An integrated genetic and proteomic approach to understanding cardiovascular disease aetiology

Lead Research Organisation: University of Cambridge
Department Name: Public Health and Primary Care

Abstract

My research programme is focussed on using genomic technologies to understand the links between inflammation and cardiovascular disease. Despite advances in prevention and treatment, cardiovascular disease is the leading cause of death worldwide, highlighting the need for new therapeutic strategies. Most deaths are due to myocardial infarction or "heart attack". Heart attacks occur when fatty deposits in the wall of an artery rupture, triggering formation of a clot which blocks the blood supply to the heart. These fatty deposits build up over many years in a process called atherosclerosis. Traditionally atherosclerosis was thought to be solely due to build-up of excess fats, but over the last few decades the importance of inflammation in this process has been increasingly recognised. Inflammation usually occurs as a protective response to infection or injury, but in certain circumstances it can be harmful. A large number of drugs have been developed to treat harmful inflammation in autoimmune diseases such as rheumatoid arthritis. This raises the possibility that these or similar treatments could be effective in cardiovascular disease. Indeed, the recent CANTOS trial showed that canakinumab, a drug (originally designed for rheumatic diseases) which targets the IL1-beta protein, reduced coronary events in patients with high inflammation levels. There are multiple proteins involved in inflammation and so the challenge is identifying those that are the most promising drug targets.

I propose to address this by integrating 'high-dimensional' genetic and proteomic data. As a result of advances in technology, it now possible to simultaneously measure large numbers of proteins (the 'proteome') in the blood in large numbers of individuals. However, simply showing an association of a protein with cardiovascular disease does not necessarily indicate that the protein is a valid drug target, as correlation does not always reflect causation. To circumvent this issue, I will integrate genetic information, utilising an approach called 'Mendelian randomisation'. This method takes advantage of the randomisation of genetic variants that occurs during reproduction, providing in effect a natural randomised trial. The first step is to identify genetic variants that affect the level of a particular protein. Then, by examining whether individuals who inherit such genetic variants are at higher or lower risk of cardiovascular disease, we can establish whether that protein is likely plays a causal role in disease and thus whether it is a valid drug target.

In a complementary strand of work in collaboration with Prof. Justin Mason (Imperial College London), I am studying patients with Takayasu arteritis, a rare disease characterised by arterial inflammation, which often results in vascular narrowing, or, less commonly, dilatation (aneurysm). Cardiovascular complications are a major cause of morbidity. The prognosis in Takayasu arteritis is very variable - some patients have a benign course, while others develop progressive vascular injury. We will examine the plasma proteome to understand factors that influence this variability and to identify signatures that could be used to distinguish high-risk patients in need of stronger immunosuppressive treatment from those in whom milder, less toxic treatments would be sufficient. In addition, we anticipate that improved understanding of an extreme form of vascular inflammation should provide insights into cardiovascular disease more generally.

Technical Summary

To gain insight into cardiovascular disease aetiology, I will use Mendelian randomisation (MR). MR exploits the natural randomisation of alleles during meiosis to disentangle causation from confounding. By providing evidence that a molecular trait is causal, MR can help identify drug targets, thus reducing the high attrition rate in pharmaceutical pipelines. A major bottleneck limiting use of MR is lack of suitable genetic instruments. To address this, we will use data from INTERVAL to perform GWAS of plasma proteins. Proteins are ideal for MR since they are under proximal genetic control and can be targeted pharmacologically. We recently performed GWAS of 3000 plasma proteins measured using the SOMAscan assay in 3301 individuals. We will extend this to 10000 individuals (providing major power gains, particularly for rarer variants) using a 5000-plex SOMAscan assay. I also co-lead the SCALLOP consortium meta-analysis of GWASs of plasma proteins measured using Olink immunoassays. A second strand of work involves -omic studies in patients with vasculitis. Prof Mason (Imperial) and I have established a multi-national collaboration studying proteomic signatures in Takayasu arteritis (TA), a vasculitis of large arteries resulting in vascular stenosis or aneurysm. Some patients have aggressive disease whereas others have a more indolent course. Prospectively identifying these subtypes would allow for tailored treatments, but current measures of disease activity inadequately predict outcome. We propose to discover novel plasma biomarkers for TA activity and prognosis through a wide-angle proteomic approach coupled with machine learning. Our results will provide a basis for stratified medicine, a key theme of HDR UK, and elucidate the pathogenic pathways underlying TA (and the role of inflammation in vascular injury more braodly), thus providing a foundation for novel therapeutic strategies. Assay/infrastructure costs associated with the TA project are separately funded.
 
Title Human plasma protein QTL dataset 
Description Genome wide association study of the human plasma proteome in 3,301 blood donors. Dataset contains 1) individual-level genotype data, 2) individual level proteomic data (n approximately 3000 proteins), and 3) summary statistics from the GWAS of each protein. 
Type Of Material Data analysis technique 
Year Produced 2018 
Provided To Others? Yes  
Impact Not aware of any as yet 
URL https://www.ebi.ac.uk/ega/studies/EGAS00001002555