14th Annual Rachmiel Levine City of Hope Symposium

March 12-15, 2014; Pasadena, CA Day #2 Highlights - Draft

Executive Highlights

We’re back with our Day #2 highlights from this cozy conference in Pasadena, CA. Today’s sessions focused on type 1 and type 2 diabetes pathophysiology and type 1 diabetes immune therapy. We bring you our top five highlights of the day here, with our full coverage of the highlights below:

  1. We found Dr. Richard Insel’s (JDRF Chief Scientific Officer) Plenary Lecture on a preliminary model for staging asymptomatic type 1 diabetes to be the highlight of the day. The model, based on numerous natural history studies in the United States and Europe, supports the idea that type 1 diabetes can be diagnosed much earlier than the onset of symptoms. Quite notably, he noted that JDRF is planning a workshop with the broader community later this year to present this model to stakeholders and make the case that this concept can help inform decisions on trials for type 1 diabetes prevention. In other words, since we can now reliably predict who will progress to symptomatic type 1 diabetes, that makes the benefit of an early pre-symptomatic intervention much more clear and certain [justifying greater safety risk], since “prevention” often implies that you may be treating someone who may never get sick. As we understand it, the staging system has been in discussion with the community for some time and we look very forward to hearing how this goes. JDRF plans to work with the clinical community (ADA, AACE, the Endocrine Society the Helmsley Charitable Trust, ISPAD, the NIH, etc.) to garner support that clinical guidelines for type 1 diabetes diagnosis should be changed to allow for diagnosis prior to onset of symptoms.
  2. Dr. Muhammad Abdul-Ghani (University of Texas, San Antonio, TX) discussed new SGLT-2 inhibitor data on glucose homeostasis that suggests SGLT-2 inhibitors may be especially effective when used in combination with drugs that reduce hepatic glucose production (e.g., incretins). He has found that a single dapagliflozin (AZ’s SGLT-2 Farxiga) administration can lead to sustained elevated glucagon secretion, and, therefore, hepatic glucose production (HGP) for 14 days, which offsets the drug’s overall glucose-lowering effects mediated through glycosuria. Therefore, drugs that can reduce the body’s compensatory increase in HGP in response to SGLT-2 inhibition may enhance the effectiveness of SGLT-2 inhibition. Dr. Abdul-Ghani’s discussion of how SGLT-1/SGLT-2 dual inhibition may also be a powerful means of enhancing SGLT-2 efficacy was also quite interesting.
  3. The renowned Dr. Mark Atkinson (University of Florida, Gainesville, FL) updated attendees on new discoveries by the nPOD studies (details on the nPOD consortium are below). These studies have provided some fundamental advances in the understanding of type 1 diabetes: for example, they have demonstrated that type 1 diabetes is a disease of autoimmunity (which is still up for debate amongst some in the field), as well as abnormalities inherent to the beta cell and pancreas. Thus, he believes one of the main therapeutic takeaways from nPOD since its founding in 2007 is that researchers should be targeting the beta cell and exocrine pancreas in the prevention and treatment of type 1 diabetes, not just autoimmunity. 

  4. Dr. Atkinson’s conclusion was also notable. He proposed several hypothetical ideas on type 1 diabetes pathophysiology: (i) stress resulting from a reduced beta cell mass/small pancreas may determine how quickly type 1 diabetes develops; (ii) stress resulting from a reduction in beta cell mass might represent a trigger for the full disease; (iii) people positive for a single autoantibody who do not have type 1 diabetes may represent an aborted or unproductive immune response leaving a person with altered beta cells but not type 1 diabetes; (iv) abnormalities in beta cells may be key drivers of the immune response; and (v) the progression from prediabetes to type 1 diabetes consists of exacerbations (or relapses) and remissions with the immune system “snipering” some beta cells rather than “blitzkrieg” attacking many beta cells at a time.

  5. Dr. Carla Greenbaum (Benaroya Research Institute at Virginia Mason, Seattle, WA) explained how antigen-based therapies have worked for allergies but not for autoimmune diseases like type 1 diabetes. Dr. Greenbaum believes that the main obstacle to discovering an effective and safe antigen-based therapy in type 1 diabetes is the lack of robust mechanistic assays and researchers’ inability to convincing demonstrate tolerance mechanisms. In particular, Dr. Greenbaum sees a need to assays that can be trusted to characterize pathogenic antigen specific cells, because such assays would help researchers better determine if a candidate therapy is causing benefit or harm. Notably, Dr. Greenbaum is doing a great deal of work now for the T1D Exchange as we understand it; we are very happy to hear this, for the sake of patients, as we believe she’ll be very smart in identifying the valuable studies that should be moved forward.

See our full coverage of each of these five highlights and more below.

Detailed Discussion and Commentary

Plenary Lecture: New Staging Classification System of Type 1 Diabetes in the At-Risk Setting

Staging Asymptomatic Type 1 Diabetes

Richard Insel, MD (Chief Scientific Officer, JDRF, New York, NY)

