American Diabetes Association 74th Scientific Sessions

June 13-17, 2014; San Francisco, CA – Type 1 Diabetes Therapies (Cure Related) – Draft

Executive Highlights

Encouraging news emerged from the hunt for a type 1 diabetes biological cure, though with it came the question of what is an appropriate level of risk for a type 1 diabetes cure. Dr. Francesca D'Addio (Harvard Medical School, Boston, MA; and San Raffaele Hospital, Milan, Italy) presented a series of studies on the controversial potential of autologous hematopoietic stem cell transplantation (AHSCT), also known as the Brazil Cocktail, to cure newly-diagnosed type 1 diabetes patients (124-OR). The therapy yielded some striking benefits: insulin independence was achieved in all patients (n=65) at least once during the trial, and persisted for an average of 18 months. However, such benefit did not come without cost - the frequency of severe side effects was fairly high (34 out of 65 patients experienced an AE), and the list included one death due to sepsis that may have been related to the therapy, as it is immunocompromising. While the results of this trial suggest that type 1 diabetes, indeed, can be cured, its unclear if the risks are palatable. At the 2013 Rachmiel Levine Symposium, the highly respected Dr. David Harlan (University of Massachusetts School of Medicine, Worcester, MA) argued that new therapies must be quite safe in order to have a positive benefit:risk profile, since the excess mortality associated with type 1 diabetes is down to ~2% over 20 years. In the hunt for an efficacious but safer “Brazil Lite” cocktail, Dr. Michael Haller (University of Florida, Gainesville, FL) presented positive phase 2 results for a combination of antithymocyte globulin (ATG) and granulocyte colony stimulating factor (GCSF), two of the four components of the Brazil Cocktail, in people with recent onset type 1 diabetes (173-OR). If this combination holds up in later studies (a phase study in people with new onset type 1 diabetes is to begin this summer), than it could offer a compromise on the high efficacy and serious risks of the Brazil Cocktail. Alternatively, Dr. Stephen Gitelman (UCSF, San Francisco, CA) presented optimistic phase 1 results for a T-regulatory cell expansion therapy – a phase 2 study for this therapy is also in development (174-OR). 

Below you’ll find detailed reports on all of these topics and more. Presentation titles highlighted in blue were not previously published in our daily highlight reports during ADA. We’ve highlighted in yellow presentations we found particularly notable.


Table of Contents 

Oral Presentations: Novel Therapeutic Agents

The Worldwide Experience with Autologous Non-Myeloablative Hematopoietic Stem Cells Transplantation in New-Onset Type 1 Diabetes (124-OR)

Francesca D’Addio, MD, PhD (Harvard Medical School, Boston, MA & San Raffaele Hospital, Milan, Italy)

Dr. Francesca D’Addio presented a series of studies conducted in Poland and China on the controversial potential of autologous hematopoietic stem cell transplantation (AHSCT) to cure newly-diagnosed type 1 diabetes patients. The tested question is a version of the Voltarelli Cocktail. The approach uses high doses of immunosuppressants (cyclophosphamide and antithymoctye globulin [ATG]) to largely wipe out the immune cells. Patients are then “rescued” with the AHSCT, and the hope is that the immune system will redevelop without islet autoimmunity. The therapy, which was tested in 65 patients in three sub-studies, yielded some fairly striking and, more importantly relatively long-lasting benefits (for context, past biological approaches to curing type 1 diabetes have only had a temporary affect, if any). Insulin independence was achieved in all patients at least once during the trial, and persisted for an average of 18 months. Over 32% of patients remained insulin independent until their last follow-up (not disclosed). This efficacy was backed up by significant decreases in insulin requirements, significant and sustained increases in C-peptide, and a massive sustained mean A1c reduction from 10.1% to 6.0%. Such benefit did not come without cost – the frequency of severe side effects was fairly high, and the list included neutropenic fever, alopecia, fever, and one death due to sepsis that may or may not have been related to the therapy, which is immunocompromising (34/65 patients experienced an adverse event).

  • Patient enrollment:  This study enrolled patients at three centers in Poland and China. People had at least one autoantibody (most were GAD-positive), no complications, and age 12-35 (mean age of 20). At the Joslin Symposium, Dr. Jay Skyler (University of Miami, FL) critiqued prior trials of this approach, stating that participants in the trials were not well characterized leaving some doubt about their disease state. In line with this observation, yesterday’s oral included relatively little details on the patients’ baseline characteristics.

Age (years)

20.4 years

% Male


BMI (kg/m2)


DKA or DK history


     - No DKA (%)

43 (66%)

     - DKA (%)

21 (32%)

     - DK (%)

1 (1.5%)


