The Peter Attia Drive - #140 - Gerald Shulman, M.D., Ph.D.： A masterclass on insulin resistance—molecular mechanisms and clinical implications

00:00:00.000 Hey, everyone. Welcome to the drive podcast. I'm your host, Peter Atiyah. This podcast,

00:00:15.480 my website, and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:19.800 into something accessible for everyone. Our goal is to provide the best content in health

00:00:24.600 and wellness, full stop. And we've assembled a great team of analysts to make this happen.

00:00:28.880 If you enjoy this podcast, we've created a membership program that brings you far more

00:00:33.280 in-depth content. If you want to take your knowledge of this space to the next level,

00:00:36.840 at the end of this episode, I'll explain what those benefits are. Or if you want to learn more now,

00:00:41.740 head over to peteratiyahmd.com forward slash subscribe. Now, without further delay,

00:00:47.740 here's today's episode. My guest this week is Dr. Gerald Shulman. Jerry's a professor of medicine

00:00:54.540 and cellular and molecular physiology at Yale. He's also the co-director of the Yale Diabetes

00:00:59.660 Research Center. In 2018, he received the Banting Medal for Scientific Achievement, arguably the most

00:01:06.020 prestigious award one can win in this field. He's pioneered the use of magnetic resonance

00:01:11.900 spectroscopy combined with mass spec to non-invasively examine cellular glucose and fat metabolism. Now,

00:01:18.060 you may ask, why does that matter? And we get to that right at the outset of this interview.

00:01:22.600 If you want to understand insulin resistance, if you want to understand hyperinsulinemia,

00:01:28.740 type 2 diabetes, non-alcoholic fatty liver disease, you have to understand the movement

00:01:34.160 of glucose and fat. A way to think about this is to think about it this way. If you go and get a

00:01:40.440 blood test, even the most fancy blood test imaginable, you're basically looking at a picture,

00:01:46.480 a snapshot, a moment in time. What these techniques that Jerry and his collaborators

00:01:52.500 have developed over many years are basically allowing you to watch videos. You can see the

00:01:59.580 flux. You can see the movement of glucose. Furthermore, they've been able to see those

00:02:04.760 things as they happen inside even human cells. So I'm going to, I guess, maybe make an apology at

00:02:12.500 the outset of this and say that this is about as technical an interview as I probably have done in

00:02:18.140 a while. There are probably only a few interviews that I've done on this podcast that get into more

00:02:23.420 technical weeds than this one. But unfortunately, that is the price one has to pay if they want to

00:02:29.200 understand arguably the most important pathologic condition in our species. And I get into what I

00:02:37.040 mean by that in the interview. So I won't elaborate on that further. We talk a lot about what it is to

00:02:44.760 be insulin resistant. This term gets thrown around with such ubiquity that you'd think everybody knows

00:02:52.020 what it means. And yet to define it is very complicated. But I think by the end of this

00:02:56.060 interview, you will undoubtedly understand how to define insulin resistance. And I think you will have

00:03:01.320 a very good sense of where it begins and what it means in a subclinical versus a clinical state

00:03:06.900 and what the sequence of events are that lead to a condition in the muscle, ultimately affecting the

00:03:13.000 liver, ultimately affecting the whole body. I'm not going to promise you that you'll get that on the

00:03:17.920 first listen. I think that might require more than one listen, and it probably requires going through

00:03:23.260 the show notes, which will be littered with fantastic figures. In fact, as Jerry and I did this

00:03:30.660 interview over Zoom for, I don't know, I would say maybe half to two thirds of the interview,

00:03:36.760 he was sharing slides as we were doing it. And keep in mind, I'm quite familiar with these concepts.

00:03:42.160 So for someone who's less familiar with these concepts, I think it will be only that much more

00:03:45.940 valuable. One really nice little bonus thing that came out of this is a beautiful discussion at the

00:03:51.300 end of Metformin, where I actually learned something really profound that was incredibly relevant to my

00:03:57.200 understanding of Metformin in my ongoing interest around the question of Metformin's suitability

00:04:03.260 as a longevity agent. I could say more about this, but I think it's just one of those things

00:04:08.700 where I ask you to sort of take a leap of faith with me that this is important, even if it feels a

00:04:14.180 little bit like drinking your cough syrup at times. But if you really want to understand longevity,

00:04:18.800 you're going to have to sort of figure out what insulin resistance means.

00:04:22.400 So without further delay, please enjoy my very in-depth conversation with Dr. Gerald Schoenberg.

00:04:34.100 Jerry, thank you so much for making time to sit down virtually with me today. As I said before we

00:04:41.400 hopped on, this is a topic that is near and dear to my heart. And frankly, all roads seem to point to

00:04:47.260 you. And that goes back to, I don't know, at least for me, probably 2011, when I became really

00:04:53.620 fascinated by this topic. And there aren't a lot of topics where I've personally experienced the

00:05:00.540 following problem, which is the more I think I understand it, the less I do. So now when someone

00:05:06.720 says to me, Peter, what's insulin resistance? You know, I can sort of give glib answers to that

00:05:12.160 question. But the reality of it is, I don't think I fully understand what it is. And I don't know that

00:05:17.840 I can represent to the listener that by the end of this, they will fully understand what insulin

00:05:21.700 resistance is. But what I think they'll understand is how maybe we can think about it through the lens

00:05:28.260 of different tissues and what may or may not be going on. And in large part, I think that's due to

00:05:32.400 the incredible work you've done over your entire career. I guess I'd like to kind of just start with a

00:05:37.820 little bit of background. You did an MD and a PhD and you're trained as an endocrinologist, correct?

