Abstract B36: Micronome-wide explorations of NSCLC: microRNA-seq discovery

Background: Recent studies have revealed that microRNAs play important regulatory roles in carcinogenesis. Some key miRNAs may regulate lung carcinogenesis and be useful for lung cancer early detection. For discovery, we used high-throughput next-gen sequencing technology to investigate miRNA expression profiles in a study of 25 lung adeno- and squamous cell carcinomas, paired with far-adjacent noncancerous lung tissues from the same donors. Method: Total RNA was isolated by Trizol from 14 pairs of lung adenocarcinomas, 11 squamous cell cancers, and paired noncancerous lung tissues, from the same patients. Small RNA library of each total RNA was constructed with Illumina's Small RNA Sample Prep Kit, followed by high-throughput sequencing using the Illumina IIa platform. In parallel, transcriptome microarrays (Affymetrix GA 1.0) were run on these same paired samples. Statistical differences in microRNA expression between groups were analyzed by GenePattern software, logistic regression, and other methods. Results: For adeno, 128 micros were differentially expressed and 63 were two-fold differentially expressed between the tumors and non-tumor tissue; for squamous, the numbers of altered microRNAs were 42 overall, and 25 >two-fold. Among adenocarcinoma, the top 10 upregulated were miRs-147b, 577, 877, 556, 449c, and others; and downregulated were miRs 516b, 1251, 486, 139, 516a, and others. Among squamous cell carcinomas, the top up-regulated miRs were 708, 1259, 494, 200a, 216b, and others; and down-regulated miRs were 184, 1251, 144, 516a, 584, and others. MiRs 1259, 135b, 200a, 21, 494, 708 were upregulated in both tumor histologies. When the highest expressed microRNAs were correlated with the most altered mRNA transcripts, putative mRNA targets were more often anti-correlated to the individual microRNAs (Chi Square, p = 4.6 × 10-6), and this anticorrelation generally held true for the top differentially expressed microRNAs and their putative mRNA targets, and within histology strata. Conclusion: The lung cancer microRNA complement differs from non-tumor tissue, adeno- differs from squamous, and many of the most dysregulated miRs appear to be anticorrelated with putative mRNA targets. Which of the putative mRNA targets are actual biological targets awaits experimental functional work. Once constructed, the combined microRNA:mRNA signature of lung cancer is likely to be more informative for biologic potential than either signature alone. [Funding source: NCI 1RC1 CA145422-01; 1K24-CA139054-01]