Levine-Riggs Diabetes Research Symposium 2017

March 28-March 31, 2017; Orlando, FL; Full Report – Draft

Executive Highlights

Greetings from the Sunshine State! The annual Levine-Riggs Diabetes Research Symposium – always a whirlwind of incredible learning related to diabetes basic science – kicked-off with the legendary Dr. Robert Sherwin’s take on the brain-glucose connection in diabetes/obesity. Subsequent sessions covered forward genetics as a means of unraveling diabetes pathophysiology (from Nobel Prize winner Dr. Bruce Beutler!), the differential advantages of mono- vs. dual-hormone approaches to closed loop (from Dr. Garry Steil), and the role LncRNAs (long, non-coding RNA strands) might play in the development of diabetes complications and in beta cell function (from City of Hope’s Dr. Rama Natarajan and University of Colorado’s Dr. Lori Sussel). This report also features many highly-respected voices in the field including Drs. Carla Greenbaum, Andrew Hattersley, and so many more.

Table of Contents 

Detailed Discussion and Commentary

Plenary Lecture

The Brain-Glucose Connection in Diabetes and Obesity: A Double-Edged Sword

Robert Sherwin, MD (Yale School of Medicine, New Haven, CT)

The Levine-Riggs Symposium kicked off with a keynote lecture from the esteemed Dr. Robert Sherwin (Yale School of Medicine, New Haven, CT) detailing the complicated interplay between the brain and the beta cell in glucose regulation. As past president of the ADA and the author of over 400 papers on diabetes (including one detailing the first-ever insulin pump study!), Dr. Sherwin is a true giant in diabetes research, not to mention a mentor to nearly every clinician in attendance at the conference. Against the backdrop of a vast and complicated literature on the relationship between the brain and glucose, Dr. Sherwin’s talk took a deep dive into two subjects in particular: (i) the neural basis of hypoglycemia unawareness; and (ii) the neural basis of obesity. Completing the “diabetes puzzle,” Dr. Sherwin noted, will be a matter of further untangling the mechanisms underlying the relationship between the brain and glucose regulation. 

  • Dr. Sherwin opened with a description of the complex neural circuitry governing the brain’s defense against hypoglycemia, and how this goes awry in cases of impaired awareness of hypoglycemia. He began with a physiological description of hypoglycemia unawareness: Under normal conditions, the brain rapidly responds to hypoglycemia when plasma glucose reaches approximately 65-70 mg/dl (before the appearance of symptoms at ~55 mg/dl), launching a counter-regulatory response involving the secretion of epinephrine and glucagon (or in diabetes, just epinephrine) to raise glucose levels. However, after intensive insulin treatment, the deployment of the epinephrine counter-regulatory response occurs much later, after plasma glucose levels have fallen to <50 mg/dl and hypoglycemia symptoms have already set in. Dr. Sherwin’s research suggests that the ventromedial hypothalamus (VMH) of the brain is the control center for the brain’s counter-regulatory response to hypoglycemia and contains two types of glucose-sensing neurons to defend against hypoglycemia: a pre-synaptic GABA-secreting neuron and a post-synaptic glutamate-secreting neuron. Under in-range glucose levels, the GABAergic neuron chronically inhibits the glutamatergic neuron. However, when glucose is low, the secretion of GABA stops, disinhibiting the glutamatergic neuron and initiating an AMP-kinase signaling cascade that culminates in the counter-regulatory release of epinephrine. Dr. Sherwin suggested that stimulating this pathway with pharmacotherapy could provide a potential therapy to circumvent hypoglycemia unawareness. Both GABA blockade and AMP-kinase activation techniques have successfully minimized hypoglycemia unawareness and restored the epinephrine counter-regulatory response to low glucose in mouse studies in Dr. Sherwin’s laboratory; the challenge will be translating these into the clinic. We heard quite a lot about the progress and knowledge gaps toward a pharmacological therapy for hypoglycemia via corrections of impaired awareness or impaired counter-regulatory response at the JDRF/HCT Preventing Hypoglycemia in T1D workshop last week – Dr. Sherwin was a co-chair and highly influential speaker at this meeting as well.
  • Dr. Sherwin went on to detail obesity-associated differences in brain circuitry and neural responsiveness. In fMRI studies, lean individuals show increased activity in brain regions associated with executive function and self-control in response to glucose ingestion (presumably in order to stop eating), whereas individuals with obesity show the opposite pattern – decreased activity in areas associated with executive function in addition to increased activation of areas associated with reward and pleasure. Fascinatingly, this pattern is more exaggerated in people with greater insulin resistance. In accordance with the rise in childhood obesity, these brain activation patterns hold true even in adolescents with obesity, suggesting that, even from a young age, individuals with obesity have an altered neural response to food that makes eating both more pleasurable and more difficult to disengage in. What causes this neural re-wiring in obesity? Dr. Sherwin suggested that the answer may lie in the polyol pathway, a highly conserved metabolic pathway (present even in the fruit fly!) responsible for converting glucose into fructose. Studies from Dr. Sherwin’s lab demonstrate that the brain, when exposed to high levels of glucose, may convert some of it into fructose. Fructose is metabolized much more slowly than glucose, and can generate an inflammatory response if present in large quantities. He hypothesizes that this plays a critical role in changing the brain’s underlying circuitry to respond differently to food in people with obesity.

Epigenetics and Non-coding RNAs

LncRNAs in Diabetic Complications, Pancreatic IncRNAs and Diabetes

Rama Natarajan, PhD (City of Hope, Duarte, CA), Lori Sussel, PhD (University of Colorado, Aurora, CO)

We were excited to learn about an emerging understanding of the role of long non-coding RNAs (lncRNAs) in beta cell function and diabetes complications from City of Hope’s Dr. Rama Natarajan and University of Colorado’s Dr. Lori Sussel. The pair of presentations highlighted growing interest in the effect of non-coding regions of the genome on gene expression – and we’re certainly intrigued as well! Dr. Natarajan discussed the potential role of lncRNAs in promoting diabetes complications and presented exciting early-stage work from her lab on the use of gapmers to counteract this effect. In diabetic nephropathy specifically, Dr. Natarajan’s team has identified key clusters of non-coding microRNAs within lncRNA regions that are upregulated in conditions of high glucose or high levels of TGF-b1 (a known biomarker of inflammation and diabetic kidney disease). Her team found that this upregulation of lncRNA was associated with endoplasmic reticulum stress, fibrotic gene expression and other markers of diabetic nephropathy. On the other hand, gapmer oligonucleotides targeted to specific lncRNA is able to reduce expression of this lncRNA and component microRNAs, inhibit expression of fibrotic genes, and confer renal protection, preventing early features of diabetic nephropathy in a mouse model. In addition, the gapmer targeting the lncRNA was also effective in human renal cells, and the particular microRNA cluster studied in mice by Dr. Natarajan’s lab was also found to be expressed in humans (in Pima Indians, specifically), suggesting the possibility of clinical translation. Immediately after Dr. Natarajan’s presentation, University of Colorado’s Dr. Lori Sussel followed up with an examination of the role of lncRNAs in pancreas development, function, and disease, focusing on the beta cell specifically. She suggested that a better understanding of the role of lncRNAs can help us create better, “true” beta cells in the lab – paving the way for better beta cell replacement therapies. Each year as we return to Levine-Riggs, it seems that the biological underpinnings of diabetes become more and more complex, from genetics to epigenetics to, now, lncRNAs. We’re so glad for the tireless work on this front from the basic science researchers in the field and we’re hopeful that these findings can inform truly disease-modifying – or even preventive or curative – therapies in the future.

