Dr Vidya Rajasekaran-Sutherland (@vidya87_) 's Twitter Profile
Dr Vidya Rajasekaran-Sutherland

@vidya87_

Scientist. Colorectal cancer genetics. Equality, Diversity and Inclusion advocate. Horror aficionado. She/Her. Own views.

ID: 1097775165516247040

calendar_today19-02-2019 08:29:29

44 Tweet

116 Followers

841 Following

Bowel Cancer UK (@bowelcanceruk) 's Twitter Profile Photo

We're wearing #Red4Research today as we celebrate the #BowelCancer research community and everyone who takes part in research. Every development in bowel cancer diagnosis and treatment starts as a research idea. We want to say a huge thank you to the amazing researchers whose

Dr Ruby Osborn 💙 (@rubytosborn) 's Twitter Profile Photo

Do you want to hear more about what's going on in bowel cancer research, and the difference this could make to patients? I'll be chatting to two bowel cancer specialists on Thursday - join live or watch the recording. tickettailor.com/events/nihrcrn…

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Delighted to share the bulk of my PhD has just been published in Gut Journal! (gut.bmj.com/content/early/…). In this multidisciplinary study I co-led with Dr Vidya Rajasekaran-Sutherland, we sought to investigate how variation at a single genetic region increased colorectal cancer (CRC) risk. A 🧵: 1/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

As with other common complex diseases, most loci that confer genetic risk for CRC lie in non-coding regions. This makes their interpretation challenging, and so we wanted to change this for one of the strongest signals identified back in 2008 (!) - 11q23.1. 2/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Along with increasing CRC risk, variation at 11q23.1 modifies expression of both nearby (cis-: POU2AF2, COLCA1 and POU2AF3) distant (trans-) genes. The immediate questions – which of these are relevant? And how do they predispose to the disease? 3/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Using bulk RNAseq, we find cis gene expression regulation to strongly overlap with CRC risk. We also consolidate the trans-effects and indicate their expression to be likely relevant to CRC risk by MR. Importantly, CRC risk increases with reduced expression of all these genes. 4/

Using bulk RNAseq, we find cis gene expression regulation to strongly overlap with CRC risk. We also consolidate the trans-effects and indicate their expression to be likely relevant to CRC risk by MR. Importantly, CRC risk increases with reduced expression of all these genes. 4/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

We replicate our previous findings (nature.com/articles/s4159…) showing many trans-genes are specifically expressed by a rare epithelial cell-type called ‘tuft cells’. Interestingly, they seem specifically correlated with POU2AF2. Perhaps this is the important local gene? 5/

We replicate our previous findings (nature.com/articles/s4159…) showing many trans-genes are specifically expressed by a rare epithelial cell-type called ‘tuft cells’. Interestingly, they seem specifically correlated with POU2AF2. Perhaps this is the important local gene? 5/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

But how could this locus cause such specific expression of distant genes? Well, recent work (nature.com/articles/s4158…) and (science.org/doi/10.1126/sc…) showed POU2AF2 and POU2AF3 encode proteins capable of transcriptional regulation…. 6/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

We used data for the former to show both POU2AF2 and POU2AF3 are capable of binding the trans-genes, in a way that overlaps with activating transcription marks. This could be a potential mechanism for the expression regulation. 7/

We used data for the former to show both POU2AF2 and POU2AF3 are capable of binding the trans-genes, in a way that overlaps with activating transcription marks. This could be a potential mechanism for the expression regulation. 7/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

So could reduced expression of these genes be indicative of tuft cell loss? Thanks to our clinicians, we got our hands on epithelial strips of normal colonic mucosa and devised a method of rolling these strips (videos in supp) to stain and count many cells in one go. 8/

So could reduced expression of these genes be indicative of tuft cell loss? Thanks to our clinicians, we got our hands on epithelial strips of normal colonic mucosa and devised a method of rolling these strips (videos in supp) to stain and count many cells in one go. 8/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Dividing samples by 11q23.1 genotype, we found individuals with the risk genotype at 11q23.1 do indeed have a reduced number of tuft cells! Something we believe is a result of the complex transcriptional relationship laid out above. 9/

Dividing samples by 11q23.1 genotype, we found individuals with the risk genotype at 11q23.1 do indeed have a reduced number of tuft cells! Something we believe is a result of the complex transcriptional relationship laid out above. 9/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

To regroup: CRC risk-associated genetic variation at 11q23.1 is correlated with the expression of several cis- and trans-genes, as well as tuft cell loss. However, one thing we can’t disentangle yet – which (or which combination) of the cis-genes is governing this effect? 10/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

For this, we went to mice. Initiated by Dr Ruby Osborn 💙, we first validated mice as a reliable model of CRC-associated molecular phenotypes at 11q23.1, and then knocked out homologs of the cis-genes independently.

For this, we went to mice. Initiated by <a href="/RubyTOsborn/">Dr Ruby Osborn 💙</a>, we first validated mice as a reliable model of CRC-associated molecular phenotypes at 11q23.1, and then knocked out homologs of the cis-genes independently.
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Remarkably, we found expression and tuft cell abundance changes to be strikingly specific to Pou2af2 genotype. While consistent with the expression we saw at the single cell level, this now experimentally decouples the function of these genes in the colon. 12/

Remarkably, we found expression and tuft cell abundance changes to be strikingly specific to Pou2af2 genotype. While consistent with the expression we saw at the single cell level, this now experimentally decouples the function of these genes in the colon. 12/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

To cement the correlation between these phenotypes and tumourigenesis, we then crossed these single knockouts onto a tumourigenic background. What is really cool? We also find only Pou2af2 genotype to be associated with changes to survival and tumour burden. 13/

To cement the correlation between these phenotypes and tumourigenesis, we then crossed these single knockouts onto a tumourigenic background. What is really cool? We also find only Pou2af2 genotype to be associated with changes to survival and tumour burden. 13/
Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Together, this study builds upon growing interest in Pou2af2 function in different contexts, and provides a body of evidence to indicate this gene as both causal of CRC risk and transcriptional regulation of colorectal tuft cells. 14/

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

The function of tuft cells in the colon is not well characterised, and so understanding how these two phenomena are linked will be difficult, but key to providing potentially actionable insights into prevention of this morbid disease. /15

Bradley Harris (@bradleyomics) 's Twitter Profile Photo

Plenty more in the paper, so do check it out! Huge thanks to all who contributed to this work: Claire Smillie, Kevin Donnelly, Marion Bacou, Edward Esiri-Bloom, Li-Yin Ooi, Morven Allan, Marion Walker, Stuart Reid, @AMeynert, Graeme Grimes, James Blackmur, Peter Vaughan-Shaw, Philip Law … 16/