Evidence that public database records for many cancer-associated genes reflect a splice form found in tumors and lack normal splice forms

摘要

Alternative splicing is widespread in the human genome, and it appears that many genes display different splice forms in cancerous tissue than in normal human tissues. However, since cDNAs for many cancer-associated genes were originally cloned from tumor samples, it is important to ask whether this repertoire of cDNAs provides a complete or representative picture of the transcript isoforms found in normal tissues. To answer this, we used bioinformatics and RT-PCR to identify novel splice forms, focusing on in-frame exonskips, for a panel of 50 cancer-associated genes in normal tissue samples. These data show that in nearly two-thirds of the genes, normal tissues expressed previously unknown splice forms, of which 40% were normally a dominant splice form. Surprisingly, the tumor-associated splice forms were twice as likely to be represented in GenBank than their normal tissue-associated splice forms, most likely because 70% of the mRNAs in GenBank for these genes were cloned from tumor samples. As an example, we describe a novel normal splice form of IKB beta, an important regulator of the NF kappa B pathway. Our data suggest that systematic re-evaluation of cancer genes' splice forms in normal tissue will yield insights into their distinct functions in normal tissues and in cancer. Our database contains 1308 novel normal splice forms, including many known cancer genes.