27 April 2019

Jane Ferguson: Hello and welcome to Getting Personal: Omics of the Heart, your podcast from Circulation: Genomic and Precision Medicine. I'm Jane Ferguson from Vanderbilt University Medical Center, and this is episode 27 from April 2019.

This month, I talk to Riyaz Patel, the first author on not one, but two articles published this issue, presenting analyses from the GENIUS-CHD consortium. But before we get to the interview, let's review what else was published this month.

First up, we have a paper from Tamiel Turley, Timothy Olson and colleagues from the Mayo Clinic, entitled Rare Missense Variants in TLN1 Are Associated With Familial and Sporadic Spontaneous Coronary Artery Dissection. In this study, the authors were interested in identifying novel susceptibility genes for spontaneous coronary artery dissection or SCAD, which predominantly affects young women who appeared otherwise healthy. They conducted whole exome sequencing in a family with three affected family members and found a rare missense variant in the TLN1, or talin 1, gene. This gene encodes the talin protein which is part of the integrin adhesion complex linking the actin cytoskeleton to the extracellular matrix. This gene and protein is highly expressed in coronary arteries. They went on to sequence additional sporadic cases of SCAD, and they found additional talin 1 variants in these individuals. While there was evidence for incomplete penetrance, these data implicate TLN1 as a disease-associated gene in both familial and sporadic SCAD.

The next paper comes from Miroslaw Lech, Jane Burns, and colleagues from UCSD School of Medicine and Momenta Pharmaceuticals and is entitled Circulating Markers of Inflammation Persist In Children And Adults With Giant Aneurysms After Kawasaki Disease. Kawasaki disease is the most common cause of acquired pediatric heart disease, but disease progression can vary a lot, and it's likely modulated by complex gene-environment interactions. Coronary artery aneurysms occur in about 25% of untreated patients, but early treatment with intravenous immunoglobulin or aspirin reduces the risk for these aneurysms to 5%, suggesting an important role for inflammation. In this study, the authors applied shotgun proteomics, transcriptomics, and glycomics on eight pediatric Kawasaki disease patients at the acute, subacute, and convalescent time points. They identified inflammatory profiles characterizing acute disease which resolved during the subacute and convalescent time points, except for in the patients who went on to develop giant coronary artery aneurysms. They went on to carry out proteomics on nine Kawasaki disease adults with giant coronary artery aneurysms and matched healthy controls, and they confirmed the inflammatory profiles in the adult samples.

In particular, calprotectin, which is composed of S100A8 and S100A9, was elevated in the plasma of patients with CAA, an association they confirmed in additional samples of pediatric and adult Kawasaki disease patients and healthy controls. These data suggest that calprotectin may serve as a biomarker of ongoing inflammation in Kawasaki disease patients following acute illness, and may be able to identify individuals at increased risk of aneurysms.

Next up, we have a research letter, Heart BioPortal: An Internet-of-Omics for Human Cardiovascular Disease Data, from Bohdan Khomtchouk, Tim Assimes, and colleagues from Stanford University. They had noticed that, in contrast to the field of cancer research, there were no open access platforms for cardiovascular disease data that offered users the ability to visualize and explore high quality data. They set out to fix this and developed the Heart BioPortal, which is accessible at www.heartbioportal.com. This portal allows the user to integrate existing CDD related omics data sets in real time and provides intuitive visualization and analyses in addition to data downloads. The primary goals are to support gene, disease, or variant-specific request, and to visualize the search results in a multi-omics context.

They currently collate gene expression, genetic association, and ancestry allele frequency information for over 23,000 human genes and almost 6,000 variants across 12 broadly defined cardiovascular diseases spanning 199 different research studies. And this is just the start, they're hoping to add more studies, more data, and functionality for querying CDD drug targets, along with lots more. This is a really great resource which will no doubt be of real value to the community. I urge you to go online, check it out, put in your favorite gene, and see what you find.

Riyaz Patel, Folkert Asselbergs, and many, many collaborators published Subsequent Event Risk in Individuals With Established Coronary Heart Disease: Design and Rationale of the GENIUS-CHD Consortium and Association of Chromosome 9p21 with Subsequent Coronary Heart Disease Events: A GENIUS-CHD Study Of Individual Participant Data. These papers present the design of the genetics of subsequent coronary heart disease, or GENIUS-CHD consortium, which was established to facilitate discovery and validation of genetic variants and biomarkers for risk of subsequent CHD events in individuals with established CHD. The consortium currently includes 57 studies from 18 countries, recruiting over 185,000 participants with either acute coronary syndrome, stable CHD, or a mixture of both at baseline. All studies collected biological samples and followed up study participants prospectively for subsequent events. Enrollment into the individual studies took place between 1985 to the present day, and the duration of follow-up ranges from nine months to 15 years. Participants have mostly European ancestry, are more likely to be male, and were recruited between 40 to 75 years of age.