  • Treatment protocol: People were treated with a three-step AHSCT regimen. The first step, hematopoietic stem cell (HSC) mobilization, involved treatment with cyclophosphamide (an immunosuppressant) and granulocyte colony stimulating factor (GCSF, which stimulates cell growth, particularly of regulatory T cells). Cells were harvested and frozen. The next step was a preconditioning process with additional cyclophosphamide and rabbit antithymocyte globulin (ATG; an immunosuppressant). Finally, the CD34+ hematopoietic stem cells (which can become many different components of the blood, including immune cells) were infused into the patient.
  • Following the therapy regimen (which includes large doses of immunosuppressants), patients’ immune systems largely recovered. Neutrophil levels stabilized following a spike, while white blood cell counts stabilized, albeit at what appeared to be a lower level than baseline.
  • Insulin independence was achieved in the majority of patients at six months post-transplant. According to Dr. D’Addio, all patients achieved insulin independence at least once, ~60% were insulin-independent at six months post-transplant, and nearly 40% of patients remained insulin independent. A substantial A1c reduction accompanied this improvement – on average, patients went from a mean baseline of ~10% down to a mean of ~6.0% at 12 months, with only a slight rebound afterwards. C-peptide levels increased significantly and stayed elevated through 48 months (a positive sign given that many other biological approaches to type 1 diabetes have only shown a temporary effect, if any). Results are to be published in Diabetes shortly.
  • The team represented by Dr. D’Addio’s categorized patients as responders or non-responders – as we understand it, responders were those who remained insulin independent at six months. Responders had a much more robust reduction in A1c (~8%) compared to non-responders (~4%). Responders also had more than double the average increase in C-peptide, and around a third the incidence of diabetic ketoacidosis. Predictors of being a responder included receiving a greater number of cells during infusion as well as a more recent diagnosis of type 1 diabetes (15 - 20 weeks between diagnosis and infusion was the average in the responder pool). Looking for a responder pool with AHSCT seems like it may be a wise idea, as the therapy involves a relatively high incidence of concerning side effects.
  • Although the efficacy demonstrated by AHSCT in this trial was certainly striking, loss to follow-up could have introduced a bias. Dr. D’Addio presented a chart representing each of the 65 patients as a horizontal bar, showing the length of their study enrollment as well as their insulin dependence status. It appeared that a number of patients did not make it to 48 months, and if patients that saw poorer results were more likely to discontinue enrollment, it could have introduced a degree of bias. However, loss to follow up is factor in any clinical trial. 
  • AHSCT was associated with a number of adverse effects, some serious. A full 34 of the 65 participants experienced at least one adverse event; in terms of serious events, the incidences of neutropenic fever, alopecia, gastric tract symptoms, and rashes were fairly high, although it was difficult to quantify the exact incidence of each from the table Dr. D’Addio presented. There was one death due to sepsis, which could be treatment related given the approach substantially weakens the immune system. During Q&A, Dr. D’Addio also acknowledged that cyclophosphamide has been linked with infertility, although this was not observed during the study. Dr. D’Addio acknowledged that the incidence of severe side effects would likely limit this therapy to selected individuals with type 1 diabetes, unless safer HSC-based options are developed. 
  • The appropriateness of the Voltarelli cocktail has been hotly debated at past meetings on type 1 diabetes cure research. At the 2013 Rachmiel Levine Diabetes and Obesity Symposium, Dr. Richard Insel (JDRF, New York, NY) stated his belief that the original Voltarelli cocktail would not be safe enough for use in children and adults (he questioned whether attendees would be willing to put their own children through the treatment), and acceptable in the eyes of regulatory agencies. Instead, Dr. Insel encouraged researchers to take the principles from the Voltarelli cocktail and learn how to apply them safely. At the Joslin Symposium in 2013, Dr. Jay Skyler (University of Miami, FL) called the Voltarelli cocktail a “heroic” attempt to fight the disease given its relatively high mortality risk.
  • As a result of this approach’s concerning safety signals, the risk/benefit assessment might limit it to very specific type 1 diabetes patients, at least until safer variations of the process are perfected. Several such approaches are already in development – for example, at 8 AM this morning, in W-2006, Dr. Michael Haller (University of Florida, Gainesville, FL) will present positive results for low-dose antithymocyte globulin (ATG) in combination with granulocyte colony stimulating factor (GCSF); an approach that has been dubbed “Brazil Lite” (in reference to it being a less intense version of the Voltarelli cocktail).

Questions and Answers:

Q: Can you tell me more about the transplant procedure?

A: Patients underwent the initial steps, and once cells reached certain numbers, cells were then collected and frozen. Patients underwent a conditioning regimen using cyclophosphamide and ATG, then the cells were thawed and infused into the patients. Patients were also treated with antibiotics, which is standard procedure for stem cell transplantation.

Q: Where do the cells go once they are infused?

A: We don’t know. We could not track the cells. We would like to know where they go – it would be interesting to check.

Q: Could you talk more about the side effects of the transplant procedure?

A: The adverse events that occurred mostly occurred in the first 12 months. The only thing I can say is that some of the compounds used in this procedure have some toxic qualities. Another trial published by a group in Brazil has showed that cyclophosphamide can induce infertility in the long term. We have not observed that so far.

Q: Did you check if antibody positivity determined who was a responder?

A: We did not check if there was a correlation, but that is a good suggestion.

Oral Presentations: Advances in the Diagnosis and Treatment of Type 1 Diabetes

Combination Low-Dose Antithymocyte Globulin (ATG) And Granulocyte Colony Stimulating Factor (GCSF) Preserves Beta-Cell Function In Patients With Established Type 1 Diabetes (T1D) (173-OR)

Michael Haller, MD (University of Florida, Gainesville, FL)

Dr. Michael Haller presented positive efficacy results for low-dose antithymocyte globulin (ATG) and granulocyte colony stimulating factor (GCSF) in patients with recent onset type 1 diabetes (four months to two years). In the single blinded trial (n=25) people were randomized to either 2:1 ATG-GCSF (2.5 mg/kg ATG over two days and 6 mg GCSF every two weeks for 12 weeks) or placebo, and subsequently followed for 12 months. Dr. Haller noted that the trial was in a later stage of type 1 diabetes, than they normally would have enrolled for such a treatment. This decision was made, because one of the trial’s sponsors, the Helmsley Charitable Trust, wanted to re-enfranchise people who fall outside the new onset window, typically tested (the other sponsor was Sanofi). In light of the trial population, the efficacy results look promising. The treatment arm’s two-hour C-peptide level was preserved over the 12 months at ~2 ng/ml (baseline was 2.14 ng/ml). In contrast, the placebo group experienced a significant C-peptide decline to ~1 ng/ml (p-value for the difference between the two groups = 0.05). Unfortunately, this preservation of C-peptide was not concomitant to either a better A1c or a lower insulin dose, as compared to the placebo group. On the safety side, Dr. Haller seemed disappointed in the rates of adverse events: 14 of the 17 participants in the experimental arm experienced cytokine release syndrome (a side effect of anti-T cell antibody infusions like ATG, producing a systemic inflammatory response) and/or serum sickness. With the presentation of preliminary positive results for both the Brazil cocktail and ATG/GCSF, an emerging question is what level of risk patients (and the FDA) is willing to accept in a potential type 1 diabetes cure. 

