Present study was approved by Institutional ethics committee of Maulana Azad Medical College & associated hospitals, New Delhi. 160 peripheral blood samples from newly diagnosed as well as therapy received NSCLC patients and gender-age (±5 years) matched 160 cancer free healthy controls were included. Patients with a history of previous cancer or metastasized cancer from other organs except lung were excluded. Both cases as well as controls belonging to north India. All subjects received a detailed description of this study and provided written informed consent. Patient follow up was obtained through of hospital records as well as by direct patient contact. The distribution of the selected characteristics for those with NSCLC and the controls are summarized in Table 1. Tumor stage and grade of NSCLC patients were determined using 7^th edition of the Lung cancer staging system (2009) and similarities between cases and controls were maintained with respect to characteristics like age, sex, or smoking history.

Blood sample (3 - 5 mL) from each participant was drawn into an EDTA vial and genomic DNA was extracted following manufacturer’s protocol of DNA sure blood mini kit (Genetix Biotech Asia Pvt. Ltd, India). Primer- introduced restriction analysis (PIRA)-polymerase chain reaction (PCR) assay (18) was used to determine the G(-248)A polymorphism (rs4645878) in the promoter region of BAX gene. PIRA-PCR method is usually used where a variant or mutation does not create a new restriction site, it may be possible to perform PCR by using a

primer containing a mismatch with the target DNA, such that the same base substitution is introduced into both normal and mutated sequences but a restriction site is formed only in the PCR product derived from the variant allele. The method introduces an artificial restriction site into a PCR product by the use of a primer with a single-base mismatch usually introduced near the 3' end of the primer that is close to the nucleotide change of interest. In the present study the mismatch (-2 G-to-C) was present in the sense primer shown in Table 2 as underlined. The mismatch introduced an artificial restriction site for MSP I restriction enzyme in the PCR product which resulted in two band sizes (89 bp + 20 bp) when G allele is present at BAX (-248, rs 4645878) position.

PCR was performed in a final volume of 50 μL containing 50 ng genomic DNA, 5 μL of 10× polymerase buffer (200 mM Tris-HCl, pH 8.0; 500 mM KCl), 3 μL of 20 mM MgCl, 5 μL of 10 mM dNTPs, 0.5 μL of 5 U Taq polymerase (Genei, Bangalore) and 0.5 μL of 25 pmol/L of sense and anti-sense primers (Table 2) and the final volume was adjusted by nuclease free ddH₂O. PCR program was started with an initial denaturation at 95˚C for 5 minutes, followed by 40 cycles of denaturation at 95˚C for 30 s, annealing at 60˚C for 45 s, extension at 72˚C for 45 s, and completed with a final elongation step at 72˚C for 5 minutes. PCR product of 109 bp (10 μL) was verified by agarose gel electrophoresis (Figure 1(a)) and the remaining PCR product was then digested by the fast digest restriction enzyme MSPI (Fermentas, USA) at 37˚C for 15 min. The genotypes were assessed as follows: a single 109 bp fragment for homozygous AA genotype, two fragments of 89 bp and 20 bp for homozygous GG genotype and three fragments of 109 bp, 89 bp and 20 bp for the heterozygous GA genotype (Figure 1(b)). Negative control with ddH₂O instead of DNA template was included in each PCR run. A blind case/control analysis was performed with approximately more than 10% random samples were selected for confirmation and the results were 100% concordant.

The results of the PIRA-PCR analysis were confirmed by sequencing of representative samples for each genotype (Figure 1(c)). PCR products were purified using NucleoSpin PCR Clean-up kit (Macherey-Nagel, Germany) and sequenced on 3730 × l DNA Analyzer (Life Technologies Corporation, California, USA) according to the manufacturer’s instructions. The sequences were compared with BAX data in the GenBank (accession number U17193 for promoter).

Follow up was calculated as per hospital records from the date of diagnosis to the death or till the end of the study and survival analysis was estimated using Kaplan-Meier plots and log-rank (Mantel-Cox) test. Statistical analysis was performed using the Graph Pad Prism 6.0 or SPSS 16.0 software. Assessment of the correlations between genetic carrier status and BAX polymorphism was carried out using the Chi-Squaredor Fisher Exact test. BAX variants and risk of NSCLC were estimated by computing the odds ratios (ORs), risk ratio (RR) and risk difference (RD) with 95% confidence intervals (CIs). Allele frequencies among cases as well as controls were evaluated by using the Chi-square Hardy-Weinberg equilibrium test. A p value < 0.05 was considered significant.

The observed genotype frequencies of BAX (GG, GA and AA) were statistically different among patients and healthy controls (Chi square = 15.1, p = 0.0005). In addition cases showed higher A allele frequency (fA = 0.43 vs 0.28) whereas controls represented with higher G allele frequency (fG = 0.72 vs 0.57) Table 3.

