GLP-1 Thematic: News Update
Parsing Recent Developments in GLP-1 Drugs
There’s been a recent development in the world of GLP-1s which has given our thematic long/short basket a boost which is now outperforming the S&P by more than 10% since inception. Surprisingly, most of the contribution this time has come from the short side in the past two days, because of this development.
I have asked my absolute favorite biology nerd to help me shed some light on what’s happened recently in GLP-1 world.
The development is this:
Why is it important?
T1DM is thought to be an autoimmune disease; essentially the immune systems mistakenly recognises pancreatic beta cells (The cells that release insulin in your body) as a foreign substance/organism, and therefore must be removed/killed from the body. The reduction in beta cells means that you are producing less and less insulin until functionally, all your beta cells have been wiped out. In this late state of T1DM, exogenous insulin, taken both after meals and constantly at a basal level, to keep blood glucose levels in check becomes a requirement. This means a demand for insulin (for the rest of their life), insulin pumps and blood glucose monitoring devices/software.
Glycated hemoglobin (HbA1c) is used as an indirect measure of blood glucose and is the primary metric that diabetics aim to control to prevent complications arising from diabetes. The study looked at 10 recently diagnosed T1DM patients and began Semaglutide injections within 3 months of their diagnosis. Weekly Semaglutide at 0.125mg, which was then scaled up to a maximum of 0.5 mg (For comparison, obese people receive Semaglutide at 2.4mg) showed improvements in HbA1c levels in the patients. Notably, after 3 months these improvements were such that all 10 patients, no longer required insulin injections after meals. After 6 months, 7 patients no longer needed even basal insulin.
Although these results are indeed worth highlighting and will evoke spirited discussion by neophytes, they are actually hardly surprising in the context of what was in fact, the intended purpose of GLP1R-agonists, namely to control blood glucose levels in diabetic patients. What provides hope and may have important investment implications, is the finding that there were patients that had an elimination for the need for all exogenous insulin.
However, among all this, the key thing to notice is actually the C-peptide levels.
C-peptide is a peptide that is released in equivalent amounts, by the pancreas, with insulin. Measuring C-peptide levels provides an indirect measure of how much residual beta cells a T1DM patient has (Insulin degrades quickly and T1DM require exogenous insulin to survive, meaning that insulin measurements are not reliable for the purpose of measuring residual beta cells).
But now, I can already hear the rallying cry of the midwits:
“buT bRo, stUDy OnLY hAs 10 PaTIenTs as A SAMplE”
No shit Sherlock, we can see that too.
(Many statistically semi-literate individuals, often lacking expertise in the relevant subject matter, tend to focus their critiques on sample sizes. This is because such comments represent the simplest and most universally applicable form of criticism, requiring little additional insight or understanding of the specific context)
To make an analogy; imagine you saw a study that looked at 5 patients who had their left arms amputated from the shoulder down. After 2 years, the 5 patients had regrown their left arms up to their elbow.
Your first thought shouldn't be " buT stUDy OnLY hAs 5 PaTIenTs as A SAMplE", it should be "That's not supposed to happen, holy shit, this is potentially a huge deal", yet here we are, speaking about
NVDA’s finances and why they’re not Enron why T1D going from “there” to “not there” maybe deserves a bit of attention as the result is peer reviewed. We feel it’s interesting there’s so much pushback on the general idea, but it’s also pretty funny to us (as we had to field a lot of questions after the GLP-1 article about how “no its def gonna melt your insides” and now maybe it’s going to help us explore a potential route for treating another disease lol).
More Science Stuff (Investment Implications at the Bottom):
T1DM is associated with the continual destruction of beta cells, meaning that in turn you would expect C-peptide levels to continue falling over time. After 12 months, you would expect C-peptide levels to have decreased in T1DM patients, or at the very least, stay flat. This is the key finding of the study.
This data implies that the effects seen here, may not only be simply due to promoting the activity of remaining beta cells inside the T1DM patients, or by only inhibiting glucagon release, but rather that the pancreas in these patients is no longer being destroyed or might even actually be regenerating.
In fact, if the statistically semi-literate had even bothered to do a cursory unpaired T-test on the given data of fasting C-peptide levels (At time of diagnosis, Mean = 00.65, SD = 0.33; after 12 months mean = 1.05, SD = 0.4, N = 10 for both, P = 0.0253).
In plainspeak, using even just quick and unsophisticated statistical testing tells you that you would only expect to see this kind of result 2.53% of the times if the treatment actually had no effect whatsoever; Ie, since 2.53% is a very low probability, it is conventional to assume that therefore the treatment probably in fact does have an effect.
There are of course caveats, since the raw data hasn't been provided, it is not possible to tell whether all individuals showed an increase in C-peptide levels, or whether the average was simply raised by a few individuals (Although the comparable variance at both time of diagnosis and after 12 months, suggests that this is unlikely to be the case).
In any case, the authors were self-aware enough to state that:
“In all, our preliminary observations support the need for prospective, randomized clinical trials with larger numbers of patients to investigate this approach further”
far from a proclamation that their study is Gospel, but rather that their entire correspondence is simply them saying: “We found a potentially interesting effect in a few T1DM patients; probably worth digging into properly to check if it’s true”.
Before future studies are completed, it is worth entertaining how these results could be possible, given our current understanding of what GLP1R agonists can do. In the context of obesity, our last Substack post was focused mainly on the central (brain) effects of GLP1R agonists in being able to control appetite. For the sake of preventing information overload, other effects were glossed over but are now worth bringing up here.