Use of 1 vs. 2 Hormone Infusions in the Artificial Pancreas

Mono-Hormone Control

Garry Steil, PhD (Harvard Medical School, Boston, MA)

In Dr. Ed Damiano’s absence (due to an airport delay), Dr. Garry Steil advocated for “equipoise” between single- and dual-hormone closed loop, arguing that we should be funding both approaches, and should avoid the bias that a dual-hormone (insulin + glucagon) system will be inherently better than an insulin-only one. In making his case, Dr. Steil highlighted affordability (glucagon is an added expense) and limited clinical trial evidence demonstrating the need for glucagon in closed loop. He pointed out that the FDA requested a study comparing the bihormonal bionic pancreas (from the Beta Bionics team) vs. an insulin-only system, but also pointed to a greater challenge: No randomized controlled trial can test for a negative, meaning there is no statistical test that could definitively establish glucagon as “essential.” When you couple this with the fact that glucagon is expensive, an insulin-only system becomes a highly-attractive option, especially from the vantage point of payers. On the other hand, at Diabetes UK, MGH’s Dr. Steven Russell positioned glucagon as the only way to fully automate glycemic control while also minimizing hypoglycemia risk. Moreover, patients who require fewer hypoglycemia treatments and achieve better glycemic control could yield substantial cost-savings over the long term – unfortunately this would be very difficult to prove, as long-term clinical trials, by definition, take a long time. In the meantime, payers would likely be more inclined to pay for single hormone closed loop (or even MDI + titration software), and the development of faster insulins and better algorithms could diminish the marginal benefit of glucagon. Importantly, Dr. Steil expressed support for the dual-hormone approach, but suggested that the promise of a bihormonal artificial pancreas shouldn’t blind us from the potential of insulin-only closed loop – single-hormone systems are getting better and safer, he explained, and should be pursued with equal fervor. Dr. Steil made no mention of amylin as a potential second hormone – Dr. Russell highlighted “promise” for the use of pramlintide as a co-secreted hormone from a single cartridge at the NIH AP Workshop in July. As a reminder, a pivotal trial of Beta Bionics’ insulin-only system is scheduled to begin by end of 2017 or early 2018, while a pivotal trial of the bihormonal bionic pancreas is slated to start mid-2018.

Questions and Answers

Q: One of the assumptions with closed loop is that patients want to skip announcing their meals. What is your impression of how important this is to people?

A: It is true that some patients and clinicians like the idea of some advanced meal bolus – they want to feel a little bit of control in advance of the meal. So we’ve played around with patient involvement. What patients can’t do well – what many nutritionists can’t even do well – is accurately count carbs off a plate. To rely on the patient to count carbs just to get a meal bolus is unrealistic, so maybe we get patients involved in other ways, with a pre-prandial meal bolus or something else. That’s still an open question.

Comment: It’s true that someone without diabetes can go into a blood sugar range around 60 mg/dl and that’s normal, but I don’t know that we want to be doing that routinely in people with diabetes, who may already have some hypoglycemia-sensing issues.

A: With hypoglycemia unawareness, is it still okay to go down <60 mg/dl? There’s certainly a fear factor here – parents of children with type 1, in particular, can be very nervous if their child is in the 60 mg/dl range. But I’d argue that if insulin has been off for a long time and your blood glucose isn’t falling, you’re okay. You don’t see a lot of seizures or other adverse effects of hypoglycemia happening in response to short bouts of glucose in the 60s. Data from Bruce Buckingham shows that hypoglycemia-induced seizures don’t happen until you’re down in the 40s for extended periods. Plus, a patient who is hypoglycemia unaware probably isn’t feeling a sense of befuddlement at 60 mg/dl. Everyone’s different, and we of course need to give patients what they want, so if they fear 70 mg/dl we should keep them away from that. We need to tailor therapy to keep people in control.

Q: Have you noticed any issues with nighttime sensor reliability? That’s been a problem in the past.

A: People are now using sensors to manage open loop therapy. I think we’re there with sensor technology.

Q: When do you expect a fully-integrated system to be available in ambulatory application?

A: I have to give Dr. Ed Damiano credit here – he and I both believe you can get to a fully automated system that doesn’t require patient involvement. Medtronic’s MiniMed 670G (hybrid closed loop) has been FDA-approved and it’s a step along the way toward closed loop, but you’ll still need to count carbs and do everything you’re doing today in diabetes management. What Ed and I are both saying is, let’s forget what products we already have and start from scratch to create the best closed loop solution. I’m not sure these systems are going to come out within the foreseeable future. I think Medtronic will keep forging ahead with incremental changes, but they’ll be small and only occur every three to five years, at best.

Presentation of Awards and Dinner Lecture

Innate Immunity in Health and Disease

Bruce Beutler, MD (UT Southwestern Medical Center, Dallas, TX)

2011 Nobel Prize in Physiology or Medicine winner Dr. Bruce Beutler (UT Southwestern Medical Center, Dallas, TX) presented a fascinating overview of the progress made in the last 17 years in forward genetics (the process of identifying genes responsible for a particular phenotype) and explored how this approach can be applied toward a better understanding of diabetes pathophysiology. He noted that, throughout the first decade of the century, tremendous progress was made in mouse genome sequencing, but the process of forward genetics continued to be limited by the need for cumbersome genetic mapping. However, Dr. Beutler’s lab has developed a process and database for high speed “positional cloning” that allows his team to identify which gene mutations are likely responsible for an observed phenotype almost instantaneously. His lab performs whole-exome sequencing on mutagenized mice to determine all possible mutations these mice may be able to pass down to their descendants – notably, his lab found that the mutagenesis process creates on average 60 mutations in the exome, far more than the 4-5 mutations previously thought. The descendants of the sequenced mice are genotyped and are also put through an 89-assay phenotype screening that includes a glucose tolerance test, several behavioral screens, tests for various biomarkers of diseases, and even an autism screen. All of these data are stored in a computer database, which determines positive associations between certain mutations and specific phenotypes. As more mice go through this process and more data are generated, researchers are able to generate an ever-clearer picture of the likely phenotypic effects of different mutations.

  • Dr. Beutler demonstrated how this tool can help inform our mechanistic understanding of diabetes. He highlighted a particular mouse phenotype observed in his lab, “teeny (tny).” Homozygotes were about half the size of normal mice, had severe hyperglycemia and insulin resistance, and exhibited a very fatty liver despite an overall lean body type. Using the computer database, the team was able to map this phenotype to a mutation in the Kbtbd2 gene. From there, a series of mechanistic studies were able to demonstrate Kbtbd2 is responsible for degradation of a specific protein, p85a, in adipose tissue – the team also found that this degradation is key to maintaining insulin sensitivity. When the Kbtbd2 gene is mutated, p85a is able to accumulate and insulin sensitivity is lost, resulting in severe diabetes. Dr. Beutler summed up that this example illustrates how forward genetics can inform our understanding of disease and, with these new advances, the time it takes to find mutations associated with specific phenotypes has greatly decreased from five years to virtually instantaneous. The next step, in his view, will be genetic screening for disease suppression. Rather than try to induce disease-causing mutations in wildtype mice, a disease suppression approach would take a complex mouse model of human disease, in which multiple genes are implicated (he listed obesity, NAFLD, and Alzheimer’s as possible examples), and subject them to the mutagenesis, sequencing, and phenotyping process in order to identify mutations that correct the disease. He suggested that this could translate into almost instant treatment targets for these disease. We’re very, very intrigued by this prospect and are eager to learn more – in particular, how far are we as a field from use of these techniques in diabetes and obesity drug discovery?

