , and 5 μ l of PCR buffer (5×).
The primers used to detect the BCL2-938C> A polymorphism were 5′ -CTGCCTTCATTT ATCCAGCA-3′ and 5′ -GGCGGCAGATGAATTACAA-3′ 11  The primers for the BAX-248G> A polymorphism were 5′ -CGGGGTTATCTCTTGGGC-3′ and 5′ -GTGAGAGCCCCGCTGAAC-3′ ' 28 . The PCR conditions consisted of an initial denaturation at 95 °C  for 5 min, followed by 40 cycles of 30 s at 94 °C, 30 s at 56 °C, and 45 s at 72 °C and a final extension at  72 °C for 5 min. The PCR products were then digested by Aci I (Aci I, which recognizes CCGC, New England BioLabs) at 37 °C for 15 min. Homozygous GG alleles (wild-type) were visualized as three major bands of 352, 256, and 96 base pair (bp) with the highest intensity for the 256-bp band under the UV light. In the heterozygous condition (AG alleles) the 256bp and 96 bp products are fainter, while the 352bp band is more intense. In contrast, the homozygous AA alleles (homozygous carrier of this SNP) showed only one 352-bp band.
Statistical Analyses. Statistical analyses were performed using SPSS (IBM SPSS Statistics 19, SPSS, Chicago, IL, USA). Differences in response to treatment or grade 3/4 toxicities were analyzed between groups with different genotypes of these SNPs (variants vs. the wild-type) by calculation of p values using the Pearson χ 2 test or the Fisher exact test. Multivariate logistic regression analyses were used to assess associations between the selected SNPs and chemotherapy response or toxicity by estimating odds ratios (ORs) and 95% confidence intervals (CIs). Survival curves for PFS and OS were determined using the Kaplan-Meier method and the Log-rank test was used to generate p values. Multivariate Cox proportional hazards models (Forward Stepwise: Likelihood Ratio) were used to estimate adjusted hazard ratios (HR) with 95% confidence intervals. Multivariable regressions analysis was adjusted for age, gender, ECOG stage, histology, disease stage, smoking status, weight loss, and chemotherapy regimens. Two-sided p-values of less than 0.05 were considered statistically significant. In order to adjust the p-value due to multiple comparisons, the Bonferroni-Holm method was used to calculate the p-value for the results of any SNP.  Table 2).
Association of Genotypes with Treatment Response and Toxicity. The overall response rate of patients to platinum-based chemotherapy was 35.7% with the following responses to the first-line treatments: one patient had CR (0.4% of evaluable patients), 83 PR (35.32%), 75 SD (31.9%), and 76 PD (32.3%). A numerically increased rate of response was observed in patients with the BAX-248GG vs. the GA or GA+ AA phenotype but the differences did not reach statistical significance after multiplicity adjustment (42.6% vs. 27.6%, p = 0.024; 42.6% vs. 27.4%, p = 0.015). No statistically significant association was observed between BCL2-938C> A and responses to chemotherapy (  After that, we combined these two BCL2-938C> A and BAX-248G> A polymorphisms for association with PFS and OS of patients. Our data showed that the -938A and -248A were classified as adverse alleles based on association with higher risk of progression and death presented in Table 4 and as the number of adverse allele increased, the median PFS and OS decreased. Patients with more than 2 adverse alleles had shorter median PFS (5 m vs. 8 m, p = 0.001) and OS (12 m vs. 18 m, p < 0.001) compared to those with 0-1 adverse alleles (Table 4, Fig. 2. tif). The multivariate analysis showed that patients with 2-4 adverse alleles had an increased risk of disease progression (HR 1.637, 95% CI 1.240 ~ 2.161, p = 0.001) and death (HR 2.365, 95% CI 1.760-3.178, p < 0.001). These significances remained after the Bonferroni-Holm method analysis.