  • Dr. Haller called the ATG/GCSF combination “Brazil Lite,” in reference to it being a (less intense) modification of the Brazil Cocktail (aka the Voltarelli Cocktail). The Brazil Cocktail uses high doses of immunosuppressants (cyclophosphamide and antithymoctye globulin [ATG]) to largely wipe out the immune cells. Patients are then “rescued” with an autologous hematopoietic stem cell transplant (AHSCT). The hope is that the immune system will redevelop without islet autoimmunity. Indeed, on Saturday, results were published on this approach showing that it can lead to sustained insulin independence. However, this treatment also have substantial risks – in the study presented on Saturday (n=65), a participant died from sepsis, potentially due to the severe immunosuppression. Thus, researchers have been working to determine which components of the Brazil cocktail are necessary and sufficient for efficacy, and which can be omitted to improve safety. Previously, the phase 2 START trial found that ATG (an anti-T cell antibody infusion) monotherapy failed to significantly delay two-hour C-peptide AUC decline over 12 months in people with recent onset type 1 diabetes, indicating that ATG alone was not sufficient (though the ATG/GCSF trial was already ongoing when these negative results were announced).
  • This trial was single blinded and 2:1 randomized 25 people with recent onset diabetes to ATG-GCSF (2.5 mg/kg ATG over two days and 6 mg GCSF every two weeks for 12 weeks) or placebo. Participants had type 1 diabetes for between four months and two years at baseline. Dr. Haller noted that patient population had a later stage of type 1 diabetes (people had type 1 diabetes for four months to two years, rather under two months as has been more commonly selected in recently trials), than they normally would have enrolled for such a treatment. This decision was made, because one of the trial’s sponsors, the Helmsley Charitable Trust, wanted to re-enfranchise people who fall outside the new onset window, typically tested (the other sponsor was Sanofi).
    • The dose of ATG used in this trial was one of the lowest seen in immunotherapy work, according to Dr. Haller. For comparison, in the START trial the ATG dose was of 6.5 mg/kg. The researchers felt they could use a lower dose of ATG, due to the concomitant administration of GCSF. In rodents, they found that they could reverse diabetes using one-third the successful monotherapy dose, when ATG was given in combination with GCSF.


ATG/GCSF arm (n=17)

Placebo (n=8)

Male (%)



Age (years)

23.64 years

23.55 years

Time from diagnosis (years)

1.04 years

0.93 years

AUC 2-hr C-peptide (ng/ml)

2.14 ng/ml

2.13 ng/ml




Daily insulin dose (U/kg)

0.44 U/kg

0.45 U/kg

  • The treatment arm’s two-hour C-peptide level was preserved over the 12 months at ~2 ng/ml (baseline was 2.14 ng/ml). In contrast, the placebo group experienced a significant C-peptide decline to ~1 ng/ml from a baseline of 2.13 ng/ml. At month 12, the p-value for the difference between the two groups was 0.05.
    • At this point, it is unclear why ATG/GCSF successfully preserved C-peptide, while ATG monotherapy did not. One hypothesis for why ATG/GCSF was successful where ATG was not is that ATG/GCSF better preserved levels of regulatory T cells (that are thought to restrain autoimmunity). At the 2014 Rachmiel Levine Diabetes and Obesity Symposium, it was hypothesized that the approach might have failed because it worsened the ratio of regulatory T cells and effector T cells (that are thought to drive autoimmunity). However, with ATG/GCSF it appeared that Treg levels were better preserved. This could potentially be the result of the lower ATG dose (such that fewer Tregs are destroyed), or that GCSF promotes the production of new Tregs.
  • Unfortunately, this preservation of C-peptide was not concomitant to either a better A1c or a lower daily insulin dose, as compared to the placebo group. The experimental group’s mean A1c increased 0.52% from a mean baseline of 6.69%.  This was not statistically significantly different from the placebo group’s mean A1c increase of 0.97% from a mean baseline of 6.03%. The only time point at which the A1c difference significantly differed (in favor of the treatment group) between the two groups was at six months. Likewise, the two groups’ course of insulin use did not significantly differ (baseline insulin doses were 0.44 U/kg and 0.45 U/kg for the treatment and placebo groups, respectively). As an attendee mentioned during Q&A, it would have been interesting to see if time in zone was improved in the treatment arm, even though A1c and insulin dose were not. Though there are many potential explanations for the discrepancy between the C-peptide, A1c, and insulin results, it is possible that the beta cell destruction was already too great at baseline for these measures of glycemic control to appreciably improve (indeed, a C-peptide level of ~2 ng/ml, while potentially clinically beneficial over a lower level, is still quite small). If this is the explanation for these results, we would be interested in seeing if a beta cell regenerative agent could be added to this combination to stimulate the growth of a new beta cells ­– of course, this first requires the identification of such an agent.
    • Participants of this trial are being followed for a total of five years – we are curious whether the C-peptide results are sustained over the long-term, and if significant differences in either A1c or insulin dose arise. All participants will have passed the two year mark in December 2014.
  • Though the rate of adverse events was better in this trial than it was in the START trial (of ATG monotherapy), Dr. Haller seemed disappointed in the relatively high rates ­– 14 of the 17 participants in the experimental arm experienced cytokine release syndrome (a side effect of anti-T cell antibody infusions like ATG, producing a systemic inflammatory response) and/or serum sickness.
  • A trial of ATG/GCSF in people (ages 12-45 years) who have been newly diagnosed with type 1 diabetes is to begin in the summer of 2014. People in this trial will have been diagnosed within 100 days. The trial is being run with TrialNet, and is sponsored by Sanofi, the Helmsley Charitable Trust, Amgen, and the NIH.

Questions and Answers

Q: Congratulations, did you have a chance to look at autoantibodies?

A: So far we have not seen any difference in autoantibody titers.

Q: Do you have CGM data in this trial?

A: I think that is a great idea for the next study. Patients certainly reported that they had less variability.

Phase I Trial Using Ex Vivo Expanded Polyclonal Tregs for Recent-Onset Type 1 Diabetes (174-OR)

Stephen Gitelman, MD (UCSF, San Francisco, CA)

Dr. Stephen Gitelman detailed data from a JDRF-funded phase 1 clinical study on the safety and tolerability of UCSF’s autologous T regulatory cell (Treg) immunotherapy for recent-onset type 1 diabetes. The study enrolled 14 patients between the ages of 18 and 45 (mean of 30 years) with disease duration of three to 24 months (mean of 10 months). The process used to harvest, grow, and purify Tregs for infusion was developed by the renowned Dr. Jeffrey Bluestone (UCSF, San Francisco CA) (Putnam et al., Diabetes 2009). The primary goal of the study was to confirm the approach’s basic safety and tolerability, which it accomplished: the Treg infusions were well tolerated, with no infusion site reactions and no significant infectious issues. The Tregs themselves peaked in circulation during the first week of infusion and diminished from then on, though they remained detectable more than six months after the initial fusion. Full two-year data was only available for the six patients who received the smallest cellular doses. In these groups, C-peptide levels and A1c remained fairly stable over the entire time period, with perhaps a slight upward trend in exogenous insulin usage. The incomplete data from patients who received larger doses of Tregs was more variable, but it was too early to tell, and the study is not sufficiently powered to examine efficacy. Dr. Gitelman believes that the strong safety and tolerability profile demonstrated in this trial (especially relative to harsher immunotherapies therapies out there) sets the stage for phase 2 testing. Notably, NeoStem has licensed the Treg technology from UCSF, and is partnering with UCSF and TrialNet  for phase 2 development.