Diabetes Complications

Mitochondrial Dysfunction in Diabetic Complications

Farhad Danesh, MD (University of Texas MD Anderson Cancer Center, Houston, TX)

Dr. Farhad Danesh (University of Texas MD Anderson Cancer Center, Houston, TX) discussed fascinating evidence suggesting that diabetic nephropathy may be a disease of mitochondrial dysfunction. He began with a description of the well-known “Warburg effect” – the observation that cancer cells tend to have excessive mitochondrial activity even if energy is plentiful. Less commonly known is the fact that this is also true of the mitochondrial activity in the podocytes of the kidney, suggesting that diabetic nephropathy, like cancer, may be due in part to abnormal mitochondrial activity. Dr. Danesh reviewed a large body of work untangling the mechanism underlying this connection between the mitochondria and diabetic nephropathy, concluding that a key player in this process is the Rho kinase (ROCK1) protein. Evidence for this comes from two complementary experiments from Dr. Danesh’s laboratory. In the first, scientists used genetic engineering techniques to knock out the Rock1 gene in db/db mice, an established model of type 2 diabetes. Db/db mice typically have high urinary albumin levels, but those lacking the Rock1 gene exhibited normal albumin levels identical to those of wild type control mice. The second experiment took the opposite approach, this time using genetic engineering techniques to overexpress the Rock1 gene in otherwise normal mice. This intervention produced excessive urinary albumin levels; notably the animals were otherwise free of diabetes and diabetic milieu, suggesting that this Rock1 mechanism is selectively involved in the pathophysiology of diabetic nephropathy only and none of the other complications of diabetes, microvascular or otherwise. How exactly does ROCK1 influence the mitochondria of the kidney cells? The story is not entirely clear, but using imaging techniques, Dr. Danesh’s group discovered that high levels of ROCK1 protein spur extensive fragmentation and cell division of the mitochondria through the ROCK1 protein’s activation of Drp-1, the protein responsible for mitochondrial fission. Exposure to high glucose produces the same effect. Knitting together these observations, Dr. Danesh’s current working hypothesis is that hyperglycemia leads to high ROCK1 activity, thus inciting Drp-1 to induce mitochondria division. Now multiplied, the mitochondria are excessively active, leading – somehow – to diabetic nephropathy. His laboratory continues to investigate this intriguing hypothesis, and we’re exciting the follow the progress of this intriguing hypothesis!

  • Dr. Danesh closed by pointing out that the standard clinical approach to managing diabetic nephropathy has not changed substantially since his days in medical school. His work, though early stage, could very well point toward new therapeutic targets for diabetic nephropathy involving ROCK1 inhibition. Indeed, huge unmet need surrounds diabetic nephropathy, and renal outcomes are becoming an increasing area of interest in recent CVOTs such as the SUSTAIN-6 and LEADER trials for the GLP-1 agonists semaglutide and liraglutide (see our coverage of this from EASD 2016, where renal outcomes were a defining theme). Dr. Danesh’s body of work also raises the intriguing possibility that a mitochondrial mechanism could be underlying the beneficial renal effects of these drugs. This is certainly an area for future exploration.

Mechanisms and Drug Targets for Protection of Ischemic Diabetic Hearts

Ravichandran Ramasamy, PhD (NYU Langone Medical Center, New York, NY)

Compared to a background population, people with diabetes not only face heightened risk for CV events, but a worse prognosis following these events. Why do diabetes-afflicted hearts go into poor recovery after MI? This was the central question of Dr. Ravichandran Ramasamy’s talk, which explored reasons for suboptimal functional and metabolic recovery post-MI in patients with diabetes, and then suggested therapeutic strategies to overcome these factors. For one, the size of the infarct in people with diabetes tends to be larger, meaning these patients lose a greater amount of myocardial tissue in an MI than people without diabetes, on average. Extent of myocardial tissue lost has been correlated with prognosis. Other common features of the diabetes-afflicted heart, as outlined by Dr. Ramasamy, include enhanced signaling for cell death (or apoptosis), contractile dysfunction in cardiac muscle, impaired energy metabolism, maladaptive remodeling of tissue that can lead to fibrosis, and elevated expression of RAGE (receptor for advanced glycation endproducts). Dr. Ramasamy explained how RAGE interacts with the FH1 domain on the Diaph1 molecule, and how knocking out genes for each of these proteins independently in rodent models of diabetes results in enhanced functional and metabolic recovery post-MI. In other words, the RAGE-Diaph1 interplay could be a viable (and novel) target for cardioprotective therapies, or for drugs that aim to improve prognosis for patients experiencing an MI. This research is early-stage, but we’re cautiously intrigued by this pathway and the prospect of a new treatment for CV complications of diabetes. Dr. Ramasamy’s group has already investigated 13 compounds to run interference between RAGE and Diaph1 – four of these demonstrated cardioprotective potential, and the research team will now refine these compounds for further experimentation.

Diabetes Impairs Atherosclerosis Regression: Mechanisms and Clinical Implications

Ed Fisher, MD (NYU School of Medicine, New York, NY)

Dr. Ed Fisher took the stage next and continued the conversation on CV outcomes in diabetes. Through a summary of data from rodent studies, he established that hyperglycemia limits the benefits of lipid-lowering therapy on atherosclerosis: Treatment with statins should cause plaque regression in the circulatory system, but this effect disappears in diabetes, meaning that atherosclerosis remains despite decrease in  LDL levels. This persistence of atherosclerosis could explain the much higher than average CV event rate for a diabetes patient population, but fortunately, the story does not stop here. Dr. Fisher discussed two possible therapeutic approaches to address atherosclerosis and hopefully reduce CV risks for people with diabetes. (i) First, he pointed to monocytosis, or an increase in circulating white blood cells, which causes a continued influx of monocytes into plaques and sustains an inflammatory state in the body. He identified monocytosis as an underlying cause for impaired atherosclerosis regression. It follows that correcting monocytosis could restore plaque regression in the presence of statins or other lipid-lowering medicines (which is the body’s natural response to statins in the absence of high blood glucose). (ii) Dr. Fisher then turned to another key characteristic of comorbid diabetes/dyslipidemia – low levels of HDL cholesterol. In mice, raising HDL concentration has successfully restored the plaque regression response to statin treatment, which Dr. Fisher suggested could have implications for both type 1 and type 2 diabetes in humans. That said, phase 3 CV outcomes trials of CETP inhibitorsthat greatly increase HDL levels like Lilly’s evacetrapib have been incredibly disappointing, which further highlights the challenge of translating treatments in mouse models to clinical therapies. He referred back to Dr. Ramasamy’s talk (which immediately preceded his), echoing a positive outlook on anti-RAGE therapies and how they might show cardioprotection for atherosclerotic CV events like MI and stroke. We loved this journey from a microscopic view of lipids and diabetes to the clinical implications of this basic science research. Treatments targeting monocytosis and RAGE are likely still a long way from real-world patients, but we’re glad to see such rigorous research happening in the lab to identify novel solutions for diabetes complications, and especially for CV complications which are the leading cause of mortality among people with diabetes.

Insulin Action and Secretion

Diabetes Impairs Exercise Capacity: Mitochondria, Blood Flow or Both?

