145: How Insulin May Be Silencing Your GLP-1

Your body's stop eating signal might be failing. Everyone's talking about GLP-1 right now, but they're asking the wrong question. It's not just what GLP-1 does to insulin, it's what insulin is doing to GLP-1. GLP-1 is your body's nutrient brake. It slows digestion, reduces appetite, and helps control blood sugar. But when insulin stays high for too long, it starts shutting that system down. That "I'm full" signal gets weaker. Now the brake is gone, food hits your system faster, blood sugar spikes higher, and your body has to release even more insulin. And now you're stuck in a cycle. More insulin leads to less GLP-1, which leads to bigger spikes, which leads to more insulin. It's not just reacting to the problem, but in this case, it's driving it. This is Lecture 145 of The Metabolic Classroom. Welcome back to The Metabolic Classroom. I'm Ben Bikman, metabolic scientist and professor of cell biology. In today's mini lecture, I hope to teach you about one of the most underappreciated relationships in all of metabolic physiology, and that is the crosstalk between insulin and GLP-1. Of course, you've heard of GLP-1 now. It's become a household name largely because of the explosion of the GLP-1 receptor agonist drugs. These are the molecules like semaglutide or tirzepatide and many, many others. If you hear that suffix -utide or something like it, then it's safe to assume you're talking about one of these types of drugs. And of course, nowadays they are used almost exclusively for weight loss. Most of the conversation, including most of the scientific discussion, focuses on one direction of this relationship— how GLP-1 affects insulin. That conversation has quieted a little bit insofar as the GLP-1 drugs don't as much— aren't as often used for diabetes therapy as much as they used to be. Now that they're used for weight loss, people don't talk about it. So there are some important questions like, does GLP-1 actually stimulate insulin secretion? Does it amplify the insulin response to a meal? That's the question or questions that most researchers and perhaps even most clinicians have been asking. But today I want to flip the direction of that question entirely because there's an equally important and actually I think more important and completely overlooked directionality here. And the question being, does insulin in turn affect GLP-1? I think that you will find the answer has some pretty important implications for understanding how metabolic disorders can develop and then even perpetuate itself. But before we get there, I of course need to lay the groundwork, starting with what GLP-1 actually does and then clearing up a common misconception about the GLP-1 and insulin relationship. That I think actually makes the inverse, the flipped question even more interesting. But as usual, let's start at the beginning. GLP-1 stands for glucagon-like peptide 1. There are multiple of these GLPs. So we're talking about the number 1. It's a 30-amino acid peptide hormone that is secreted primarily from some specialized cells in the intestines. These are called the L cells, and they sense the arrival of nutrients and then in turn will release GLP-1. So as nutrients and foodstuff is moving through the gut, the L cells will be activated and then will release GLP-1 into the blood. GLP-1 was first identified and characterized in the mid-1980s, and it was quickly recognized as an incretin hormone. This refers to the incretin effect, and the incretin effect is the amplification of insulin secretion that occurs after an oral glucose load compared to the same load of glucose if it's given intravenously. Now, let me say that another way. There are gut-derived signals. So, so hormones that come from the gut that potentiate the insulin response to a meal. In fact, I'm going to say that another way. Scientists basically found that we can more rapidly clear the glucose from our blood when we eat that glucose compared to when the glucose is directly injected. In other words, there was something that happened when that glucose came from the gut compared to when it just directly found itself in the blood. Through an intravenous infusion. GLP-1 and some other incretin signals or hormones are this signal, and they're estimated to account for somewhere between 50 to 70% of the, the reason and the drive for the insulin response after an oral glucose load in a healthy person. And critically, this effect is greatly reduced in someone with type 2 diabetes. We're going to come back to that idea in a lot of detail in a moment. However, this view is more nuanced than it seems, but also GLP-1's actions do extend well beyond just insulin alone. GLP-1 also suppresses