ӳ����ý

In the early 2010s, Jose Florez and his collaborators knew that a disproportionate number of people from Latin America suffered from type 2 diabetes, but they didn’t know why. The researchers pored through genetic databases looking for markers unique to these patients, but most information in these databases is from people of European ancestry, which limited the ability of researchers to find meaningful clues.

Jose Florez. Credit: Gretchen Ertl.

To broaden this search, Florez, an institute member at the ӳ����ý of MIT and Harvard, chief of the endocrine division and the diabetes unit at Massachusetts General Hospital, and director of ӳ����ý’s Diabetes Research Group, collaborated with colleagues in Mexico. They analyzed DNA from nearly 9,000 people from Latin America, with and without diabetes. One of the studies from this work, which was part of the Slim Initiative for Genomic Medicine in the Americas (SIGMA), identified several variants in the gene SLC16A11 that dramatically increase the risk of type 2 diabetes.

That finding, in 2014, helped kick off several ongoing research projects at ӳ����ý that are revealing what SLC16A11 and other genetic markers are doing in the body at a cellular and molecular level to cause diabetes. ӳ����ý researchers have since found, for example, that the SLC16A11 variants they’ve identified decrease expression of SLC16A11 in the liver. They are now exploring how this may contribute to diabetes.

Building on these and other insights, ӳ����ý scientists are pursuing new drug strategies for treating the underlying causes of diabetes. Their goal is to find a cure, since most current medications only treat hyperglycemia, or high blood sugar, which is the end result of multiple pathophysiological processes. Other ӳ����ý efforts are using genetic information to define more subtypes of diabetes, which could help doctors better tailor existing drugs for individual patients. Florez, who is also a professor of medicine at Harvard Medical School, says that the future of diabetes treatment will be determined by the ability of researchers to leverage genetic data. “When I retire, I want to see that the practice of diabetes changed because we had access to the genome and used this information in smart ways,” he said.

In all these efforts, ӳ����ý researchers are using a variety of skill sets and expertise to attack the problem from many angles in a collaborative way. This, Florez says, is a key strength of the Diabetes Research Group. This team has roots in the earliest days of the ӳ����ý thanks to diabetes genetics research from David Altshuler, who was a founding core member of the ӳ����ý and is now at Vertex Pharmaceuticals, Joel Hirschhorn, co-director of the ӳ����ý’s Metabolism Program and a professor of pediatrics and genetics at Boston Children’s Hospital, and others.

“It’s very enriching to be a physician working together with molecular biologists, statisticians, and basic bench scientists,” Florez said. “We all benefit from working closely with each other.”

A multifaceted foe

Diabetes affects more than 10 percent of the American population, according to the American Diabetes Association, and was the seventh-leading cause of death in 2017. It is marked by the body’s inability to regulate blood sugar levels, which can happen in two key ways. In type 1 diabetes, immune cells attack insulin-producing beta cells in the pancreas. In type 2 diabetes, a combination of mechanisms that include insufficient insulin production by beta cells and changes in other tissues make the body less responsive to insulin. Diabetes can cause issues in other parts of the body, such as the kidneys, eyes, blood vessels, and nerves. While diet, exercise, and a variety of medications can manage symptoms, none of them are cures. 

Miriam Udler. Credit: Gretchen Ertl.

While doctors generally prescribe the same treatments for patients diagnosed with the same subtype, either type 1 or type 2, people in each category can have widely varying symptoms and some don’t fit into either group. “Every time I see a new patient, I have to try to determine what type of diabetes they have. It’s not always obvious,” said Miriam Udler, an associate member at the ӳ����ý in the Diabetes Research Group, as well as an assistant professor at Harvard Medical School and an endocrinologist at Massachusetts General Hospital, where she directs the MGH Diabetes Genetics Clinic. “The way we think of diabetes right now is very simplistic.”

Udler and her colleagues are looking for other possible subtypes of diabetes based on genetics. They hope that grouping patients into more and better-defined subtypes will help doctors prescribe more effective medication for their patients. 

To do this, the researchers are looking for gene variants known to increase diabetes risk in patients who have a range of diabetes-related traits and symptoms. They are searching for correlations between certain high-risk gene variants and traits such as high blood sugar, insulin insensitivity, and obesity. If a variant known to increase risk of diabetes also increases risk of one of these traits, that provides clues about what the variant might be doing in the body. By , Udler and her team can pinpoint cellular pathways that might underlie diabetes and potentially be targeted by new drugs.