Dr. Richard Insel presented a preliminary model for staging asymptomatic type 1 diabetes that supports the idea that type 1 diabetes can be diagnosed much earlier than the onset of symptoms.  JDRF plans to work with the broader community (ADA, AACE, the Endocrine Society, the Helmsley Charitable Trust, ISPAD, the NIH, etc.) to garner support to change the diagnostic criteria for type 1 diabetes. Exact diagnostic criteria unspecified, but we imagine that presence of ≥two autoantibodies plus dysglycemia would be fairly robust criteria. In the meantime, Dr. Insel indicated that JDRF is also including FDA in the discussion to better inform decisions on trials for type 1 diabetes prevention and is planning a workshop with stakeholders later this year. This would be an outstanding result for patients, particularly since FDA has not focused on prevention or given any official guidance that we know of. Data from the TEDDY and Pre-TEDDY studies and DPT-1 and the TrialNet Pathway to Prevention study powerfully demonstrate that once someone has reached a certain stage (e.g., multiple autoantibodies) in the asymptomatic phase of the disease, progression to symptomatic type 1 diabetes essentially becomes inevitable. Onset of dysglycemia in the face of multiple autoantibodies is associated with an accelerated rate of progression to symptomatic disease.  Thus, diagnosing the disease prior to the onset of symptoms may offer a window of opportunity for preventing progression to symptomatic type 1 diabetes and, correspondingly, the complications of diabetes. The proposed staging system consisting of three stages: (i) beta cell autoimmunity with no dysglycemia and no symptoms; (ii) beta cell autoimmunity with dysglycemia (e.g., impaired glucose tolerance) and no symptoms; and (iii) overt symptomatic type 1 diabetes.

  • What are the implications for changing the diagnostic criteria for type 1 diabetes? As Dr. Insel noted, if one can diagnose the disease before the onset of symptoms, then one has the opportunity to intervene earlier and potentially (with future therapies) even prevent the onset of symptomatic disease. (In the meantime, we still support earlier diagnosis even in the absence of therapeutic interventions since we think the diagnosis experience would be easier on families if the child or adult is not in DKA at the time of diagnosis.) Furthermore, classifying asymptomatic type 1 diabetes as a disease would help interventions that could prevent the onset of symptoms reach the market, because it would change the risk/benefit ratio of interventions potentially making regulatory agencies see more value in allowing such trials. Also, the disease designation would  ultimately likely lead to payers, patients, and providers seeing more value in adopting such interventions. Another astute remark we heard from Dr. Insel was that classifying the asymptomatic diabetes as a disease also changes how industry can position therapeutic products: products to treat the asymptomatic phase of type 1 diabetes could then be billed as interventional therapies rather than therapies for prevention (addressing the age old question of “why treat someone who is not yet sick?”).
  • Stage 1 of the proposed three-stage system is based on the number of islet autoantibodies (aAbs) a person has. Stratification in this manner has been supported by studies demonstrating the ability to detect aAbs and predict someone’s risk of progressing to symptomatic type 1 diabetes. HLA-high risk children screened at birth who develop h two or more aAbs have a 44% and 70% risk of symptomatic T1D in 5 and 10 years and with a lifetime risk nearly 100% (JAMA 2013).  Similar risk estimates are observed in relatives (children and adults) of people with type 1 diabetes with multiple aAbs, with a faster rate of progression in the children. The presence of 3 or 4 aAbs is associated with a faster rate of progression versus 2 aAbs. While the type of aAb can also make a difference (those developing the IA-2 autoantibody progress faster, even if this is the only aAb they have), the Staging system does not yet account for that (more details in the second to last bullet).
  • Stage 2 builds upon stage one by characterizing the onset of dysglycemia on top of multiple autoantibodies accelerates progression – in the DPT-1 trial, people with dysglycemia had an approximately  75-80% risk of progressing to symptomatic type 1 diabetes over five years. In addition, in TrialNet’s Pathway to Prevention, a 10% rise in A1c has been associated with an 84% rate of three-year progression to symptomatic disease (Diabetologia 2013).
  • Stage 3 is the current definition of diagnosed type 1 diabetes based on the onset of symptoms.
  • Dr. Insel emphasized that this preliminary staging system still needs further optimization and refinement. For example, additional stratification prior to Stage 1 (onset of the autoimmune process) may be possible since characteristics such as interferon signature patterns, innate immunity, metabolomics, or the microbiome may have predictive value for who progresses to developing aAbs. As another example, single aAb status is currently associated with a higher rate of progression to type 1 diabetes than having zero aAbs (~10% over 20 years for one aAb compared to nearly 0% for zero aAbs), but a much lower rate of progression than having two or more aAbs. As Stage 1 currently stands, it takes into account that people with two or more aAbs are very likely to progress to symptomatic disease, but it does not deal with the risk, though low, of someone with one aAb. A qualitative analysis (e.g. affinity, specificity) of the type of single aAb and studies of the T-cell phenotypes and responses may allow for more refined stratification of Stage 1 to provide more nuanced predictions on risk of progressing to type 1 diabetes for people with a single aAb.
  • As raised in Q&A, sociological and clinical challenges remain to changing the diagnostic paradigm for type 1 diabetes. For example one must ask, how do families or adult individuals cope with the diagnosis before any kind of intervention is available for the asymptomatic phase of the disease? Another outstanding question is if it will be acceptable to say that your daughter has a disease that, for now, has no symptoms and requires no treatment? With proper awareness and education, we see these challenges as addressable, and we are curious if insulinization when blood glucose levels first start to rise (versus when a person has DKA or severe hyperglycemia for the first time) could prolong the onset of symptomatic disease and the honeymoon phase when it does occur. We would draw a parallel to HIV/AIDS – a person can contract HIV with the possibility that he may or may not advance to the symptomatic AIDS, but at the same time still claim to have a disease and be treated for it (an exception to this analogy, of course, is that a person with HIV can transmit the disease to another person [and do so more readily without treatment] while a person with type 1 diabetes cannot).

Questions and Answers

Q [Dr. Bart Roep, Leiden University Medical Center, The Netherlands]: I’m a little concerned that we are drawing classifications based on substudies in biased populations. Ninety percent of patients do not have a family member with diabetes, but we only have studies on relatives and family members. Half of patients get disease outside the age window that has been studied (20 years and younger). Forty percent of patients are not genetically different from the healthy population. Half of patients from the TRIGR study that developed diabetes were not selected for risk and were in the group they left out as controls. What is your thought?

