Genetic polymorphisms of genes on the retinoic acid pathway and risk of head and neck cancer

[[abstract]]Background: Although most cases of head and neck cancer (HNC) can be attributed to three established risk factors, including alcohol, betel quid, and cigarette, the underlying genetic variations among individuals may be important in modulating the risk of HNC. Previous studies showed that the metabolism or function of retinoic acid may be modified by the use of alcohol, betel quid, or cigarette, resulting in the dysregulation of cell growth, differentiation, and apoptosis. The current analysis examined the association between HNC risk and genetic polymorphisms of genes on the retinoic acid pathway. Methods: 408 incident cases of HNC and 473 sex- and age- frequency matched controls were recruited from the department of otolaryngology and department of stomatology. Information on the use of alcohol, betel quid, and cigarette was collected by in-person interviews. In the discovery phase, 223 cases and 220 controls were genotyped for 368 single nucleotide polymorphisms (SNPs) of 37 retinoic acid genes. Unconditional logistic regression was performed to estimate the log-additive odds ratio (OR) and 95% confidence interval (CI) of HNC risk associated with each SNP. Those that had a P < 0.05 were further genotyped for 185 cases and 253 controls in stage 2. Additional analyses were performed stratified by alcohol, betel quid, and cigarette to evaluate gene-environment interaction. Results: In the stage 1, 25 SNPs were associated with the risk of HNC with a P < 0.05. With additional genotyping performed for these 25 SNPs in stage 2, only one SNP of RARB showed a decreasing P value with additional samples (stage 1 OR = 1.43, 95% CI: 1.02-2.00, P = 0.04; stage 2 OR = 1.33, 95% CI: 0.94-1.89, P = 0.1; combined OR = 1.34, 95% CI: 1.06-1.69, P = 0.01). In addition, the association between this RARB SNP and HNC risk appeared to be modified by alcohol drinking status. A positive association between the minor allele of the RARB SNP and HNC was observed among never drinkers (OR = 1.96, 95% CI: 1.30-2.95) but not among ever drinkers (OR = 1.09, 95% CI: 0.81-1.47) (P for interaction = 0.01). Conclusion: The risk of HNC was associated with a SNP of RARB and this association appeared to depend on alcohol drinking status. Further investigations are needed to determine the causal SNP and its functional significance