Discussion
Cisplatin resistance occurs through a variety of mechanisms, such as changes in cellular uptake and efflux of the drug, increased expression of detoxification enzymes, increased DNA repair or inhibition of apoptosis 29 . In this study, we investigated whether the occurrence of SNPs, located in the promoter regions of two apoptosis-related genes are associated with responses and/or outcomes in patients with advanced NSCLC that are treated with cisplatin-based chemotherapy. A numerically increased rate of response was observed in patients with the BAX-248GG vs. the GA or GA+ AA phenotype but the difference did not reach statistical significance after multiplicity adjustment. We found no significant association between the BCL2-938C> A polymorphism and chemotherapy response. There was no statistical association between BCL2-938C/A or BAX-248G> A polymorphism and grade 3/4 hematologic or gastrointestinal toxicity in these patients. Furthermore, patients with the BCL2-938C> A variant genotype (A allele) or BAX-248G> A variant genotype (A allele) associated with poor PFS and OS. The combined BCL2-938C> A and BAX-248G> A were also associated with PFS and OS of the patients. The multivariate analysis showed that ECOG PS, BCL2-938C> A, and BAX-248G> A were all independent predictors for OS of these NSCLC patients. To the best of our knowledge, this is the first study of this kind to demonstrate an association between the BAX-248G> A or the combination of BCL2-938C> A and BAX-248G> A with outcome of advanced NSCLC patients to cisplatin-based chemotherapy.
BCL2 is localized to chromosome 18q21.3 30 , coding a protein with three exons and two gene promoters (P1 and P2). These two promoters have different functions in regulation of BCL2 expression, i.e.,   20 . Expression of the anti-apoptotic BCL2 protein should contribute to resistance to cisplatin-induced tumor cell apoptosis and thus lead to a poor response and outcome. Indeed, BCL2 overexpression has been shown to be a marker of chemotherapy resistance in both SCLC and NSCLC 39,40 . However, Fontanini et al. 41 reported that survival probability was higher in patients with BCL2-expressing, resected NSCLC because of the less aggressive behavior of NSCLC with BCL2 overexpression. A meta-analysis of 28 studies revealed that BCL2 overexpression had a positive influence on survival of NSCLC 42 . However, further study is needed to clarify this discrepancy. BAX is localized to chromosome 19q13.3, coding a protein with six exons and a promoter 43 . The BAX promoter binds to different transcription factors or proteins, such as p53 response elements, the TATA box, canonical E-boxes, and the NF-κ B binding site to regulate BAX expression 44 . BAX was extensively studied in different types of cancer such as pancreatic cancer 45 , colon cancer 46,47 , esophageal cancer 48 , lung cancer 49,50 , squamous cell carcinoma of the head and neck 11 , prostate carcinoma 51 , ovarian carcinoma 52 , and breast cancer 53 . Recently, BAX-248G> A (rs4645878) was reported to be associated with reduced expression of BAX protein and altered susceptibility to chronic lymphocytic leukemia 25,26 , although a meta-analysis of seven independent studies with 1772 cases and 1708 controls revealed that neither allele frequency nor genotype of BAX-248G> A associated with risk of human cancer using different genetic models 54 . Several studies reported that patients with BAX low expression had a significantly longer median survival in NSCLC 50 , esophageal squamous cell carcinoma 48 , and colon cancer 47 . Moreover, Paola Perego et al. 55 showed that p53 mutations developed cisplatin resistance in ovarian cancer as a consequence of the loss of p53 transactivation of BAX expression. Our current SNP data further confirm the role of BAX in the regulation of cisplatin resistance and shorter PFS and OS of patients with advanced NSCLC.
Indeed, previous studies showed that BAX was able to heterodimerize with BCL2 and Mcl-1 and that overexpression of BCL2 and Mcl-1 proteins compromised the proapoptotic capacity of BAX 56,57 . Therefore, we determined to assess whether BCL2 and BAX interact synergistically to contribute to cisplatin resistance and alter the PFS and OS of patients with advanced NSCLC. We found that as the number of variant alleles increased, the median PFS and OS were decreased accordingly. Patients with more than 2 variant alleles had a much shorter median PFS and OS compared to those carrying 0-1 variant alleles.
Our current study is just the first of this kind for proof-of-principle. Although we showed that the BCL2-938C> A and BAX-248G> A SNPs significantly associated with platinum-based chemotherapy response, PFS and OS of patients with advanced NSCLC, we did not show the association of these SNPs with chemotherapy-related toxicities. However, the mechanism by which BCL2 SNPs may influence the clinical outcomes of patients with advanced NSCLC to platinum-based chemotherapy is unclear. This raises the question of whether a patient with such polymorphisms would have better survival than those without these polymorphisms in the absence of treatment. Do the polymorphisms actually affect the effectiveness of the regimen, or do they simply afford a better prognosis to patients? Hence, further work is necessary; for example, a larger sample size and a prospective study would be able to confirm our current data. Variation of BCL2 and BAX expression associates with an altered sensitivity and clinical outcome of NSCLC patients to chemotherapy.