| dc.description.abstract | Despite the increased molecular understanding of lung cancer, advances in therapy, and reduced smoking rates in developed countries, lung cancer remains the deadliest cancer type in the world today. Lung adenocarcinoma (LUAD) is the most common histological subtype of non-small cell lung cancer (NSCLC) and accounts for about 40% of all lung cancer incidents. LUAD can be characterized based on mutation status of specific genes and divided into three molecular subtypes based on gene expression patterns. Further improvement in molecular characterization and tumor classification is required to deepen the understanding of LUAD heterogeneity and to improve molecular diagnosis and treatment decisions.
Cancer can be characterized by alterations in gene expression and the epigenetic landscape, potentially leading to repression of tumor-suppressor genes and activation of oncogenes. DNA methylation at CpG sites is an essential epigenetic mechanism for normal development and maintaining cell phenotype. Alteration in DNA methylation patterns associated with alterations in the expression of specific genes has been identified in many cancer types, including lung cancer, and is considered a hallmark feature of cancer. DNA methylation exerts different influences on gene expression depending on where modifications occur in relation to its target gene. Based on the conventional model, methylation in cis-transcriptional regulatory elements and gene expression are negatively correlated, e.g., methylation in local promoter represses the expression of a target gene. Multiple studies in cancer have demonstrated the importance of methylation at distal cis-regulatory elements located in intergenic regions (enhancers), following the conventional model of regulation. Enhancer regions include binding sites for sequence-specific transcription factor (TF) proteins, which mediate the regulatory mechanism between enhancers and target promoters. The interplay between DNA methylation at enhancers, TF binding, and target expression regulation in lung cancer is relatively unexplored and poorly understood.
We performed expression-methylation quantitative trait loci (emQTL) analysis on patient data from The Cancer Genome Atlas (TCGA). The emQTL is a method for identifying potentially epigenetically regulated pathways in cancer and involves genome-wide in cis and in trans correlation analyses between DNA methylation and gene expression. Grouping of the significant correlations (emQTLs) and characterization of the identified emQTL-genes led to the discovery of three main biological processes related to immune cell infiltration, proliferation, and hormone signaling. To functionally characterize the identified CpGs linked to each process, we assessed the enrichment of CpGs in regions related to different chromatin states, transcription factor binding sites, and chromatin interaction loops. Of particular interest, we found the hormone-related CpGs to be enriched in enhancer regions and in the binding sites for TFs linked to hormone-related signaling such as FOXA1/2, FOSL1/2, SMAD3, and JUN. The hormone-related emQTL-CpGs and -genes were also enriched in chromatin interaction loops, further suggesting epigenetic regulation of the hormone-related genes through enhancer-promoter interactions. Moreover, our results suggest that putative epigenetic regulation of immune-related genes occurs in the tumor-infiltrated immune cells; however, future studies are required. Also, we did not find evidence for proliferation processes in LUAD being under putative epigenetic regulation.
Unsupervised hierarchical clustering of expression and DNA methylation of the identified genes and CpGs was performed to identify subgroups of patients with potential clinical impact. Through unsupervised hierarchical clustering, we identified subclasses of patients with different clinical outcomes in all three biological programs based on the expression pattern. Findings were validated for the hormone-related program in a Norwegian cohort for LUAD, highlighting the potential clinical importance of hormone-related pathways in LUAD patients. The identified molecular subclasses partially overlapped with earlier defined LUAD expression subtypes, providing additional information in tumor subtyping not captured by previous studies using transcriptomic data independently from DNA methylation. Overall, the results of our computational work demonstrate the link between enhancer methylation and hormone-related gene expression, where expression patterns constitute prognostically relevant subtypes in LUAD patients. | en_US |