A: We can either stick our head in the sand, ignore the data we have, and keep muddling around, or we can put a stake in the ground and say this is where we start, knowing it’s not perfect. We can’t let perfection be the enemy of the good. So where do we go? From the pre-TEDDY and TEDDY studies we are gaining insight into the pathogenesis of childhood-onset type 1 diabetes. We really do not understand adult-onset type 1 diabetes with respect to when the process starts but we can use the clinical and laboratory data from DPT-1 and TrialNet from first and second degree relatives as a starting point to develop a staging classification. And as Bart points out only about 10% of people diagnosed have a relative. So we start here.

We also have to realize that this classification is only going to be relevant for individuals who are being screened.

Dr. Roep: Ah yes, that is a good point. I concede.

A: So today who can we screen? We have been screening based on HLA type and relatives. Another screening initiative we’re looking at is the concept of going into the childhood population and screening for autoantibodies, which would take HLA out of the picture. If we do that, we can use that data to refine the model. But you’ve got to start somewhere.

Q: What are your and JDRF’s thoughts are about moving beyond the relatives, because all the data that has been out there in the general population doesn’t say that the risk is low in the multiple aAb positive general population.

A: The art of this is having access to ample subjects, so it is complicated. As we think about screening we have to think about how cost-effective screening can be. Ultimately I would like to think we can get to a day where we can have universal childhood immunization with a vaccine to prevent childhood type 1 diabetes without having to use any diagnostics. Although that sounds futuristic, we need a public health based approach for what I consider a public health problem. But to get there, we will need to go through the steps of initially testing interventions in high–risk populations so we need to find more cost-effective ways of screening.

Dr. David Harlan (University of Massachusetts, Worcester, MA): I still see patients with diabetes every day. I spend time every day so I know how tough this is but I will also say that as a physician you get to be very humble. There are many things we thought were safe that then turned out not to be when tested them in people. If we’re testing new things with people that don’t have a disease, we better be damn sure they’re safe. It’s really a sticky wicket. To a certain extent I think we have done this – we’ve tried insulin and nicotinamide in people without disease.

A: I agree nobody wants to accelerate lifelong insulin-dependent disease. At the same time, we know these individuals are going to become lifelong insulin dependent once they reach a specific stage. Based on that we should use this opportunity. I’m speaking on behalf of our stakeholders – they are urging us to push forward and find a safe and effective way of intervening. Safety is first. .

Q: I think this is a really important discussion that you’re starting. From a scientific standpoint I can see the value in that. From a societal standpoint how this will be viewed? Does the 10 year old little girl in third grade with three aAbs have diabetes?

A: Yes, in an asymptomatic form of course.

Q: I wonder how the families of people would be affected by this and trying to explain that she has diabetes, but she doesn’t take insulin and doesn’t need any treatment, but we don’t have any therapy for it yet. What do the families say about this approach?

A: That’s a really important question. As we think about disease and the opportunities to intervene earlier and earlier, how we communicate to patients and families will be absolutely critical. We do know that it is not uncommon for individuals with type 1 diabetes to undergo a honeymoon period where some individuals have negligible insulin requirements. Do those individuals have diabetes? As we look to the future, I would posit we have to think about disease a little differently. In the last three to four centuries, disease diagnosis has l been primarily clinically based. We’re going to have to embrace more of a molecular pathogenesis approach to disease if we want to practice predictive, personalized medicine. But at the same time, not ignoring the ethical, sociological, and clinical care aspects. Those will be challenges for sure. I appreciate the importance of your question.

 

New Insights into Diabetes Pathophysiology

Emerging Role of the Kidney in the Regulation of Glucose Homeostasis and the Treamtent of Type 2 Diabetes

Muhammad Abdul-Ghani, MD (University of Texas, San Antonio, TX)

Dr. Muhammad Abdul-Ghani, who is part of Dr. Ralph DeFronzo’s team, reviewed the rationale and mechanisms behind SGLT-2 inhibition in type 2 diabetes. What was most interesting about his presentation was his discussion of two treatment strategies that could produce synergistic effects by enhancing the efficacy of SGLT-2 inhibitors. He presented data suggesting that patients respond to SGLT-2 inhibition with a compensatory increase in glucagon secretion and, therefore, hepatic glucose production (HGP). This compensatory pathway offsets the overall glucose-lowering effect of SGLT-2 inhibition mediated by glycosuria. Thus, pairing an SGLT-2 inhibitor with another drug that suppresses HGP could greatly enhance SGLT-2 inhibitor efficacy (incretins, for example, suppress glucagon production and, therefore, HGP). In addition, he discussed how SGLT-1 inhibition could easily increase the efficacy of SGLT-2 inhibitors by two-to-three fold (see below for the physiological rationale). This was exciting to hear; while of course companies who are themselves developing SGLT-1/SGLT-2 dual inhibitors (Lexicon’s phase 3-ready LX4211 and Novartis’ phase 2 LIK066) have expressed enthusiasm, we have yet to hear as many opinions from academics.