A multivariate analysis based on logistic regression like odds ratio, risk ratio and risk difference with 95% confidence intervals were calculated for each group to estimate the association between the BAX (-248 G/A) genotypes and risk of NSCLC in Indian patients (Table 4). It was found that an increased risk of NSCLC was associated with the BAX (A) allele in an allele dosage-dependent manner. Compared to the GG genotype, odds ratio, risk ratio and risk difference for the heterozygous GA and homozygous AA genotypes were OR = 1.08 (0.65 - 1.80) RR = 1.04 (0.82 - 1.30) RD = 1.9 (−10.8 - 14.7) and OR = 3.18 (1.71 - 5.89) RR = 1.95 (1.31 - 2.92) RD = 27.4 (14.1 - 40.8) respectively, suggesting a possible dominant of this polymorphism with 3.2 fold increased risk associated with homozygous AA genotype. Upon comparison with the subtypes of NSCLC significantly increased risks were observed in squamous-cell carcinoma patients, OR = 5.26 (2.6 - 10.8) RR = 1.98 (1.4 - 2.9), and RD = 38.1 (22.5 - 53.8) Table 5. The increased risk associated with the BAX polymorphism appeared to be more pronounced in male cases and who were >45 years at diagnosis (OR 4.3 (2.09 - 8.91) and 5.4 (2.5 - 11.8 respectively) for the AA genotype (Table 5).

^*OR: Odd ratio; ^#RR: Risk ratio; ^{^}RD: Risk difference.

Among the lung cancer patients, 81 (50.6%) were diagnosed with adenocarcinoma and 79 (49.4%) with squamous-cell carcinoma. Majority of the cases were in advanced TNM stage 100 (62.5%) including more than 26% of patients with metastasis to distant organs. Grade 1 SCC 43 (54.4%) and grade 3 differentiated ADC 45 (55.5%) represented a major portion in histological grade of NSCLC cases.

Clinicopathological characteristics including histological type (p = 0.02), TNM Stage (p 0.03) and first degree family history of any cancer (p = 0.05) showed statistically significant associations with respect to BAX G(-248)A genotypes. SCC patients represented higher A allele frequency (0.52 vs 0.34) whereas higher G allele frequency was observed in ADC patients (0.66 vs 0.48). In addition higher prevalence of f(A) was among cases with advanced stage (0.49 vs 0.33) as well as in cases with metastasis to distant organs (0.52 vs 0.39) like brain, bone or adrenal etc. Cytological differentiation among ADC as well as SCC showed a continuous increasing trend of BAX A allele frequency from grade 1 to grade 3 differentiations. In relation to parameters like age and gender the BAX promoter polymorphism doesn’t presented any significant correlation, however, males and patients of >45 years showed higher f(A) allele (Table 6).

Healthy controls with similar smoking history as those of patients were included in the study in order to find the genetic susceptibility and risk of NSCLC in relation to BAX promoter polymorphism. Smokers constitute more than 70 percent, wherein bidi (61.3%) and huka (62.3%) were the major smoking source and about 57% had left smoking from more than 6 months among both cases as well as controls (Table 1). We observed that although an increased risk of NSCLC cancer was associated with BAX AA genotype in both smokers and nonsmokers, the OR for the AA genotype in smokers OR 4.3 (2.04 - 9.17, p = 0.0000) was significantly higher than those of non- smokers OR 1.5 (0.49 - 4.58, p = 0.48). Moreover, when the risk associated with the BAX polymorphism was further evaluated within the strata of pack-years smoked, risk of development of NSCLC was increasing as the number of pack-years increases (Mild; OR 1.5, Moderate; OR 5.5 and Heavy; OR 7.2). In addition interaction between susceptible genotypes and source of smoking was mostly evident among cigarette (OR 6.7, CI: 1.6 - 28.9) and huka smokers (OR 6.5, CI: 1.1 - 36.6) (Table 7).

Survival analyses were based on 160 NSCLC patients with mean follow-up time of 11.95 months (median 10.50; range, 0.5 - 127.5 months) for the overall survival. Out of 160 NSCLC patients, 87 (54.4%) died with mean

follow up time of 10.46 months ( median 10.00; range, 0.5 - 40.0 months) and for the patients who survived the follow up period (censored patients), the follow up time was 14.24 months (median 13.00; range, 0.5 - 127.5 months). Patients homozygous for A allele exhibited a significant poor overall survival (p < 0.0001) Figure 2(a), Figure 2(b). Median survival time for patients with BAX (-248) GG, GA, GA + AA and AA genotype were 30.0, 11.0, 9.0 and 6.0 months respectively. Significant poor overall survival was alsoobserved in patients presented A allele homology with ADC (p < 0.0001) or SCC (p < 0.0001) histology (Figure 2(c), Figure 2(d)) as well as with early (p < 0.0001) or advanced (p < 0.0001) TNM stage (Figure 2(e), Figure 2(f)). In addition patients of ADC and advanced stage with AA genotype showed lower median survival time than SCC and Early stage NSCLC patients (Median 3.0 and 5.0 vs 8.0 and 9.0 months respectively).