Comprehensive gene expression analysis of 5′-end of mRNA identified novel intronic transcripts associated with hepatocellular carcinoma

To elucidate the molecular feature of human hepatocellular carcinoma (HCC), we performed 5′-end serial analysis of gene expression (5′SAGE), which allows genome-wide identification of transcription start sites in addition to quantification of mRNA transcripts. Three 5′SAGE libraries were generated from normal human liver (NL), non-B, non-C HCC tumor (T), and background non-tumor tissues (NT). We obtained 226,834 tags from these libraries and mapped them to the genomic sequences of a total of 8,410 genes using RefSeq database. We identified several novel transcripts specifically expressed in HCC including those mapped to the intronic regions. Among them, we confirmed the transcripts initiated from the introns of a gene encoding acyl-coenzyme A oxidase 2 (ACOX2). The expression of these transcript variants were up-regulated in HCC and showed a different pattern compared with that of ordinary ACOX2 mRNA. The present results indicate that the transcription initiation of a subset of genes may be distinctively altered in HCC, which may suggest the utility of intronic RNAs as surrogate tumor markers.