I'm not sure why, but I wrote this blog while listening to The Amorettes "Everything I Learned, I Learned from Rock and Roll:
This is a treatment with a long history, and this is a very long posting. You will see some marks like [d1] showing that there is more discussion at the bottom. This blog posting benefited from a lot of external review, so I want to thank all of the reviewers. Of course, any remaining mistakes are my own.
Strong Results From a Phase-II Teplizumab Prevention Trial
Teplizumab is a humanized monoclonal antibody [d1] which targets CD3 cells in the immune system in order to lower (or stop) the body's autoimmune response. This treatment tries to prevent type-1, or lessen it's severity, by "turning down" [d2] the immune system's attack on the body's own pancreas cells. This basic approach has resulted in treatments (but not cures) for other autoimmune diseases.
The Clinical Trial
This clinical trial was aimed at a prevention. It enrolled people who had a family history of type-1 diabetes, already had two (out of five) autoantibodies, and were showing abnormal blood glucose measurements after a meal. These are people who have a high chance of showing the classic symptoms of type-1 diabetes within a few years.
The treatment was a 14 day regimen of Teplizumab, given intravenously at a clinic (this is not something you can do yourself at home), and the people were followed for a long time (2-9 years) to see how many were diagnosed with type-1 diabetes.
The results (in terms of preventing/delaying the diagnosis of type-1 diabetes) were very strong, in my opinion:
- 19 (43%) of participants treated with Teplizumab were diagnosed with T1D, versus 23 (73%) of participants on placebo.
- On average, people who got the treatment were diagnosed with type-1 after 4 years. People who got the placebo were diagnosed with type-1 after 2 years.
The graph below is the key finding of the study. The blue lines represent people in the study who got Teplizumab, the red line represents people who did not. In this chart, the higher the line, the better; as the line drops more people are getting type-1 diabetes. Notice that the red lines drops more quickly than the blue line, especially in the first few years.
Graph is from the published paper, and is presented for educational purposes only.
Journal Article: https://www.nejm.org/doi/10.1056/NEJMoa1902226
Human Interest: https://endpts.com/we-kept-at-it-jeffrey-bluestone-plots-late-stage-comeback-after-teplizumab-shown-to-delay-type-1-diabetes/
Clinical Trial Registry: http://clinicaltrials.gov/show/NCT01030861
A little history: in 2018 a company called Provention Bio acquired Teplizumab from MacroGenics, which had previously owned the rights to develop the treatment. As I discuss more in the timeline section, MacroGenics's phase-III trial was unsuccessful, but Provention Bio reanalyzed data from several clinical trials and felt it was worth developing. In March 2019 they started [d3] a phase-III clinical trial. This was all before publication of the data from the "Phase-II Teplizumab Prevention Trial" that I described above.
Provention Bio's trial is a phase-III trial in people with honeymoon type-1 diabetes. This is a large (300 person) trial recruiting people aged 8-17 who have been diagnosed within the last 6 weeks. They will get two treatments, six months apart. Each treatment is 12 daily infusions of Teplizumab. People will be followed for 20 months although the primary end point is at 18 months. They plan to have results in May 2022.
The primary end point is C-peptide production which measures the body's ability to generate insulin. There are several secondary end points for safety, immunology, and indirect measures of insulin production.
It is recruiting now at 15 sites across the US. Too many to list here, but if you live in the US, check out the list in the clinical trial record, give them a call, or send them email. Contact info is 908-356-0514 or firstname.lastname@example.org, and they have plans to start recruiting in 16 more places in the future.
Provention Bio's web site: https://www.proventionbio.com
The clinical trial record for this is here: https://clinicaltrials.gov/ct2/show/NCT03875729
Provention Bio's Plan Moving Forward
Provention plans to commercialize Teplizumab as quickly as possible, in a two step process.
First, the data for the completed phase II prevention trial will be submitted for approval for use in pre-symptomatics (people who test positive for two or more autoantibodies, but who don't yet have the symptoms of T1D). The company calls this "the at-risk population". They plan to do the submission to the FDA by the end of 2020.
Second, if the phase III Honeymoon trial is successful, its data will also be submitted for an additional approval for honeymooners (which the company calls "the newly diagnosed population").
Both steps assume that the FDA will approve a treatment that raises insulin production, even if it does not remove the need for injected insulin. It is generally understood that people with T1D who produce more of their own insulin have all kinds of good benefits. Their BG is "in range" more often, they have fewer DKA episodes, and they have fewer of the serious long term complications of T1D which we all fear later in life. But will the FDA approve a new treatment on that basis? And if they will, how much improvement in insulin production would be required for approval? No one knows, because it has never (to my knowledge) come up before.