00:05:42.160 Yeah, that's correct. And then I did residency in medicine at Duke, fellowship in endocrinology at

00:05:47.900 the Mass General Harvard. And then I've always been interested in metabolism, diabetes. I guess

00:05:53.760 probably my father was a diabetologist, went to summer camp. He was the doctor for type 1 diabetics. At

00:06:02.600 an early age, I was exposed to problems, type 1 diabetes in my peers. I was just a camper and saw my

00:06:09.900 peers getting hypoglycemic or getting into issues with ketoacidosis. So I think I was exposed to

00:06:16.300 metabolism at an early age. I'm sure it left an impression on me. My father wanted me to become a

00:06:21.180 radiologist because of my physics background, but I ended up staying in metabolism and doing

00:06:25.840 endocrinology. I'm sure you would have done great things in radiology, but I also think we're far better

00:06:30.480 off for the contributions you've made in this field. When did this particular question of

00:06:36.600 understanding what insulin resistance meant and actually starting to differentiate between

00:06:43.080 some of these phenotypes of what is the fate of glucose in a person with normal metabolism versus

00:06:50.060 what is the fate of ingested glucose in someone with type 2 diabetes? When did that question begin

00:06:54.920 to obsess you? And specifically, now that's a sort of change from the patients that you grew up with,

00:06:59.640 type 1 diabetes. Studying as an undergraduate medical school, I was always interested in

00:07:04.120 biochemistry, physiology. I had an experience. I was visiting a medical student at Vanderbilt in the

00:07:10.000 70s and got interested in in vivo metabolism, studying metabolism in awake animals, looking

00:07:17.100 specifically at glucose and fatty acid turnover, using tracers to actually measure how fast things are

00:07:24.060 being made, glucose is made, how fast fatty acids are being made in the body and metabolized.

00:07:28.420 In medical training, you go back to medical school, you learn how to become a good doctor,

00:07:32.840 take care of patients. But then in your fellowship years, you're back in the lab.

00:07:37.060 And I really wanted to get back to understand metabolism by looking inside the cell. So everything

00:07:43.280 I had done prior to that, and most people studying biochemistry, physiology would, to understand.

00:07:49.620 So diabetes, metabolic disease, I was interested in this question. It's an important disease,

00:07:54.420 leading cause of blindness, end-stage renal disease, leading cause of non-traumatic loss of limb. The

00:08:00.340 cost to U.S. society is huge impact. And now it's becoming a global problem as they adapt to westernized

00:08:07.860 diets and things. And I wanted to look inside the cell, metabolism inside the cell. And so that took me

00:08:13.060 into the world of nuclear magnetic resonance spectroscopy and actually brought me down to New Haven,

00:08:18.100 where they were just setting up methods, this technique to actually look inside living yeast

00:08:23.680 cells. And I said, gosh, this, we can adapt this to humans and look inside metabolism in humans,

00:08:29.780 in liver and muscle and other organs. To specifically get at your question, I think it's such an important

00:08:35.160 metabolic disease, the most common metabolic disease. And so someone who's interested in metabolism,

00:08:39.440 it's a natural segue.

00:08:40.600 I sometimes describe it to patients as the foundation upon which the major three chronic

00:08:46.200 diseases sit. So you described some ways in which patients with type 2 diabetes die,

00:08:52.100 specifically through amputations or complications of amputations, such as infections, and obviously

00:08:56.940 through end-stage renal disease. But I would argue that the majority of the mortality through diabetes

00:09:01.980 comes not so much through diabetes, but through its amplification of atherosclerotic disease,

00:09:07.100 cancer, and dementia, all of which are force multiplied in spades by type 2 diabetes.

00:09:13.220 So the way I explain it to people, and I hope that by the end of this, you'll help me refine this

00:09:18.160 because it may not be accurate, but I describe to patients that there is a continuum from hyperinsulinemia

00:09:23.900 to impaired glucose disposal to NAFLD and NASH to type 2 diabetes. And that continuum makes up a plane

00:09:31.900 upon which all chronic disease get worse. If we're going to be serious about the business

00:09:36.600 of delaying the onset of death, we have to be serious about the business of delaying the onset

00:09:41.160 of chronic disease. And if we want to do that, we must fix our metabolisms. That's my thesis.

00:09:47.580 Total agreement. You're spot on. So insulin resistance is the main factor which leads to type 2 diabetes,

00:09:55.660 but it also, and again, this is give credit to Jerry Riven, who in his 1988 Banting lecture first

00:10:01.900 got everyone's interest in basically saying insulin resistance is not only leading to diabetes, but

00:10:09.120 as you say, atherosclerosis, basically hyperlipidemia associated with inflammation,

00:10:15.960 high uric acid, polycystic ovarian disease. Now we can kind of add to that. Now we can talk about

00:10:21.880 NAFLD, or I prefer the term metabolic associated fatty liver disease, MAPLD. That's going to be the most

00:10:28.440 common cause of liver disease, liver inflammation, end-stage liver disease, and liver cancer.