Jane Reusch, MD (University of Colorado, Denver, CO)

The great Dr. Jane Reusch reviewed a working model of how type 2 diabetes makes exercise more difficult and discussed how GLP-1 agonists could intervene to promote exercise in type 2 patients. Together, cardiac dysfunction, muscle dysfunction, and vascular dysfunction culminate in decreased functional exercise capacity. Cardiac dysfunction introduces abnormal blood pressure, while muscle dysfunction affects mitochondria and energy reserves. All three categories of dysfunction impact blood perfusion, simultaneously decreasing blood flow and limiting oxygen extraction into muscles – it’s easy to see how this diabetes double whammy would restrict physical activity. Luckily, there are therapeutic targets within this working model that could break the cycle and sustain functional exercise capacity. Dr. Reusch presented data demonstrating that DPP-4 inhibitor saxagliptin (AZ’s Onglyza) nearly doubles endurance and running capacity in diabetes-induced mice. Saxagliptin’s critical mechanism of action here is increasing GLP-1 (by preventing degradation), so Dr. Reusch next investigated the differential effects of exenatide (AZ’s GLP-1 agonist Bydureon) and exendin 9-39 (a GLP-1 antagonist) on exercise. In line with the saxagliptin results, exenatide treatment extended the distance a mouse could run while exendin 9-39 decreased running distance. Exendin 9-39 also interfered with exercise training, so Dr. Reusch suggested that GLP-1 may play a role in an adaptive exercise training response. She presented data demonstrating TZD rosiglitazone’s positive effect on physical fitness as well, which she attributed to the drug being an insulin sensitizer, but this response was accompanied by a mean five lb weight gain (and given the CV harms associated with rosiglitazone, we doubt it’ll be prescribed no matter the benefits to endurance). We wonder if pioglitazone, an alternative TZD that has actually shown CV benefit at low doses, could similarly boost functional exercise capacity – presumably yes but this would have to be tested. Suffice it to say, our curiosity is piqued about how different diabetes drug classes affect functional exercise capacity. Dr. Reusch drew attention to an important facet in diabetes care – that physical fitness is independently correlated with CV and all-cause mortality, and that HCPs should consider ways to maintain or expand functional exercise capacity in their patients with type 2 diabetes. Of course, drugs like GLP-1 agonists that can pack a triple punch of glucose-lowering, weight loss, and increased exercise capacity, and independently confer cardioprotection on top of that (as demonstrated in LEADER and SUSTAIN 6), could be very attractive therapeutic options indeed.

Vascular and Cellular Mechanisms Responsible for INcreased Insulin Sensitivity after Exercise

Erik A. Richter, MD (University of Copenhagen, Copenhagen, Denmark)

All the way from Copenhagen, Dr. Erik Richter provided a comprehensive overview of how exercise improves insulin sensitivity, not only during physical activity but for several hours thereafter. In a study recently published online in Diabetes, young males were instructed to exercise one leg while keeping the other at rest. A subsequent insulin clamp showed that glucose uptake was 50% higher in the active leg vs. the rested one (p<0.05), and notably, this statistically significant difference lasted into hours three and four post-exercise. Dr. Richter explained how this glucose response might be mediated by microvascular perfusion, which rose 65% in the exercised leg and only 25% in the rested leg (p<0.05) – greater blood flow and priming of the tissue for glucose extraction could promote insulin sensitivity in the muscles, which has major implications given that a majority of insulin resistance in type 2 diabetes resides in skeletal muscle. In Dr. Richter’s words, “if you can rectify skeletal muscle insulin resistance via exercise, you can do something quite clinically important for these individuals.”

  • Dr. Richter described exercise as “a powerful signaling event in the muscle,” emphasizing that even a single bout of exercise can have profound effects: 10 minutes of physical activity changed 1,004 of 8,511 phosphocites in a human muscle biopsy, and 90% of these changed phosphocites previously seemed entirely unrelated to exercise. He underscored that physical activity cannot be ignored in approaches to diabetes care – we absolutely agree and look forward to seeing future ways in which to address this.

Immunity and Measures of Immunity in T1D

The Pathology of Human Type 1 Diabetes - Unique Features and Similarities to Other Autoimmune Diseases

Mark Atkinson, PhD (University of Florida, Gainesville, FL)

In describing our understanding of how type 1 diabetes develops, University of Florida’s Dr. Mark Atkinson claimed that “too many dogmas are dogs,” and that certain aspects to the traditional model of type 1 pathophysiology need revision. For decades, scientists were extrapolating learnings on insulitis from mouse models to humans, when in reality, there are stark differences in how insulitis presents in mice vs. humans. Where insulitis occurs on a sliding scale of sorts in rodents, Dr. Atkinson explained that it’s more all-or-nothing in humans – insulitis is either present or absent. He noted that the decades long practice of researchers using therapies that target rodent insulitis as an indicator or success (where 100’s of drugs prove effective) may be one of the primary reasons translation of such results to humans has been disappointing.  Similarly, the predominant notion for many years was that type 1 diabetes arises solely due to autoimmunity. Rather, Dr. Atkinson pointed to a wide variety of emerging data suggesting a key role for stressed beta cells, a state that can be induced by viral infections, hypoxia, high glucose, long-chain fatty acids, to name a few. He also presented data revealing high variance in number of beta cells among adult populations without diabetes, suggesting that beta cell number in and of itself might be a risk factor for type 1 diabetes development. Ultimately, Dr. Atkinson emphasized that the field is in a state of transition, where better scientific tools can increasingly power nuanced insights into type 1 diabetes progression. One of the biggest looming questions, which he urged researchers to approach without too many preconceptions, is “the chicken or the egg” problem: Which comes first, beta cell abnormalities that elicit an autoimmune response, or an autoimmune attack that stresses beta cells and forces abnormalities? We very much appreciated the call for open-mindedness, since in our view, it seems that every year there are more new questions than answers on the contributing causes to type 1 diabetes. Importantly, this isn’t to say we haven’t noticed remarkable advances in the field – only that we recognize the complexity of type 1 progression and the likelihood of unexpected findings. We’ll be keenly following further advances to better understand how this disease develops, and how it might be decelerated or even stopped. For more on the incredible insights Dr. Atkinson and his colleagues are building through the JDRF Network for the Pancreatic Organ donors with Diabetes (nPOD), see our coverage of the 2016 nPOD meeting as well as Dr. Atkinson’s presentation at IDF 2015.

  • Dr. Atkinson also highlighted untapped insights from autoimmune diseases that parallel type 1 diabetes. While access to human pancreatic tissue is limited, researchers can more feasibly examine hair follicles (affected by alopecia) or even the small intestine (affected by celiac disease), and can apply some of the same principles to autoimmune attack on the beta cells. Interestingly, Dr. Atkinson mentioned that none of these autoimmune diseases show a sex bias, presenting with very similar frequency in males and females.

Type 1 Diabetes Autoimmunity and the Islets

Systems Immunology Approaches to the Progression and Treatment of Type 1 Diabetes

Matthew Dufort, PhD (Benaroya Research Institute, Seattle, WA)

Dr. Matthew Dufort (Benaroya Research Institute, Seattle, WA) presented a fascinating overview of what he terms his “systems immunology approach” to type 1 diabetes, identifying what distinguishes responders from non-responders to certain experimental type 1 diabetes therapies. In particular, he reviewed his laboratory’s most recent study investigating clinical response to teplizumab, an anti-CD3 agent. Interestingly, the subset of patients with the best response to teplizumab (i.e. the greatest degree of C-peptide stabilization) showed an accumulation of a distinct population of CD8 T cells expressing high levels of the transcription factor EOMES and the inhibitory receptors TIGIT and KLRG1 – a phenotype normally associated with immune exhaustion. The rise of this “exhausted” cell type following teplizumab therapy suggests that some patients with type 1 diabetes may benefit from the dampening of T-cell activity, a counterintuitive result since the traditional paradigm in other immunotherapies is to instead awaken exhausted T-cells. We were certainly intrigued by these provocative findings, which contribute to an emerging body of literature to support that the pathophysiology of type 1 diabetes may be far more complicated than just a beta cell-targeting autoimmune response. Of course, it is important to note that Dr. Dufort’s detection of exhausted T-cells was only relevant to a subset of teplizumab-treated patients (a roughly equal number experienced virtually no response to this intervention.) This underscores the importance of individualized therapy for a disease as heterogeneous as type 1 diabetes. Determining not only which interventions are effective but who they will be effective for will be a defining challenge in the future of type 1 diabetes care – an especially critical issue given the importance of intervening as swiftly and early as possible to preserve residual beta cell mass in T1D. Dr. Dufort’s work also provided exciting data demonstrating important differences between T1D in children as compared with adults.   While clinical studies from TrialNet and others have emphasized differences in the rate of disease progression, Dr. Dufort highlighted that clinical results and immune markers associated with response to Rituximab therapy were in children. This type of accumulating evidence is essential for industry and academic investigators in discussion with regulators regarding conducting clinical trials in children with T1D.