It has been shown that Exendin-4 (A GLP1R agonist) can improve glycemic levels, even in T1DM patients who do not have any residual beta cells (Dupré et al, 2004). Similar results have also been seen in T1DM patients who received Liraglutide (A GLP1R agonist) in conjunction with insulin therapy (Kielgast et al, 2011; Varanasi et al, 2011), suggesting that GLP1R expressed in cells, outside of beta cells, may also have a role in improving glycemic levels even in people who functionally lack beta cells and would ultimately mean lower need for insulin in T1DM patients.
However, in the context of increased C-peptide levels, there are some other effects that are more relevant:
Continuous infusion of GLP1 for 2 days, has been shown to be able to help prevent the death of beta cells in a rat-model of T2DM (Zucker Diabetic rats) (Farilla et al, 2002). Human islet cells (Alpha, beta, PP, delta and epsilon cells) in cultured in the lab with GLP1 do not undergo cell death as often as control cells (Farilla et al, 2003). Such results suggest that GLP1 may be able to delay or offset the death of beta cells in conditions such as T1DM.
In mice that were subjected to streptozotocin (A chemical that destroys beta cells), treatment with Exendin-4 (A GLP1R agonist) leads to an increased number of insulin-producing cells in the pancreas, compared to control mice (Xu et al, 2006). Such data suggests that GLP1R may be able to stimulate the proliferation of existing beta cells. Such proliferation has been shown to be mediated by the secretion of growth factors such as IGF2 (Insulin-growth factor 2) (Cornu et al, 2010) and EGFR (Epidermal growth factor receptor) (Buteau et al, 2003).
GLP1 has also been able to promote the expression of IDX-1 (Stoffers et al, 2000). In more accessible language, it implies that GLP1 may be able to convert existing pancreatic cells into new beta cells.
In the fashion of any other insufferable macro-analyst, I summarise these effects as the 3 P’s
Prevention (of cell death), proliferation (of beta cells), progenesis (of new beta cells).
These mechanisms are worth mentioning since if the results in this latest correspondence in the NEJM are indeed replicated, there is a large difference if these effects are occurring either
a) Because pancreatic beta cells are dying less
b) Because pancreatic beta cells are growing back
c) Because pancreatic beta cells are not being targeted by the immune system anymore
If replicated, I find a) and b) to be the most likely reasons; in this case, T1DM patients will merely have to substitute insulin with GLP1R agonists, but not be able to come off GLP1R agonists. Since T1DM patients still harbor immune cells that are targeting their beta cells, coming off GLP1R agonists means that these immune cells would once again be able to take wreak havoc on pancreas all over again. In the case of a), this would mean that this treatment really only works when you catch T1DM early, before remaining beta cells die off. In the case of b) however, this would suggest that it may even be possible to treat T1DM patients who are already in late stage and no longer show any C-peptide levels.
I bring up c), only because I assume that some people may try to jump to this conclusion, but frankly find this to be very unlikely. If c) is the cause, this would represent an actual cure, ie you would be able to stop taking GLP1R agonists at some timepoint. Although GLP1 has been shown to have effects on the immune system, such as in cardiomyocytes (Noyn-Ashraf et al, 2009) or endothelial cells (Shiraki, et al, 2012), many of these studies have shown little relevance to the organ of interest, namely the pancreas. What few studies have been done in the pancreas (Koehler et al, 2009), while they did show some downregulation of specific cytokines (By itself, downregulation of specific cytokines is a rather unremarkable effect on the immune system, plenty of substances, that don’t cure anything, can have similar effects), it did not show any actual improvements in outcomes in a rodent model of pancreatic inflammation.
Autoimmunity is caused by the actions of a few specific cells in the immune system, meaning that broad-based anti-inflammatory/immunosuppressant drugs that are capable of reducing autoimmune effects, will inevitably do this at the expense of a normal functional immune system. There is little reason to believe that GLP1R agonists are somehow specifically targeting only autoimmune cells, while sparing the rest since molecularly there is no reason for only autoimmune cells to express GLP1R, but for other normal immune cells to not express GLP1R. In the outlandish (In my opinion) scenario that c) is indeed the cause, this would mean that GLP1R agonists are in fact not just a cure for T1DM, but for autoimmune diseases in general; but again, I find this to be very unlikely.
Past studies exploring the use of GLP1R agonists for T1DM have indeed been done; but normally in the context of either in conjunction with normal insulin therapy, or with immunosuppressants (The reasoning is to help prevent further autoimmune destruction of beta cells). For example, Exanatide alone, or in combination with insulin therapy or immunosuppressants, such as daclizumab, but did not improve C-peptide secretion (Rother et al, 2009). Most studies did show very modest improvements in the insulin required by T1DM patients, but were hampered by side effects, namely GI and nausea. Given this, why is Semaglutide able to increase C-peptide levels all of a sudden? The simplest explanation may be that Semaglutide has a longer life-time and more favourable side-effect profile; allowing patients to stay on it chronically and have more sustained GLP1R agonism. The astute reader may have noticed that a lot of these past studies were conducted using GLP1R agonists with much shorter half-lives (GLP1 = 5 minutes, Exenatide = 2 hours, Liraglutide = 13 hours). If this is the case, there is no reason to believe that Semaglutide is the only drug capable of achieving these effects, since other GLP1R agonists exist with comparable or even better half lives and side effect profiles.