  • Dr. Abdul-Ghani presented new data from a small study (n=18) suggesting that SGLT-2 inhibition elicits a compensatory response from the liver. For two weeks after a single administration of dapagliflozin, patients experienced elevated hepatic glucose production (HGP) compared to placebo and baseline pre-drug administration (the rise in HGP was accompanied by a 35% increase in glucagon secretion). The patients still experienced an overall 20% decrease in fasting plasma glucose despite the elevations in glucagon and HGP, but the implication is that if one can inhibit this compensatory HGP response, then one could greatly increase the efficacy of SGLT-2 inhibitors. Dr. Abdul-Ghani suggested pairing SGLT-2 inhibitors with incretins, which are known to suppress glucagon production (and, therefore, HGP) to enhance the efficacy of the SGLT-2 inhibitor.
  • Dr. Abdul-Ghani also noted that the effectiveness of SGLT-2 inhibition could also be enhanced by pairing it with SGLT-1 inhibition. This was exciting to hear, as we have rarely heard support for this idea outside of the setting of companies who are themselves developing SGLT-1/SGLT-2 dual inhibitors (Lexicon’s phase 3-ready LX4211 and Novartis’ phase 2 LIK066). The physiological rational here is that in the renal system, SGLT-2 is positioned in the proximal part of the proximal tubule while SGLT-1 is positioned in the distal part of the proximal tubule. Under normal physiologic conditions, SGLT-2 accounts for about 90% of glucose re-absorption in the kidney, and SGLT-1, accounts for only about 10% of glucose re-absorption. However, when SGLT-2 is being blocked, all the glucose that would have been reabsorbed by SGLT-2 then flows further downstream and encounters SGLT-1. A greater percentage of SGLT-1 transporters are occupied at this higher concentration of glucose in the filtrate, so SGLT-1 ends up reabsorbing about 50% of the glucose that SGLT-2 would have absorbed upstream had it not been blocked. Dr. Abdul-Ghani suggested that a two-to-three fold increase in efficacy of SGLT-2 inhibition could be achieved by a 40-50% inhibition of SGLT-1.
    • As background, researchers previously thought that SGLT-1 inhibitors caused GI effects, because the original SGLT-1 phloridzin was associated with these adverse events and SGLT-1 is present in the intestines. However, research with other SGLT-1 inhibitors suggests this was an agent effect not a class effect.
    • Of the selective SGLT-2 inhibitors, J&J’s Invokana (canagliflozin) is the least selective for SGLT-2 over SGLT-1 (so it has some transient activity on SGLT-1), and canagliflozin is arguably the most potent selective SGLT-2 inhibitor (no head to head studies have been conducted, but it seems to produce more urinary glucose excretion compared to AZ’s Forxiga [dapagliflozin] or Lilly’s empagliflozin). This would be consistent with the idea that SGLT-1 inhibition enhances the efficacy of SGLT-2 inhibition.
    • Lexicon (phase 3-ready) and Novartis (phase 2) are both developing SGLT-1/SGLT-2 dual inhibitors, and Lexicon has indeed presented phase 2 data suggesting that its candidate (LX4211) may have better efficacy than selective SGLT-2 inhibitors (while at the same time requiring less action on the part of SGLT-2 than the selective SGLT-2 inhibitors).

 

Pathophysiology in Type 1 Diabetes

New Discoveries by nPOD Studies in the Development of Type 1 Diabetes

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

Dr. Mark Atkinson highlighted several key lessons that nPOD (description of nPOD below) has provided since its founding in 2007: (i) type 1 diabetes is an autoimmune disease (which still remains controversial amongst some in the field); (ii) type 1 diabetes is a disease of the beta cells; (iii) type 1 diabetes is also a disease of the pancreas; (iv) in persons without type 1 diabetes, the potential for beta cell regeneration is highest in the first decade of life and falls off considerably thereafter; (v) adult beta cell mass in those without type 1 diabetes is highly heterogeneous; and (vi) perhaps as many as eighty percent of type 1 diabetes patients, regardless of disease duration, have at least some beta cells in their pancreas. The seeming simplicity behind these conclusions underscores the fundamental uncertainty that has existed on the mechanisms driving type 1 diabetes and how complex the disease is (both of which make it strikingly challenging to find a biological cure for the disease). Indeed, Dr. Atkinson suggested that the involvement of beta cell and pancreas abnormalities in the disease means that researchers should consider beta cell and pancreatic targets for preventing and treating the disease, rather than just focusing on autoimmunity. Dr. Atkinson concluded his presentation by proposing several hypothetical ideas on type 1 diabetes pathophysiology, which spurred a lively panel discussion following the session: 1) stress resulting from a reduced beta cell mass/small pancreas may determine how quickly type 1 diabetes develops, 2) stress resulting from a reduction in beta cell mass might represent a trigger for the full disease, 3) people positive for a single autoantibody who do not have type 1 diabetes may represent an aborted or unproductive immune response leaving a person with altered beta cells but not type 1 diabetes, 4) abnormalities in beta cells may be key drivers of the immune response, and 5) the progression from prediabetes to type 1 diabetes consists of exacerbations (or relapses) and remissions with the immune system “snipering” some beta cells rather than “blitzkrieg” attacking many beta cells at a time. 