Insights Into Type 1 Diabetes Heterogeneity from Doing Monogenic Research

Andrew Hattersley, MD (University of Exeter Medical School, Exeter, UK)

This year’s Rachmiel Levine Award winner Dr. Andrew Hattersley acknowledged publicly for the first time that he may have been wrong with his oft-repeated dogma that diabetes diagnosed before six months always neonatal monogenic diabetes and is never type 1 diabetes. Dr. Hattersley is famed for his work in monogenic diabetes (winning him the EASD/Novo Nordisk Foundation prize at EASD 2016 and the Albert Renold award at EASD 2015), but his Rachmiel Levine Award Lecture focused on the insights he inadvertently gained into type 1 diabetes through his decades of work. Dr. Hattersley described a variety of screening strategies and tools to differentiate between monogenic, type 1, and type 2 diabetes that have been developed, validated, and employed in clinical practice. Methods to differentiate between monogenic and type 1 diabetes include the mail-in urine C peptide analysis and universal antibody screening in pediatric diabetes patients to select people for genetic screening of the genes known to cause monogenic diabetes. Based on large-scale screening studies conducted with each of these methods, Dr. Hattersley’s team gradually built an understanding of type 1 diabetes. For instance, Dr. Hattersley’s team found that age of diagnosis can predict long term C peptide levels with those diagnosed under 10 year mostly having un recordable (<3pmol) C peptide and many of those diagnosed after 10 years having levels greater than 100pmol/l, this goes with the different immune findings in the pancreas associated with age of diagnosis.  The team also developed a type 1 diabetes genetic risk score (GRS) based on known genes associated with type 1 diabetes, which is good at determining type 1 vs. type 2 diabetes in a population of people with diabetes much better than genetics can predict type 1 in the general population without diabetes.  Based on all of this accumulated knowledge, his team have produced a Diabetes Diagnostics app (available for iOS and Android) that combines information on BMI and age at diagnosis to help determine whether an individual with diabetes has monogenic diabetes. This will soon be modified to include presence of autoantibodies, and GRS to help classification of Type 1 and Type 2 diabetes. Coming full circle, Dr. Hattersley team decided to apply the type 1 diabetes GRS to babies diagnosed with diabetes at six months of age or younger who did not screen positive for any of the 24 known monogenic causes of diabetes – this applied to about 20% of those diagnosed before six months. The pooled results of the GRS suggest that a portion of these babies have a yet-unidentified form on monogenic diabetes – but, surprisingly, a substantial portion very likely had type 1 diabetes instead. Even more unexpectedly, these babies exhibited low birth weight, suggesting that their type 1 diabetes had begun to develop in utero. Prior to these findings, Dr. Hattersley had frequently, unequivocally lectured that all diabetes diagnosed at this young age is monogenic and not type 1 diabetes – he attributed this easy public health message with a clear age cutoff to the increase in knowledge of and screening for monogenic diabetes in neonates globally. Thus, he somewhat facetiously expressed a little nervousness about changing his message now – nonetheless, Dr. Hattersley emphasized that this very young onset type 1 diabetes is extremely rare, occurring in <8% of those diagnosed at under six months or less than one child per 6,000 total children under 18 with diabetes. Overall, we were absolutely fascinated by the insights Dr. Hattersley was able to glean about type 1 diabetes from his work and it’s clear that we need much more research to better understand type 1 diabetes in this very young population.

Obesity and Metabolism

Mitch Lazar, MD, PhD (University of Pennsylvania Perelman School of Medicine, Philadelphia, PA)

Dr. Mitch Lazar (University of Pennsylvania, Philadelphia, PA) described his laboratory’s fascinating work on the circadian biology, which has surprisingly led to a potential new drug target for NASH. Dr. Lazar’s laboratory has described two complementary pathways that regulate liver metabolism: a so-called “reactive” pathway by which insulin activates hepatic lipid metabolism in response to food intake (mediated by the SCAP and SREBP proteins), and a so-called “anticipatory” pathway by which the circadian clock dampens hepatic lipid metabolism during the night when less energy is required (mediated by the by Rev-erbα/HDAC3 protein complex). These two pathways together maintain healthy levels of fat in the liver. Evidence for this comes from findings that mice with genetic knockout of the HDAC3 protein (thus preventing the formation of the Rev-erbα/HDAC3 complex) develop NASH, and those in which the SCAP protein is additionally eliminated exhibit an extremely severe form of NASH that is lethal within a few weeks.  Remarkably, these mice can be rescued from NASH with genetic engineering techniques that cause the selective expression of SREBP1c protein in liver cells. This is surprising since SREBP1c is involved in lipogenesis and triglyceride synthesis – processes that intuitively would appear to exacerbate liver fat level rather than normalizing it. Future studies continue to dive deeper into these counterintuitive results, and we are eager to learn how the story develops. If the surprisingly beneficial effects of SREBP1c persist in future investigations, we can’t help but wonder whether this may become a new target for NASH drug development. NASH is a therapeutic area of severe unmet need, and the drug competitive landscape features a number of promising candidates – though none of these have a basis in circadian biology like Dr. Lazar’s work. This is certainly a very different take on NASH. For more information on these findings, check out the Lazar laboratory’s recent publication, out just a few months ago in Cell Metabolism. Many more learnings about NASH are to come – we will be headed to the first-ever NASH Summit in just a few days!

Obesity and Metabolism

Perturbing the Early Life Microbiome and Its Consequences

Martin Blaser, MD (University of Langone Medical Center, New York, NY)

Levine-Riggs kicked off with a lively morning symposium on the metabolic dynamics underlying obesity. Perhaps unsurprisingly, the discussion focused largely on the hottest topic in obesity basic science: the microbiome. Adding to the microbiome’s sheer complexity, this symposium left us with plenty more to think about – including the relationship between the microbiome and everything from antibiotics to circadian rhythms. Esteemed and very compelling microbiology expert Dr. Martin Blaser (New York University, New York, NY) argued that population-wide shifts in the microbiome may be underlying the global rise of obesity. In particular, he pointed to antibiotics as a ubiquitously present environmental exposure that has major consequences for the microbiome (which is, after all, composed of bacteria). Consistent with observational studies in the literature reporting an association between antibiotic exposure and type 2 diabetes incidence (Mikkelsen et al., 2015), Dr. Blaser’s team has reported that antibiotic exposure in mice increases adiposity and alters the regulation of hepatic lipogenesis genes, an effect that is exacerbated when mice are exposed to a high-fat diet in addition to antibiotics.


