Skip to main content
Premium Trial:

Request an Annual Quote

Q&A: Regeneron's Aris Baras and Alan Shuldiner on Large-Scale Sequencing for Drug Discovery

Alan Shuldiner, Aris Baras

NEW YORK (GenomeWeb) – In early 2014, Regeneron Pharmaceuticals announced a new human genetics initiative, along with a collaboration with the Geisinger Health System to sequence more than 100,000 patients, that aims to define new disease targets and improve the drug development process.

As part of the effort, the company, which has three biologic drugs on the market and 17 product candidates in clinical development, built its own sequencing center, the Regeneron Genetics Center, which has sequenced more than 30,000 exomes to date (see other story).

On a recent visit to Regeneron's campus in Tarrytown, NY, Alan Shuldiner, vice president of translational genetics for the RGC, and Aris Baras, the center's executive director, spoke with GenomeWeb about the company's strategy for employing sequencing-based human genetics in drug discovery and development. Below is an edited transcript of the conversation.

When did Regeneron realize that human genetics studies have potential for drug discovery and development?

AB: Regeneron has long been a huge advocate of human genetics informing drug discovery and development. For example, Arcalyst (rilonacept), one of our therapeutics, came from initial studies in rare disease genetics involving a disease called CAPS (cryopyrin-associated periodic syndromes) that pinpointed the mutation and also helped understand the downstream consequences, with IL1 overproduction and associated autoimmune disease. We were able to develop a therapeutic for that rare disease that was approved in 2008.

Subsequently, we have had another success in terms of human genetics discoveries leading to therapeutics, the PCSK9 program. The basic discovery was that this gene is involved in hypercholesterolemia. Another study identified loss-of-function mutations that led to very low levels of LDL or 'bad' cholesterol and also conferred great protection from cardiovascular disease. Regeneron moved quickly, as did some others in the field, to develop antibodies to block that protein, and we are amongst some companies now that have filed for FDA review of those products and hope to market it and get it to patients in the second half of the year.

Those types of examples showed how much power there is in rare variation − large, functional effects on phenotypes caused by very rare mutations that exist in fewer than one in 100 or one in 1,000 individuals. Obviously, it requires sequencing thousands or tens of thousands of individuals to find those mutations and correlating them to phenotypes and diseases.

Why did you decide to build your own genome center, rather than outsource sequencing to a third party, as some of your competitors have done? Also, why did you make the RGC a separate subsidiary?

AB: Regeneron has a track record in terms of innovation in automation and technology. We did want to be involved in generating the data — all the nuances of that certainly proved to be valuable in terms of how we analyze the data downstream. It also turned out to be very cost-efficient and -effective to do it that way. With regards to the separate entity, that was important in terms of working with genomic data and with a lot of collaborators in this highly academic type of environment and culture — that facilitated a lot of these efforts.

Why did you decide to focus initially on the exome rather than the whole genome? Some of your competitors seem to be going a different route — Genentech, for example, recently announced two large-scale whole-genome sequencing projects, with 23andMe and Human Longevity.

AS: The field seems to dichotomize exomes vs. genomes, but we don't really consider it an either-or. But if you think about where the biggest bang for the buck is, in terms of being able to identify rare variants that have functional consequences that a) you can identify and b) you can make some biological sense out of, it's the exome. And even those who are sequencing full genomes, for the most part, analyze the exome sequence data first. And when you get into study design and cost considerations, if you ask anybody, 'Would you rather sequence 10 or 50 individuals' exomes vs. one genome?' — most would agree that there is much bigger bang for the buck from sequencing more exomes.

What diseases with unmet need do you see as most promising for this kind of work?

AB: We see this as a tool that can be broadly applied across clinical and therapeutic areas. As we've already seen, there have been many successes in rare diseases, in cardiovascular disease, ophthalmology, neurology, so I'm not sure this is an approach that's specific to any one disease area. At Regeneron, we are certainly applying these types of approaches across all of our programs and therapeutic areas.

AS: What we're doing here essentially is casting a net across the whole genome, letting what we discover inform biology, and then through those basic discoveries, we then identify therapeutic areas that they would be most relevant to, which I think in many ways is very different from conventional pharma discoveries, where you start with the disease and try to figure out from there what approaches to use to find therapeutics.

You have a number of collaborations, including with the Geisinger Health System, Columbia University, the Clinic for Special Children, Baylor College of Medicine, and the NHGRI's Undiagnosed Diseases Program. How do they differ?

AB: Some are focused on family-based approaches, the Geisinger collaboration is population-based, and for some we are working with founder populations, like the Amish.

Combining these, we can expand upon findings, and we can give ourselves more confidence. For example, when we're looking at one individual family, we can reference the entire Geisinger database and see that [a couple of mutations we considered as possibly disease-causing] are, in fact, somewhat common in the rest of the population and don't cause disease.