  • nPOD collects tissues of interest (e.g., pancreases, bone marrow, peripheral blood) from organ donors with type 1 diabetes, persons without the disease, as well as disorders of related interest (e.g., type 2 diabetes) and distributes samples to researchers free of charge. Since its foundation in 2007, nPOD has collected tissues from almost 300 individuals. Over 175 investigators from around the world have worked with nPOD on over 120 approved projects. Dr. Atkinson noted that nPOD was the result of the revered late Dr. George Eisenbarth’s vision. We think nPOD has the potential to greatly benefit the lives of people with type 1 diabetes. This is because nPOD is providing researchers with a means of understanding type 1 diabetes pathophysiology in humans rather than in mouse models, which have offered some valuable insights but have also notoriously led research astray due to the great differences that exist between mice and humans. Indeed, Dr. Atkinson highlighted key differences in the quantity and quality of the autoimmune islet infiltrate, as well as the pancreas (in general) and islet cells (in particular), between NOD mice and humans with type 1 diabetes.
    • nPOD projects have looked at many different topics related to type 1 diabetes pathophysiology including but not limited to (i) organic pollutants; (ii) progenitor cells; (iii) bone marrow; (iv) proteomics; (v) beta cell dedifferentiation; (vi) stem cells; (vii) the potential tie to celiac disease; (viii) cytokines; (ix) inflammation; (x) imaging; (xi) insulitis; (xii) cell conversion; (xiii) ER stress; (xiv) autoreactive T cells; (xv) viruses; (xvi) epigenetics; (xvii) regulatory T cells; (xviii) beta cell vascularization; (xix) degranulated beta cells; (xx) glucose transporters; (xxi) adhesion molecules; (xxii) beta cell regeneration; (xxiii) viruses; (xxiv) islet cell transplantation; (xxv) autoreactive B cells, and more.
    • nPOD has formed several working groups and provided these organizations with software that enables them to share and collaborate on data. Existing working groups include one for European researchers, one on transplantation, and one on viruses. Groups are forming that are dedicated to studying the extracellular matrix, beta cells, and genomics.  This form of research is quite unique in the type 1 diabetes research community as investigators actively disclose their research findings in real time (i.e. prior to publication), share valuable reagents, and work as “teams” on projects determined by group thought.

 

Panel Discussion

Dr. Matthias von Herrath (VP and Head, Seattle Diabetes R&D Center, Novo Nordisk, Seattle, WA/La Jolla Institute for Allergy and Immunology, La Jolla, CA): Should we be using anti-inflammatory therapies right at diagnosis? Should we be trying to diagnose the disease earlier in the natural history?

Dr. Mark Atkinson (University of Florida, Gainesville, FL): The main takeaways from nPOD in terms of therapies is that there are more targets than the immune-based targets. We may want to think about beta cell therapies and those targeting the exocrine pancreas. I also think that the widely held notion that a person has lost about 90% of their beta cell mass at the time of diagnosis is probably not true. Depending on the person’s age, there are more beta cells than once thought. I think that makes a case for early immune therapy very near the time of diagnosis. As far as disease prevention, we need to do something that accounts for what appears to be an on again, off again process of destroying beta cells. If you don't time things right, it could be you will administer a drug when they don't have insulitis and when the drug wears off they will have insulitis again.

Dr. Peter In’t Veld (Brussels Free University-VUB, Brussels, Belgium): Many of the immune interventions we have used were based on the NOD mouse. We have to be careful about converting those findings to the patient population. One of the things that struck me was that patients with autoantibodies did not show clear signs of insulin pathological damage. Another thing that struck me was the high level of beta cell replication we see in some cases; some islets have an unexpectedly high capacity for replication. If we can stimulate the cells to overcome the initial resistance to replication, that might be helpful.

Dr. Carla Greenbaum (Benaroya Research Institute at Virginia Mason, Seattle, WA): In general do you see the percentages of patients experiencing insulitis being related to HLA type or other genetics?

Dr. In’t Veld: One of the problems of looking at retrospective material is that we don't have enough clinical data on all of these cases. One of the great things about nPOD is that we have all of the elements you are referring to and can now do these studies.

Dr. Atkinson: We are working on that in nPOD. We have almost reached 300 cases. The numbers aren’t quite there yet to make certain assertions. We are seeing some interesting cases though and do believe, with time, nPOD will help better decipher the contribution of genetics to type 1 diabetes.

Q: Acute pancreatitis is probably a more common association with diabetes than we have previously appreciated. I wonder if in nPOD you see evidence that initial inflammation can be associated with a later autoimmune response?

Dr. In’t Veld: We discussed this at the Exeter nPOD working group meeting. We do not observe that in the historical material. The changes are relatively light. There is some interlobular fibrosis that might indicate a low level of infiltration but its not dramatic.

Dr. Atkinson: The notion that you do remodel – or see a major turnover in cellular number – your beta cells in early life and then are left with quotient X – that is something people are now very interested in. This could be a very important to understand why some persons develop type 1 diabetes early in life while in others, it takes far more time. As a result, we really want to establish what is happening in that first decade of life in terms of beta cell numbers.

Q: What is the beta cell mass at the time of onset of diabetes?

Dr. In’t Veld: That is a very important question.

Dr. Atkinson: It is not known.

Q: Is it 10%?

Dr. In’t Veld: As far as I know there is only one study looking at this in people with type 1 diabetes and it was in five patients with chronic type 1 diabetes. We don't have anything in recent onset diabetes.

Q: I was struck by how many of the antibody positive people that you found who were nondiabetic who did not have insulitis. You focused on those that did have insulitis but many did not. Mark, you said that type 1 diabetes is an autoimmune disease but I am wondering if we are looking for an autoimmune disease and if a lot of what is happening in the beta cell is not autoimmune. Maybe we are looking in the wrong place.

Dr. Atkinson: I think this an example of what is called the Streetlight Effect. In that story, a man who lost his keys at night and looked for them under a street light. When a police officer came and asked him what he was doing and he responded that he was looking for his keys, the police officer asked if he remembered where he lost it. The man said no so the officer asked why he was looking where he was. The man said “this is where the light is.” This could be analogous to what we have been doing. I think we have been following the animal models and George’s model too closely and have not been thinking outside of the box. I think we are now practicing, with nPOD, much more out of the box thinking. I want to emphasize though that while type 1 diabetes is an autoimmune disease there is more to type 1 diabetes than just T cells.

Comment: I think the pancreas is a fatty tissue and it would be more appropriate rather than to report the absolute size of the pancreas to normalize that with the BMI of the individual.