glucagon, which is the counter-regulatory hormone that drives glucose production from the liver. It also slows gastric emptying dramatically. Controlling the rate at which the nutrients leave from the gut and in turn enter the bloodstream. That's probably the most important of all, frankly. And then it also signals to the brain to reduce appetite and cravings. It has this protective effect on the pancreas and the beta cells itself, actually. So GLP-1 is a master coordinator of the meal response, and it works simultaneously at the gut and the pancreas and the brain. Because endogenous GLP-1, or the GLP-1 that we make ourselves, has a very quick half-life, just about 2 minutes. So it's degraded very rapidly by the enzyme DPP-4. Um, the pharmaceutical industry has developed GLP-1 receptor agonist drugs, so a version of the GLP-1 molecule that is more resistant to that degradation in order to try to provide more of a sustained signaling. Now that's the drug story, but the hormone story is what I want to focus on today. Now, before I get to the flipped question that I alluded to earlier, namely how insulin affects GLP-1, I need to spend a few minutes on the conventional framing because getting it right actually makes the rest of today's lesson more interesting. The textbook says that GLP-1 stimulates insulin secretion. You'll read it. You'll hear it everywhere, and it is partly true, but the conditions under which it's true matter enormously. When researchers infuse GLP-1 intravenously and hold blood glucose elevated with what's called a glucose clamp— so the, the clinical team here, the biomedical group, would be forcefully bumping up the glucose and then infusing GLP-1. So not touching on any kind of gastric component here. Then in this situation, GLP-1 does amplify insulin secretion from the beta cell. That property of the GLP-1 receptor is real and it is well confirmed in humans. So if you are artificially keeping the glucose elevated and you're infusing GLP-1, again, to say all this another way, the insulin will be higher. Now, but what happens when you test GLP-1 under physiological conditions, like during an actual meal with none of this artificially elevated glucose being stuck, this hyperglycemic clamp? The answer is the opposite. A publication by Flint and colleagues found this directly in 2001. They infused GLP-1 into 18 obese men who then ate a fixed breakfast. The GLP-1 infusion slowed gastric emptying, like I said, is maybe the most important of all the effects, but it did so, so powerfully that it reduced the nutrient absorption by about a third, to the point that blood glucose barely rose above fasting levels. And now what about the insulin? That's because, of course, the central part of what I'm looking at here— well, the insulin peak was less than half of what it was in the placebo group. So after they infused the GLP-1, the insulin response in the GLP-1 group was less than half of what it was in the placebo group. So the GLP-1 did not stimulate insulin. In fact, it reduced it, probably because it was kind of limiting the rate of nutrient entry into the body. So it largely eliminated the glucose stimulus that insulin would have otherwise had to respond to. That same pattern has since been confirmed with GLP-1 receptor agonist drugs. Hirota and colleagues in 2019 gave healthy volunteers an oral glucose tolerance test after a GLP-1 drug, an early version of it called exenatide, a short-acting GLP-1 drug. And they found that both glucose and insulin peaks were remarkably delayed and reduced. The authors described this explicitly as an insulin-sparing effect. So far from GLP-1 being an insulin secretagogue or activator, they were describing it as the opposite. It spares the body's need to produce more insulin. So the more precise statement could perhaps be something like this: GLP-1 does have the ability to stimulate insulin release under artificial conditions, but in a meal context, the dominant action is slowing gastric emptying, which means less glucose arrives, which means less insulin is needed. GLP-1 in a physiological context is an insulin-sparing hormone, to use the word they did earlier. It's a brake on postprandial insulin, not an accelerant as it's often viewed. Now, here is why that matters for what comes next. If GLP-1's primary job is to moderate the meal response, to apply the brakes on how fast nutrients enter the blood system, then what happens when GLP-1 is lost? Well, then nutrients will flood in faster. Glucose is going to spike higher, and there's going to be a subsequent much higher rise in insulin to try to account for this glucose rise. So the brakes are gone, and it's all just gas pedal. And this brings us to the question