00:10:34.040 And finally, another arm, you know, for Jerry's circle of insulin resistance and all these arms

00:10:39.060 butting off of them, heart disease, as we talked about, high uric acid, high triglycerides, and high

00:10:44.740 cholesterol, is cancer. So we're now, as you know, seeing huge increases in many forms of cancers,

00:10:51.920 which are associated with obesity, breast cancer, colon cancer, pancreatic cancer, liver cancer.

00:10:57.840 And by the bed, it's insulin resistance that's driving the increase in all of these cancers. Now

00:11:03.800 it's not causing them necessarily, but it's promoting the growth. And again, we have very strong

00:11:08.320 preclinical evidence for this in animals. You can take animals, Rachel Perry, who was in my group and

00:11:14.120 now starting her own lab, has taken breast cancer models, human breast cancer models, colon cancer, 0.62

00:11:19.220 put them into mice, and just giving them insulin, putting in insulin pumps just simply. And that

00:11:24.580 rate of tumor growth is accelerated, and you treat them with an insulin-sensitizing agents,

00:11:29.960 and you can slow down tumor growth. So I think you're spot on, Peter. Insulin resistance is driving

00:11:35.700 a lot of disease, and you're also spot on in that that's what's killing our patients with type 2

00:11:41.180 diabetes. It is heart disease. These other things are the chronic complications of hyperglycemia,

00:11:45.880 the blindness, the end-stage renal disease, and the small vessel disease leading to non-traumatic

00:11:51.080 loss of limb, also hyperglycemia. But insulin resistance, which is very common, it's probably

00:11:56.920 one quarter of our population and one half of our population has it perfectly asymptomatic. You

00:12:03.080 don't know you have it. We can test for it using sophisticated tools that we can talk about, but

00:12:07.960 it's a very common phenomenon. So before we launch into what I think is an important discussion around

00:12:15.560 the fate of glucose under normal conditions, which is the backdrop against which I think

00:12:20.760 everything we are going to talk about has to be laid out, I'd like you to spend a moment doing

00:12:25.720 something you're probably not asked to do often, which is at least explain to some extent what the

00:12:31.000 NMR technique allows you to do. Because so much of what we're going to talk about today requires

00:12:39.240 either a leap of faith that you know what you're talking about, or at least some sense of how a

00:12:44.920 scientist is able to actually look at substrates and substrate utilization and substrate movement in the

00:12:54.600 ways that we have to be able to talk about them at a molecular and cellular level to make sense.

00:12:59.480 So I know it's a bit complicated, but because it is such a cornerstone of your work, can you explain

00:13:05.800 what labeling means and how you can measure those labeled molecules in vivo?

00:13:12.760 In metabolism, the traditional methods since going back to dates maybe 50 years ago, when you wanted to

00:13:19.640 measure more than just concentration of a metabolite, you go to your doctor, you measure blood sugar,

00:13:25.560 cholesterol, and it's a static concentration. And what we know is what's much more important than just

00:13:31.080 measuring concentration is flux. And that's basically production versus uptake by a tissue and know where

00:13:38.680 something's being made, where it's going. And the traditional approach has been to put a label on that,

00:13:45.480 whatever you're interested in tracing, glucose. And so you used to basically, with the advent of

00:13:51.000 cyclotrons, it really started in California, in Berkeley, they started, you know, had cyclotrons,

00:13:56.600 they're interested in nuclear theory. The side product is you can make isotopes. So you can make

00:14:02.440 carbon radio labeled, so it's an emitter, and put that carbon onto a glucose molecule and then trace

00:14:08.440 it. So for more than 50 years, we've been able to buy radio labeled isotopes and put a carbon,

00:14:15.320 C14, which is radioactive, low dose radiation or tritium, which is a form of hydrogen, and then

00:14:22.040 give it to a person, animal, and do blood sampling and actually measure then turnover of that metabolite.

00:14:29.240 So that's telling us very important information. Many, many important studies have used this,

00:14:33.800 and to date we still use this, to track production and clearance of whatever we're labeling. What you

00:14:40.440 can't get from that, though, is really what's happening inside the cell, which is really where

00:14:45.640 I wanted to go. So we've been measuring turnover of metabolites. And again, that's what I did many

00:14:50.360 years ago where I first started my interest in metabolism. To do that, you need to get inside

00:14:55.320 and look at the cell. So the approaches have been traditionally something called positron emission

00:15:00.120 tomography, which is now used clinically sometimes to track tumor growth because tumors take up glucose.

00:15:06.120 And you can give a PET emitter of glucose and then see if the tumor is taking it up. That's

00:15:11.320 radioactive. And again, I'm a clinical physiologist. I'd prefer not to give radio

00:15:15.880 labeled substrates, radioactive substrates to volunteers who volunteer for my study.

00:15:21.240 The other approach was nuclear magnetic resin spectroscopy. There were two groups that were

00:15:27.000 pioneering this kind of work. There was one group in George Rada at Oxford, and this was phosphorus NMR.

