Five Questions for Stanley Shaw

Part of what we do at the Broad involves unraveling the genomes of humans, dogs, and dozens of other creatures as part of a larger mission to improve human health. In addition to finding genes or mutations linked with disease, Broad scientists seek to learn what the genes are doing functionally, and how to use that information to devise new therapies.

Toward that end, Broad associate researchers are using small molecules to probe the functionality of certain disease genes associated with diabetes. A paper describing the work was recently published in the Proceedings of the National Academy of Sciences. In this work, the researchers studied how specific alleles, one of the two or more forms of a gene, function in response to interactions with small molecule chemical compounds.

“The approach is at the heart of Broad aspirations of linking human genetics to future therapeutics,” says Stuart Schreiber, a core member of the Broad and co-author of the paper. To learn more about the project, I spoke with first author Stanley Shaw of the Center for Systems Biology at MGH and a Broad associate researcher who was a postdoctoral fellow in Schreiber’s lab.

Q1. How did the success of genome-wide association studies (GWAS) spark interest in this work?

Stanley Shaw: This work was definitely motivated by the success of GWAS in finding disease alleles, or gene loci, associated with disease susceptibility. GWAS successes highlighted the need for a generalizable way to assign function or some biological context to disease susceptibility alleles, and in turn, get a better understanding of how a particular allele can lead to disease. We needed to find a way to leverage what we know about a disease allele to find ways of identifying new therapeutics.

Q2. Describe how your team applied the blood cell/small molecule approach as an interactive screen to learn more about how disease gene alleles function.

SS: Probing with small molecules is a very powerful way to find the functional properties of your gene of interest. We wanted to apply the logic of these screens to study disease alleles in human cells.

We used blood lymphoblast cells derived from 18 patients from one family for our model. In contrast to pancreatic beta cells that are hard to obtain from live patients, B-lymphoblast cells are used all over the world as a renewable, easily accessible source of DNA.

Half of the family members studied had a disease mutation in the transcription factor HNF4α, which led to an early onset form of diabetes. The others lacked the mutation and did not have the disease. All of the 18 lymphoblast cell lines were probed with a library of 4,000 small molecules. These small molecules, including hundreds of FDA-approved drugs, were well-annotated with known functions and targets. The idea was to find small molecules where the cells’ response to the small molecule was different in cells from diabetic individuals compared with individuals without diabetes.

We then used a statistical analysis (developed by our Broad Cancer Genomics colleagues) that looked for sets of related compounds that seemed to have a slightly different effect in mutant diabetic cells compared to nondiabetic cells. This allowed us to find a signal directly in patient cells, so that the approach could potentially be applied more broadly for a variety of disease alleles of interest.

Q3. How does probing lymphoblasts with the small molecules shed light on the disease allele?

SS: This method was a way to apply small molecules to elucidate differences in physiology in cells that have a disease mutation and those that do not. The small molecules act as targeted perturbations to the different cells. How cells respond to the molecules tells you something about functional differences between mutant versus normal cells. We used small molecules to try to uncover how cells are wired to respond to different stimuli.

In this case, we found a few classes of FDA approved drugs or fatty acids that did seem to have a different effect on the cells depending on whether or not they had the disease mutation.

Q4. What did you learn about diabetes?

SS: Specifically, we found that some of the compounds that came out of this analysis could modulate insulin secretion from pancreatic beta cells, the key problem in these diabetic patients. Interestingly, many of these compounds are FDA-approved drugs but not for diabetes. Some are anti-arrhythmic drugs and diuretics, but they emerged from our screen as having some interaction with the diabetes gene and they seem to have some ability to modulate insulin secretion.

Q5. What are the next steps for using a small molecule screen to learn more about how other genes function?

SS: Going forward, we would like to apply this small molecule approach to other disease alleles implicated by GWAS. We are working on identifying these now. As an example, we are collaborating with Ramnik Xavier, Professor of Medicine at Harvard Medical School and the Chief of Gastroenterology at MGH, as well as the Director of the Broad “Genes to Drugs” Initiative, to study the functional significance of disease alleles in inflammatory bowel disease. More broadly, we have an interest in trying to look at patient-derived cells as a new way to phenotype patients and use these surrogate model cell systems to study disease biology.

It is encouraging to think of ways that this could be applied to emerging genetic discoveries to help identify genes that work together with disease alleles. We hope to be able to get a sense of the cellular consequences of disease alleles and perhaps get some clues into pathways that could be modulated for therapeutic effect.

Paper cited:

Shaw, SY, et al. Disease allele-dependent small-molecule sensitivities in blood cells from monogenic diabetes. December 23, 2010, doi: 10.1073/pnas.1016789108