Amir Zarrinpar, PhD (University of California, San Diego, CA)

Dr. Amir Zarrinpar – a veritable rising star in the microbiome world – provided an overview of his laboratory’s research on the circadian dynamics of the gut microbiome and its implications for obesity and metabolism. Dr. Zarrinpar opened with a surprising finding from the time he was a post-doc in Stchin Panda’s laboratory: simply restricting the time interval during which food consumption occurred could prevent obesity from emerging in mice fed a high-fat diet. The researchers exposed mice to either normal or high-fat chow and let them eat either ad libitum or in a time-restricted manner, where food was only available for certain hours of the day. Regardless of diet or eating pattern, all animals consumed the same number of calories per day. However, those on an ad libitum high-fat diet developed a phenotype resembling obesity and type 2 diabetes – they weighted over 30% more than the other mice, and exhibited high blood glucose levels and an elevated insulin response to glucose. Simply switching mice from ad libitum to time-restricted feeding reversed this. According to Dr. Zarrinpar, these findings support the benefit of a “high amplitude” feeding pattern for metabolism. That is, metabolic rhythms respond best to defined periods of fasting and feeding (as reflected in the time-restricted feeding condition) rather than constant feeding (as reflected in the ad libitum condition). What is the mechanism underlying this phenomenon? Increasing evidence suggests that the microbiome may play an important role. In a follow-up study, Dr. Zarrinpar and his colleagues discovered that the composition of the gut microbiome exhibits daily patterns of fluctuation which are disrupted in mice with diet-induced obesity and can be partially restored after time-restricted feeding. Preliminary evidence suggests that these microbiome fluctuations ensure the appropriate expression of liver genes such as FXR and FGF-15 to mediate bile acid secretion. For more details on these findings, check out Dr. Zarrinpar’s recent Cell Metabolism publication

The Molecular Identity of the Beta Cell and Islet Cell Subtypes

Flattop and Beta Cell Subtypes

Heiko Lickert, PhD (Helmholtz Zentrum Munich, Munich, Germany)

The Institute of Diabetes Regeneration Research’s (Munich) Dr. Heiko Lickert reviewed his recent Nature paper, explaining the contrast between proliferative and mature beta cells in the islet. The chief finding was that Flattop, a Wnt/planar cell polarity effector gene, subdivides the two populations, expressing highly in mature, vessel-proximal beta cells, while Flattop negative cells show increased proliferative capacity and are more frequently found at the periphery. The study also found that the formation of 3-D islet matrices upregulates Flattop expression, leading to increased beta cell maturation. Meanwhile, in a 2-D matrix, less mature beta cells dominate the structure. As all of this work was done in vitro with animal islets, the Dr. Lickert’s lab is in the process of generating insulin/Flattop reporters in cultured human cells. In addition, just last week, the lab implanted Flattop reporter knock-in embryos into a pig for the first time. Flattop illuminates yet another example of beta cell heterogeneity and plasticity, and may offer a therapeutic target for the regeneration of specific subpopulations.

Mutant Ins-Gene Induced Diabetes of Youth (MIDY)

Peter Arvan, MD (University of Michigan, Ann Arbor, MI)

Michigan’s Dr. Peter Arvan presented three potential therapeutic approaches to the proinsulin misfolding that occurs in MIDY (mutant INS-gene induced diabetes of youth). Excitingly, he also suggested that these therapeutic targets might apply in type 2 diabetes, where an accumulation of misfolded proinsulin is thought to contribute to insulin deficiency over time. Dr. Arvan characterized the power of misfolded proinsulin in their ability to “attack innocent bystanders,” blocking non-mutated proinsulin molecules from the secretory pathway that leads to functioning insulin and glycemic control. Interventions for MIDY, and perhaps for type 2 diabetes, should therefore focus on clearing misfolded proinsulin as well as preserving correctly-folded proinsulin and pushing it along to become insulin – this was the overarching thesis of Dr. Arvan’s talk. He dove deeper into a few specific treatment strategies based in this line of thinking, though it’s important to keep in mind that these are all in very early stages of research still:

  • (i) ERO1 agonists could accelerate proper folding of unaffected proinsulin molecules, making them more resistant to attack from misfolded mutants. ERO1 is an enzyme that facilitates disulfide bond formation – three of these bonds are involved in the folding process for proinsulin as it exits the endoplasmic reticulum (ER) of a beta cell and proceeds toward the insulin secretory pathway. Dr. Arvan explained that we wouldn’t expect an ERO1 agonist to have any impact on mutant proinsulin. Rather, this therapeutic agent would work by accelerating bond formation/folding of non-mutated proinsulin, ideally to a level where proper folding outpaces the “attack on innocent bystanders.”
  • (ii) A therapy might focus on ER-associated protein degradation, helping beta cells more quickly identify and trash misfolded proinsulin so that it doesn’t “attack the innocent bystanders.” This strategy takes an opposite but complementary approach to an ERO1 agonist, with a primary aim to clear misfolded proinsulin as fast as possible. Dr. Arvan reminded us that all beta cells – “yours and mine” – make a small amount of misfolded proinsulin, but at a slow enough rate that it doesn’t accumulate. Thus, there’s hope that an intervention could speed up clearance so that it’s no longer outpaced by misfolded proinsulin accumulation in diabetes.
  • (iii) In mouse models, the Grp170 protein has shown a diminishing effect on misfolded proinsulin. Dr. Arvan presented data to demonstrate how Grp170 associates with and lowers levels of mutated proinsulin. The protein’s mechanism involves direct recognition of the misfolded molecules. In Grp170 knockout mice, misfolded proinsulin accumulates readily.

Advances and Obstacles in Beta Cell Proliferation and Regeneration

Novel Drugs for Human Beta Cell Proliferation

Justin Annes, MD, PhD (Stanford University School of Medicine, Stanford, CA)

Dr. Justin Annes (Stanford University, Stanford, CA) gave an intriguing overview of his cutting-edge work in type 1 diabetes drug development. His goal? Finding molecules with the potential to restore pancreatic function by promoting residual beta cell proliferation. He hypothesized that some existing, already FDA-approved medications that increase cAMP levels may have unacknowledged potential to stimulate beta cell growth and function. A screening project in his laboratory revealed the antidepressant mirtazapine and the blood thinner dipyridamole as potential candidates: in preclinical in vitro studies, human islet insulin secretion was boosted both by mirtazapine in combination with norepinephrine and by dipyridamole in combination with exendin-4. Of course this work is very early stage, but the use of existing drugs as a component of a novel type 1 diabetes therapy offers huge advantages for savings in terms of both development time and costs. Such drugs are already proven to be safe, thus reducing the number of trials that would need to be conducted (and funded) until the drug could reach the market – while traditional manufacturers may have lower motivation to pursue further clinical development of a generic drug, the lower development costs could make it possible for a government, academic, or nonprofit institution to drive the program forward. Furthermore, the generic status of these drugs suggests that they would be priced affordably – a win for patients. We will be keeping close track of how this exciting work progresses.

  • In the second half of his presentation, Dr. Annes went on to describe another approach his laboratory is using in pursuit of a beta cell proliferating drug: computational drug design. Using Novartis’ now-discontinued GNF4877 (previously a candidate drug for leukemia) as a starting point because of its action on the diabetes-implicated Dyrk1A pathway, the researchers are using computational models to predict what modifications could be made to the structure of this molecule to increase its action on Dyrk1A to make it a more potent diabetes drug. The next step will be synthesizing these derivatives of GNF4877 in order to test which structural modifications, if any, improve efficacy to the extent that the candidate could become a dedicated glucose-lowering agent. Another challenge, Dr. Annes pointed out, is determining a way to ensure that these drugs are delivered selectively to the pancreas – this is certainly no small feat. We left Dr. Annes’ talk with an appreciation for the massive scientific challenges required to develop pharmacotherapy for type 1 diabetes, and also a sense of enormous inspiration by the innovation and creativity we’ve seen so far surrounding pursuit. 

Lessons on Beta Cell Proliferation from Human Insulinomas

Andrew Stewart, MD (Mount Sinai School of Medicine, New York, NY)

Dr. Andrew Stewart (Director, Diabetes & Metabolism Institute, Icahn School of Medicine) explained how a small molecule, harmine, is able to induce replication in human beta cells. His lab zeroed in on the compound by testing 100,000 compounds on a “mitogenic sensor,” a cMyc promoter. Myc is a ubiquitous growth mediator, so any compound that leads to its activation in human beta cells could be expected to induce proliferation. Harmine acts (at least partially) by inhibiting DYRK1A, a downstream target that inhibits cell cycle advancement. Dr. Steward presented mouse studies in which harmine induces beta cell proliferation and glycemic control in the cases of partial pancreatectomy and human islet graft into a diabetic mouse model, respectively. The lab is now working with computational biochemists to generate “harmalogs” – versions of harmine with varied backbones – that may bind to and inhibit DYRK1A more potently. One major issue? There is currently no way to specifically shuttle harmine to human beta cells in vivo, and off target effects are certainly a concern. In fact, the molecule is found in ayahuasca, a traditional hallucinogenic South American tea, so we wouldn’t necessarily recommend oral ingestion [note: an online search shows that harmine itself isn’t hallucinogenic, but it potentiates the action of the tea’s psychedelic ingredient].