00:15:33.160 And so what George was doing, so NMR takes advantage of the fact that the nuclei of certain

00:15:40.440 atoms have spin properties. And I won't get into all the physics behind this, but they make them

00:15:46.120 behave like tiny bar magnets. And so when you put them in a strong magnetic field, they tend to line

00:15:51.000 up or against this magnetic field. And because they have spin properties, they will actually precess

00:15:56.840 in this magnetic field at a set frequency. If you pulse them at the frequency that they're precessing,

00:16:03.480 they tip out of this field. And then when they relax, they emit energy that you can pick up with

00:16:08.360 a little radio, with an antenna and basically get chemical information about where that label is

00:16:15.080 within the molecule. So everything I just said, all you need to understand is you can use this method to

00:16:21.960 basically measure the amount of the metabolite. More importantly, which, for example, carbon

00:16:29.080 atom within that glucose molecule is labeled. It has something called chemical shift experiences

00:16:34.280 a slightly different magnetic field, depending where it is within that glucose molecule.

00:16:39.960 So for the listeners, all you need to understand is using this method, we're able to get

00:16:45.320 biochemical information of not only measuring a metabolite, but then using the power of, for example,

00:16:51.880 C13 NMR, track the label as it's being metabolized inside the cell. So that's carbon NMR. So in our

00:16:59.240 bodies, 99% of the carbon in our body is C12, which is NMR invisible, but 1% is C13, which is NMR visible,

00:17:08.680 has this precession properties. You can use a label, for example, C1 label glucose, and then track

00:17:15.960 that C1 glucose as it gets into the, say, a muscle cell or liver cell and gets metabolized and finds its

00:17:22.840 way into glycogen. And then you can measure flux. You can actually, for the first time in humans,

00:17:28.360 non-invasively, without any ionizing radiation, measure how much is going in through, measure

00:17:35.320 intracellular pathway flux. Phosphorocentomar, as getting back to George Roddy, George pioneered

00:17:41.640 phosphorus NMR. There you don't have to give any isotopes. There you actually see P31, phosphorus

00:17:47.800 31 is 100% natural abundant. You see all the phosphorus that's in solution in our bodies. So,

00:17:54.280 for example, when our volunteers go inside a magnet and we put a leg or arm into the magnet,

00:17:59.480 we can see all the high energy phosphates in, for example, ATP, adenosine triphosphate. There are three

00:18:06.840 phosphates. And you can actually see each one of those phosphates. You can see phosphocreatine has

00:18:13.640 a different chemical shift. You can see your inorganic phosphate. And we developed methods,

00:18:18.520 Doug Rothman and others at Yale, who I worked with, were able to develop methods to measure glucose 6

00:18:23.640 phosphate. So we can actually look at one, a key intermediate, getting glucose from outside,

00:18:27.560 inside. Another method we developed was we can measure intracellular glucose inside human muscle,

00:18:33.160 non-invasively. So by measuring these metabolites, measuring flux, we can actually then ask the very

00:18:39.880 simple questions, which this is how we started out. In humans, as you say, you know, again, diabetes

00:18:46.120 is an abnormality of metabolism. Glucose is the metabolite. And we were able to basically ask very

00:18:51.880 simple questions when a person, a human, which is my favorite model because it's the one most relevant

00:18:57.720 understanding diabetes and metabolic disease. When we ingest carbohydrate, how much of that

00:19:04.280 carbohydrate ends up in glycogen versus oxidation into carbon dioxide or gets converted to lactate

00:19:11.880 through glycolysis? And so, and then more importantly in the patient with, or the volunteer with diabetes,

00:19:17.880 how important is that pathway glucose to glycogen accounting for their insulin resistance?

00:19:23.320 This story is very short. Before you go there, let's demonstrate clinically a difference between

00:19:30.360 these two people. So let's take the normal patient without type 2 diabetes, and then let's contrast them

00:19:37.640 with a very similar person of similar size who has type 2 diabetes. We will feed them both a high

00:19:44.920 carbohydrate meal in the evening. Let's just assume that that meal contains 100 to 200 grams of

00:19:51.560 carbohydrate. They will digest their food. We won't really have much insight into what's happening

00:19:57.640 overnight. You will tell us. But at the next morning, we do a fasting blood glucose level on

00:20:03.560 them. This is now 12 hours after their meal. The patient who does not have type 2 diabetes might arrive

00:20:10.840 with a blood sugar of 100 milligrams per deciliter, which we will say is normal. His counterpart with type 2

00:20:18.040 diabetes may actually, at that time, may actually have a blood sugar of 200 milligrams per deciliter,

00:20:24.200 which is obviously abnormal and consistent with the diagnosis of type 2 diabetes. Now, of course,

00:20:29.960 that only represents about an extra 5 grams of glucose in the circulation that is the difference

00:20:35.960 between the 100 and the 200 milligrams per deciliter, which is a small fraction of the, call it,

00:20:41.880 1 to 200 milligrams, pardon me, grams rather, 5 gram difference. So it's a small fraction of what was

00:20:48.600 ingested the night before. What is the difference between those two people? Why does one of them have

00:20:55.160 such a hard time with that extra 5 grams of glucose? What was the fate of glucose in the

00:21:02.200 healthy person to begin with? How did the body treat it? The body, when you take in, and again,

00:21:07.320 this is what we were able to demonstrate by actually measuring glycogen flux in liver and muscle,

00:21:12.920 that ingested in a healthy individual ends up as mostly liver and muscle glycogen. It takes up muscle

00:21:19.960 and depends on the size of the meal and how it's being administered, the proportionality between

