We demonstrate that the cytogenetically de®ned translocation t(X;1)(p11.2;p34) observed in papillary renal cell carcinomas results in the fusion of the splicing factor gene PSF located at 1p34 to the TFE3 helix ± loop ± helix transcription factor gene at Xp11.2. In addition we de®ne an X chromosome inversion inv(X)(p11.2;q12) that results in the fusion of the NonO (p54 nrb ) gene to TFE3. NonO (p54 nrb ), the human homologue of the Drosophila gene NonA diss which controls the male courtship song, is closely related to PSF and also believed to be involved in RNA splicing. In each case the rearrangement results in the fusion of almost the entire splicing factor protein to the TFE3 DNA-binding domain. These observations suggest the possibility of intriguing links between the processes of RNA splicing, DNA transcription and oncogenesis.

In papillary renal cell carcinomas the TFE3 transcription factor becomes fused to the PSF and NonO pre-mRNA splicing factors and most commonly to a protein of unknown function designated PRCC. In this study we have examined the ability of the resulting PRCC ± TFE3 and NonO ± TFE3 fusions to activate transcription from the plasminogen activator inhibitor-1 (PAI-1) promoter. The results show that only fusion to PRCC enhanced transcriptional activation, indicating that the ability to enhance the level of transcription from endogenous TFE3 promoters is not a consistent feature of TFE3 fusions. In investigations of the normal function of PRCC we observed that PRCC expressed as a green¯uorescent fusion protein colocalizes within the nucleus with Sm premRNA splicing factors. It was also found that endogenous PRCC is coimmunoprecipitated by antibodies that recognize a variety of pre-mRNA splicing factors including SC35, PRL1 and CDC5. Association with the cellular splicing machinery is therefore, a common feature of the proteins that become fused to TFE3 in papillary renal cell carcinomas. Oncogene (2001) 20, 178 ± 187.

The developmental gene PAX 3 is expressed in the early embryo in developing muscle and elements of the nervous system, including the brain. Since no one has investigated the expression of the isoforms of PAX 3 in the neuroectodermal tumors melanoma and small cell lung cancer (SCLC), we have carried out a comprehensive screening for the expression of the isoforms PAX 3a-e using RT-PCR in human melanoma cell lines, primary human ocular and secondary cutaneous melanomas. We have identified 2 new isoforms of PAX 3, g and h, which we have isolated, cloned and sequenced. Sets of primers for each isoform were designed and their specificity was confirmed by sequence analysis of the products. The isoforms PAX 3a-e were detected in all human cutaneous melanoma cell lines (8/8), but only PAX 3c (1/2) and PAX 3d (2/2) in ocular melanoma cell lines. The same PAX 3 isoforms were detected in more than 80% of human cutaneous melanomas: PAX 3a and b (15/17), PAX 3c (14/17), PAX 3d (16/17) and PAX 3e (15/17). In contrast the results for 7 SCLC cell lines were PAX 3a (0/7), PAX 3b (1/7), PAX 3c (3/7), PAX 3d (6/7), PAX 3e (2/7); 8/8 cutaneous melanoma cell lines and 8/8 ocular melanoma tissues, together with 14/17 cutaneous melanoma tissues screened, expressed the new isoform PAX 3g. All 8 cutaneous melanoma cell lines expressed PAX 3h, but it was not detectable in any of the tumor tissues (0/20). Neither of the 2 ocular melanoma cell lines expressed the 2 new isoforms. Comparison of the different amplicon staining intensities on a gel suggests that PAX 3c and PAX 3d are the predominant transcripts expressed, with relatively low expression of PAX 3e and PAX 3h. We propose that these and the 2 new isoforms we have discovered may be important in oncogenesis and differential diagnosis of melanomas or SCLC.

No abstract

No abstract