C-peptide is an accepted biomarker for insulin production, so the FDA accepts it as a way to measure progress towards a cure. But will they approve a treatment just because it improves C-peptide numbers without going as far as a cure, or without showing specific improvement in long term complications? That is unknown.
Also, the second step assumes that the FDA will approve Teplizumab even though Provention Bio will only conduct one phase-III trial. Usually, two such studies are done. Provention is assuming that the data from previous studies means that the FDA will not require them to do another phase-III trial.
Provention has already started a dialog with the FDA to determine if the data they have is enough for eventual approval. The next major step in that process is a meeting later this year, where they expect that the FDA will agree "in principal" that the amount of data they have is enough. After that, Provention makes the formal submission, and the FDA reviews it, and makes the final decision based on the data actually submitted.
If the first part of the plan works, then Teplizumab would be available in 2021 (that's my best guess). It would be approved for people who have 2 autoantibodies (but doctors might proscribe it in other situations).
Normally, when a person says they hope their treatment will be on the market in 20xx, what they really are saying is if various clinical trials are successful, then they hope it will be on the market in 20xx. However, in this case, the clinical trials are already done. Here, the question is one of FDA approval. In the last 15 years, I've never blogged on a cure focused treatment that had completed all the trials and was waiting for approval in the US. It is a big difference between "scientific risk" (the risk that it just doesn't work), and "approval risk" (the risk that it will not get approved).
Now, this treatment does have a higher than average "approval risk". Those are the risks I discussed above, but it is still a big milestone: to get to a point where a company thinks it already has enough data to get FDA approval for a treatment which preserves beta cells. Now we all just need to wait and see if the FDA agrees.
Delay Or Prevention?
A big, obvious question is this: Is this treatment delaying the onset of type-1 diabetes, or preventing it? That's a big difference, and there is no way to know which it is. The people in the trial will continue to be followed, so we will have more data which might resolve the issue. In the mean time, my opinion (and remember, I'm not a researcher) is that the graph shows delay rather than prevention [d4].
Is More Better?
If you look at the "survival curve" (ie. the number of people who do not have type-1 diabetes) in the study as time progresses you can see that the treatment is most effective in the first year. After one year only about 5% of the treated group had been diagnosed with type-1, while about 45% of the untreated group had been. The next two years were pretty good as well, but in the last 2 years, the effect seemed much smaller, both groups dropping at about the same rate.
So one obvious follow on study, is to repeat the treatment once a year and see what effect that has on the results. If this creates the same effect seen in that first year, but repeatedly over many years, that would be a clear improvement.
The phase-III study is clearly trying to learn from earlier results. Not only do they repeat the treatment once, but their primary end point is 18 months after the first treatment and 12 months after the second. This is when the largest differences between treated and placebo groups were seen in earlier trials, so it is the best time to compare the two.
For many years, it has been generally understood that one possible path to a cure in people with established T1D is to combine a treatment which stopped/lowered the autoimmune attack with a second treatment which encouraged beta cell growth or regrowth. But researching these "combo cures" is harder than you might expect, because there are two unknowns [d5].
However, this result, coupled with previous Teplizumab results, provides strong evidence that Teplizumab is lowering the autoimmune attack in pre-symtomatics enough to have a major impact on the progression of T1D. When combined with Teplizumab's good safety results, this could open the doors to a bunch of "combo trials" in people with established T1D where Teplizumab is paired with various drugs believed to help grow or regrow beta cells. There are several such drugs out there: Harmine, Verapamil, the whole class of DPP-4 inhibitors, and possibly a whole class of non-prescription drugs called PPIs (proton-pump inhibitors) [d6].
On the other hand, in pre-symptomatics, a different type of combo treatment may be effective. Pre-symptomatics have enough beta-cells, so for them Teplizumab could be combined with a second immune modulator so that the combination would be more effective than these results. Treatments like Rituximab, Abatacept, and Alefacept have all shown some beta cell preservation during the honeymoon phase, but not enough to delay/prevent/cure by themselves. However combining them with Teplizumab might result in a stronger effect.
If Provention Bio succeeds in getting Teplizumab approved by the FDA, all of these combo trials will become much easier to run. Instead of two unapproved medicines being tested at once, it will be a standard treatment plus an experimental one, and that is a much more common situation. It is less risky for all involved, and leads to faster approvals.
Discussion Of Prevention (vs. Cure)
I know many people touched by type-1 diabetes don't care about prevention. And I understand that; after all, in a sense prevention is too late for them. But I will continue to care about prevention, even though it will not help my daughter. First, because I think we might learn something from a prevention that might lead to a cure in the future. That is not always true, of course, but it certainly is possible. Second, because someday my daughter may have kids of her own, and a prevention for my grandchildren would be worth a lot to me. Third, just in general, I want the world to be a better place even if my family doesn't personally benefit. And prevention will surely help many people in the future. (Of course, there is a counter argument: that prevention will cause research money aimed at a cure to "dry up", because there will be less and less need of it in the future.)