00:21:26.440 liver and muscle. But bottom line, it's 80, 90% is stored as glycogen. In the diabetic contrast is there's

00:21:35.160 two processes that have gone awry. One is that the liver is geared up to make more glucose than it

00:21:44.840 should be through a process called gluconeogenesis, the conversion of non-glucose precursors like amino

00:21:51.320 acids, alanine, and lactate to glucose. And that process is accelerated. So the liver is making twice the

00:21:58.600 amount of glucose as it should. And then you have a block in the periphery where the glucose, same

00:22:05.400 amount of insulin is not causing the glucose to be taken up by the muscle. Again, in terms of flux,

00:22:10.920 what I care most about, production is up and clearance or disappearance is down. And besides this,

00:22:16.920 also even in some diabetics, insulin is inappropriately low because we know if we give more insulin,

00:22:23.000 we can overcome these abnormalities. And so the beta cell has also become abnormal in the established

00:22:29.320 diabetic where it's not making enough insulin. That can be secondary to these other issues, glucose

00:22:34.040 toxicity and other factors that have caused this beta cell impairment. Because we know,

00:22:39.320 most importantly, when we reverse the insulin resistance, this is a very important study,

00:22:43.640 is we've taken these type 2 diabetics and short-term hypocaloric feeding, 1200 calories a day,

00:22:49.560 we basically can reverse all these abnormalities through reduction in ectopic lipid, which we can

00:22:55.880 get into molecular mechanisms and reverse their diabetes. And this has now been shown by many,

00:23:00.600 many other investigators. And most recently, Roy Taylor, my colleague who trained with us,

00:23:05.560 is now doing this in the primary care clinic back in the UK. But usually, you've asked the question,

00:23:11.560 usually when we talk about diabetes, actually, it may be easier to understand when you start

00:23:18.200 in the young, lean 20-year-old who already has insulin resistance. These are the young, lean

00:23:24.360 college students that we study. It's actually easier, I think, for your listeners to understand

00:23:28.840 if we start with just pure insulin resistance, which we see is the most common thing. As I said,

00:23:33.880 probably half the people in the US actually have insulin resistance, don't know it because they're

00:23:38.600 asymptomatic. And we even see this in young, lean 20-year-olds, Yale undergraduates who volunteer for

00:23:45.080 our studies, profound insulin resistance in muscle, no problems in liver, and then take you through the

00:23:50.680 progression from how you just go from insulin resistance in muscle to fatty liver and insulin

00:23:57.000 resistance in the liver, and then progress to type 2 diabetes. That's something we can actually go

00:24:02.120 through if that would be of interest. It would because it actually kind of fits with the way

00:24:07.800 I was going to try to temporally split this, which would look as follows. When we take a patient who

00:24:14.680 has normal fasting glucose and normal fasting insulin, and we challenge them with an oral glycemic load,

00:24:22.680 and then measure insulin and glucose in 30-minute intervals, a lot of times we expose a problem that

00:24:30.360 seems most easily explained by the muscle's inability to assimilate glycogen. So a person shows up and they

00:24:38.840 have a normal fasting insulin, say it's 5, and their fasting glucose is say 90. You challenge them with

00:24:45.240 75 to 100 grams of glucose, but say 30 or 60 minutes later, their fasting glucose is 200, their insulin

00:24:52.840 is 70. We call that insulin resistance. And we impute from that that something has broken down in the

00:25:01.880 pathway that prevents their muscle from taking in glucose. Now, you've done very elegant work to

00:25:08.920 examine all of the possible places that failure could have taken place. Did it take place at the

00:25:15.000 GLUT4 transporter or one of the mechanisms which we should discuss how the GLUT4 transporter gets across

00:25:20.280 the cell membrane? Is it a problem not of bringing glucose in, but really is the problem downstream

00:25:27.400 at hexokinase or glycogen synthase, things like that. So is that sort of what you're saying,

00:25:32.040 which is can we start with postprandial hyperglycemia? Yeah, I think we're not even

00:25:36.280 hyperglycemia. This is before any abnormality, just insulin resistance. What I like about the question

00:25:42.200 you asked and how you pointed out, insulin resistance is the root cause for not only diabetes,

00:25:48.520 but it's going to be the root cause for all these other abnormalities, fatty liver disease,

00:25:54.360 make us prone, makes a lot of cancers worse. Heart disease, and again, that's the number one killer

00:26:00.600 in this country. It's insulin resistance that's driving all these things. And not even talking

00:26:05.000 about, even though I'm a diabetologist, I, of course, care. I want to fix diabetes. But even before

00:26:11.000 blood sugar goes up, which is how we define diabetes, let's understand insulin resistance. Because

00:26:16.440 if we can understand insulin resistance, then that's going to be the best way to fix diabetes,

00:26:20.920 type 2 diabetes. Heart disease is going to make a big impact there. Fatty liver disease and slow

00:26:26.760 down cancers. So let's start with insulin resistance. Okay. What is insulin resistance? So we define it by

00:26:32.920 giving insulin. And we know insulin normally does some effects, makes glucose being taken up by liver

00:26:39.720 and muscle. And when that same amount of insulin is not doing these things, we say there's insulin

00:26:45.640 resistance. So you need more insulin than to cause muscle to take up glucose or the liver to turn off

00:26:52.360 glucose production or take up glucose. And the same thing, again, in the fat cell. What insulin does in

00:26:57.640 the fat cell is that puts the breakdown of fat. It's called that lipolysis or take up glucose to esterify

00:27:05.080 fatty acids into glucose. So these are the three key insulin responsive organs. And when insulin is not

00:27:12.120 doing that properly, we call that insulin resistance. And again, keep emphasizing, I think for your

00:27:18.280 listeners, this is probably every other person in this country or in Western Europe are insulin

00:27:24.680 resistant. Your doctor won't even know this unless they do careful maybe studies to assess insulin

00:27:30.360 resistance because you won't pick this up as the simple plasma glucose test. So what causes resistance?

