We have established that BRD4 contributes to patterns of splicing both in normal cells in vivo and in cancer cells. Through its direct interaction with the splicing machinery on the genome, BRD4 acts preferentially by modulating exon usage.These findings lead to the model that BRD4 functions to integrate the processes of chromatin structure, transcription and splicing to ensure proper regulation of gene expression. The role of BRD4 in regulating alternative splicing in vivo was documented in the thymus where deletion of BRD4 resulted in marked changes in the patterns of alternative splicing. Furthermore, we have found that the transitions from one developmental stage to another - DN to ISP to DP to SP - are accompanied by changes in the patterns of splicing, indicating that splicing is developmentally regulated in the thymus. BRD4, which is expressed at approximately equal levels in each of these stages, contributes to this regulation since deletion of BRD4 results in changes in the splicing patterns at each stage of differentiation, preferentially affecting skipped exon events, but not affecting the expression of splicing factors. Consistent with its role in cancer, BRD4 also contributes to patterns of alternative splicing in T-ALL cells, largely affecting the genes involved in cell cycle regulation. Interestingly, degradation of BRD4 results in much larger changes in splicing patterns than the inhibition of BRD4 binding to chromatin through its bromodomains. This suggests that BRD4 mediated regulation of alternative splicing does not depend entirely on its binding to chromatin through its bromodomains. Further support for this perspective comes from the finding that BRD4's co-localization with splicing factors is not perturbed by JQ1, which prevents binding to chromatin. In addition, whereas degradation of BRD4 targets both promoters and gene bodies, blocking of BRD4 bromodomain-mediated interactions with chromatin by JQ1 is largely confined to the promoter regions of genes, lending further support to the conclusion that BRD4's interaction with the elongation and splicing machineries does not depend on its bromodomains. Whether it binds to chromatin at gene bodies directly by a bromodomain-independent mechanism or indirectly through its binding to either the transcription elongation factors or splicing machinery remains to be determined. Our finding that BRD4 directly interacts with both FUS and HnRNPM, suggests that it can participate as an integral component of the splicing machinery. The precise mechanism(s) by which BRD4 contributes to the regulation of splicing remain to be determined.
