The results of the AbATE clinical trial to slow the progression of type 1 diabetes have been trickling out at conferences and talks for a while now, and so I was pleased to see the whole story finally published last month in Diabetes. The clinical trial, led by Dr. Kevan Herold, was designed to test whether two courses of treatment with a drug called teplizumab, spaced a year apart, were able to prevent the progression of type 1 diabetes in patients who were newly diagnosed.
Teplizumab is an anti-CD3 antibody– that is, a large protein that is able to bind to the CD3 receptor which is located on the outside of T cells. Binding to CD3 is necessary to pass activation signals into T cells, and thus what teplizumab is doing is activating T cells. This may seem counter-intuitive; after all, activated T cells are responsible for the destruction of beta cells that causes type 1 diabetes! Paradoxically, though, instead of resulting in widespread autoimmunity, the net effect of this activation is actually immunosuppression, perhaps because the widespread binding of CD3 without other corroborating immune signals results in the death of many T cells.
However, unlike other immunosuppressive drugs that have been used in clinical trials to treat type 1 diabetes, the main action of teplizumab does not seem to be simply silencing T cells. Instead, animal studies imply that teplizumab somehow induces the activation of regulatory T cells, a subpopulation of T cells that modulate the immune system, perhaps because of the chemicals that are released by T cells when CD3 is widely bound. The regulatory T cells, researchers hypothesize, are able to maintain stability in the pancreas even when teplizumab treatment is ended.
This presumed regulatory activity initiated by teplizumab would explain why teplizumab has proved more successful and promising than other immunosuppressive treatments. Other immunosuppressants like cyclosporin A or CTLA4-Ig have to be continuously given to patients to have lasting effects, resulting in constant immunosuppression and all its unacceptable side effects. Plus, these drugs have a diminishing effect over time; the immune system seems to find a way around direct suppression, and gradually the T cells attack the pancreas again. Mouse studies and previous human trials, however, showed that a single course of anti-CD3 treatment delayed the progression of the autoimmune destruction of beta cells for a year after treatment.
After a year or so, treated subjects caught up with the controls, but the initial success of a short treatment with anti-CD3 antibodies gave researches hope that progression of type 1 diabetes could be delayed. So the researchers began a new study, the AbATE clinical trial, this time with two courses of teplizumab, spaced one year apart.
In the AbATE trial, a total of 77 type 1 diabetics, diagnosed within 8 weeks of treatment, were enrolled and completed enough of treatment to be included in downstream analysis. Patients aged 8 – 30 were recruited, but, notably, the vast majority (94%) were under 18 years of age, with the average age being about 12.5. Of the 77, 52 received a 14-day course of teplizumab initially, and then a second course a year later. Only 40 of the 52 treated patients received the second treatment, with the remaining 12 either having no detectable C-peptide, a peptide that is produced simultaneously with insulin in beta cells, and therefore serves as a proxy for insulin production, or having other complications that prevented administration of the second treatment. Notably, the trial was randomized, but open-label, meaning the patients and the researchers knew who was in the treatment arm and who was in the control. Thus, the control arm did not receive a placebo treatment, but did receive the same intensive clinical support for blood glucose management throughout the two-year course of the trial. Both groups met with certified diabetes educators with the aim of maintaining an HbA1c below 7.5%.
The baseline characteristics (BMI, insulin use, autoantibody presence, HbA1c) of the treatment and control arms were well matched. It is worth noting, however, that after treatment, there were higher rates of adverse events among the treated group, as was expected. The most common adverse event was cytokine release syndrome, a common side effect of anti-CD3 treatment that results from immune signals that the activated T cells release. In cytokine release syndrome, the body essentially flares up, as if responding to a systemic infection, even though none exists. If it gets out of hand, this syndrome can be very severe and even lethal, but in the case of the AbATE trial, all reactions were transient, and passed after the completion of the teplizumab course.
Was an increase in adverse events the only difference between the treatment and control groups? Happily, the answer there is a clear no. And with that, we get to the rather tantalizing results of the trial:
Two years after treatment, teplizumab-treated patients had a 75% higher C-peptide level over four hours (indicating a higher function of beta cells) than the control group. Figure 2B from the paper shows this result, which translates to a remarkable 16–month delay in the decline of insulin production. Accordingly, even though HbA1c values after two years were not significantly different between the two groups, the treatment arm used less insulin to achieve this control, as shown in Figures 3A and 3B from the paper.
Left alone, these results are already extremely exciting. Even though treatment does not prevent diabetes, what diabetic wouldn’t love an extra 16 months of honeymooning? And, more importantly, the fact that we can push back beta cell destruction a little bit opens up the possibility that with continued treatments or combination therapies, we might be able to extend that window of protection.