00:27:36.840 Let's start with muscle. And the reason I like to start with muscle is when we study our young

00:27:43.480 volunteers, again, I like them because they're perfectly healthy. They're 20 years of age, 19 to 20.

00:27:50.840 They're lean because we know everyone who's overweight or obese probably has insulin resistance. There's so

00:27:56.040 many confounding factors that happen in overweight obesity. These are lean 22, 23 BMI, lean. Non-smoking,

00:28:04.520 so we screen out smoking. No medication, no drugs. And sedentary because we know people who exercise,

00:28:11.640 we can reverse insulin resistance and we can talk about how that happens. So you give these young

00:28:16.920 20-year-olds, let's say you screen, we screen to this date probably 1,000, but you get a distribution,

00:28:22.040 given a drink of glucose tolerance, 75 grams, you measure insulin, and you can calculate insulin

00:28:29.080 sensitivity index. It's a crude index, and it's kind of a bell-shaped curve.

00:28:33.400 And you have people in the bottom quartile who are insulin resistant by definition in the top quartile.

00:28:40.680 Then you ask, why are those people in the bottom quartile insulin resistant? And you measure

00:28:45.800 glycogen synthesis using the methods we talked about briefly, carbon, NMR, give C1 glucose, measure

00:28:52.680 flux into glycogen. And it's already down by 50% compared to the sensitive ones under matched insulin

00:29:00.040 and glucose concentrations. So they're resistant because they can't get glucose in the glycogen.

00:29:05.800 That's the major pathway. It's not glucose to lactate, not glucose oxidation. So that's your

00:29:11.400 pathway. Then you want to know where the block in that pathway is. With phosphorus NMR, we can measure

00:29:17.160 glucose 6-phosphate inside the cell. With a carbon NMR method, we can measure glucose inside the muscle

00:29:23.160 cell. The reason this is important is we can see where the biochemical block is. So your listeners

00:29:29.560 all just probably, you know, get into a car and they're on the road. And if there's construction

00:29:33.560 going on, we all know construction piles up after that, wherever that roadblock is,

00:29:39.080 where the construction is happening. Biochemistry is the same thing. You have a roadblock

00:29:43.160 and traffic builds up behind it. So we measure G6P to argue, you mentioned about three steps, synthase,

00:29:51.800 hexokinase, and transport, glucose transport. And they had all been implicated to be the roadblock,

00:29:59.160 the step response for the insulin resistance. And so we were able to sort out which was rate

00:30:04.040 controlling by measuring these intermediates. So if the block is at synthase, glucose 6-phosphate

00:30:10.520 should build up and glucose should build up. If the block is at hexokinase, you should basically

00:30:16.200 have lower G6P and a buildup of glucose. And if the block is at transport, there should be

00:30:21.240 reductions in both glucose 6-phosphate and glucose. Through a series of studies, we found in not only

00:30:27.560 these young lean insulin resistant offspring, but obese insulin resistant individuals, as well as

00:30:33.080 individuals with poorly controlled diabetes, G6P, glucose 6-phosphate and glucose are both reduced

00:30:39.960 in the muscle cell in vivo, in humans, implicating transport. That's where your biochemical block is.

00:30:47.000 So the block is at transport. That's your target to fix if you want to fix muscle insulin resistance.

00:30:53.640 And the corollary is these other steps are not good targets, drug targets to go after to fix insulin

00:30:59.720 resistance in muscle. This is the first abnormality we found in its transport and in these young,

00:31:06.760 healthy 20-year-olds. And then the question is, what's wrong with the transport mechanism?

00:31:12.280 That led us into the world of lipids. Again, it's been known for decades that obesity associated

00:31:17.640 with insulin resistance. That's why virtually every obese adult or child have insulin resistance. There

00:31:23.800 are rare exceptions. And then we basically found, we developed a method to measure fat inside the muscle

00:31:30.120 cell. And that was the best predictor for insulin resistance in the muscle and this block and

00:31:35.160 translocation. Let's give people a quick primer on normal glucose disposal into a cell. So when the

00:31:45.320 insulin molecule hits the insulin receptor on the surface, I believe it autophosphorylates itself,

00:31:52.600 correct? That then signals to insulin receptor substrate one, IRS one inside the cell. So that

00:32:00.920 sends a signal inside the cell, which also leads to a phosphorylation, which then signals PI3 kinase.