Why were previous trials unsuccessful (or weakly successful [d7]) but this one strongly successful: Definitions Of Success and Honeymooners vs. Pre-symptomatics
This is a complicated subject, which I cannot do justice in a few paragraphs, but I will try.
The MacroGenics phase-III trial was unsuccessful in that it did not meet its primary end point.
However, this phase-II prevention trial was successful because it did meet its primary end point.
So a central question is: will the next trial (the Provention Bio phase-III trial) be successful? And, central to answering that question is: why was one unsuccessful and the other successful?
One theory is that MacroGenics failed because they chose a bad primary end point. Modern cure focused studies use C-peptide as their primary end point but the MacroGenics study was designed at least 13 year ago, and they used a different primary outcome. If this is the cause of the previous failure, then the Provention Bio study is in good shape because they are using C-peptide. The MacroGenics study did measure C-peptide (it was a secondary end point) and Provention Bio did look at that data, and they do think that it shows success, and bodes well for their study.
An alternate theory is that MacroGenics was unsuccessful because they targeted honeymooners, while the phase-II prevention trial was successful because they targeted pre-symptomatics. If this is the cause of the previous failure, then the Prevention Bio study has a problem, because they are also are targeting honeymooners. The researchers involved believe that there is no fundamental difference between honeymooners and pre-symptomatics, except fewer beta cells. (And I think it is fair to say that this is the mainstream, consensus view of researchers in general, not just the people working on this project.) Fewer beta cells cause the symptoms seen in honeymooners, but the disease process is fundamentally the same in both. So these researchers believe that because the basic mechanism is the same in both, the success of the phase-II prevention study supports the future success of Provention Bio's phase-III study.
History Of Research
Teplizumab has a long, complex history and I cannot give more than a short overview in the space that I have, but here goes. (Note that the dates are based on trial registration and publication dates [d3].)
- Teplizumab is developed in the early 1990s.
- An early clinical trial ran from 1999-2005. It showed that Teplizumab could preserve beta cells. I would describe these results as good, but weak. There is a longer discussion about why I think that here: [d7].
- From 2005-2013 there was a follow up which had similar results. [d8]
- In 2008 Tolerx (in partnership with GlaxoSmithKline) starts two phase-III clinical trials on Otelixizumab in honeymooners which was a similar anti-CD3 monoclonal autoantibody.
- In 2009, MacroGenics (in partnership with Eli Lilly) starts two phase-III clinical trials on Teplizumab.
- In 2009 TrialNet registers the phase-II prevention trial which I'm reporting on here, but it doesn't start recruiting until 2011.
- In 2010, Tolerx publishes long term results from their phase-II trial. They are successful (but not strong).
- Obviously, there was a lot of excitement at this point. Phase-III studies are usually the last step before marketing approval, and there are now four such studies in progress.
- In 2010, disaster strikes: The MacroGenics phase-III finishes and the result is failure of the primary end point. The company cancels the other ongoing studies they are running.
- Also in 2010, Eli Lilly drops Teplizumab.
- In 2011, the other shoe drops: Tolerx's phase-III finishes and it also fails primary end point, and they stop their second phase-III trial.
- Later in 2011 Tolerx goes out of business after spending $150 million trying to bring Otelixizumab to market.
- In 2018, Provention buys rights for Teplizumab from MacroGenics.
- In 2019, Provention starts a phase-III trial in honeymooners.
- Later in 2019, TrialNet reports results from their phase-II prevention trial.
- The research proponents of Teplizumab did not give up even in the face of unsuccessful trials.
- Research takes a long time.
Teplizumab is a new treatment, which has not been FDA approved for use in any other disease, so we do not have any real-world safety data for it (good or bad). On the other hand, it has been tested in 9 different clinical trials, and no serious safety issues have surfaced to date. Teplizumab is an anti-CD3 monoclonal antibody. I know of two other anti-CD3 monoclonal antibodies, Otelixizumab and NI-0401. Both are experimental treatments which never progressed out of development (so no real world experience). Otelixizumab has been tested in multiple clinical trials [d9].
It is important to remember that treatments with names that end in -mab are all monoclonal antibodies. But "monoclonal antibody" refers more to the method used to create them than their effect in the body. They all target different cells within the body. So if you hear of a safety issue for one -mab drug that does not mean that a different one is likely to have a similar problem. It is far more important which type of cell the -mab is targeting. If a -mab targets cell type X and is found to have a bad side effect, that does suggest that other -mabs that also target cell type X might also have the same bad side effect, but it does not tell us anything about -mabs in general.