Dr. Atkinson: We have a very skilled statistician working on the pancreatic weight project. We correct for everything and no matter which way you cut it, it is a smaller pancreas in those with type 1 diabetes. This has been known for a while, we just have not appreciated it.  

 

Insulitis in Type 1 Diabetes

Peter In’t Veld, PhD (Brussels Free University-VUB, Brussels, Belgium)

Dr. Peter In’t Veld detailed the mechanisms of insulitis in pre-type 1 diabetes. Insulitis is the infiltration of immune cells into the islet of Langerhans, and onset of insulitis can precede onset of type 1 diabetes. One of our main takeaways from this presentation was that the differential prevalence of insulitis between young and old-onset patient subgroups underscores how heterogeneous the natural history of type 1 diabetes is. Dr. In’t Veld emphasized that our understanding of insulitis is limited because between 1902 and 2013 only 165 cases of insulitis in type 1 diabetes have been available for study. What we do know is that insulitis tends to be only seen in islets still containing beta calls. Therefore, the prevalence of insulitis in the pancreas of a person with type 1 diabetes tends to decline as her disease progresses (since those with more advanced disease progression have fewer beta cells). Curiously, the incidence of insulitis also appears to decrease as a person’s age of type 1 diabetes onset increases. The main explanation we have heard for this difference is that the onset of type 1 diabetes is less intense and dramatic in people who are diagnosed later in life (indeed, the drop in C-peptide is known to be steeper in younger patients than older). This means that fewer islets would be affected by insulitis at any given time in an adult with new onset type 1 diabetes, such that it would be seen less often in a cross-sectional study. Or it could be that adults have a different form of type 1 diabetes that is not characterized by insulitis. Either way, the contrasting prevalence of insulitis between new onset pediatric and adults patients underscores how heterogeneous the disease’s natural history is across age groups.

Prevalence of insulitis and insulin-containing islets in patients with type 1 diabetes for one year or less

Age of onset

Prevalence of insulitis

Prevalence of insulin-containing islets

0-14 years

68%

37.9%

15-19 years

29%

56.4%

 

Panel Discussion

Dr. Matthias von Herrath (VP and Head, Seattle Diabetes R&D Center, Novo Nordisk, Seattle, WA/La Jolla Institute for Allergy and Immunology, La Jolla, CA): Should we be using anti-inflammatory therapies right at diagnosis? Should we be trying to diagnose the disease earlier in the natural history?

Dr. Mark Atkinson (University of Florida, Gainesville, FL): The main takeaways from nPOD in terms of therapies is that there are more targets than the immune-based targets. We may want to think about beta cell therapies and those targeting the exocrine pancreas. I do think that the idea that a person has lost about 90% of their beta cell mass at the time of diagnosis is probably not true. Depending on the person’s age there are more beta cells than once thought. I think that makes a case for early immune therapy. We need to do something that accounts for the randomness in the longevity of some beta cells. If you don't time things right, it could be you will administer a drug when they don't have insulitis and when the drug wears off they will have insulitis again.

Dr. Peter In’t Veld (Brussels Free University-VUB, Brussels, Belgium): Many of the immune interventions we have used were based on the NOD mouse. We have to be careful about converting those findings to the patient population. One of the things that struck me was that patients with autoantibodies did not show clear signs of insulin pathological damage. Another thing that struck me was the high level of beta cell replication we see in some cases; some islets have an unexpectedly high capacity for replication. If we can stimulate the cells to overcome the initial resistance to replication, that might be helpful.

Dr. Carla Greenbaum (Benaroya Research Institute at Virginia Mason, Seattle, WA): In general do you see the percentages of patients experiencing insulitis being related to HLA type or other genetics?

Dr. In’t Veld: One of the problems of looking at retrospective material is that we don't have enough clinical data on all of these cases. One of the great things about nPOD is that we have all of the elements you are referring to and can now do these studies.

Dr. Atkinson: We are working on that in nPOD. We have almost reached 300 cases. The numbers aren’t quite there yet to make certain assertions. We are seeing some interesting cases though.

Q: Acute pancreatitis is probably a more common association with diabetes than we have previously appreciated. I wonder if in nPOD you see evidence that initial inflammation can be associated with a later autoimmune response?

Dr. In’t Veld: We discussed this at the Exeter nPOD working group meeting. We do not observe that in the historical material. The changes are relatively light. There is some interlobular fibrosis that might indicate a low level of infiltration but its not dramatic.

Dr. Atkinson: The notion that you do remodel in early life and then are left with quotient X – that is something people are very interested in. It could be important. We want to establish what is happening in that first decade of life.

Q: What is the beta cell mass at the time of onset of diabetes?

Dr. In’t Veld: That is a very important question.

Dr. Atkinson: It is not known.

Q: Is it 10%?

Dr. In’t Veld: As far as I know there is only one study looking at this in people with type 1 diabetes and it was in five patients with chronic type 1 diabetes. We don't have anything in recent onset diabetes.

Q: I was struck by how many of the antibody positive people that you found who were nondiabetic who did not have insulitis. You focused on those that did have insulitis but many did not. Mark, you said that type 1 diabetes is an autoimmune disease but I am wondering if we are looking for an autoimmune disease and if a lot of what is happening in the beta cell is not autoimmune. Maybe we are looking in the wrong place.

Dr. Atkinson: There is a story of a man who lost his keys at night and looked for them under a street light. When a police officer came and asked him what he was doing and he responded that he was looking for his keys, the police officer asked if he remembered where he lost it. The man said no so the officer asked why he was looking where he was. The man said “this is where the light is.” This could be analogous to what we have been doing. I think we have been following the animal models and George’s model and have not been thinking outside of the box. I think we are now hearing out of the box thinking. I want to emphasize though that while type 1 diabetes is an autoimmune disease that there is more to type 1 diabetes than just T cells.