00:32:07.960 It upregulates PI3 kinase. And that basically leads to a GLUT4 transporter, which you can think of as like a big

00:32:16.360 tube being shoved up to the surface of the cell across its membrane. And that basically passively

00:32:24.040 allows glucose in. It is not an active transporter, correct? That's correct. Everything you said is

00:32:29.080 spot on. Basically up until now, we don't know where the breakdown is in that whole process. All we know

00:32:35.240 is that something is impaired in getting glucose in the cell. But in terms of, is it, there's not enough

00:32:42.520 insulin that hits insulin receptor. Is there something wrong with IRS one with PI3 K? Is there

00:32:47.720 something blocking the transporter? We're going to have to figure that out still, but you've already

00:32:51.720 taken two thirds of this puzzle off the plate by saying, we know it's not downstream of that.

00:32:58.680 That's correct. If you fix the transporter, that's where the roadblock is. And that's the target.

00:33:04.440 The next set of questions becomes, why isn't insulin causing? And as you point out,

00:33:10.280 this translocation of the group four transporter to the membrane to allow glucose to come into the cell

00:33:18.360 through facilitated transport down a gradient. So that's what we can talk about next, if you want to.

00:33:23.400 That's perfect.

00:33:24.440 Can I share my screen with you at this point?

00:33:26.520 You can. And what we will do, Jerry, is we are going to take everything that you are sharing with me,

00:33:33.560 and we're going to turn these into show notes that will be time stamped to this part of the discussion.

00:33:39.880 Because while I guess people like you and I do tend to picture these things in our head easily,

00:33:46.600 I think for many people, it is going to be incredibly helpful to be able to actually

00:33:50.920 look at some biochemical drawings. I benefit from this greatly. It's still not purely second nature to me.

00:33:57.160 I like to think in pictures too. So as much as we can help the audience out with graphics,

00:34:01.960 I think it will be beneficial. So here's a cartoon. I'll walk you through this and stop me if you have

00:34:07.560 questions. This is a cartoon of a muscle cell. We went through how insulin normally works. Insulin

00:34:13.800 binds to the receptor and everything, as you said, we're going to actually show this in this cartoon,

00:34:18.520 binds to the receptor. The receptor autophosphorylates, becomes a kinase. The key substrate for this

00:34:24.120 kinase, this receptor kinase in muscle's insulin receptor substrate one, which undergoes tyrosine

00:34:30.120 phosphorylation, allows it to bind and activate this other protein, PI3 kinase, which Lou Cantley

00:34:36.280 discovered. And that's a required step for translocation. So that's all been worked out.

00:34:43.080 And somehow this is not working. This is broken in the insulin resistant individual. And again,

00:34:48.920 these young 20 year olds, the patient with diabetes, the obese insulin resistant individual,

00:34:54.120 and the question is, what's wrong? So I'm going to share with you at least my view,

00:34:58.680 which would explain insulin resistance in most situations of lipid induced insulin resistance,

00:35:04.520 which I think accounts for, I would say the majority of these patients I see who have type 2 diabetes,

00:35:11.960 or who are obese and insulin resistant, or even these young lean insulin resistant offspring. And so this is

00:35:17.560 the picture. So here, and it relates to fatty ass fat metabolism. Before I told you the other MR method 0.95

00:35:25.240 that we developed is actually something called proton NMR. And this is actually most of your listeners

00:35:30.440 are very familiar with. Everyone knows about MRI, magnetic resonance imaging. This is people go into a

00:35:35.480 scanner and they get very pretty pictures of, of an organ brain or some other organ for diagnostic

00:35:41.560 reasons. And it's the same biophysical principles. You're, you're basically getting this NMR signal

00:35:49.080 from protons and protons are the most abundant NMR visible nucleus in the body. And it's mostly water

00:35:55.880 we're looking at. So when you're basically getting the same signal from protons and mostly protons are

00:36:01.640 water and fat. And so an imager gives you this three dimensional reconstruction of proton density

00:36:08.120 in water and fat. And that's what gives you the images. And again, we're doing biochemistry. So

00:36:12.280 we're getting, taking that same kind of information, but actually looking at individual carbon atoms or

00:36:18.520 phosphorus atoms, or in this case, protons lining the carbons and triglyceride. It's fat. So what I said,

00:36:25.000 using proton NMR to measure fat inside the cell, this is different from fat outside the cell. So if you look

00:36:32.280 at a steak and you see the marbling of fat in a steak, that's fat outside the muscle cell. What you don't

00:36:37.800 see if you look at a steak is the fat inside the muscle cell. And using NMR, we can actually discern

00:36:44.600 fat outside the cell versus fat inside the cell. We can do this in many organs and muscle started in

00:36:50.840 muscle. And using this approach, we found fat inside the muscle was the best predictor for this block and

00:36:57.080 transport in all of our volunteers, young people, old people, children, sedentary individuals, sedentary

00:37:04.040 individuals, fat inside the cell is the best predictor for insulin resistance. And so this

00:37:08.360 led us into the world of lipids. We're keen to understand then finding the lipid intermediate

00:37:15.000 that might do this. And in studies where we took healthy individuals, perfectly normal sensitivity,

00:37:22.600 we gave them an intralipid infusion, just raised plasma fatty acids for three to four hours, and found

00:37:30.760 that after three to four hours, we can make them as insulin resistant as anyone with type 2 diabetes.