  • Dr. Gitelman first provided background on the Treg expansion therapy studied in the trial. The progression of type 1 diabetes is believed to stem from a relative overabundance of effector T cells (Tems) that lead the autoimmune attack and a lack of Tregs, which suppress Tems activity. The idea behind the therapy is that a Treg “boost” could restore immune balance; results in preclinical models have been promising. Dr. Gitelman suggested that this approach could avoid the toxicity and side effects of other T-cell-targeted therapies such as ATG/GCSF and anti-CD3 antibodies.
  • The ability to successfully grow and purify Tregs was a prerequisite to the successful implementation of this therapy. Using a protocol developed by Dr. Bluestone’s group (Putnam et al., Diabetes 2009), Dr. Gitelman’s team was able to achieve over 500x expansion of the cell population in two weeks. The approach uses flow cytometry to sort and purify the sample. 
  • The phase 1 trial set out to establish the basic safety and tolerability of a single dose of the Treg infusion therapy. Enrollment was limited to type 1 diabetes patients who were three to 24 months from diagnosis, age 18 to 45 years, had at least minimal levels of C-peptide (>0.1 pmol/ml following a meal test), and did not have any diabetes complications. The group of 14 patients was divided into four cohorts, each of which was given a different cell dose (ranging from 5x106 to 2.6x109 cells, ordered from cohort #1 through #4 in order of increasing dose). The study is scheduled to last five years, and the results presented yesterday were for two years follow-up (for cohorts 1 & 2) and one year follow-up (for cohorts 3 & 4).
  • The study results confirmed the therapy’s safety across the range of doses. The infusion was well tolerated, with no infusion site reactions. There were no significant infectious issues, and the majority of events were not related to the treatment.
  • The injected Tregs peaked in concentration approximately three to seven days following administration in the two higher dose cohorts, and were still detectable six months following injection (albeit at greatly reduced concentrations). Deuterated glucose was used to track the Tregs. The gradual loss of signal might be due to a decrease in Tregs, which would mean that the therapy would need to be regularly readministered. However, the signal disappearance could also be due to cell division, cell trafficking out of circulation, or another process that might not necessarily lead to a cessation of Treg action.
  • Dr. Gitelman next presented C-peptide data – it was hard to conclude much from this data, as the study size is quite small. In cohorts 1 and 2, C-peptide levels and A1c appeared to remain fairly stable over two years. Overall subjects in these groups were well-controlled, with perhaps a slight uptick in insulin dose. One patient in each of the higher dose cohorts had poor glycemic control, which made it harder to interpret the one-year results from those cohorts, especially given the small sample size and short duration.
  • Cells were isolated and expanded at UCSF then shipped to Yale. This speaks to the stability of the cell population, and could make it both easier to conduct large trials and contribute to the therapy’s commercial viability (if the therapy reaches the market down the road). 
  • The safety data, Dr. Gitelman stated, paves the way for a phase 2 study, which will be a partnered effort with TrialNet and product licenser NeoStem. Dr. Gitelman mentioned in his conclusion that at least one other group (Trzonkowski et al., Diabetes Care 2012, Clin Immunol 2014) has been working on Treg therapy in pediatric type 1 diabetes patients, and has found similarly acceptable tolerability as well as hints of efficacy.

Questions and Answers:

Q: So far, we have been disappointed with findings from cord blood transfusion studies. Why would you think that this would work better, especially if you are not creating space by ablating before you infuse?

A: One limitation of cord blood is the number of Tregs that are infused. We need to think of another way to enrich it before it goes in vivo. We’ve discussed this at UCSF.

Dr. Michael Haller: We are working on expanding studies of Tregs in cord blood now. The hope is to do a pilot study, with results that could be at least as good or even better than these.

Oral Presentations: Clinical Pancreas and Islet Transplantation

Continuous Glucose Monitoring Shows Superiority of Islet Transplantation to Intensive Insulin Therapy in Type 1 Diabetes (80-OR)

Peter Senior, MD, PhD (University of Alberta, Edmonton, Canada)

Dr. Peter Senior presented the results of a cohort study comparing glycemic variability following islet cell transplantation versus conventional intensive insulin therapy. In the study, 17 patients with type 1 diabetes after islet cell transplantation and 12 patients on intensive insulin therapy underwent one-week of blinded continuous glucose monitoring. Results indicated improved glycemic variability in patients following transplant, with significantly reduced percent time above (>180 mg/dl) and below (<54 mg/dl) the target range. Notably, patients undergoing transplant were older (mean age 55 vs. 28 years) with longer diabetes duration (37 vs. 13 years), suggestive of islet cell transplantation conferring a benefit even when a person has worse baseline disease.

  • In this prospective cohort study, glycemic variability was compared between patients with type 1 diabetes following islet cell transplant (n=17) and patients on intensive insulin therapy (n=12; n=4 on MDI and n=8 on CSII). Patients underwent blinded continuous glucose monitoring for one week with instruction to continue usual diet and activities. Overall at baseline, patients following islet cell transplant were older (mean age 55 vs. 28 years) and with longer diabetes duration (37 vs. 13 years) compared to patients on intensive insulin therapy. Patients additionally had reduced insulin requirements (0.23 units/kg/day vs. 0.59 units/kg/day) and lower baseline A1c levels (6.3% vs. 7.4%), though with no significant difference in Clarke hypoglycemia scores. Mean time since transplant was 65 months.
  • Results indicated no significant difference in mean sensor glucose values in patients following islet cell transplant versus patients on intensive insulin therapy, though with reduced percent time above (>180 mg/dl) and below (<54 mg/dl) target range. Additionally, glycemic variability was significantly reduced as measured via standard deviation from the mean, coefficient of variation, glycemic lability index, and mean amplitude of glucose excursions (MAGE), all measures of glycemic variability. Subgroup analyses by insulin independence of patients in the islet cell transplantation group (n=8 insulin-independent vs. n=9 insulin dependent) indicated no difference in percent time above or below target range though with significant improvement in glycemic variability as measured by standard deviation, coefficient of variation, and MAGE.
  • Dr. Senior suggested these results demonstrated superior glucose control in patients following islet cell transplant versus conventional intensive insulin therapy, regardless of increased age and diabetes duration. He further noted that insulin independence following transplant might not be a prerequisite for these benefits in control. As candidates for transplant typically have poorly controlled baseline disease, these results do imply strong improvement in control for these patients following transplant – though the impact of glycemic variability alone on clinical outcomes remains unclear. Of course, a major unsolved problem is a shortage of islets for transplantation, which would prevent most people with type 1 diabetes from being able to undergo the procedure even if it were ready for primetime.