Comment: I think the pancreas is a fatty tissue and it would be more appropriate rather than to report the absolute size of the pancreas to normalize that with the BMI of the individual.

Dr. Atkinson: We have a statistician working on this. We correct for everything and no matter which way you cut it, it is a smaller pancreas. This has been known for a while, we just have not appreciated it. 

 

Immune Therapy

Antigen-Specific Immunotherapy

Carla Greenbaum, MD (Benaroya Research Institute at Virginia Mason, Seattle, WA)

Dr. Carla Greenbaum explained how antigen therapies’ success in allergies has provided researchers with hope – but not certainty – that the approach could safely and effectively address autoimmunity. Unfortunately, however, antigen therapies have been unsuccessful for type 1 diabetes and other autoimmune diseases like multiple sclerosis. Dr. Greenbaum believes that the main obstacle to discovering an effective and safe antigen-based therapy in type 1 diabetes is the lack of robust mechanistic assays and researchers’ inability to convincing demonstrate tolerance mechanisms. Thus, she pressed that more robust, clinically validated, and standardized mechanistic assays are. In particular, Dr. Greenbaum sees a need for assay that can be trusted to characterize pathogenic antigen specific cells. Such assays would help researchers better determine if a candidate therapy is causing benefit or harm. We find antigen-specific immunotherapies to be particularly interesting because Dr. Matthias von Herrath (VP and Head, Seattle Diabetes R&D Center, Novo Nordisk, Seattle, WA/La Jolla Institute for Allergy and Immunology, La Jolla, CA) has previously indicated that the Novo Nordisk Type 1 Diabetes R&D Center (which he directs in Seattle) is focused on antigen-specific candidates (as well as combination therapies). However, Dr. Jay Skyler (University of Miami, FL) has questioned the antigen-specific approach, citing the failure of this strategy in the past.

  • Ongoing antigen trials in type 1 diabetes, according to Dr. Greenbaum, are studying 1) oral insulin to prevent type 1 diabetes (conducted by TrialNet), 2) nasal/oral insulin for prevention (INIT II and Pre-POINT), 3) GAD-alum for prevention, 4) proinsulin peptide in people with type 1 diabetes, and 5) autologous dendritic cell therapy with antisense oligonucleotides in people with type 1 diabetes. Additionally, the study DIABGAD is looking at GAD-alum, vitamin D, and ibuprofen in people with type 1 diabetes.
  • Antigen therapy in multiple sclerosis (an autoimmune disease) has been associated with exacerbation of the disease and in diabetes it has no or unclear benefit. The generally negative studies of antigen therapies in diabetes have looked at oral insulin in at-risk individuals to prevent the disease, GAD-alum (Diamyd), altered peptide ligand (Insulin B9-23), DiaPep-277 heat shock protein, parenteral insulin in at risk people to prevent the disease, and nasal insulin to prevent.
  • In contrast, Dr. Greenbaum explained that allergen-specific immunotherapy has been effectively used to reduce allergy symptoms for 100 years. For background, specific immunotherapy reduces the sensitivity of a person to an allergen by administering progressively larger doses of an allergen extract. It has been successfully performed subcutaneously and sublingually (under the tongue). Antigen specific cells are present in people with and without an allergy to that antigen, however, pathogenic antigen specific cells in people with the allergy are terminally differentiated CD4+ T cells (CD4+ T cells are generally considered to be helper T cells) while these CD4+ T cells are absent in people without the allergy. Successful antigen therapy in people with an allergy to grass or trees results in the elimination of these allergen specific terminally differentiated cells.
  • Dr. Greenbaum explained that one difference between antigen therapies for allergies and those for diabetes is that it is unclear if antigen therapy has deleted or anergized (made functionally incapacitated) pathogenic cells in type 1 diabetes. She stated that some evidence exists that parenteral insulin therapy in people at risk for type 1 diabetes decreased proliferation of pathogenic cells to islet antigens. Additionally, some evidence suggests that autologous myelin peptide coupled cells have decreased antigen specific proliferation.

 

Primary Prevention with High Dose Oral Insulin

Ezio Bonifacio, PhD (Technical University of Dresden, Dresden, Germany)

Dr. Ezio Bonifacio presented preliminary results from an exploratory study on using high dose oral insulin as a primary prevention technique for type 1 diabetes. His study, called the pre-POINT study, found that the highest dose used, 67.5 mg, elicited antibody and T-cell activity, suggesting that it was having “some” immune effect. However, it was conducted in an age range (two to seven years) that was older than optimal since most seroconversion (development of autoantibodies) occurs around the six-month to two-year mark. More pre-POINT studies are to be conducted prior to initiating a POINT study, which would aim to be an efficacy study on the prevention of seroconversion in children as young as six months. It doesn’t seem like much progress has been made in the field of using insulin as a primary prevention method, as the results Dr. Bonifacio presented are still quite early stage and exploratory.

 

Panel Discussion

Dr. Matthias von Herrath (VP and Head, Seattle Diabetes R&D Center, Novo Nordisk, Seattle, WA/La Jolla Institute for Allergy and Immunology, La Jolla, CA): Can we arrive at some form of a consensus on what we need to do for type 1 diabetes with antigens that might be different from allergies?

Dr. Bart Roep (Leiden University Medical Center, Leiden, The Netherlands): There is no single way that you can say. Carla showed the difference between type 1 diabetes and allergy. With diabetes we have the allergen in the body. We are continuously reminding the immune system of that allergen. I think you have to take the immune response signature you see in patients to see what you can change with the various strategies that you select.