that nobody really thinks about. Most researchers, when they talk about GLP-1 and insulin, are asking, what does GLP-1 do to insulin? Now, we've briefly covered that, I hope pretty well, but again, albeit briefly. In a real meal, GLP-1 spares insulin by slowing gastric emptying. But the direction nobody focuses on is the reverse. What does insulin do to GLP-1? This question helps us better understand how GLP-1 agonists might be used most effectively in clinical settings. And that's what the rest of the lesson today is going to be about. A consistent and robust finding across the scientific literature is that people with insulin resistance, obesity, prediabetes, type 2 diabetes, they secrete significantly less GLP-1 after a meal than metabolically healthy people. In fact, in some cases it is substantial and it follows a dose-response gradient that tracks perfectly with the insulin resistance. Now, let me walk you through some of the key findings, and I'm just going to highlight a few. There are many more that I could have cited. In 2001, a group of scientists published a paper in the Journal of Clinical Endocrinology and Metabolism. That's a good journal. and they gave a standardized mixed breakfast meal to 54 patients with type 2 diabetes, 33 matched healthy controls. So people matched for, uh, sex and age and weight, and then 15 people with impaired glucose tolerance. Of course, that's that insulin-resistant pre-diabetic state. They measured GLP-1 secretion over 4 hours. The GLP-1 response was significantly lower in the type 2 diabetic patients compared to healthy controls. and the impaired glucose tolerance group was right in between them. The gradient was quite apparent that the worse the metabolic state, the less GLP-1 was secreted. Critically, when the research team ran their statistical modeling and controlling for BMI and sex, the diabetic state remained really an independent predictor of reduced GLP-1. So it was not just a matter of body size. And I want to flag this now because it becomes important later. They found that insulin area under the curve— so how much insulin was produced— was the negative independent predictor of GLP-1 area under the curve. In other words, the more insulin rose after the meal, the less GLP-1 was secreted. That inverse relationship is exactly what you'd expect if insulin were suppressing the GLP-1-secreting cells. A paper in 2001 looked at a spectrum of insulin sensitivity in non-diabetic men measured by the gold standard hyperinsulinemic euglycemic clamp. So keeping glucose in a normal level after dividing the subjects into tertiles based on insulin sensitivity and then feeding them all a mixed meal. The finding was remarkable. At just 15 minutes after the meal, The GLP-1 level in the most insulin-resistant men was only about half of the level seen in the most insulin-sensitive men. And multiple linear regression confirmed that insulin resistance, independent of obesity, was the key predictor of that blunted GLP-1 response. Fat mass alone was not a, a relevant variable here. It was the metabolic dysfunction. In other words, the insulin resistance. A 2008 paper pushed this further across 51 subjects spanning an enormous range of BMI from 20 to 61 and a glucose tolerance range from normal, healthy glucose-tolerant people to very glucose-intolerant type 2 diabetics. They found that obesity and glucose intolerance each independently and additively reduced the GLP-1 response. So separate contributors in this case, but— and they found that neither fully explained it. Then there is the twin study published in Diabetes Care in 2014, some of the most compelling human evidence on this question. They studied 35 monozygotic and 75 dizygotic twin pairs, and some of them were discordant for weight. In other words, across these identical and non-identical twins, They had a group, they had groups that were in some instances lean and overweight. In pairs where the heavier twin had higher insulin resistance, the GLP-1 response was blunted in every single co-twin, every single one. Finally, I want to come, let's go further back in time to one of my favorite papers published in 1996 in the journal Gut. Because I think it's one of the most telling observations in the entire literature on this topic. And it's the first paper I remember personally as a young scientist seeing in, in this topic that some people have much lower GLP-1 responses than others. They studied 6 lean and 6 obese premenopausal women, and they gave both groups a high carbohydrate meal and a high fat meal. After the high-fat meal, there was no real difference between the groups. Both had a modest, um, but small GLP-1 response. But after the high-carbohydrate meal, the GLP-1 