Transcription start site-level expression of thyroid transcription factor 1 isoforms in lung adenocarcinoma and its clinicopathological significance.

There are multiple transcription start sites (TSSs) in agreement with multiple transcript variants encoding different isoforms of NKX2-1/TTF-1 (thyroid transcription factor 1); however, the clinicopathological significance of each transcript isoform of NKX2-1/TTF-1 in lung adenocarcinoma (LAD) is unknown. Herein, TSS-level expression of NKX2-1/TTF-1 isoforms was evaluated in 71 LADs using bioinformatic analysis of cap analysis of gene expression (CAGE)-sequencing data, which provides genome-wide expression levels of the 5'-untranslated regions and the TSSs of different isoforms. Results of CAGE were further validated in 664 LADs using in situ hybridisation. Fourteen of 17 TSSs in NKX2-1/TTF-1 (80% of known TSSs in FANTOM5, an atlas of mammalian promoters) were identified in LADs, including TSSs 1-13 and 15; four isoforms of NKX2-1/TTF-1 transcripts (NKX2-1_001, NKX2-1_002, NKX2-1_004, and NKX2-1_005) were expressed in LADs, although NKX2-1_005 did not contain a homeodomain. Among those, six TSSs regulated NKX2-1_004 and NKX2-1_005, both of which contain exon 1. LADs with low expression of isoforms from TSS region 11 regulating exon 1 were significantly associated with poor prognosis in the CAGE data set. In the validation set, 62 tumours (9.3%) showed no expression of NKX2-1/TTF-1 exon 1; such tumours were significantly associated with older age, EGFR wild-type tumours, and poor prognosis. In contrast, 94 tumours, including 22 of 30 pulmonary invasive mucinous adenocarcinomas (IMAs) exhibited exon 1 expression without immunohistochemical TTF-1 protein expression. Furthermore, IMAs commonly exhibited higher exon 1 expression relative to that of exon 4/5, which contained a homeodomain in comparison with EGFR-mutated LADs. These transcriptome and clinicopathological results reveal that LAD use at least 80% of NKX2-1 TSSs and expression of the NKX2-1/TTF-1 transcript isoform without exon 1 (NKX2-1_004 and NKX2-1_005) defines a distinct subset of LAD characterised by aggressive behaviour in elder patients. Moreover, usage of alternative TSSs regions regulating NKX2-1_005 may occur in subsets of LADs.