CYP2A6 Activity and Cigarette Consumption Interact in Smoking-Related Lung Cancer Susceptibility

Abstract

Cigarette smoke, containing both nicotine and carcinogens, causes lung cancer. However, not all smokers develop lung cancer, highlighting the importance of the interaction between host susceptibility and environmental exposure in tumorigenesis. Here, we aimed to delineate the interaction between metabolizing ability of tobacco carcinogens and smoking intensity in mediating genetic susceptibility to smoking-related lung tumorigenesis. Single-variant and gene-based associations of 43 tobacco carcinogen-metabolizing genes with lung cancer were analyzed using summary statistics and individual-level genetic data, followed by causal inference of Mendelian randomization, mediation analysis, and structural equation modeling. Cigarette smoke-exposed cell models were used to detect gene expression patterns in relation to specific alleles. Data from the International Lung Cancer Consortium (29,266 cases and 56,450 controls) and UK Biobank (2,155 cases and 376,329 controls) indicated that the genetic variant rs56113850 C>T located in intron 4 of CYP2A6 was significantly associated with decreased lung cancer risk among smokers (OR = 0.88, 95% confidence interval = 0.85-0.91, P = 2.18 × 10-16), which might interact (Pinteraction = 0.028) with and partially be mediated (ORindirect = 0.987) by smoking status. Smoking intensity accounted for 82.3% of the effect of CYP2A6 activity on lung cancer risk but entirely mediated the genetic effect of rs56113850. Mechanistically, the rs56113850 T allele rescued the downregulation of CYP2A6 caused by cigarette smoke exposure, potentially through preferential recruitment of transcription factor helicase-like transcription factor. Together, this study provides additional insights into the interplay between host susceptibility and carcinogen exposure in smoking-related lung tumorigenesis.

Significance: The causal pathway connecting CYP2A6 genetic variability and activity, cigarette consumption, and lung cancer susceptibility in smokers highlights the need for behavior modification interventions based on host susceptibility for cancer prevention.

Figure 1:

Decomposition of the genetic effect of rs56113850 in CYP2A6 on lung cancer risk by smoking status. A, Manhattan plot for genetic effects of tobacco carcinogen metabolic genes on lung cancer risk stratified by smoking status. The x-axis represents each chromosome, with different colors assigned to each gene; the y-axis represents association P values (−log₁₀ transformed) with lung cancer risk, derived from lung cancer GWAS summary statistics in subgroups of smoking populations deposited in ILCCO. Red dashed horizontal line indicates a P value equal to GWAS significance at 5 × 10⁻⁸. B, Interaction effects between rs56113850 in CYP2A6 and smoking status on lung cancer risk. Genotyping data of rs56113850 were acquired from ILCCO for 14,803 cases and 12,262 controls with individual smoking information. OR, odds ratio, calculated via logistic regression model underlying joint analysis approach. C, Four-way decomposition analysis of the rs56113850 effect on lung cancer risk potentially mediated by smoking status. Y is the outcome: lung cancer; A is the exposure: rs56113850 genotypes obtained from ILCCO; M is the potential mediator: smoking status. The OR and corresponding 95% CI were calculated by mediation analysis, with causal effects estimated for exposure and at the mean level of covariates. The CDE and the reference interaction were computed by fixing smoking status as never (M=0) or ever (M=1).

Figure 2:

Causal inference for the causal pathway of CYP2A6 activity, cigarette consumption, and lung cancer risk. A-C, Scatter plots for genetic associations across CYP2A6 activity, CigDay, and lung cancer risk in the smoking population. The x-axis represents the per allele association of exposure-relevant SNPs and assigned outcomes, with the likelihood-based MR estimate for genetic IVs. D, Directed acyclic graph for the causal mediation pathway of CYP2A6 activity, CigDay, and lung cancer risk in the smoking population. E, exposure; M, mediator; O, outcome. IE, indirect effect. The Sobel test was used to evaluate the mediation effect upon the causal effect derived from the MR estimate.

Figure 3:

Expression pattern of CYP2A6 across tissues, smoking status, and allele-specific manners. A, CYP2A6 expression pattern in the top 10 tissues using the consensus normalized expression value (NX) derived from HPA, GTEx, and FANTOM5. NX_liver = 199.5; NX_lung = 0.2; NX_others = 0. B, CYP2A6 expression pattern in normal tissues derived from TCGA PANCAN. The x-axis is assigned to tumor type, including BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; CHOL, cholangiocarcinoma; COAD, colon adenocarcinoma; ESCA, esophageal carcinoma; HNSC, head and neck squamous cell carcinoma; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; KIRP, kidney renal papillary cell carcinoma; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; NA, not available; PAAD, pancreatic adenocarcinoma; PCPG, pheochromocytoma and paraganglioma; PRAD, prostate adenocarcinoma; READ, rectum adenocarcinoma; SARC, sarcoma; SKCM, skin cutaneous melanoma; STAD, stomach adenocarcinoma; THCA, thyroid carcinoma; THYM, thymoma; UCEC, uterine corpus endometrial carcinoma. The y-axis represents CYP2A6 normalized expression. C, Differential expression analyses of CYP2A6 between tumor and normal tissues derived from publicly available TCGA PANCAN (lung and liver cancers; left) and Harvard Biobank (lung cancer; right). An unpaired t test was applied for comparison of CYP2A6 expression between tumor and normal tissues. D, Forest plot for the effect of smoking status on CYP2A6 pulmonary gene expression. The effect size of smoking status (ever-smoker vs. non-smoker) on CYP2A6 expression was calculated via linear regression model, accompanied by the 95% CI. The size of the square is proportional to the weight, which is estimated by the standard “inverse-variance” method for random-effects models in meta-analysis. E, CYP2A6 expression pattern at levels of RNA, protein, and activity after 2% CSE exposure in HBE cell line. Gene expression was normalized to that in cells treated with DMSO. An unpaired t test was applied for the group comparison. Both RNA and activity detection were performed in three biological replicates, with three technical replicates each, and protein was in three biological replicates. F, Allele-specific effect of rs56113850 on CYP2A6 expression pattern at levels of RNA, protein, and activity after 2% CSE exposure in HBE cell line. Allele-specific constructs containing the putative activity region flanking rs56113850 were cloned into the pcDNA3.1-basic vector and transfected into HBE cells. Gene expression was normalized to that in cells treated with DMSO. Both RNA and activity detection were performed in three biological replicates, with three technical replicates each, and protein was in three biological replicates. G, Allele-specific effect of rs56113850 on TF HLTF binding affinity through super-electrophoretic mobility shift assays. H, Graphical representation of the findings of this study. In smokers, a causal pathway model for relationships among CYP2A6 variants (rs56113850 C>T included), CYP2A6 activity, smoking intensity, and lung cancer risk exists, which may be biologically interpreted by nicotine metabolism (indicated by NMR) and TSNA activation after CYP2A6 activity induced by cigarette smoke exposure in lung tumorigenesis.