“If we find a druggable pathway, that could potentially benefit everybody, not just patients with the variant,” Udler said.

Grouping variants together can also help delineate new subtypes of diabetes. “Sometimes patients develop a condition and wonder why this happened to them,” Udler said. “It’s really nice to help them understand the pathways they inherited and use that to better direct their treatment.”

Drug combos and data sharing

Florez leads a parallel initiative to discern how genetics can help physicians prescribe the most effective medication. (GRADE) is a nationwide NIH-funded clinical trial that has recruited patients who are taking a medication called metformin for type 2 diabetes. Patients received one of four other diabetes drugs to take in combination with metformin and are being monitored for how well the combinations are working.

Florez leads the genetic research arm of this trial. Using the genomic sequences of each patient, he and his colleagues hope to find genetic variants that correlate with better response to certain drugs, as well as variants that might signal potential resistance to other drugs.

Jason Flannick. 

Udler and Florez also support the (RADIANT), a program that studies patients whose disease does not fit the typical characteristics of type 1 or type 2 diabetes. Udler leads clinical recruitment at Mass General, while Florez directs sequencing efforts within ӳ����ý’s Genomics Platform. 

“It’s a matter of time before every person’s sequence is part of their medical record,” Florez said. “At that point, we should be able to decide what drug a patient should get or what complications to watch for, because of their genetic profile.”

Access to such genetic data is critical for this kind of diabetes research, and ӳ����ý researchers have spearheaded projects to collect, organize, and freely share sequencing data about diabetes and related diseases. Associate member Jason Flannick, an assistant professor of pediatrics at Harvard Medical School and Boston Children’s Hospital, and Noel Burtt, director of operations and development for diabetes research and knowledge portals at ӳ����ý, helped found the (AMP-CMDKP).

Noel Burtt. Credit: Maria Nemchuk.

The portal grew from an effort to catalogue variants related to type 2 diabetes and . “It’s a place where all the genomic data has been deposited in ways that it can be analyzed by anyone in the world,” Florez said. 

"Our goal has been to enable people to ask questions about the genetic data that go beyond what's normally published about them,” said Flannick.

Turning models into medicines

While existing therapies can dramatically improve quality of life for many diabetes patients, the lack of a cure creates a need for new drugs that address the underlying molecular mechanisms of the disease. “We don’t have anything that targets the core pathophysiology,” Florez said. “We hope that through genomic discovery, we can identify targets for disease-modifying therapy, as opposed to disease-treatment therapy.”

New medications could be hiding in plain sight — drugs that have already been studied for use in other diseases but have never been tested in diabetes. Institute scientist Bridget Wagner, director of pancreatic cell biology and metabolic disease at ӳ����ý, is using the ӳ����ý’s Drug Repurposing Hub — a library of over 7,000 compounds that have been previously tested for safety in humans — and other similar collections to find potential new diabetes drugs.

Bridget Wagner. Credit: Gretchen Ertl.

For example, in collaboration with Florez, Suzanne Jacobs, an institute scientist and associate scientific director of the Diabetes Research Group, and other ӳ����ý colleagues, Wagner and her group are searching for drugs that target SLC16A11. Variants of this gene that increase diabetes risk . So Wagner’s group has screened several thousand compounds from the Drug Repurposing Hub for ones that boost expression of SLC16A11 and narrowed the list to just a few, which they are now further characterizing and developing. They hope that such a drug will help researchers better understand how SLC16A11 drives diabetes risk and could even be a potential treatment for patients with SLC16A11 variants.

Wagner has also applied the drug collection to the insulin-making beta cells of the pancreas, which are often damaged or not working well in patients with diabetes. She and her team are looking for compounds that can increase insulin secretion in beta cells. Because evaluating insulin secretion in cells is time-consuming, Wagner and colleagues have come up with a way to measure it at scale. They’ve . This allows the scientists to quickly screen for drugs that make the cells light up as a sign that the compounds are also boosting insulin production in the beta cells.

“It was a big technical advance,” Wagner said. “This gives us the ability to screen for insulin production in cells in a very high-throughput way.”

Paul Clemons. Credit: Gretchen Ertl.

This work of reviving beta cell function also involves a collaboration with institute scientist Paul Clemons, director of computational chemical biology research at ӳ����ý. For decades, scientists have believed that an adult human pancreas cannot make new beta cells, but Clemons and Wagner are challenging that notion by searching the Drug Repurposing Hub for compounds that help beta cells proliferate. 

