Gene expression signature of non-involved lung tissue associated with survival in lung adenocarcinoma patients

Lung adenocarcinoma patients of similar clinical stage and undergoing the same treatments often have marked interindividual variations in prognosis. These clinical discrepancies may be due to the genetic background modulating an individual's predisposition to fighting cancer. Herein, we hypothesized that the lung microenvironment, as reflected by its expression profile, may affect lung adenocarcinoma patients' survival. The transcriptome of noninvolved lung tissue, excised from a discovery series of 204 lung adenocarcinoma patients, was evaluated using whole-genome expression microarrays (with probes corresponding to 28 688 wellannotated coding sequences). Genes associated with survival status at 60 months were identified by Cox regression analysis (adjusted for gender, age and clinical stage) and retested in a validation series of 78 additional cases. RNA-Seq analysis from non-involved lung tissue of 12 patients was performed to characterize the different isoforms of candidate genes. Ten genes for which the loge- transformed hazard ratios expressed the same direction of effect in the discovery (P <1.0 × 10^-3) and validation series comprised the gene expression signature associated with survival: CNTNAP1, PKNOX1, FAM156A, FRMD8, GALNTL1, TXNDC12, SNTB1, PPP3R1, SNX10 and SERPINH1. RNA sequencing highlighted the complex expression pattern of these genes in non-involved lung tissue from different patients and permitted the detection of a read-through gene fusion between PPP3R1 and the flanking gene (CNRIP1) as well as a novel isoform of CNTNAP1. Our findings support the hypothesis that individual genetic characteristics, evidenced by the expression pattern of non-involved tissue, influence the outcome of lung adenocarcinoma patients.