In their first analysis using these data, they investigated whether the established 9p21 locus associated with subsequent events in individuals with established coronary heart disease. Confirming previous smaller studies, they showed that while genotype at 9p21 is associated with coronary disease when compared to healthy controls, 9p21 genotype is not associated with a risk of future events in people who already have coronary disease. Dr. Patel joins me to tell me more about the GENIUS-CHD consortium and the analyses described in these papers.

Today, I'm joined by Dr. Riyaz Patel, who's an associate professor at University College London and a cardiologist at the Barts Heart Centre in London. Dr. Patel, thank you so much for joining me.

Dr. Riyaz Patel: Pleasure to be on, thanks.

Jane Ferguson: So, as we're going to discuss, you are the lead author on two back-to-back publications that were published in Circ Gen this month exploring genetic predictors of coronary heart disease as part of the GENIUS-CHD consortium. Before we delve fully into them, could you tell us a little bit about your background and how you got into this research field?

Dr. Riyaz Patel: Yes. I'm an academic cardiologist, as you know, and I first got into genetics of coronary disease about 12-13 years ago, now, around the time that genome wide association studies were about to take off, or were taking off. I studied, I worked at Emory University, in fact, in Atlanta, in the US. We had a very big cohort of patients who had coronary disease, who were undergoing coronary angiography. At that time, we were doing quite a lot of genetic association studies and biomarker work in patients with heart disease. One of the key problems we often encountered was sort of looking for replication cohorts and trying to do things at a bigger scale than what we had available. So that kind of really was the initial driver for trying to bring together a bigger collaboration to take that sort of work to the next level.

Jane Ferguson: It sounds like you've got valuable expertise, because looking at the author list for these papers, I think it's one of the longest author lists I've ever seen. It's a huge endeavor. I'd love to hear more about how that got started and how you managed to build this consortium, and you know, and tell us what the consortium actually is.

Dr. Riyaz Patel: Yeah, it's been a labor of love. And essentially, I started when I returned back to the UK and we were looking to develop this further. We had already collaborated with several colleagues in the US and abroad from my time at Emory. So, we pulled together a small group of people who we were already working together with and then we did predicts of systematic searches of literature to identify cohorts who were also doing similar things. Again, investigating people with heart disease and looking at subsequent event risk. So, we did that and then we systematically approached, very much, as many people as we could find and over the course of the last, maybe 3 or 4 years, we've brought together a small community of collaborators around the world, and as you rightly said, it's a very long list. In total, we're counting around 180 or so investigators. But, in a way, that also speaks to how this consortium is not just a collection of studies. It is a collection of people and a lot of expertise was brought to the table because of that. People have been thinking about these questions for many, many years and this platform essentially is an opportunity for everyone to share that knowledge.

Dr. Riyaz Patel: So that's kind of how the consortium started and is being pulled together. We operate on a sort of loose memorandum of understanding where every member of the consortium is free to participate in studies as they wish. We run analysis in a federated way which means that [inaudible 00:10:50] scripts are shared and people standardize their data and then they run analyses locally and they only share summary level data so that obviously overcomes the big governance hurdle. So, that's pretty much how the consortium works at moment.

Jane Ferguson: Yeah. I'm sure there was probably a lot of challenges along the way in figuring this out and getting scripts that work for everybody, dealing with all the people, so how do you do this? Do you have regular phone calls with 180 people on it? Do you have lots and lots of emails?

Dr. Riyaz Patel: (laughs)

Jane Ferguson: How's it actually working?

Dr. Riyaz Patel: So, we have a steering committee which is represented by at least one person from each study. So, that limits the number of people down to about, a more manageable number, about 50 or 60. And we do have regular teleconferences, particularly in the early days when we were still pulling everything together. Now, we try and meet at least once a year, if not twice at year at the major conferences, at the European Site of Cardiology and one of the big American meetings, ACC or AHA, so that's usually a good face to face meeting that we have with everyone and then as with all consortia, we have regular email lists and contact through that means.

