Lung adenocarcinoma (LUAD) has several histologically distinct subtypes that differ by a number of clinical features including patient survival. Molecular mechanisms underlying histological and clinical differences between subtypes remain poorly understood.

We conducted a comparative analyses of gene expression in acinar, lepidic, papillary and solid subtypes, as well as mucinous adenocarcinoma. We used a novel, more efficient approach to identify subtype-specific genes. We compared the mean gene expression level separately for tumors and adjacent normal tissue with pure or a highly represented (≥ 75%) subtype of interest to the mean expression in tumors where the subtype of interest was not present. We also performed tumor to adjacent normal tissue comparisons and identified genes differentially expressed between tumor and adjacent normal tissues for each subtype.

The number of subtype-specific genes varied from 1 for the acinar to 482 for the papillary subtype. Comparative analysis of gene expression in adjacent normal tissues also identified subtype-specific genes, 38 in total. Gene set enrichment analysis identified oxidative phosphorylation as a biological function associated with papillary, and immune response - with solid subtype. Using data on differential expression between tumor and adjacent normal tissue among the subtype-specific genes and existing evidence for association with lung carcinogenesis, we have identified several candidate subtype-specific driver genes.

n We identified subtype-specific genes, biological functions, and potential drivers of subtype-specific carcinogenesis for LUAD subtypes. The study showed importance of gene expression in adjacent normal tissue for subtype-specific tumorigenesis.