In eukaryotic genomes, ~10% of genes share a bidirectional promoter, whereby the transcription start sites (TSS) of a gene pair are located within 1kb of each other on opposite strands. These promoters are regulated to permit two scenarios of gene activity: 1) both genes are ‘on’ or 2) asymmetric transcription where one gene is ‘on’, while the other is ‘off’. Maintenance of asymmetric gene expression at bidirectional promoters is essential as ‘switching on’ of the antisense transcript can alter the expression of the sense gene. This a common occurrence at oncogenes for e.g. BRCA1 and TP53 are both located at bidirectional promoters shared with an antisense non-coding RNA (ncRNA) gene (NBR2 and WRAP53, respectively), whose up-regulation is implicated in cancer progression.
We have recently found that CTCF binding is enriched at bidirectional promoters (p-value 1.6e-217 (Fisher’s exact test)). As CTCF is a well-characterised insulator we hypothesise that CTCF acts as an insulating barrier to prevent aberrant bidirectional transcription.
We have now generated the following preliminary epigenome data in prostate cancer cells after CTCF siRNA treatment: 1) CTCF ChIP-seq, 2) H3K4me3 ChIP-seq and 3) RNA- seq. We found that 10% of genes that lost CTCF following knockdown shared the same three characteristics: 1) up-regulation of a ncRNA, 2) bidirectional with asymmetric transcription (one gene ‘off’/one gene ‘on’) & 3) displayed ‘spreading’ of H3K4me3 away from the TSS of the expressed gene into the up-regulated gene. Using data from the International Cancer Genome Consortium we have also shown that the CTCF sites at these bidirectional promoters are enriched for mutations in both breast and prostate cancer.
This data provides preliminary evidence that CTCF may be a regulator of bidirectional transcription and we are now exploring whether mutations of these sites are events that promote the development of breast and prostate cancer.