Questions and Answers

Q: What was the relationship of glycemic variability to C-peptide levels?

A: All the islet cell transplant patients were C-peptide positive. I don’t think there was a relevant relationship of C-peptide levels to glycemic variability. I don’t know the correlation yet – it would require more exploration.

Q: Could you explain the discrepancy between your results for mean glucose values on CGM and A1c?

A: Remember this is a six-day window of data. So I suspect that for instance with the time spent high, if you bill that over three months there would be a significant difference in mean glucose. There’s also probably also an observation effect in the trial such that people know they’re being monitored and improved their habits.

Q: When you put people on a continuous monitor, it takes people a while to get used to it. Were these subjects given any training?

A: This was blinded CGM. The subjects were getting no real-time feedback at all. We wanted to know what was happening without any feedback or change in behavior.

Q: Did the subjects in your intensive insulin therapy group have any residual C-peptide?

A: We do not have that data.

Oral Presentations: ADA Presidents Oral Session

Phase 3 Trial of Transplantation of Human Islets in Type 1 Diabetes (T1D) Complicated by Severe Hypoglycemia (388-OR)

Bernhard Hering, MD (Clinical Islet Transplantation Consortium, Minneapolis, MN)

Dr. Bernhard Hering presented the results of a phase 3 trial on the efficacy of human islet cells transplants to treat type 1 diabetes patients. The prospective, open label study occurred in eight centers and examined 48 adults with reduced awareness of hypoglycemia and who had experienced at least one episode of severe hypoglycemia in the prior year. Patients were 18 to 65 years old, had been diagnosed with type 1 diabetes for ≥5 years, and had no stimulated C-peptide (<0.3 ng/ml). One year after the first islet transplantation, only one incident of hypoglycemia had occurred, and over 50% of subjects were insulin-independent. Insulin use decreased from a baseline median of 0.49 U/kg/day to 0.00 U/kg/day (ending range of 0.00-0.43 U/kg/day). No deaths or permanent morbidities associated with the islet infusion occurred in the study. However, the immunosuppression regime used to protect the graft reduced kidney function, lowering the mean baseline GFR 102 ml/min by 8.1 ml/min. Dr. Hering concluded that islet transplantation could prevent severe hypoglycemia in type 1 diabetes patients for which other interventions have failed.

  • Of the 48 patients, 26 people received two islet transplants, and one person received three transplants. Overall, the mean islet dose for all subjects was 806,586 IEQ, and it was 589,101 IEQ for those with only one transplant. Immediately before or after transplant, etanercept and antithymocyte globulin (ATG) were used for immunosuppression. Sirolimus and tacrolimus were taken as maintenance immunotherapy.
  • Islet function was maintained through the end of the year. At day 75, almost all patients had islet graft function as measured by C-peptide activity (~95%), and islet graft function was similar at day 365 (94%). Serum glucose decreased over time, and at day 365 patients spent an average of 90-95% of their time in range. 
  • There was only one incident of hypoglycemia in the year following the islet transplantation. Measures of glucose variability and hypoglycemia severity, including the Clarke score, Ryan score, and glycemic lability index all decreased (p <0.0002).  Specifically, the median Clarke score decreased from six to zero.
  • No adverse events resulted in death, disability, or permanent harm. In total, there were 19 study-related severe adverse events, of which five were procedure-related and 13 were immunosuppression-related.
  • Dr. Hering concluded that islet transplantation could prevent severe hypoglycemia in type 1 diabetes patients for which other interventions have failed. While multiple interventions, such as continuous glucose monitoring and sensor-augmented pump therapy, are available, the T1D Exchange Registry reports that 35% of type 1 diabetes patients still have at least one severe hypoglycemia event each year. Key obstacles to making islet transplantation more available to people with diabetes are (i) long-term immunosuppression can be dangerous (leading to serious infections and cancers), and (ii) not enough islets are available to perform the procedure on many people with diabetes. Thus, researchers are working to induce immunotolerance (such that immunosuppression would not be required, or would be reduced in intensity) or encapsulate islets so that are protected from immune attack, and to identify new cell sources (i.e., from stem cells or pig islets).

Questions and Answers

Q: Can you explain the Clarke score and the Ryan score?

A: The Clarke score takes into consideration the ability of a patient to recognize hypoglycemia, and assesses the awareness of patients during hypoglycemia. The Ryan score takes into consideration the patient’s ability to diagnose and treat hypoglycemia.

Q: Was there any effect on patients’ awareness during hypoglycemia?

A: We didn’t study awareness restoration in this study, but the Clarke score suggests that awareness improved.

Q: You stated that the number of patients at A1c <7% and the incidence of severe hypoglycemia were the primary endpoints of the study, but you did not report the number of patients who reached A1c <7%. How many reached this target?

A: Sorry, we don’t want to disclose it yet because the study is under review. 

Q: Given that the C-peptide levels at one year are almost normal, why were half the patients still insulin-dependent?

A: We don’t know, but insulin action did improve after islet transportation. I cannot fully explain why a substantial proportion of patients still needed insulin. We were reluctant to discontinue insulin, since it was not a primary endpoint. Also, some previous research has shown that if we don’t stress the transplant too early, we may be able to prolong its function.

Q: Why does it take >100 days before you get a C-peptide response? Is it because you need further transplants?

A: It takes a while for the islets to work. We did some studies at 75 days to see if islet function was restored, and that was before we had time to do the second and third transplants.

Q: Hasn’t it been shown that TNF is needed for proper immunoregulation in type 1 diabetes? Why did you use TNF-alpha blockers?

A: TNF-alpha blockade is important. Previous studies have shown that this induction protocol promotes engraftment, and the insulin and C-peptide secretory response was two to three times higher with this protocol than in those that don’t include T-cell antibodies and TNF-alpha blockade.