secretion was severely blunted in the obese women compared to the lean women. In fact, there was no significant response of GLP-1 to the dietary carbs throughout the entire time span. Think about this. The carbohydrates drive, of course, a sharp insulin spike, but dietary fat does not. The obese women showed blunted GLP-1 only after the meal that most aggressively drives insulin, the high-carb meal, not after the one that doesn't, the high-fat meal. This is a very specific pattern, and I'm going to come back to explain exactly why it fits the mechanism so precisely. Okay, moving on. In 2009, a group at the University of Toronto published a very important paper in the journal Endocrinology. This study went directly to the question: does the L cell itself respond to insulin? And does chronic hyperinsulinemia impair GLP-1 secretion? The first finding was fascinating. Using multiple L-cell models. So L-cells from mice, from rats, and from humans, but grown in little petri dishes. The researchers confirmed that L-cells express the insulin receptor, so the GLP-1-producing cell can sense and respond to insulin. Now, this alone was a significant and novel discovery. The gut hormone cell that play— that can tell the pancreas what to do is itself listening to the pancreas. So that it established this crosstalk. As much as the prevailing paradigm was that L cells make GLP-1 and affect the beta cells and insulin production, they found the opposite is also true, that beta cells and their product insulin affect the L cells and its product, um, GLP-1. When they acutely treated L cells with insulin, GLP-1 secretion increased by a lot. All right, so that supported the traditional paradigm here that insulin is kind of— did this crosstalk going both directions. So this makes some physiological sense here. But let me say that again in the event I sort of bumbled it. When they acutely treated the cells with insulin, so they gave the cells a small dose of insulin, then the GLP-1 went up. So this is supporting that idea of the cross-talk going this other direction. So that makes some sense. So in a healthy meal response, nutrients will arrive, GLP-1 goes up, insulin rises in response, and that insulin signal echoes back to the L cells to help sustain that GLP-1 production throughout the meal. So insulin and GLP-1 under normal conditions are cooperating, but the researchers took it further. They asked, well, what happens when L cells are chronically exposed to high insulin? Of course, those are the conditions that exist in sustained hyperinsulinemia, the conditions that exist in anyone with insulin resistance. In fact, dare we say, that is the default state of most humans in the world as they eat 6 to 7 times a day, and it's constantly consuming the consumption of refined starches and sugars. So again, they increased the insulin and kept the insulin elevated activated in the culture, in the cell culture. They induced insulin resistance in the L cells by exposing them to high insulin, something I've done in my lab before with other cell types. The results were remarkable. Insulin receptor expression dropped significantly. Downstream signaling on the insulin pathway was severely blunted, and the L cells could no longer mount a normal GLP-1 secretory response. Not to insulin and not even to other stimulators that they tried in the culture. So in other words, the more the insulin was elevated, the less ability insulin had to induce a GLP-1 response. And they actually confirmed this in a particular mouse model, which is an insulin-resistant mouse with chronically elevated insulin. These mice had elevated basal GLP-1 levels, but when challenged with an oral glucose load, So giving the animals a big swallow of glucose, their GLP-1 secretory response was markedly blunted compared to the normal mice. The chronically elevated insulin had rendered the L cell unresponsive to the meal stimulus, even though the meal was the very thing that should trigger GLP-1 secretion. This is a pivotal finding. Chronic hyperinsulinemia doesn't just cause insulin resistance in the muscle and liver and brain, et cetera. The tissues that we classically think about, but it also reduces the gut's GLP-1-producing cells, their ability to work and produce GLP-1. So the L cell goes quiet precisely when it should be the loudest. Now, let's connect this line of thinking back to the study I mentioned in the 1996 paper published in the journal Gut. Recall that the obese women showed a severely blunted GLP-1 response to the high carbohydrate meal, but not to the high fat meal. And other incretins were also, you know, changed, but not as much as the GLP-1. And that's what I want to focus on. Why would GLP-1 be selectively