Jane Ferguson: So, now that you've got everybody together, you have over 185,000 participants as part of this from 18 different countries. So, how have you been able to use all of these different data and harmonize the different phenotypes and sort of put everything together to actually run the analyses.

Dr. Riyaz Patel: The way we started off is by asking everyone to share almost an inventory of what they have collected. We then sought to try and standardize all of the core variables: age, sex, smoking and so forth. Once we were happy about the key variables had been standardized, units were the same and so forth, we then created, effectively a GENIUS-CHD data set that each cohort had curated. So, this was the main way of harmonizing the data set. Now, obviously, there are a lot of other differences between each of these studies. So, we have within the consortium a combination of different studies. We have randomized clinical trials, we have cohort studies, we have nested cohorts from larger population studies and we try and, in all of the analyses, we have pre specified subgroup analyses to try and look out and check for any heterogeneity that is introduced because of all of this. But the biggest, sort of, difference that we have factored in is that each of these studies collects patients with different types of coronary heart disease.

Dr. Riyaz Patel: So, there are about ... 40% or so are acute coronary syndrome recruited patients, where these people are recruited at the time or after their acute event. And a similar proportion are recruited when they're much more stable. So, in all of our analyses we do try and factor in the differences in terms of the type of CHD patients are enrolled with but everything else, as best as we can, we have tried to standardize including all of the outcomes. So, for example, we share the ICD codes that would define a particular type of outcome across all the different cohorts, so even if you're in a different country, they will generally be reasonably well standardized.

Jane Ferguson: Mm-hmm (affirmative), yeah, yeah. I think it's important and I can see the pros and the cons, you know, you have more diversity and you're representing a broader spectrum of disease by including everybody but then, of course, it's hard to figure it out, but I'd say it gives you a lot of versatility with the types of analyses you can do.

Jane Ferguson: As we mentioned, there's two papers so people can go online and read those two papers. And the first one, is sort of the design and goes really into detail of how you guys set this up and I think is a really nice, sort of, example of, if anybody else was trying to (laughs) do something like this, of how to follow it. But then you also did, sort of, an initial analysis, right, to show what this consortium can actually do. I looked at 9p21, so I'd love to hear more about those analyses.

Dr. Riyaz Patel: Yeah, so 9p21 is one of the most reproduced variants with coronary disease across the world. And it's remarkable how well replicated it's been in all sorts of settings in different countries. But the key thing is that it's been associated mostly in case controlled studies or in first event type of studies. And when we looked at this question some years ago now, at whether a variation of chromosome 9p21 is also associated with subsequent events, IE., we could test in people who've already had a heart attack or coronary disease, does it predict a worse outcome for them. We found that it hadn't.

Dr. Riyaz Patel: [inaudible 00:16:06] was in the literature metro analyses and, sort of, all the caveats that come with that. So, we thought that as a feasibility analysis within the consortium, "why don't we also look at 9p21," which we did and this time around, we were able to identify that 93,000 people with coronary heart disease who had our primary endpoint of coronary heart disease death or MI subsequent to other index events. Again, we confirmed our previously met analyses findings that in this particular setting, 9p21 doesn't seem to associate with risk of subsequent events. And that sort of fits with our understanding of 9p21 so far. And interestingly, in one of our analyses, we identified that it does associate with risk of repeat revascularization. And from what we know about 9p21 so far, it seems to associate with risk of atheroma development or progression as opposed to perhaps plaque vulnerability or rupture which might give you an acute coronary event.

Dr. Riyaz Patel: So, it's been a good example, I think, and really an illustration of how this consortium can work at scale. We have a lot of flexibility in terms of different subgroups that we can look at. And we really drilled down in this paper at all the possible reasons why a neutral finding may have occurred. We've looked at selection bias, we've looked at all the different subgroups which was can do because of the scale of the analysis. So, yeah, so that's kind ... it's really, the findings are not particularly novel in their own right but it is a very good example of feasibility of a consortium.

Jane Ferguson: Yeah, I agree. Because it is, so often, if you get, sort of, a negative finding, you keep wondering, "Well, was it just the power? Do we not ... are we not able to find it?" But, I think, with the scale that you have, you're really able to drill down and say, "Look, we really think there's nothing here. It's a true negative finding." You know, 9p21 is not associated with subsequent events, although, I think the revascularization is interesting and that can, sort of, inform, I guess, more basic research into the the mechanisms of 9p21.

Dr. Riyaz Patel: Exactly. Exactly.