Q: Is there an increase in the effectiveness of the implant because the cells are dividing? How long will you follow-up with the subjects?

A: We don’t know yet if the cells are dividing or replicating. In this study, the transplant may function longer, which is how we are interpreting results. We do have a long-term follow-up protocol in place. In this study, islet function seems to improve over time, unlike in other studies.

Oral Presentations: Beta Cell Development and Postnatal Growth

Long-term Immunologically Unrecognized Reprogrammed Beta Cells as a Promising Cure for Autoimmune Diabetes (199-OR)

Xiangwei Xiao, PhD (University of Pittsburgh, Pittsburgh, PA)

Dr. Xiangwei Xiao described a series of experiments whereby alpha cells in mice were converted to beta cells in vivo by the direct infusion of a retrovirus into the biliary duct. In this model, the virus enforced Pdx1 and MafA expression in infected cells. Many cells were subsequently found to be positive for both glucagon and insulin, supporting the interpretation of trans-differentiation. While a control virus did not reduce blood glucose, the Pdx1/MafA virus (AAV8-PM) significantly reduced blood glucose in about a third of the mice, an effect that lasted for three months after treatment. During this period, insulin-positive cells were found in AAV8-PM but not in control virus treated mice. However, after three months, blood glucose rapidly increased again in all mice, possibly as a result of a delayed autoimmune response. Dr. Xiao suggested that the immune attack might be delayed, because the newly developed beta cells have different markers than original beta cells, preventing the autoreactive cells from recognizing them initially. Dr. Xiao and colleagues are working on experiments to further characterize the immune response in this model.

  • An emerging approach to generate insulin-producing cells in the treatment of both type 1 and type 2 diabetes is to reprogram existing duct cells and alpha cells to produce insulin. Conversion of both acinar duct cells and alpha cells to insulin-producing cells has been achieved with forced expression of transcription factors such as Pax4 or Pdx1. Some attention has been focused on alpha cells in particular, as they more closely resemble beta cells than duct cells. Dr. Xiao’s group found that a combination of Pdx1 and MafA expression could shift alpha cells from glucagon to insulin expression in a way that involves down-regulation of several key genes and thus likely represents a true trans-differentiation event.
  • Intraductal infusion of a retroviral vector provides a method for enforcing Pdx1 and MafA in vivo. In this approach, micro-clamps are placed on the common bile duct to block off the liver, and a retrovirus-containing suspension is infused into the pancreatic duct. In mice, a GFP-expressing retrovirus confirmed that pancreatic cells of different types were infected by the method. Other imaging strategies revealed that some formerly glucagon-containing cells had begun to express insulin, and some cells co-expressed insulin and glucagon, possibly indicating cells in transition. Dr. Xiao expressed his belief that this approach may also be feasible in humans.
  • Retroviral treatment increased beta cell mass and lowered blood glucose in mice, and this effect lasted for about three months. Using both an alloxan-induced hyperglycemia model and a non-obese diabetic mouse strain, Dr. Xiao’s group demonstrated that AAV8-PM retroviral treatment partially reversed the loss of beta-cell mass observed in these models. Concurrently, blood glucose decreased to a normal level in about one-third of the treated mice, and this normalization persisted for about three months. However, after three months, blood glucose rapidly increased, which may indicate a late autoimmune response that eliminated newly formed beta cells.

Questions and Answers

Q: Can you explain why the correction fails and blood glucose shoots up again after three months?

A: After three months, beta cells disappear again. I think this means that there is a delayed attack, rather than an absence of attack. There are two aspects to this phenomenon: newly formed beta cells are not exactly the same, and there may be a lack of some antigens that can be recognized by T lymphocytes. Secondly, many internal antigens are adopted and not the same as the original cell’s antigens. Both of these differences can lead to this attack.

Oral Presentations: Beta Cell Development and Postnatal Growth

T3 Has Dual Effects Upon Pancreatic Beta Cells Inducing Markers of Both Maturation and Aging (203-OR)

Susan Bonner-Weir, PhD (Harvard University, Boston, MA)

Dr. Susan Bonner-Weir presented a series of elegant experiments from her lab demonstrating that the thyroid hormone T3 promotes both maturation and aging in beta cells, but through different receptor isoforms. These findings have some implications for manipulation of beta cells, as maturation would tend to increase beta cell function and senescence (deterioration due to aging) would tend to decrease it. Expression of both MafA, an effector and marker of maturation, and p16Ink4a, an effector and marker of senescence, increased in mouse islets, rat islets, human embryonic stem cells, and relevant cell lines upon stimulation with T3. Dr. Bonner-Weir’s group next discovered that the gene for MafA only had a binding site for the THRB isoform, not the THRA isoform, and MafA had more induction by a THRB agonist than a THRA agonist. In contrast, p16Ink4a had more induction with the THRA agonist. Finally, they demonstrated a shift in isoform predominance from THRA in fetal and neonatal stages to THRB in adult mice. Overall, these data show that T3 acts through THRA to promote senescence but acts through THRB to promote maturation, and the distribution of these isoforms may help to determine the response to T3.

  • The thyroid hormone T3 increases expression of the important maturation marker MafA and senescence marker p16Ink4a. T3 is known to be important for the growth and maturation of many tissues, but higher T3 level also correlates with shorter lifespan in humans and other mammals. Questions and Answers

Q: Have you done a global ChIP-Seq to see what other genes may be regulated by the different TRH isoforms?

A: We haven’t gotten there yet, but we are definitely interested in pursuing the question.

Q: Do hypothyroidism and hyperthyroidism affect the loss of proliferation that happens in aging?

A: Christina [in my lab] is really looking at the effects that happen with aging. We don’t have the data, but we’re interested in this too.

Q: How do the receptors target different genes?

A: The thyroid receptors are normally bound to DNA, either with co-repressors or co-activators. We don’t know yet what the differences are between the isoforms, but it probably has to do with the availability of these factors.