suppressed after the carbohydrate load, but not the fat load, but only in the obese group? Well, one of the other papers I cited earlier gives the answer that the carbohydrate drives a sharp insulin spike. Fat does not. If the L cells of the obese women had become insulin resistant due to years of chronically elevated insulin, their GLP-1 secretory machinery would be impaired. But that impairment would only become fully apparent under conditions that demand a robust GLP-1 response. Carbohydrate is exactly that demanding stimulus. Fat isn't. So the obese women's impaired L cells could still produce a modest GLP-1 response. To the dietary fat, which was admittedly the weaker challenge, but they failed when confronted with carbohydrate, the stronger stimulus. This is likely because of the elevated insulin. So it is speculative because the authors did not measure insulin resistance or insulin directly. So it's a little circumstantial, but it does fit. There's also corroborating evidence from some animal studies, a 1994 paper found that in, in rats, that exogenous hyperinsulinemia— so if they gave them a lot of insulin during a glucose challenge, it suppressed secretion of GLP-1 and other incretins from the gut. And to return to a study I mentioned earlier, across 102 human subjects eating a real mixed macronutrient meal, the more insulin rose after that meal, the less GLP-1 was secreted. That inverse relationship in humans in a real post-meal context is exactly what you would expect, um, if this mechanism is functioning in people, that elevated insulin is reducing the body's ability to make GLP-1. Now put it all together. Insulin and GLP-1 are obviously in a very dynamic relationship. In a healthy metabolic state, they cooperate. Insulin can briefly amplify GLP-1 during a meal. But when insulin is chronically elevated, the L cells pay the price. It becomes insulin resistant and it loses its ability to respond to meals. So the GLP-1 secretion just kind of falls apart. Now, let's summarize what we've discussed. In early insulin resistance, whether driven by excessive carbohydrate intake, obesity and hypertrophic fat cells or poor sleep or other metabolic stressors, insulin levels will rise and be elevated chronically. The body compensates by secreting more insulin to overcome the resistance and to try to maintain blood glucose. But that chronically elevated insulin begins to act on the L cells of the gut, and chronic hyperinsulinemia then renders the L cell to be insulin resistant. Just like it would to other cells. And now the L cells can no longer mount a robust GLP-1 response to a meal. Now, remember the GLP-1, how— what it does in a real meal context. Remember, it breaks or slows down the gastric emptying. It suppresses glucagon, and it consequently spares insulin. When GLP-1 is lost, that brake is lost, that, that, that limiter. Now nutrients will flood into the system much faster. Glucagon is no longer suppressed like it normally would be. And so the liver keeps producing glucose even when it shouldn't. The appetite signal also is diminished and the postprandial glucose excursion is larger, demanding more insulin. More insulin will further suppress the L-cell. and you can see the cycle, and the cycle just keeps repeating itself and feeding on itself. This explains the gradient we see across the human studies. GLP-1 impairment is most severe in type 2 diabetes. It's intermediate in people with prediabetes, and it's absent in healthy controls, tracking exactly with the progression of hyperinsulinemia. The twin study showed that obese individuals with preserved insulin sensitivity had normal GLP-1. It was only the metabolically healthy elevated insulin obese twin, um, where the GLP-1 was blunted. And back to that 1996 Gut paper published in the journal Gut, finding the carbohydrate-specific blunting effect, that's consistent with a system where chronically elevated insulin has silenced the L cells so that when carbohydrate drives insulin even higher, the L cells simply cannot respond. Think about what this means for a moment in practical terms. Every high carbohydrate meal drives insulin up. Chronically elevated insulin gradually renders the L cells insulin resistant. In this state, GLP-1 secretion falls. Without GLP-1's gastric brake or slowing down, nutrients are entering faster. Glucose is spiking more sharply and the insulin demand is much more dramatic, which can suppress GLP-1 even further. You could see how this can compound over the years. And here is perhaps the most important clinical implication. When GLP-1 receptor agonist drugs work, and they clearly do work, they work very well. They work precisely because they bypass this broken system. They deliver