In another example, our colleagues found what they think is an obesity-causing gene in a family. Of course we have to validate that and replicate it in other families. But what they were able to do is work with our Geisinger colleagues and look at the database and find that different mutations in the same gene have an effect on obesity and BMI in the general population. These are the types of things the team here can do on the order of weeks and months, and put together different lines of evidence to strengthen our conclusions or hypotheses.

AS: We have large populations with rich phenotype data from electronic health records; we have collaborations on rare Mendelian disorders that we're doing new gene discovery in; and we have family populations, like the Amish, that kind of sit somewhere in between. And with the sequence data generated from the same platform, there is this exchangeability of data, the ability to analyze across different study designs that provides a terrific opportunity to go from a gene variant that we think might be connected to a phenotype to not only replication, but to get to a better feel for the genetic architecture of the phenotype, from very rare, large-effect, and sometimes maybe more common variants that have smaller effects in large populations.

Not all of the RGC's collaborations involve your sequencing facility, right?

AB: We have many collaborations where we provide large amounts of sequencing data. We also have some collaborations where our partners contribute sequence data to our joint downstream analysis. The Baylor collaboration as well as the NIH Undiagnosed Diseases Program collaboration are really focused on identifying genes, trying to confirm those, and maybe doing some functional studies as well. To date, our focus in those collaborations has been to help them with some of the downstream functional work. This ties into Regeneron's experience in mouse genetics and other biochemical and molecular biology expertise.

What types of additional collaborations are you looking for?

AB: We think the family-based approach is a very high-yield approach. We only need to sequence a family or cohort of families to get us some really interesting gene discoveries. We like that approach and would like to apply it to many phenotypes that are of interest or that we think are therapeutically relevant. You can imagine dozens of phenotypes and collaborators that we would love to engage through that approach.

AS: We try to target some of our collaborations towards phenotypes that we have a particular interest in. And then on the other side of the coin, we are also opportunistic in that there are thousands of experiments in nature that have occurred across the planet in our species. These usually manifest as rare so-called Mendelian disorders or occur in founder populations, a population of families that have a higher incidence of these disorders. Those are the kinds of collaborations we're looking for.

How do you prioritize the tens of thousands of samples you receive for analysis?

AB: A lot of it is operational and logistics. We do have some initiatives in terms of our existing drug targets where we prioritize some of the analyses. But we also have a pretty agnostic pipeline for new gene discovery. We tend to work through projects and complete them by a certain time, so we don't have too many things ongoing.

Can you expand on the projects with your existing drug targets?

AB: We have a list of genes that correspond to our drug targets and other genes of interest, and we constantly scan those for interesting mutations in our database, for example loss-of-function mutations. In many of those genes, we found carriers of these loss-of-function mutations. Now that we have done this on a few occasions, we can look at the phenotypic consequences and try to make associations to certain phenotypes for clinical outcomes. In collaboration with Geisinger, we are also working on calling back some of these individuals by genotype and conducting clinical research studies, deeper phenotyping, to understand what those mutations may be doing.

How are you planning to integrate results from the RGC with Regeneron's discovery platforms in mouse genetics and monoclonal antibody technology to get to new therapeutic candidates?

AS: We have constant interactions with all of our therapeutic heads and all of our clinical development teams. They are part and parcel of us developing outside collaborations in specific areas, and also taking any genetic and genomic findings and extending them into biological contexts. That then brings in our functional modeling groups, particularly cell models and our powerful mouse transgenic models.

We're also very excited about understanding biology in humans, and to do these call-back studies where we can perform clinical studies in people of pre-defined genotype for deep hypothesis-driven phenotyping. And to design those studies, we work very closely with our therapeutic heads and clinical development areas.

We have a number of studies that are being designed now where we want to understand what that genetic mutation does in humans. You can then recruit people that either have that mutation or don't and bring them into a clinical research center for deep phenotyping.

In my prior life at the University of Maryland, where we worked a lot with the older Amish, that was exactly what we did. We used high-throughput genomic approaches to identify people with mutations of interest, then brought them back for phenotyping. One of our stories included a null mutation in APOC3, reported in Science back in 2008, that had a cardioprotective effect. We then brought these people back, did biopsies, very sophisticated lipid turnover studies, which allowed us to understand in great detail what APOC3 did in humans. That can then lead to the identification of drug targets, and, ultimately, a therapeutic.

George Yancopoulos, Regeneron's founding scientist and CSO, attended President Obama's recent announcement of the Precision Medicine Initiative. How does the company plan to contribute?

AS: It's still in the early planning phases. We really applaud the administration's investment in this initiative; we think it's great anytime there is a big commitment to precision medicine and biomedical research in general. Obviously, we are huge believers. George has been involved in discussions, certainly Regeneron colleagues have been involved, and we're excited to follow up and see what the opportunities are.