Oral Presentations: Advances in the Diagnosis and Treatment of Type 1 Diabetes

ATP/P2X7R Engagement Facilitates the Onset of Type 1 Diabetes (180-OR)

Sara Tezza (Boston Children’s Hospital, Boston, MA)

Ms. Sara Tezza presented ATP/P2X7R as a novel therapeutic approach for delaying the onset of type 1 diabetes. Her research group examined the role of ATP/P2X7R in NOD mice and humans. In vitro studies showed that ATP was released from islet cells following their stimulation by high levels of glucose and cytokines. ATP induces P2X7R expression on T cells, and P2X7R was found to be upregulated in lymphocytes infiltrating islets in hyperglycemic NOD mice. Her group targeted ATP/P2X7R in mice using oxidized-ATP, and found that doing so in young NOD mice can reduce P2X7R expression and delay diabetes. However, this approach does not affect hyperglycemia after the disease has already been established.

  • Islet ATP release is increased in hyperglycemic mice, before they have type 1 diabetes. Upon the onset of type 1 diabetes, this signal disappears. 

Questions and Answers

Q: Does extracellular-ATP (eATP) rise with age?

A: We expect that it does, although we haven’t tested it.

Q: Do other cells besides islets produce eATP upon physiologic stress?

A: Yes.

Q: How specific is the upregulation in memory cells? Is it specific to islet reactive cells?

A: We think it’s islet-specific.


Patients with T1DM Show Reduced but Clinicall Significant Elevations in Glucose in Response to Glucagon Injection in the Presence of LY2409021

CM Kazda, P Garhyan, Y Ding, RP Kelly, TA Hardy, C Kapitza

Lilly presented a study showing that type 1 diabetes patients (n=20 males) receiving a glucagon receptor antagonist (LY2409021 100 mg or 300 mg) were still responsive to exogenously administered glucagon. For background, glucagon receptor antagonists are under investigation as a potential treatment for type 1 or type 2 diabetes. One safety concern about the class is that it may hinder a patient’s ability to recover from hypoglycemia, and this study aimed to test whether glucagon would still be an effective treatment for hypoglycemia in the presence of a glucagon receptor antagonist. The study found that one day after a single dose of LY2409021, intramuscular administration of 1 mg of glucagon in a fasting state resulted in a blunted but clinically significant elevation in blood glucose. As would be expected, the glucose rise in the placebo group was greatest (peaking at 202 mg/dl) followed by the 100 mg group (155 mg/dl) and the 300 mg group (141 mg/dl) from a baseline of ~85 mg/dl. The study did not examine the role of endogenous glucagon response, nor did it test the response during hypoglycemia. We hope to see those studies conducted as well.

Symposium: Immune Cell Modulation of Beta Cell Survival and Regeneration

Macrophages in Islet Injury and Regeneration

Alvin Powers, MD (Vanderbilt University, Nashville, TN)

Dr. Alvin Powers described a model of beta cell regeneration where beta cell loss results from overexpression of an endothelial cell-stimulating factor (VEFF-A), and beta cells regenerate upon removal of this factor. Dr. Powers went on to show that the stimulus responsible for regeneration was a local pancreatic factor, not a circulating factor, and that regeneration depended crucially on the presence of bone marrow-derived macrophages. Gene expression profiling of the recruited macrophages revealed a regenerative expression profile characteristic of alternatively activated macrophages. Furthermore, beta cells stained positive for the proliferation marker Ki67 during regeneration, suggesting that beta cell regeneration was due to replication, not neogenesis, in this model.

Questions and Answers

Q: I was curious as to what you think is attracting the macrophages, for example apoptotic signals, etc. Is there a real clue?

A: First of all, we don’t know. VEGF-A itself can act as a macrophage attractant. As you say, there are dying beta cells, and that could be a stimulus, or endothelial cells might be releasing factors to attract the macrophages. We are working on developing a model without the initial endothelial cell expansion.

Q: The insulin promoter is 1,000 times stronger than the VEGF promoter, so do you think beta cell death could be due to problems related to that?

A: Indeed, death could be due to ER stress and the unfolded protein response. However, there’s a mosaic appearance of VEGF-A expression among beta cells, and we observe that both high- and low-expressing cells die and replicate in our model, so we think the cell death has other causes.

Q: I imagine that your macrophages were producing IL-13. We have discovered that IL-13 has anti-apoptotic properties on beta cells. Have you measured apoptosis?

A: We’ve measured apoptosis by TUNEL, but it’s hard to see any positive staining. We do think the cells are dying from apoptosis based on other observations, though.

Q: What you presented is what is happening in the islet, but what about in the exocrine pancreas?

A: There is very little change because VEGF-A is tightly bound and stays in the local microenvironment.

Q: I was wondering whether your transgene also included the 5’ regulatory region of VEGF-A.

A: No, it did not.

Q: What we have found is that this region is sensitive to glucose regulation.

A: In our model, the mice have normal glucose tolerance and euglycemia, so we don’t think the effects are being driven by glucose levels.

Symposium: Recurrent Autoimmunity and Alloimmune Rejection after Pancreas or Islet Transplantation

Autoimmune Diabetes Recurrence In Pancreas Transplantation

George Burke, MD (University of Miami Miller School of Medicine, Miami, FL)

Dr. George Burke presented specific cases of pancreas-kidney transplant (PKT). Autoantibody conversions occur post-transplant. Dr. Burke observed memory T-cell infiltration in peripancreatic tissue, and insulitis that did not respond to acute rejection treatment. This phenomena is called Type 1 Diabetes Recurrence (T1DR), and it was previously thought that immunosuppression used for preventing graft rejection was sufficient to prevent this condition from occurring. Dr. Burke emphasized that if we are creative in obtaining tissue staining in transplant biopsy, we will get the appropriate lead in understanding the inflammatory process.

  • Dr. Burke introduced the University of Miami Kidney-Pancreas Transplant Program. The program consists of about 500 patients, transplanted from 1990 to present. Of these patients, 95% of them received a PKT. A person having type 1 diabetes is confirmed by C-peptide level pre-transplant. All patients have bladder drainage. The patients’ sera are stored pre-transplant and on follow-up. The program also monitors islet autoimmunity. The rate of acute rejection is very low, and there is very low incidence of viral infection and Post Transplant Lymphoproliferative Disease (PTLD). Ten-year patient survival is 70% (25% with cardiovascular disease), and ten-year pancreas graft survival is 91%.
  • Observations of the program show that T1DR occurs in about 5% of SPK patients. T1DR can occur years after transplantation despite immunosuppression to prevent rejection. T1DR is preceded by autoantibody conversion, and ZnT8 autoantibodies have the strongest association.
  • Antibody conversions in patients with Simultaneous Pancreas Kidney Transplant (SPK) lead to T1DR. Before transplant, 60% of people are autoantibody negative. Dr. Burke concluded that 22% of patients are converted from autoantibody negative to positive following transplant, and that 25% of patients are persistently autoantibody positive (i.e., are autoantibody positive before and after transplant). The presence of autoantibody ZnT8 is associated with the shortest time for development of T1DR.
  • Dr. Burke reviewed approaches to treatment. The anti-T cell approach did not work. The combined anti-T and anti-B cell approach also did not work. Possible treatments to treat T1DR include the usage of anti-IL-17 receptor antibody (brodalumab), anti-IL-12/IL-23 p40 (ustekinumab), anti-IL-6 receptor antibody (tocilizumab), and the usage of IL-1beta receptor blockade (canakinumab).