But the researchers didn’t stop there. Though the treatment arm taken as a single group had a four-hour C-peptide that was 75% higher than the control arm overall, subsequent analysis of individual patient responses indicated that not all treated patients responded equally well to the treatment. The scientists therefore split the treatment group (after the trial was completed, in analysis only): they designated those patients who lost less than 40% of their baseline C-peptide after two years as “responders,” and those patients who lost more than 40% as “non-responders.” Importantly, these were not binary groups; the line at 40% is drawn somewhat arbitrarily, but nonetheless highlights a phenomenon seen in many clinical trials in which some patients respond much better than others, and allows the researchers to begin to figure out what might differentiate good responders from non-responders.
When divided in this manner, the results look even more intriguing. While non-responders as a group (~55% of the drug-treated arm) had declines in C-peptide over the course of the two year trial almost identical to the control arm, the responders (~45% of the drug-treated arm) actually regained some C-peptide at 18 months, and by 24 months, had lost only 6% of C-peptide as compared to their baseline. This is shown in Figure 4B from the paper, where we also see that after two years, responders had three times as much C-peptide as either non-responders or controls.
Why did responders respond whereas non-responders did not? What made responders special? The researchers evaluated a number of different attributes of the patients to try to determine whether there were any systematic differences between the two groups. Responders and non-responders could not be distinguished by age, sex, BMI, duration of disease, autoantibody levels, or variation in treatment protocol. However, insulin use and HbA1c at the start of the trial were both lower for the responders than for non-responders, as can be seen in Figures 5B and 5C from the paper. This implies that though the two groups had been diagnosed for the same amount of time on average, the responders, on average, had tighter metabolic control at the time they started the trial, and this control may have affected the efficiency of the immunomodulatory treatment. (The authors additionally note some differences in the distribution of different types of T cells in the responders peripheral blood versus that of the non-responders, but the differences and corresponding statistical significances are not entirely convincing. The authors mention the cell distribution differences, but do not put too much weight in them.)
This post-hoc analysis of responders leaves many open questions; are these real differences, or just chance trends? If metabolic control really mattered, why did it matter? Were there genetic differences between responders and non-responders? Did responders also maintain tighter metabolic control after the trial, independent of the fact that they were responding better to treatment? Would the same trends hold true if participants were not children? However, given the data available, it is difficult to answer these questions at this time, and we may have to wait for subsequent trials to acquire more data. However, now that researchers have observed what seems to be responders and non-responders, they will be more prepared to test various hypotheses as to what differentiates the two groups.
In any case, the fact that the responders responded so well is thrilling. Two years of delay in the progression of type 1 diabetes! As a woman who eventually plans to have children, that’s huge. And it opens up so many new questions– would continued teplizumab treatment extend that period? Would combining teplizumab treatment with other immunomodulatory agents help? If teplizumab is increasing the numbers of regulatory T cells, can this effect be augmented by low-dose IL-2 treatment , which would deliver an signaling molecule that may instruct these regulatory cells expand? How can we turn non-responders into responders?
All these questions will require further research and trials, both in animals and in humans. And that, of course, requires time. It is important to note, however, that researchers are making progress, incremental though it may be.
That said, if you are like me, you are asking– okay, but what can I do today? There is no single right answer to that, but I do have some recommendations:
* This research project and others like it have a number of different funding sources, with the National Institutes of Health (NIH) often leading the charge. The NIH is in dire straits right now due to severe budget cuts and the sequester. As a result, many researchers are struggling to find funding to study type 1 diabetes or any number of other devastating diseases. Write to your representatives, and tell them how important you think medical funding, and especially diabetes research funding, is! There are many resources, including template letters out there- send your letter today!
* This is the most important thing I will say in this entire article: if you are a diabetic, or the relative of a diabetic, encourage your nieces, nephews, brothers, sisters, and especially your children to start getting screened for type 1 diabetes susceptibility early. TrialNet runs a free national screening program for people with relatives with diabetes, and clinical trials like the one I have discussed here make it clear that there is a huge advantage to knowing you or your children are at risk earlier rather than later. It’s no longer just a depressing but unavoidable factoid; we are reaching the point where knowing a patient is at risk means we can do something about it. So get screened! Get your children screened! Don’t wait until it’s too late!
* And, if you or your loved ones are at risk, ask about clinical trials you can be involved in. The JDRF is a great resource for looking up diabetes-related clinical trials. Part of what makes research take so long is that it takes a long time to recruit qualified patients for clinical trials. Do your part!
Herold KC, Gitelman SE, Ehlers MR, Gottlieb PA, Greenbaum CJ, Hagopian W, Boyle KD, Keyes-Elstein L, Aggarwal S, Phippard D, Sayre PH, McNamara J,Bluestone JA; the AbATE Study Team. Teplizumab (anti-CD3 mAb) treatment preserves C-peptide responses in patients with new-onset type 1 diabetes in a randomized controlled trial: Metabolic and immunologic features at baseline identify a subgroup of responders. Diabetes. 2013 Jul 8.