Jane Ferguson: So, what's next? I'm sure there's a lot more papers and analyses that are, sort of, to come out of this. So, can you give us, sort of, a sneak peek of what you're working on now?

Dr. Riyaz Patel: Yeah, so, like with 9p21, we did have a selection of variants to answer important questions. So, for example, we were looking at the role of PCSK9 variation to try and see how that relates in this particular setting, given that trials have already reported on the effective drug. And similarly, we're also looking at interlinking six receptor blockade as a, sort of, similar sort of [inaudible 00:19:11] randomization study to look at the validity of a drug target in a secondary prevention setting.

Dr. Riyaz Patel: Beyond that, we are looking at genome wide association studies and, hopefully, once that is done, the consortium will be in a position to do lots of quick look-ups or all sorts of different questions in genetic variation to inform drug target analyses. So, those are immediate priorities, but we are also, in parallel, looking at non-genetic analyses, so, once again, there are lots of standard clinical risk factors that we need to explore a bit more thoroughly in this setting. So as you're aware, there are various paradoxes that keep creeping up in studies where patients have coronary heart disease already, so the obesity paradox is a good example. And what we're hoping to do, is we're hoping to drill down into many of these observational findings in this particular setting, which hasn't really been done, simply again, because the lack of available resources of anything at this scale.

Jane Ferguson: It's exciting and it sounds like you have a really powerful set of different data sets to be able to ask a lot of interesting questions. So, I'm excited to see what's gonna come out next.

Dr. Riyaz Patel: The other key thing we're working on is also about risk prediction. So, again, one of the things we're missing in the clinical community is good risk prediction tools for subsequent event risk among patients with heart disease. We are working with various colleagues to try and develop better risk prediction algorithms for people who've survived coronary event or have coronary disease.

Jane Ferguson: Alright, that's really interesting and that feeds in really nicely then to, sort of, the precision medicine approach. Well, congratulations on building this. I think that's a huge effort in itself and then also in these two papers that were published this month. I think it's really, really, really great work.

Dr. Riyaz Patel: Well thank you. And a key message here is that we want to build and expand this community of investigators around the world who are looking at risk question because individually, I think, we've all struggled with various, sort of, issues. But collectively, I think we have so much more potential to really address some big questions. And the consortium, as I mentioned, is not just investigator led in terms of what we're doing. We're also very open to collaboration and for people wishing to replicate their own findings and are looking for similar cohorts or larger scale validation opportunities so that is also another key advantage in benefit or risk consortium.

Jane Ferguson: Well, that's wonderful. So, if anyone has either data sets that they want to contribute, are you still, sort of, accepting new investigators?

Dr. Riyaz Patel: Absolutely. Very much so. I mean, in the paper, we do mention that we are limited, particularly in terms of cohorts that are enriched for female patients as well as cohorts enriched for patients who are non-Caucasian, in terms of ethnicity. Because, again, those are important patient groups that we need to address. But, generally speaking, we are absolutely open to including anyone who's interested and who meets the inclusion criteria which is collecting people with coronary heart disease, have got genotyping or examples stored for future analysis and have prospective outcomes connected.

Jane Ferguson: And is there a minimum size of sample that somebody needs to participate?

Dr. Riyaz Patel: Ideally, we'd like to, sort of, set that level at about 1,000 recruited patients. But again, if someone has a very deeply phenotyped cohort and that are interested, we'd be more than happy to discuss that and take that to the steering group.

Jane Ferguson: Okay, wonderful. So, people can just email you if they wanna contact-

Dr. Riyaz Patel: Absolutely.

Jane Ferguson: You any further.

Dr. Riyaz Patel: We also have a website, which is for the consortium, which also has contact details on there.

Jane Ferguson: Okay, perfect. Alright, so let me see. Your email is riyaz.patel@ucl.ac.uk-

Dr. Riyaz Patel: Right.

Jane Ferguson: And then the website for the consortium?

Dr. Riyaz Patel: Www.genius-chd.org

Jane Ferguson: Okay. Perfect. Thank you. So, any listeners that are interested, we'll urge them to either go to the website, read some more, go read the papers, email you directly to talk more. Thank you so much for joining me and for talking about this work.

Dr. Riyaz Patel: Thank you for having me.

Jane Ferguson: That's it for April. Come back in May for the next issue. And thank you for listening.

This podcast was brought to you by Circulation: Genomic and Precision Medicine and the American Heart Association Council on Genomic Precision Medicine. This program is copyright American Heart Association 2019.