An Interview with Joslin's George King, MD
George King, MD, devoted his professional life to unraveling the mysteries of diabetes, in large part because he wanted to figure out why several Asian populations, unique among the peoples of the world, are prone to developing the disease without first becoming obese.
Decades of research have yet to bring him closer to answering that question, but King has no regrets about his career choice. He is now a professor of medicine at Harvard University and the chief scientific officer at Joslin Diabetes Center. His long-term partnership with Lloyd Paul Aiello, MD, PhD, led directly to the anti-VEGF treatments, which inhibit vascular endothelial growth factor and can almost eliminate blindness caused by diabetic retinopathy and diabetic macular edema. Better still, King believes, another long-term project may eventually pay off with diagnostic tests that will predict which patients face the greatest risk of diabetic complications and treatments that will prevent some of those complications.
“I started off with big questions about the fundamental nature of diabetes and, thanks to the unexpected complexity of the disease, I still have most of those very same questions,” said King. “On the bright side, thanks to great progress in basic science, my colleagues and I found ways using this basic information to discover some factors that may predict complications and some factors that protect against them.”
It has long been apparent that some patients fare far better with diabetes than others, even after adjusting for glycated hemoglobin levels and other known risks. A patient who eats well, exercises, and controls blood sugar can lose kidney function 10 years after diagnosis and die shortly thereafter. Another patient can eat doughnuts, be on 1 or 2 insulin shots a day, and live 50 healthy years. Researchers have always looked for hidden factors that explain such differences, but King recognized years ago that an initiative begun in 1948, along with advances in information technology (IT), might allow him to make connections that others had missed.
The initiative, dubbed the Joslin Medalist Program, attempts to improve self-management among patients with type 1 diabetes (T1D) by giving medals to long-time survivors; the first medals went to people who lived with the disease for 25 years. However, survival periods grew so much with the passing of time that Joslin presented its first 80-year medal in 2013. The program effectively gave Joslin a very large list of people who fared extraordinarily well with T1D, a list that just begged to be converted into a long-term study cohort.
King and his team started a decade ago with about 550 people who had survived at least 50 years beyond their initial diagnosis.1 Since then, the researchers have performed yearly physicals on each patient and taken samples of blood, urine, and DNA, with follow-up visits every 3 years. They have also tried to convince cohort members to donate their organs to Joslin.
“Large biopsies would be ideal, but patients are understandably reluctant to give us significant portions of organs they’re still using, so this program has been the primary source of invaluable tissue samples,” said King. “It’s critical, though, that we get them as soon as possible after they die. Ideally, either the patient or a family member will contact us when death seems imminent, so we can make arrangements to preserve the tissue.”
Not every patient has undergone every checkup and only about half have agreed to donate their organs, but Joslin has still collected a huge amount of biological information about how T1D affects a highly resistant population over long periods of time. In years past, this data would have overwhelmed researchers, obscuring all but the most obvious connections, but IT has made it possible to take several million data points for each patient and compare hundreds of patients in the search for patterns that explain (or at least predict) outcomes.
“Much of the value in this cohort lies in the fact that although its members have all done better than most type 1 patients, they’re still a heterogeneous group. Different members eventually suffer different complications. Some have kidney problems but perfect eyes, while others have eye problems but perfect kidneys. That allows us to compare and contrast patients who do and don’t suffer a particular complication,” said King. “Of course, it would be impossible to separate relevant data from noise if we had just a few patients to compare, but when you have hundreds of subjects in each group, you can narrow down the possibilities enough to test them.”
Once King and his team think they may have identified a protective factor against a particular complication, they move to the lab to evaluate the hypothesis. If, for example, they find a protein that may protect the eye, they will culture eye cells, overexpress the protein in some batches of eye cells, and see if the enhanced eye cells fare better than control cells when they simulate the effects of T1D, such as hyperglycemia.
If a possible protective factor works in the test tube, the researchers move on to animal models. Typically, they will engineer a mouse to express the relevant factor (and to develop T1D) and then see if that factor protects mice against the associated complication. If such experiments produce positive results—and the vast majority of potential protective factors fail before this point—then Joslin looks for partners in pharmaceuticals or biotech to help translate this basic research into actual treatments.
King and his colleagues have been using this basic protocol to test potential protective factors for many years now. It has yet to produce a commercial treatment for patients with diabetes, but it has identified a number of possible drug targets. Joslin’s partners are currently using animals to test compounds that bind with 1 eye-related target and 1 kidney-related target. If all goes well, King says, efforts could progress to clinical trials of entirely novel treatments in 3 to 5 years.
“When we first started the process, we hoped that we would find a single factor that protected against all diabetic complications. In reality, we’ve found different factors that seem to protect against different complications, which is logical given that different people develop different complications,” King said. “On the bright side, most of the potential protective factors we have identified are proteins and metabolons, which often lend themselves to drug development, rather than bits of genetic code, which do not.”