[d1] Monoclonal antibodies is a method of creating therapies by cloning a single cell that attacks the cell you don't want. You end up with a vast number of identical cells, all of which attack the cell you don't want. By carefully choosing the starting cell, you can "target" the monoclonal antibody to attack a very specific type of immune cell. Because of how they are produced, early monoclonal antibodies were targeted at mouse cells, rather than human cells, and this sometimes caused problems. Later techniques were developed to make the cells more human like, but still not completely human. These were called "humanized" or "chimeric" monoclonal antibodies. Finally, techniques were developed to create fully human monoclonal antibodies, and these are called "human" or "fully humanized". Teplizumab is a fully humanized monoclonal antibody.
There are scores of monoclonal antibodies (of all three types) approved for use in the United States, for a wide variety of illnesses. Different monoclonal antibodies have different safety profiles. Most have names ending in "mab".
[d2] The medical term for this is "modulate". Teplizumab is said to modulate the immune system. But it is important to realize that this modulation is narrowly targeted at one type of immune cell. It is not modulating the whole immune system, but just the broken part. An important part of these clinical trials is to make sure the treatment is stopping the bad parts without hurting the good parts.
[d3] This is the date the clinical trial was registered with the FDA's Clinical Trial registry (clinicaltrials.gov). The the dates in the timeline are the same: first registered dates. Note that researchers have usually been working on the trial for months, even years, before this date to set up the study, but also that patients typically start treatment weeks or months after this date.
[d4] My opinion is based on this: if you look at the first 18 months of data, the placebo group had about 45% diagnosed with T1D, but the treated group only had about 5%. So that means about 40% of the people were helped by the treatment. If those people had been prevented from getting T1D, then the treated curve (the blue one) would never drop below 40%, even years later. However, it does go lower (although just bearly). Therefore, my guess is delay rather than prevention. I would love for future data to prove me wrong.
[d5] These studies are hard to start, because researchers must do work to justificaty each drug and also must consider the safety of the combination of the two drugs. That is more than twice the work of testing a single drug. Also, the results are harder to interpret if the study is not successful. You don't know if it failed because the first drug did not do its job, or the second drug failed, or something about the combination did not work.
[d6] More details of Harmine are here:
[d7] "Weak" Results
This is another long discussion, which I'm trying to make short:
There are three definitions of "success" when it comes to clinical trials. A clinical trial is said to be successful if there are good results from the primary outcome. However, researchers often refer to a study as successful if they learned something new or if they could measure some change. On the other hand, people effected by T1D often consider research successful only if it leads to something useful to them: a cure, prevention, delay of onset, an improved treatment, lower rates of complications, easier management, something tangible. So there is a grey area if a clinical trial shows some change that researchers think might be important, but doesn't really impact people with the disease.
In this blog I try hard to use the term "success" if there is "good data from the primary end point" and "unsuccessful" if there is not. However, if the primary end point is something more of interest to researchers than to patients, I might say it is a "weak success" to show my opinion that the result is important more to researchers than to patients. Similarly, if the size of the change is small, so that a patient might not notice the difference, again I might refer to those results as "weak". Researchers might refer to the second type of results as "not clinically significant" meaning "we can measure a difference, but it doesn't really matter to the person with the disease".
Looking at this study specifically, it found that insulin production stayed constant during the honeymoon phase, or dropped more slowly than in untreated people. But the people all started out with T1D and they all ended up with T1D. Is that a success or a week success? They used less insulin and had slightly better A1c numbers. So is that a success or a weak success? We now believe that this saved insulin production leads to fewer long term complications, although how many fewer is still unknown. Is it a success or a weak success? Different people will have different opinions, but mine, at least right now, is "weak success".
The earlier study is reported here:
[d8] The AbATE study:
[d9] Teplizumab and NI-0401 have not had significant safety issues. Otelixizumab is a little more complex. One of the trials showed increased Epstein-Barr viral loads. However, that study did not have a control group, so there was no way to compare people who got Otelixizumabto people who did not. I eye-balled the data, and there was no obvious relationship between Otelixizumab dose and increased Epstein-Barr loads. The data looked pretty random to me. It appeared that the change in viral loads did not impact the people in the study long term, but most did have flu-like symptoms for a week or two. In any case, that was a phase-II study, and when the larger phase-III study was run (which did have a control group), there was not a safety issue.
publicjoshualevy at gmail dot com
All the views expressed here are those of Joshua Levy, and nothing here is official JDRF or JDCA news, views, policies or opinions. My daughter has type-1 diabetes and participates in clinical trials, which might be discussed here. My blog contains a more complete non-conflict of interest statement. Thanks to everyone who helps with the blog.