Last week, diabetes headlines were dominated by a new study from the Faustman Lab at Massachusetts General Hospital, published on Wednesday, August 8th on PLoS One, suggesting that a 90-year-old tuberculosis vaccine called BCG might hold promise for people living with type 1 diabetes. “Human Study Reignites Debate Over Controversial Diabetes ‘Cure,’” wrote Reuters. “Diabetes May Be Reversed By Long-Used Vaccine for TB,” proclaimed Bloomberg news.
It sounds intriguing. But as people with type 1 diabetes (and their families) know, it’s best not to trust a headline – according to the news, type 1 has been “cured” numerous times since I was diagnosed in 2001. So I decided to go straight to the study’s author — Denise Faustman, M.D., Ph.D., controversial researcher and director of the Faustman Lab – to learn more about the study from the person who designed it. How did it work? What did she find? And how excited should we actually be?
First, the study itself had a more subdued title than the headlines devoted to it: “Proof-of-Concept, Randomized, Controlled Clinical Trial of Bacillus-Calmette-Guerin for Treatment of Long-Term Type 1 Diabetes.” It was inspired by trials in non-obese diabetic (NOD) mice models suggesting that a 90-something-year-old tuberculosis vaccine called Bacillus-Calmette-Guerin (BCG for short) might contribute to the elimination of insulin-killing immune cells in NOD mice. Faustman and her team hoped this might also be true for humans with type 1 diabetes.
Killing these cells would be significant because, in the past decade or so, researchers have begun to believe that the human body is actually able to regenerate insulin-producing beta cells. (Previously, researchers believed that once you killed off your beta cells, they were totally gone – kind of how they used to think that you couldn’t grow new brain neurons.) One current theory is that if we could stop the immune system’s attack on beta cells, the body would be able to grow new ones – thereby curing (or at least greatly improving the lives of) people with type 1 diabetes.
Interestingly, in both mice and humans, BCG itself doesn’t appear to be doing anything to the white blood cells – known as insulin-autoreactive T cells — that are responsible for destroying insulin-producing cells. Instead, BCG increases the body’s production of a different type of molecule called Tumor Necrosis Factor – TNF for short. It’s TNF, not BCG, that appears to kill the insulin-autoreactive T cells.
Faustman had originally hoped to administer TNF directly, but TNF is not approved as a drug; the approval process would have taken years, and no drug companies were interested in spending millions of dollars developing an experimental treatment that, if successful, would reduce the market for their other diabetes products. Instead, Faustman and her colleagues searched for an approved drug with a great safety profile that increased levels of TNF. BCG fit the bill.
It’s worth noting that this is not the first time BCG has been studied – several previous trials have failed — and several members of the diabetes research community that I spoke with were extremely skeptical about BCG in general, as well as this particular trial. (Faustman believes previous trials failed because of a lack of knowledge about mechanism and incorrect dosing.)
With that said, here’s how this most recent trial worked: the study was a proof-of-concept trial, designed to set the stage for what Faustman hopes will be a much larger intervention. The clinical group, therefore, was really small: she got FDA approval to treat three people with BCG, and used three more as double-blind, randomized controls (all six had had type 1 for an average of just over 15 years). She also included a larger “reference” population of people with and without type 1 diabetes, which acted as non-blinded, non-placebo-treated controls. The treated group got two doses of BCG, similar to the dose used for tuberculosis vaccination: one at the trial’s start, and one four weeks later. Researchers drew participants’ blood once a week for twenty weeks; in total, they studied more than 1,073 blood samples.
They were looking for four distinct biomarkers: the insulin-autoreactive T cells, regulatory T cells (a type of T cell that’s thought to act sort of like a police force, keeping more destructive T cells in check), an autoantibody called glutamic acid carboxylase (GAD – its existence is often used in early stages of the disease to confirm that you have type 1 or LADA rather than type 2), and C-peptide, a protein molecule, produced along with insulin, that indicates how much insulin your body has produced. (It’s impossible to measure insulin itself, both because you use it up quickly after you make it, and because there’s no way to distinguish between the artificial insulin you take and the insulin your body actually makes. C-peptide is a better marker on both counts: artificial insulin doesn’t have C-peptide, and C-peptide hangs around in your blood stream even after the insulin has been used up.)
