Functional lung adenocarcinoma risk SNPs identified through positional integration with human alveolar epithelial cell epigenomes

Abstract

Motivation: Lung cancer is the leading cause of cancer death, and lung adenocarcinoma (LUAD) is its predominant histological subtype. Fifteen single nucleotide polymorphisms (SNPs) have been significantly associated with LUAD risk in genome-wide association studies (GWAS). However, because most GWAS SNPs reside in non-coding regions and are co-inherited with hundreds of SNPs in linkage disequilibrium (LD), which SNPs play a causal role in disease development remains largely unknown. We hypothesized that some of these SNPs affect oncogenic transcriptional programs by modulating the activity of gene enhancers in alveolar epithelial cells (AECs), the purported cells of origin for LUAD. Methods: To test this hypothesis, we overlaid epigenomic features of primary human AECs over the locations of index LUAD risk SNPs and associated high LD SNPs. Luciferase assays for enhancer activity were performed for candidate SNPs that were predicted to disrupt transcription factor (TF) binding sites in loci marked by features of active enhancers. Expression quantitative trait loci (eQTL) analysis was also performed using The Cancer Genome Atlas (TCGA) dataset and the online Genotype-Tissue Expression (GTEx) Portal to identify potential target genes of each SNP-enhancer pair. Results: Thirty-three LUAD risk-associated SNPs mapped to putative AEC enhancer regions. TF binding site prediction suggests that numerous SNPs might alter the binding affinity of TFs implicated in lung cancer, including RXRA and NKX2-1. Luciferase assays indicate that two of the SNPs significantly affect enhancer activity in lung cancer cell lines. eQTL analyses link each of the putative enhancers to candidate target genes, including both known oncogenes and genes not previously associated with lung cancer. Conclusions: Taken together, our analyses provide new mechanistic insight into long-known associations between non-coding SNPs and LUAD outcomes, and may ultimately yield more effective and personalized strategies for lung cancer risk assessment, prevention, and treatment. Magnetic nanocubes for selective capture of circulating tumor cells in NSCLC