  • In the second part of his talk, Dr. Stewart explained how his lab’s work investigating the genomics of insulinomas (beta cell-derived pancreatic tumors) also led him to the DYRK1A pathway. Dr. Stewart’s philosophy is that if he can unearth the factors driving beta cell proliferation in the case of tumor formation, then he can try to recapitulate these processes in diabetes. For the past 10 years, his lab has collected insulinomas, arriving at a total of 78 – the tumors are exceedingly rare. In a study of 38 of them (half were excluded because they presented with MEN1 mutations, canonically known to cause insulinomas), the lab found 278 somatically mutated genes, 92 of which were key driver variants. However, only six of the mutations were recurrent. The analysis also revealed recurrent copy number variants on chromosomes 7 and 11, a rare presentation in oncology. Most strikingly, a differential mRNA expression analysis vs. normal purified human beta cells came back with ~80,000 exons within ~14,000 genes that are differentially expressed. The number one hit? DYRK1A (p=6.4 x 10-126). Dr. Stewart added that other genes have dethroned DYRK1A in more recent experiments, but it is still high up, and represents a solid therapeutic target.

Updates on Stem Cell-Derived Islet Cells for Diabetes

Doug Melton, PhD (Harvard Stem Cell Institute, Boston, MA)

This year’s Arthur Riggs Award winner, Harvard and Semma Therapeutics’ Dr. Doug Melton, provided a fascinating update on the evolving goals and future challenges of creating stem cell-derived human pancreatic islet cells for the treatment of diabetes. Dr. Melton and his lab are acclaimed for their protocol to quickly and accurately derive billions of human insulin-producing beta cells from progenitor stem cells – this work was published to great fanfare in 2015 and, since then, these cells have demonstrated promising results in an immune-protective alginate encapsulation device (in partnership with MIT’s Dr. Dan Anderson) and biotech firm created to commercialize the process, Semma Therapeutics, has partnered with Defymed’s MAILPAN macroencapsulation system and received investment from the JDRF T1D Fund. Notably, Dr. Melton pointed out that these cells are mature beta cells and differ from the progenitor cells used in ViaCyte’s PEC-Encap and PEC-Direct encapsulation therapies – from a morphological perspective, Dr. Melton’s cells exhibit the same crystallized insulin granules observed in human cadaveric beta cells, while ViaCyte’s immature beta-like cells exhibit mixed glucagon and insulin granules. Based on this and several other comparative parameters, Dr. Melton expressed confidence that we can consider these cells a “serious source of human beta cells.” That said, he spent most of his presentation highlighting the remaining questions and challenges going forward. First and foremost, he shared that these stem cell-derived cells do not perfectly recapitulate the insulin secretion response to glucose that human cadaveric cells exhibit in vitro. Dr. Melton’s cells secrete far less insulin than human donor cells in response to high levels of glucose stimulation (about 1/3 the amount of insulin at best) and the stem cell-derived cells appear to have little to no response to lower levels of glucose stimulation. Dr. Melton’s lab is currently investigating the cause of this subdued response in order to adjust the production protocol – there’s some evidence so far supporting a potential hypothesis that there may be a mitochondrial dysfunction in how these cells produce ATP. Another challenge highlighted by Dr. Melton is the fact that the process is not optimally efficient right now – only about 20%-30% of cells currently produced in each “batch” are beta cells; the others are other kinds of islet cells. The team is currently attempting to identify transcription factors that promote differentiation into these non-beta cells types – ideally, they will eventually be able to knock out the specific genes that are necessary for differentiation into these non-beta cell types, which will essentially “force” differentiation of all of the progenitor cells into beta cells. Despite these challenges, Dr. Melton suggested that his lab is essentially at the “goal line” for beta cell creation.

  • The next big challenge is preventing autoimmune attack of these beta cells, according to Dr. Melton. He noted that his lab is willing to engage in a wide variety of partnerships with encapsulation device teams, like Dr. Anderson’s MIT lab and Defymed, for a first-generation therapy. More ambitiously, Dr. Melton suggested that an unlimited beta cell source offers the possibility of genetic screens to determine gene targets to confer protection from autoimmunity to these cells. Essentially, Dr. Melton suggested that his team can use the CRISPR-Cas9 system to precisely knock out specific genes that produce proteins recognized by the immune system in autoimmune attack. Through this modification, transplanted stem cell-derived beta cells can thus be protected by the immune system in type 1 diabetes. Dr. Melton acknowledged that this work is in its very infancy and that he needs help from immunologists to identify promising genes to modify in this approach. The great Dr. Carla Greenbaum pointed out in Q&A that this kind of modification may have unforeseen consequences, which Dr. Melton readily agreed with. Nonetheless, we’re intrigued by this tantalizing approach and glad to see the field moving forward to confront the next great challenge toward a biological cure for type 1 diabetes.
  • Dr. Melton shared that his lab has had some early success in “tweaking” his islet cell creation protocol to create glucagon-producing alpha cells and somatostatin-producing delta cells. He acknowledged that it’s theoretically possible to product pancreatic polypeptide-producing gamma cells and ghrelin-producing epsilon cells as well, but his team is less interested in these cell types as their application to diabetes is more limited. Dr. Melton first discussed his work on alpha and delta cells at ADA 2015 and we’re excited to hear that he has made substantial progress on both fronts. He acknowledged that this work is “a year or two” behind the beta cells work, but predicted that within a few years his team will be able to produce islets with a specified ratio of beta to alpha to delta cells on demand. We’re curious if these stem cell-derived alpha cells can correct the glucagon dysfunction observed in type 1 diabetes – this would be a huge win in terms of managing postprandial glucose (and the lows associated with rapid-acting insulin). We’re absolutely thrilled that Dr. Melton and his team are thinking so ambitiously about the “next step” in islet replacement therapy and type 1 diabetes cures – he’s a powerhouse in the field to be sure and seems most deserving of the 2017 Arthur Riggs Award!
  • As a reminder, Dr. Melton also received quite a bit of press recently when he retracted his high-profile 2013 study outlining the betatrophin hypothesis after several attempts were unable to replicate the findings. Notably, this was related to his beta cell protocol and the retracted study does not affect his current work.

Genetics and Personalized Medicine

Personalized Medicine for Diabetes: From the Perspective of the NIDDK, Genetics of Type 1 Diabetes (And TEDDY)

Judith Fradkin, MD (NIDDK, Bethesda, MD), Stephen Rich, PhD (University of Virginia, Charlottesville, VA), Joel Hirschhorn, MD, PhD (Harvard University and Broad Institute, Boston, MA), Erwin Bottinger, MD (Berlin Institute of Health, Berlin, Germany)

The final day of the Levine-Riggs Diabetes Research Symposium featured a riveting morning symposium on genetics and personalized medicine approaches to diabetes therapy. The wide-ranging discussion covered personalized medicine at the molecular and clinical levels alike, and we noticed an interesting dichotomy between high-level treatment individualization approaches designed to illuminate the factors most correlated with better clinical outcomes and deep data dives to investigate which components of the genome (or epigenome!) most contribute to disease risk and progression. We have previously expressed frustration over the lack of visible progress on the need for and promise of personalized medicine in diabetes, but this symposium reinvigorated our optimism that this future is on its way, and renewed our appreciation for the massive challenges – particularly in  data analysis – posed by the pursuit of precision medicine.