The duration of Joslin’s Medalist study has also allowed King to undertake a Big Data approach to developing tests that predict which patients are likely to develop particular complications at particular times. The process starts by taking all cohort patients who developed a certain complication after they joined the study and then analyzing data collected from them before that diagnosis. If, say, there are common elements in blood tests taken from most study group members 5 years before kidney failure (elements that are rare in tests from patients whose kidneys will still be working in 5 years), software can sometimes point them out to researchers who can then test the predictive power of those biomarkers in other patients.
Of course, validating the predictive powers of biomarkers would be impractical if it had to be done in real time. Any prospective trial that would validate a diagnostic tool’s ability to predict kidney failure 10 years in advance would take 10 years or longer. Fortunately, King can demonstrate the predictive power of patterns from the long-term records of the Medalist patients by showing that they could predict outcomes using long-term records of other long-term studies which have been following a group of T1D patients for many years and collecting blood and urine for over 20 to 30 years.
King and his team believe this technique has enabled them to spot a number of biomarkers that make earlier and more accurate predictions about which patients will develop different complications. Indeed, King and his Joslin colleague Hillary Keenan, PhD, filed for a patent on a test that would gauge a patient’s risk for diabetic nephropathy by testing urine, blood, or kidney tissue for the profile of proteins and metaboloms and comparing the results with those from patients who lived 25 or more years without such a complication.2 Such a test could help doctors personalize the courses of treatments for particular patients. It could also be used to help evaluate the effectiveness of treatments designed to prevent diabetic nephropathy.
The predictive tools that emerge from Joslin’s labs could eventually power commercial testing products, but it would take years of testing to secure regulatory approval for any such offering.
King expects to publish information about the biomarkers discovered by his team long before that, however. He hopes that caregivers can find ways to start harnessing some of their predictive power in the next few years.
If either the diagnostic tools or therapeutic treatments that emerge from the Medalist study do transform the treatment of diabetes, it would not be the first such triumph for King. That first transformational achievement came more than 20 years ago when he and Aiello hypothesized that a compound cloned by scientists at Genentech was the “Factor X” that significantly contributed to diabetic eye disease.
The VEGF molecule had all the properties that researchers expected in Factor X. It was made in the eye, its expression increased as oxygen levels declined, and it made cells in the retina grow and leak. Starting with this observation, King and Aiello led the research that proved VEGF spurred diabetic eye disease. They subsequently worked with Genentech on the creation of a VEGF inhibitor that eventually translated into a new class of drugs to prevent vision loss from diabetic reti-nopathy that has benefited hundreds of thousands of patients worldwide.3,4
such that it would lower the risk of cardiovascular disease rather than increasing the risk [in] the way current formulations are thought do in many patients,” said King. “We don’t know anywhere near as much as we’d like to know about the connection between diabetes and cardiovascular disease. However, we think we’ve made real strides toward an insulin formulation that can decrease arteriosclerosis in mice. It’s a long-term project, but it’s very exciting.”
Looking back on his career, King says that he has always focused most of his effort on discovering the causes for different diabetic complications. When he started, there was still hope of finding a single underlying cause for all diabetic organ failure (and thus, quite possibly, a single treatment that would protect all organs). He has come to believe (like most of his colleagues) that there is no single underlying cause, that the disease may attack each organ in a different way, and that researchers like him will have to hunt them down one by one. “Everything has proven to be more complicated than expected, which has been frustrating, but we still made real progress in the treatment of patients,” said King. “Patients survive longer and enjoy higher quality of life. VEGF inhibitors alone have already saved more than 200,000 patients from losing their vision. We have also developed a set of very powerful tools that should enable continuous progress toward better treatments.”
King naturally hopes that the targets identified through the Medalist study prove just as useful, but they’re not the only project he’s excited about. “We’re currently working to redesign insulinReferences
1. Sun JK, Keenan HA, Cavallerano JD, et al. Protection from retinopathy and other complications in patients with type 1 diabetes of extreme duration: the Joslin 50-year medalist study. Diabetes Care. 2011;34(4):968-974.
2. Bartlett J. Joslin develops a way to predict diabetes complications. Boston Business Journal website. http://www.bizjournals.com/boston/blog/health-care/2015/07/joslindevelops-a-way-to-predict-diabetes.html. Published July 30, 2015. Accessed November 5, 2015.
3. Aiello LP, Pierce EA, Foley ED, et al. Suppression of retinal neovascularization in vivo by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF-receptor chimeric proteins. Proc Natl Acad Sci USA. 1995;92(23):10457-10461.
4. Caffrey MK. NEJM study: Aflibercept offers benefits over rivals for DME if vision loss is worse at outset. Am J Manag Care. 2015;21(SP11):SP370.