a pharmacological load of GLP-1 and activate that— those receptors from the outside. Well, as you inject it to the inside, but that can override the dysfunctional L cell. But they do not repair the L cell. They do not address the chronic hyperinsulinemia that's made the L cells insulin resistant. The drug is treating a downstream consequence of a problem whose upstream cause, the hyperinsulinemia, is still there. But it also suggests that as insulin settles down with improved diet, the sensitivity could return. That would be a fantastically relevant research question to ask in the future. Take those individuals who are having a diminished GLP-1 response, use perhaps a GLP-1 drug to help them improve their insulin sensitivity and lose weight, etc. And then after they wean off the drug, see whether the GLP-1 response has returned. Did you restore it because of this intervention? Now, what does this mean for you and maybe some people you care about? First, there are several points here to make. In conclusion, it does reframe how we think about GLP-1. The evidence does indicate that in many people, GLP-1 deficiency is an acquired condition, probably because of the chronic hyperinsulinemia that has progressively silence the L cells. Second, it suggests that interventions that lower chronic insulin levels, like reducing your consumption of refined starches and sugars, losing weight, you know, shrinking those fat cells, improving stress, all of those interventions that improve insulin levels, they may over time restore L cell function. We don't know yet, of course, from larger trials. So it's very speculative at this point. But the mechanistic logic is sound. Third, these findings, it helps contextualize the GLP-1 drug class, I think, more honestly. GLP-1 receptor agonists are very powerful drugs, and their mechanism in a real meal context works primarily through slowing gastric emptying, not through amplifying insulin, as is commonly assumed. That is actually a more elegant mechanism than the textbook would conventionally suggest, but they do not address the root cause. Addressing that root cause through lifestyle and dietary changes that genuinely lower chronic insulin, I submit, remains the most fundamental approach. A fourth point, um, the research illustrates a unifying principle that runs through much of my own work, which is that chronically elevated insulin is not just a response to dietary insult, it actively perpetuates and deepens itself through mechanisms that we're still really working out. Insulin resistance begets more insulin, which begets more insulin resistance, which through the L cell begets a GLP-1 deficiency, which can then in turn remove a critical brake or a reduced postprandial glucose surge. Which demands, of course, still more insulin. So let's wrap this up. What we're seeing may be a vicious cycle. Hyperinsulinemia silences GLP-1. Lost GLP-1 releases the gastric brake. Now nutrients flood into the system faster with more glucose and more insulin, which further suppresses the L cell, which further reduces GLP-1. At the same time, a person who eats carbohydrates may have a diminished satiety signal to those carbohydrates. So this is, I think, a very important point, and it's a good one to, as we wrap up, imagine thinking back to that 1996 paper published in the journal Gut. The lead author is a guy named, the last name was Ranganath. Imagine those two groups. The lean group eats that high carbohydrate meal and has a huge GLP-1 response, whereas the obese group eats that same meal and has essentially no GLP-1 response, then consider that in light of the satiety effect that comes from GLP-1, where one person eats that high-carb meal and with a big GLP-1 response has quite a significant feeling of satiety. They feel full. They don't need to eat anymore. But the other person who doesn't have any GLP-1 response eats that same carbohydrate meal and does not get that satiety signal from GLP-1, they're not making a lot of GLP-1. They may then go back to that plate and go back for seconds and thirds. So driving this overconsumption, GLP-1 receptor agonist drugs are powerful, I think, and they work because they bypass this disrupted system. They don't repair it necessarily, but they perhaps create the metabolic milieu that can repair it. If a person is able to use these GLP-1 drugs and learn to control their cravings for carbohydrates, then their insulin levels will come down. Perhaps that will restore the function of the L cells and allow them to begin producing a normal GLP-1 response. Something to think about. Regardless, we've come to the end. Thanks for listening. I will see you next time. Until then, remember, more knowledge, better health.