Combining ‘omics and epidemiology to uncover mechanisms linking aspirin treatment to reduced risk of colorectal cancer


Year:

Session type:

Theme:

Aayah Nounu1,Gemma Sharp1,Alexander Greenhough1,Caroline Relton1,Ann Williams1
1The University of Bristol

Abstract

Background

Colorectal cancer (CRC) is the second leading cause of cancer deaths both in the UK and Europe. Epidemiological studies have shown that aspirin is beneficial in primary, secondary and tertiary prevention of colorectal cancer. Aspirin is a well-established inhibitor of the (COX)/(PGE2) signalling pathway. A study carried out in mice showed that NCK-4016, an aspirin derivative that does not inhibit COX, showed a higher chemopreventative effect compared to aspirin in a model of colon cancer indicating that aspirin may act through other mechanisms not yet known. Using a multiple ‘omics approach, we set out to identify new targets of aspirin.

Method

Colorectal adenoma cells (RG/C2s) were treated with 0mM, 2mM and 4mM aspirin for 24 hours before protein was collected for a SILAC approach and DNA was collected for methylation micro-array analysis. Changes in methylation were identified using the Infinium HumanMethylation450 BeadChip array. We used previously published transcriptomic data Sabates-Bellver et al. (2007) comparing normal and adenoma colons and combined the proteomic, transcriptomic and methylomic results to identify novel aspirin targets.

Results

Our results show that in general, aspirin decreases the expression of proteins in adenoma cells. After setting a threshold, our SILAC approach identified 129 proteins affected by aspirin treatment. After combining with methylomic and transcriptomic data, information of aspirins effect was available for 114 proteins. We set a threshold to identify the number of proteins consistently downregulated with aspirin and identified 7 proteins.

Conclusion

We have potentially identified 7 novel targets of aspirin and these may be useful in trying understand the alternative mechanisms of its action. We plan to find suitable instrumental variables for these proteins for use in a Mendelian Randomization approach.