Dr. Carla Greenbaum (Benaroya Research Institute at Virginia Mason, Seattle, WA): The challenge of where we are now as opposed to a decade ago is that we know a lot about the heterogeneity of the disease. If we are going to do small proof of concept studies on the mechanism we have to start with the robustness of the biomarker. We are seeing how reproducible a favorite assay is. Once we understand that we can power these studies. Otherwise I think we are at risk of going down the wrong path.

Dr. Ezio Bonifacio (DFC-Center for Regenerative Therapies Dresden, Dresden, Germany): Mark Peakman can do the ELISPOT reproducibly; he is experienced with it. If I do it, I will not get it to work; I don't have the experience with it. This is a difficult thing. You chose your assays because you know you can do them and you have faith in them. This makes it hard to get them reproducible across different studies.

Dr. Von Herrath: Overcoming these kinds of things can depend on the assay and how easy it is to do. But as Carla mentioned some of these things change over time. It takes an intensive community-wide approach. It is difficult to see this type of thing through. It is difficult to validate these assays without substantial resources. It is not a particularly easy situation.

Q: My question is for Professor Bruschard – there have been recent studies that show that changes in diet rapidly and reproducibly change the gut microbiota toward cardiometabolic health and that administration of the associated protective gut strains can sometimes achieve the same effect. So how do you examine the gut microbiota of type 1 diabetes patients on a gluten free diet?

Dr. Karsten Bruschard (Copenhagen University, Denmark): We haven’t done it in patients, but of course it can be done, and I know they are starting with sequencing the material and trying to figure out what is popping up and down.

Dr. Roep: We are doing fecal flora transplantation in the Netherlands in type 1 diabetes now where we actually replace the flora from diabetic patients with that of a healthy individual, and that is the most quick and dirty way to do this I guess, but I’ll keep you informed on that.

Q: How many do you do and when will it be available?

Dr. Roep: We’ve now randomized five. It’s slow.

Q: What is the outcome you are looking for?

Dr. Roep: The outcome? What are you looking for? [laughter] We are measuring all kinds of immune responses and beta cell function, but it is just an exploratory study.

Dr. Steve Gitelman (University of California, San Francisco, CA): Most of us think of this as mostly a T cell mediated disease. I am wondering if you think there is some inherent defect in the dendritic cell?

Dr. Roep: There have been some studies addressing this and there is some controversy. In humans some studies say there are some defects and some say there are not. The truth of the matter is you have to look at the function of those cells and that is a bit of a challenge.

Dr. Marian Rewers (University of Colorado, Aurora, CO): Dr. Roep you showed a single patient that was slow to progress to diabetes. Is anyone working on reverse engineering from observations like this outlier to see how Mother Nature is protecting people? I think maybe we could use reverse engineering instead of seeing if some new cocktail works.

Dr. Roep: Patients differ and disease differs and if we can understand these differences we will then understand what is happening. You have to take the baseline signatures in mind and see what the effect is of an intervention and how it correlates with the disease. But you are right, we have to pay attention to the outliers and learn why they are outliers.

Dr. Greenbaum: There is a lot to be learned from single antibody people who do not progress to diabetes and double antibody people. I think it will be important to see how markers change over time.

Dr. Bonifacio: The genes associated with the age of sero-conversiont are the ones that are stratifying this more than anything else.

Q: A question on the pre-POINT trial – you showed that seroconversion really occurs before age two in may individuals. Is it feasible to do your trial again, which you conducted in children age two to seven, at an earlier age? Secondly, when are you getting to the POINT trial rather than pre-POINT?

Dr. Bonifacio: Yes we’ve been talking to regulatory authorities, pediatricians, etc. regarding going to age 6 months for 6 months of treatment. There is general consensus that we can move the oral insulin to that as the next step. We are working with them to define what that mechanistic validation protocol will be for safety etc., first off. For POINT, the study would be an efficacy trial at that younger age to prevent or reduce the seroconversion incidence. We’ve modeled that on 1,200 kids so it’s rather a major trial. What that will be we don’t know. We want to look at other ways of delivery to see if we can get a more robust immune response, and we are also planning to do more pre-POINT studies.

Q: What did the dose escalation show?

Dr. Bonifacio: It looks like there is a significantly higher response on the 67.5 mg highest dose.

Q: Is the conversion time course of the group that is followed from birth and the group that is followed from TrialNet parallel?

Dr. Greenbaum: One of the challenges of TrialNet is they are screening relatives at the cross-sectional time of when they come in. We don't know if they are positive from birth. We don't have that data.

Comment: I am asking you to speculate.

Dr. Greenbaum: We don't see many people convert after the age of 13 or 14 year. If I was guessing, I would say that there is a wave that seroconvert as babies who get diabetes pre-pubertal and then there is a second wave who seroconvert in childhood and get diabetes later. That is hand waving though.

Dr. Bonafacio: I think there is a period in the early years where something is happening relatively often and then you can get it at any age thereafter but at a lower frequency. Anytime that you take a child up to the age of 15 years and you see multiple antibodies there is always a 11% risk they will get diabetes each year that often. What that means is that it is not 11% risk plus 11% risk. It is every year there is another 11% of that population who get diabetes, so it never gets to 100% of them getting diabetes.

Dr. Greenbaum: We don't know if the people we have identified with a single antibody at 16 years are the ones coming in with diabetes at 55 years. We don't have that data. But it does appear to be a constant risk even if you are taking them at a cross-sectional period of time.

 

-- by Hannah Deming, Jessica Dong, and Kelly Close