Their teams study beta-cell growth using human pancreatic islets — parts of the pancreas that house the beta cells. As islets contain a variety of other cell types, it can be difficult to evaluate the growth of only the beta cells. Wagner’s group stains cells from islets with fluorescent dyes to distinguish beta cells from other cell types and to highlight proliferating cells. The team uses fluorescence microscopy to capture images of the cells, and Clemons is working to automate the analysis of the images and screen them for signs of increased beta-cell growth. 

“You can’t afford to limit yourself to one approach,” Clemons said. “You need multiple approaches to triangulate what’s going on. It’s a puzzle to figure out how a compound elicits its action.”

Solving gene variant mysteries

Other members of ӳ����ý’s Diabetes Research Group are using genetic information to better understand how cells are malfunctioning at a molecular level in patients with diabetes. Jacobs, in collaboration with Florez and other ӳ����ý scientists, that variants in SLC16A11 decreased its expression in the liver and prevented the protein encoded by the gene from localizing to the cell membrane, which is necessary for it to function properly. They linked this defect to a change in the way lipids are metabolized in the liver cells.

Suzanne Jacobs. Credit: Gretchen Ertl.

Jacobs and her team are now delving deeper into the effects of SLC16A11 risk variants on liver cell metabolism and physiology. They that decreased expression of this gene leads to altered glucose metabolism in liver cells and lab animals. These animals also show changes in energy metabolism and liver physiology that are often seen in individuals with type 2 diabetes. 

“We need to understand what genes are affected by the variants and in which cell types they’re acting. Then we can test how a change in expression or function of that gene affects metabolism and disease risk,” Jacobs said. “We’ve now taken this story from genetics to mechanism and to an initial understanding of physiology.”

Figuring out what genetic variants do in cells is even more challenging when variants occur in regions of the genome that don’t code for proteins, but rather regulate expression of other genes.

To streamline and scale up this approach, institute member Melina Claussnitzer, an institute member at the ӳ����ý and an assistant professor at Harvard Medical School at the Center for Genomic Medicine and the Endocrinology Division at Massachusetts General Hospital, and her group have developed a computational and experimental toolkit that predicts and tests how variants in noncoding regions affect molecular and cellular function.

Claussnitzer, who has co-lead the ӳ����ý’s Variant-to-Function initiative, and her team recently used the toolkit to figure out how variants in a noncoding region of a gene called ADCY5 disrupt metabolism in two cell types, fat cells and bone cells. This altered metabolism is associated with increased risk of both high blood sugar and dense but fracture-prone bones often seen in patients with type 2 diabetes.

Melina Claussnitzer. Credit: Gretchen Ertl.

Claussnitzer and her colleagues now want to use the framework to explore how variants affect entire networks of gene expression, and therefore cellular programs in fat cells. She and Florez are collaborating on a study analyzing fat cells, which contribute to insulin resistance in type 2 diabetes, from a large cohort of people with and without diabetes. They’ve stratified the patients according to their polygenic risk score for diabetes, which is a calculation of a patient’s cumulative genetic risk for the disease based on their many different genetic variants. They have also applied pathway-specific polygenic scores for type 2 diabetes using the approach . The team’s goal is to determine how cell morphology and gene expression differ between fat cells from individuals with high and low polygenic risk. The researchers hope their analysis will provide clues about the molecular and cellular mechanisms in fat cells that could be driving diabetes risk, which can better inform the development of future treatments.

The diabetes researchers say this and other projects wouldn’t be possible without the types of collaborative relationships they have cultivated. “It’s been really transformative to work closely with collaborators, which has allowed us to study patient information in this way,” Claussnitzer said. “It’s really driven our research forward.”

Other members of the diabetes group echo her feelings. 

“We have experts who span the whole spectrum from genetic discovery and mechanistic insight through chemical screening and clinical application, and we’re able to bring it together,” Jacobs said. “The opportunity to work with and learn from each other has been a very rewarding experience.”

Support for the work described in this article was provided in part by the Carlos Slim Foundation, National Institute of Diabetes and Digestive and Kidney Diseases, National Institute of General Medical Sciences, National Human Genome Research Institute, National Heart Lung Blood Institute, Massachusetts General Hospital Research Scholars Program, and the Massachusetts General Hospital Transformative Scholars Program.