Implications of trans-splicing of RNA for tumor diagnosis and detection

Abstract

Abstract Chromosomal rearrangements causing head to tail fusions of genes separately encoded in germline DNA are frequent findings in human cancers. The specificity of gene fusions for different types of cancers has made these alterations in DNA valuable markers for the diagnosis, classification, and monitoring of human cancers. A variety of methods have been used to detect fusions within tumors. Among these have been conventional cytogenetics, Southern blot, FISH, and RT-PCR. In particular, RT-PCR has become popular because of its relative simplicity, low cost, and sensitivity for finding small numbers of malignant cells within tissue samples. RT-PCR was originally developed for detecting gene fusions because the wide distribution of breakpoints within DNA for most chromosomal rearrangements makes detection difficult by standard PCR or Southern blotting. Most breakpoints lie within introns. Therefore, despite the variability in the sites of recombination that joins the two genes involved in the fusion, the structure of the mature, chimeric RNA transcripts produced from these fusions is consistent from tumor to tumor. This fact permits detection of the RNA products of gene fusions using a single pair of PCR primers, or a small number of such pairs, in amplifications of cDNA generated from tumor RNA by reverse transcription. In theory, successful RT-PCR of chimeric RNA should be entirely specific for tumors. However, low levels of amplification products are sometimes generated from tissues of healthy individuals. The presumption has been that chromosomal rearrangements occur in normal tissues at some low frequency and that these rearrangements produce chimeric RNA detectable by sensitive RT-PCR assays. By themselves, in the absence of cooperating mutations and epigenetic changes, these gene fusions are insufficient to cause neoplastic transformation of cells and the cells bearing these gene fusions have limited proliferative and survival capacity. We have recently discovered a chimeric RNA that joins sequence from the 5’ end of the JAZF1 gene to the 3’ end of the JJAZ1 (also called SUZ12) gene in normal endometrial cells. This RNA is identical to that transcribed from the JAZF1-JJAZ1 gene fusion produced by the (7;17)(p15;q21) chromosomal translocation found in about 50% of endometrial stromal sarcomas (ESSs). In normal endometrial cells, JAZF1-JJAZ1 RNA is generated not from a gene fusion, but from trans-splicing between precursors mRNAs of the two genes. Trans-splicing in these pre-mRNAs seems to be at least partially under control of hormones. The chimeric RNA is translated into chimeric JAZF1-JJAZ1 protein, which may function to increase cell survival in hypoxic environments. Although the level of chimeric JAZF1-JJAZ1 RNA in normal endometrial cells is lower than that seen in tumor cells containing the t(7;17)(p15;q21), the presence of chimeric RNA in normal cells has clear implications for the specificity of JAZF1-JJAZ1 RNA as a marker for ESSs and for the diagnosis of ESSs by RT-PCR. Furthermore, the finding of a chimeric RNA in normal cells identical to that transcribed from a chromosomal rearrangement suggests that other chromosomal rearrangements exert their pro-neoplastic effects by giving rise to constitutive production of chimeric RNAs that are normally produced over restricted intervals during tissue differentiation or under the control of defined regulatory factors. If true, this scenario would explain the frequent finding of low levels of chimeric RNA for other chromosomal rearrangements in healthy individuals. Another possible ramification of our finding is that abnormal, unregulated trans-splicing of RNA may produce chimeric RNAs that have a pro-neoplastic effect in tumors lacking any corresponding gene fusion.