Faustman’s results indicated that BCG was doing something. First, even though none of the six clinical participants had significant levels of insulin-autoreactive T cells at the study’s start, Faustman’s team noticed significant numbers of dead insulin-autoreactive T cells in circulation after each dose of the vaccine. This confused me when I read it – where were the autoreactive T cells coming from if they weren’t there at the beginning of the study? And, uh, don’t you not want new autoreactive T cells? Faustman explained that, first of all, she doesn’t think the vaccine is making any new ones. Rather, she believes that the BCG-induced TNF is killing off old stores of insulin-autoreactive cells — while no one yet knows exactly where they’re located, there appear to be reservoirs of these insulin-autoreactive T cells in the bodies of type 1 diabetics, which are ready to pounce upon any new insulin-producing cell that dares show its face. Faustman believes that TNF is attacking the reservoirs of these cells and killing off part of their population; their dead carcasses then show up in circulation. Since she observed an increase the number of these dead cells after both doses of the vaccine, she also believes that the dose they used in the study is probably too low.
As for the results of the other biomarkers, GAD was confusing – the levels in one treated participant went up, another’s went down, and the third turned out never to have had GAD to begin with. (Faustman said it’s relatively common for a long-term type 1 diabetic to not have GAD, and that researchers working on islet-cell transplants have observed similarly confusing GAD results.) Two of the three treated patients showed an increase in regulatory T cells. And, lastly, two out of the three treated patients showed an increase in C-peptide, indicating that they were beginning to make a teensy bit of their own insulin again – an intriguing finding, considering that the study’s subjects, on average, had had type 1 for more than 15 years.
If these results are correct, Faustman’s hope is that BCG could eventually be used as a therapy to kill off the reservoirs of insulin-autoreactive cells – via TNF — in people with type 1 diabetes. By eliminating the T cells that are destroying the insulin-producing cells, her hope is that you could give the body a chance to regenerate some of its ability to produce its own insulin.
Here, however, is where the huge caveats come in: “It’s not like anyone is throwing away their insulin syringes,” says Faustman. The elevation of C-peptide was small, and it did not last long.
What’s more, even if BCG treatment could eliminate enough insulin-autoreactive T cells to allow for the regeneration of beta cells – which will take much more research to definitively prove – there would be many, many questions to answer, like what the ideal dose would be, how often it would need to be given (Faustman thinks it would be a series of “booster”-like shots), whether any other “good” cells are being destroyed (Faustman believes they are not) and what the long-term effects of these repeated booster shots of BCG — and ensuing spikes in TNF – might be.
“Please make sure they don’t think we’re curing type 1,” Faustman emphasized, when I asked her what readers should take away from the study results. Allow me, therefore, to repeat that in all caps: BCG IS NOT A CURE FOR TYPE 1 DIABETES. And no, it is not a good idea to go ask your doctor to vaccinate you against tuberculosis. When I asked her what headline she would write for the study, Faustman paused for a moment before offering something more subdued than most of last week’s teasers: “Promise for Disease Treatment for People With Advanced Diabetes Seen in Generic TB Drug.”
Her other take-home message? “We’re just happy that we’re optimistic enough to think we can do something at a stage of diabetes where people weren’t even willing to risk doing trials,” she said. Long-term diabetics “have been told they’re too old to be cured,” she continued. “Wouldn’t it be great if we could figure out new things for people with all stages of the disease? I think it opens a window for people to think about the fact that maybe, if we pushed hard enough, we could develop something for them.”
In other words, it’s a type of study that, by not limiting itself to newly diagnosed diabetics, might actually apply to me. That would be something to be excited about.
If you’re interested in finding out more about Faustman’s work, or indicating interest in future studies, visit faustmanlab.org