  • On the clinical front, the highly respected Dr. Judith Fradkin (NIDDK, Bethesda, MD) highlighted the NIH’s efforts to support precision medicine in diabetes. Very exciting is the GRADE Study, the first-ever head-to-head comparison of the long-term effectiveness of several major classes of diabetes drugs. The study will compare the sulfonylurea Amaryl (glimepiride), the DPP-4 inhibitor Januvia (sitagliptin), the GLP-1 agonist Victoza (liraglutide), and the basal insulin Lantus (insulin glargine), each added on to metformin over four years – SGLT-2 inhibitors are not included because the trial was designed and initiated before these agents became available on the market (we continue to believe they should have been added though we realize this is challenging from a design perspective). The study aims to determine the patient characteristics associated with different responses to the different drug classes – in short, why do certain drugs work particularly well for certain people. This study has massive potential to support individualization of diabetes therapy and we are waiting for the results with baited breath – that said we unfortunately won’t be able to draw any conclusions about the comparative effectiveness of SGLT-2 inhibitors from this trial, which is disappointing given how attractive their clinical profile is shaping up to be. The fact that monotherapy is the only thing tested is also problematic since so many patients now start on combination therapy. The study began in 2012 and, according to ClinicalTrials.gov, is anticipated to report in 2020. Notably, the NIDDK has also established the Accelerating Medicines Partnership in Type 2 Diabetes, an effort at greater collaboration between government, academia, industry, and nonprofits with the shared goal of understanding safe and effective novel therapeutic targets for type diabetes. The cornerstone of this is a type 2 diabetes “knowledge portal” where scientists can have shared access to large clinical and scientific data sets. Dr. Fradkin noted that the partnership hopes to eventually move into the realm of therapies for diabetes complications as well. She noted that diabetes was specifically highlighted by President Obama (in a State of the Union Address, no less!) as a disease that could benefit from his Precision Medicine Initiative, and given these efforts by NIDDK, we are hopeful that this is indeed the case though overall, of course, we are extremely concerned about lower NIH funding happening.
  • On the molecular front, Drs. Stephen Rich (University of Virginia, Charlottesville, VA), Joel Hirschhorn (Harvard University and Broad Institute, Boston, MA), and Erwin Bottinger (Berlin Institute of Health, Berlin, Germany) discussed their respective work understanding the complicated basis of genetic risk for type 1 diabetes, obesity, and hypertension. In type 1 diabetes, Dr. Rich has identified several gene clusters correlated with disease risk, but there is so much heterogeneity at these sites that he hypothesizes that understanding the whole story will require an examination of the overlying epigenetic landscape. (“If there are any bioinformaticists in the audience, we need about 100 of you to analyze this data” he jokingly concluded.) In obesity, Dr. Hirschhorn’s research team has discovered a that the major gene clusters associated with obesity risk correspond to proteins expressed in the nervous system (particularly glutamate signaling and synaptic plasticity), as well as liver and brown fat regulation. This knowledge could help inform the targets of future obesity drug development – indeed many candidates in the current competitive landscape have neural targets. Finally, in hypertension, Dr. Bottinger discovered that the APOL1 allele of the apolipoprotein gene is highly correlated with hypertension: one copy of this allele produces a ~1.5 mmHg increase in systolic blood pressure, whereas two copies produce a nearly 3 mmHg increase, plus a younger age of onset for hypertension. 14% of the US African American population is estimated to have two risk copies of this gene, making these findings hugely relevant for hypertension screening.

Critical Assessment of Cell Therapeutics

Preclinical Studies to Consider for Stem Cell Derived Reta-Like Cells Moving to the Clinic

David Harlan, MD (University of Massachusetts, Worcester, MA)

UMass’s Dr. David Harlan gave a highly practical lecture outlining precautions and steps that the stem cell field must take in order to maximize the chances of getting functional stem cell-derived beta cells safely into humans for the treatment of diabetes. The International Society for Stem Cell Research (ISSCR) released most of these recommendations in a document containing guidelines for stem cell research and clinical translation last May. Some important caveats included: (i) Do not market therapies until they are tested in preclinical models and then clinical trials – and avoid over promising; (ii) Strive to insure that the cell type going into humans is the same one that has been thoroughly investigated at the bench/in animals (Dr. Harlan referenced a StemCells study that failed at least in part because they violated this principle); (iii) Compare the potential benefits of the therapy to the current standard of care – this bar gets higher as better therapies become available; (iv) Ask patients to consent to an autopsy in the event of a death –researchers need to know exactly what went wrong. Not only must we strive to “do no harm:, but when something bad happens in a nascent domain, the field can be derailed (e.g. the Christa McAuliffe and Jesse Gelsinger stories); and (v) Develop the therapy with economic value in mind – e.g. an autologous transplant, which may cost millions of dollars per patient, will not be as valuable as a standard cell line that everybody gets. Dr. Harlan then provided a list of questions that must be addressed in research (i.e. what is the best site for implantation? What are the effects of encapsulation? See below for a full list). While many of the questions he posed can be addressed in a normo- or dysglycemic mouse model, clinical trials are obviously a must. For these, Dr. Harlan said openly asked to community to consider “initially study[ing] these therapies in people with limited survival prognoses – ALS, incurable CNS malignancies – and diabetes. When they die one to two years later, we can look at how the cells behaved.” As he put it, these people can volunteer to be “astronauts,” and they may even get a short term glycemic benefit if the therapy works. The ethical and practical considerations are many, but similar to automated insulin delivery systems – you only get one chance to launch a stem cell therapy, so it is critical that it goes right. The diabetes patient and provider community, as well as the public, has very high hopes for stem cell therapies for type 1 diabetes – researchers in this field should take extra care to ensure that expectations are managed and that therapies proceed through the clinical development process with the most rigorous standards.

  • Despite these considerations, Dr. Harlan is optimistic that the 21st Century Cures Act will offer pathways to streamline beta-like cells into the clinic. ViaCyte and Sernova both have implantable beta cell encapsulation devices in phase 1/2 trials. To our knowledge, there are no ongoing trials involving implantation without encapsulation. See our beta cell encapsulation competitive landscape.

Dr. Harlan’s (/ISSCR’s) Top Stem Cell Research/Translation Considerations

  1. Do not widely market therapies until they are being tested in the clinic
  2. If you’re going to put a cell derived in lab into patient, know exactly what you’re putting in. Have a good manufacturing practice. Know every single reagent that went into making the cells, be able to track where it came from, how it was stored, etc.
  3. Have strict criteria for when these cells should be put in
  4. Rigorously test the therapy in preclinical studies with clear endpoints
  5. Make sure that the cell you test in the lab is the same cell that goes into the patient
  6. Assess the risk for tumorigenicity
  7. Assess the effects of encapsulation
  8. Compare therapy to the current standard of care and ensure there’s a benefit
  9. Ask patients to consent to autopsy in event of death. We need to know what went wrong in cells. If something bad happens, it really derails the field and of course, we all strive to prevent lives cut short 
  10. Stem cell based interventions should be developed with an eye toward delivering economic value to patients, payers, and healthcare systems. For example, a standard cell line could be used for all people.

Questions to Address in Research

  1. What is the proper/best site for implantation?
  2. What is the cell-dose response?
  3. What are the glycemic/metabolic effects on beta-like cell differentiation, function, and survival?
  4. How long are the cells alive and functional?
  5. Are the cells stable over time?
  6. What is the tolerance for teratomas or malignant transformations?
  7. Do the cells migrate after implantation?
  8. What are the effects of encapsulation?
  9. What are the immunological effects (autoimmune and alloimmune)?
  10. If needed, what are the effects of immunosuppressive agents?


--by Abigail Dove, Helen Gao, Brian Levine, Payal Marathe, and Kelly Close