Impact of Autoimmune Responses on Islet Transplantation Outcomes

Bart Roep, MD, PhD (University of Leiden, Leiden, Netherlands)

Dr. Bart Roep gave an emphatic talk that detailed reasons why he thought reactivation of autoreactive memory T cells (which attack the hosts own cells), and not alloreactivity (the attack of foreign cells), posed the greatest challenge to successful pancreatic islet transplantation. Specifically, autoreactive cytotoxic T lymphocytes (CTLs) targeting beta cell antigens posed the greatest threat. In a comparison of patients who achieved insulin independence after transplantation and those who remained insulin-dependent, the rate of pre-transplant and post-transplant autoreactivity, not alloreactivity, best predicted the outcome. These findings suggest that in order for islet transplantation to be more successful, the autoimmunity of people with diabetes (rather than just the immune response against foreign cells) must be addressed. Dr. Roep went on to say that most drugs focus on avoiding alloreactivity, but a few drugs, like the T-cell depleting antibody alemtuzumab, would also counter autoreactivity. As evidence, Dr. Roep described studies where alemtuzumab yielded favorable results over other immunosuppressants. He encouraged the discovery of synergistic combinations of T-cell targeting agents to achieve more complete suppression of autoreactive memory cells. Finally, he mentioned encapsulation as another way to avoid cell-mediated autoimmunity, a strategy which has been employed in studying human embryonic stem cell-derived allografts.

  • The biggest challenge to successful islet transplantation for type 1 diabetes, according to Dr. Roep, is not alloreactivity, but rather reactivation of autoreactive memory cytotoxic T lymphocytes. To illustrate this point, Dr. Roep presented an image of a new-onset type 1 diabetes patient’s pancreas where insulin was still abundantly present but conspicuously absent adjacent to CTLs. He stated that T cells from the peripheral blood of type 1 diabetes patients are capable of recognizing and destroying HLA-humanized mouse islets. Furthermore, in a study that looked at insulin independence after islet transplant, a low rate of pre- and post-transplant autoreactivity, not alloreactivity, best predicted successful insulin independence. Indeed, in patients where T-cells reactive to the beta cell antigens GAD and IA-2 were identified before transplant, 0% achieved insulin independence, compared to 93% of patients without any T-cell autoimmunity. Dr. Roep even described examples where HLA-mismatched transplants (which would be more prone to alloreactivity but less prone to autoreactivity) actually showed better short-term success. A fascinating experiment in mice provides further evidence that autoreactive CTLs may mediate beta-cell destruction – genetically deleting a single islet-reactive epitope allows implanted islets to avoid CTL reactivation and graft failure.
  • Dr. Roep thinks that alloreactivity must also be considered as a source of graft failure, but he warned that its manifestations can be hard to predict. Classically, alloreactivity presents the most vexing problem with allotransplantation, and it plays a key role in islet transplant. If a proliferative response is observed in a mixed lymphocyte reaction between donor and recipient, then the recipient rarely achieves insulin independence. Moreover, in the case of multiple allografts, a new HLA-mismatched islet can actually precipitate rejection of all the previously unaffected grafts. However, it is also possible for new grafts to promote immunological tolerance and IL-10 induction, leading to a successful outcome.
  • T-cell depletion therapies and encapsulation are some strategies that may address autoimmune memory. Most immunosuppressants typically used for transplantation only deal with alloreactivity, not autoimmune memory. Alemtuzumab is a T-cell targeting monoclonal antibody that has shown some promise in decreasing transplant failure and increasing IL-10 post-transplantation. The overall CD3 counts, including both CD4 and CD8 cells, decrease significantly with alemtuzumab, but effector memory T-cells specifically may not decrease. The T-cell targeting agent alefacept is being tested in the T1DAL trial, and adding other immunosuppressive agents like the IL-1 blocker anakinra may have synergistic effects. Another strategy is to encapsulate the islet cells to prevent access by CTLs. However, complement, antibodies, and cytokines can still reach the implanted beta cells, and inflammatory signals can make them better targets for alloreactive T cells.

Questions and Answers

Q: Assuming the proportion of patients with autoreactive T cells is similar between islet and pancreas transplants, why does it take so long for autoimmunity to appear in pancreas transplant recipients?

A: I fear it is the immediate exposure to circulating autoreactive CD8 T cells when you inject islets into the bloodstream that predisposes them to immune memory activation. If you instead implant a vascularized graft, it takes longer for the cells to reach the beta cells and reactivate. I believe the early events are the most important for determining the immune response.

Q: When we studied CD4+ T-cells, they weren’t matched, but we still saw killing. I don’t think it’s as simple as matched or unmatched.

A: You’re right, CD4+ T cells don’t care about direct recognition. Unmatched transplants may give them a better start, but of course, once the CD4 response kicks in, it doesn’t matter.

Q: With alemtuzumab, there is a much higher incidence recurrence of other autoimmune diseases, more so than with other immunosuppressants in use.

A: Yes, let us not forget about all those side effects, there is no free lunch. Once you target immune memory, you do deal with it, but of course it is not a holy grail.

Q: About the alemtuzumab story, I know that it’s been tried in multiple sclerosis. Do you have any ideas about how complete the effector memory cell suppression has to be for success? Are the failures in the Edmonton trial dependent on what happened with effector memory cells?

A: One thing I would like to point out is that the doses for multiple sclerosis are much, much lower than what is used in islet transplantation. So one would wonder to what level it could just be modulation, or what other role it may play in that. In any case, it would be worthwhile to investigate.

-- by Eric Chang, Hannah Deming, Derek Pham, Wilbur Song, Manu Venkat, Michelle Xie, and Kelly Close