00:37:36.280 And others had shown that in addition to us. I mean, this is, we weren't the first to show this, but

00:37:40.920 what we were the first to show is it's due to this block in glycogen synthesis. And it's the same block,

00:37:47.080 it's that block in transport. Just to be clear, when you deliver intralipid, that's intravenous lipid,

00:37:54.440 as a triacylglycerol or diacylglycerol? No, this is a triglyceride. This is an emulsion,

00:38:00.920 a triglyceride emulsion. It's often given to patients for hyperalimentation. When they can't

00:38:06.520 eat, you give us energy rich infusions. Just like TPN or something like that?

00:38:11.080 It's TPN. It's used in TPN often. But what we also do is just a little low dose of heparin to

00:38:17.320 activate lipoprotein lipase. So all of a sudden then you can artificially raise fatty acids twofold,

00:38:24.440 something, you know, up to about, you know, one and a half millimolar and ask the question,

00:38:28.760 what does this do? What does this have to do with, does it ultra metabolism? And it has profound

00:38:33.240 effects. So by increasing LPL expression... Not expression, activity. I did not know that

00:38:40.120 heparin activated LPL. So by activating LPL with heparin, cool trick to know, I'll keep that in mind,

00:38:46.440 you're going to get more of that lipid into the muscle cell. You will raise fatty... So what the

00:38:52.680 heparin does is it causes activation of lipoprotein lipase, and that will then break down the triglycerides

00:39:00.040 to raise fatty acids and more deliver fatty acids to all cells in the body. Yeah.

00:39:04.360 Okay. So this becomes basically a quick vehicle by which you can deliver lipid directly into the

00:39:12.360 muscle cell. Exactly. Where you can acutely change that. And again, you can't do this just by getting

00:39:18.440 fatty acids. Fatty acid turnover is so fast, you can't just infuse fatty acids to significantly

00:39:24.440 raise. So this is a way we're able to raise fatty acids specifically in vivo, in humans,

00:39:29.560 and we do this in animals. And so it's a nice pharmacological way of asking the question,

00:39:35.320 what impact does just simply raising fatty acids for a few hours have on metabolism? And it's profound.

00:39:41.240 It takes three to four hours before you see this, and then boom, you get very profound insulin

00:39:46.680 resistance. And in our early studies, again, we showed using the same methods I told you about,

00:39:51.960 measuring glucose six phosphate, measuring intracellular glucose, measuring glycogen

00:39:56.280 synthesis. We found simply raising fatty acids for three to four hours blocks glycogen synthesis,

00:40:03.080 profound insulin resistance, as I say, as anyone with obesity or type two diabetes. And it's due to the

00:40:09.160 same, an acquired block in transport, insulin activation of transport, both G6P and glucose

00:40:14.440 are down. So that to us was a very important lesson because it basically changed the paradigm,

00:40:20.840 because prior to this, people, workers, biochemists, you may know the name Philip Randall,

00:40:26.600 who did some pioneering studies in the 60s at University of Bristol, and was really the first to say,

00:40:32.760 hey, fatty acids may be toxic, may be causing insulin resistance,

00:40:37.080 and did studies in rat tissue, cells, heart tissue, diaphragm muscle taken from rats in vitro,

00:40:44.520 incubated it with fatty acids, and in vitro in the test tube induced insulin resistance.

00:40:50.600 The mechanism that they postulated was that it was altering basically oxidation, the TCA cycle,

00:40:57.720 citrate levels would build up and lead to inhibition of phosphofructokinase, which is a key glycolytic

00:41:03.960 enzyme. The prediction that Randall made was glucose 6-phosphate should increase leading to

00:41:11.400 inhibition of hexokinase. We were interested in that because we said, oh, fat in our hands is

00:41:16.920 important. We're raising fatty acids and causing resistance. And we see this really strong relationship

00:41:23.000 between fat in the muscle cell and insulin resistance in all of our subjects, obese, diabetic,

00:41:28.680 young insulin resistant individuals. And so we wanted to see if his mechanism, Randall's postulated

00:41:34.920 mechanism, translates to humans, because these were all in vitro studies done in tissues taken from

00:41:40.760 animals. So in a series of studies, we took, again, the healthy individuals, raised fatty acids through

00:41:46.600 this triglyceride and little dose of heparin infusion, and found just the opposite to what Randall predicted.

00:41:53.560 They got insulin resistance, which is what he would have said, but not through his mechanism. He said,

00:41:59.560 G6P should go up. We saw it go down, and we saw glucose go down. So it wasn't through inhibition of

00:42:05.880 glycolysis, as he said. It's somehow interfering with the insulin activation of transport. So, and again,

00:42:12.920 same rate-controlling step we saw in all of our diabetics and obese individuals and pre-diabetic

00:42:19.320 individuals. But just to be clear, Jerry, it caused hypoglycemia. The intralipid dropped glucose?

00:42:27.480 No. Raising the fatty acids caused insulin resistance. Inability of insulin to stimulate

00:42:33.880 glucose transport. Okay. Okay. Yep. I may have misheard you, but okay. I'm going to now fast forward. We

00:42:40.200 then took these observations back to the bench. We're able to replicate this in rodents, rats, and mice. And

00:42:46.200 the power, even though I'm most passionate about our human studies, I'm a clinical physiologist and

00:42:53.240 I care most about understanding what's happening in humans, the animal models allow you to really

00:42:58.680 interrogate a biochemical process. There we can get tissue out. In humans, I like to be non-invasive

00:43:04.680 with our MR methodology. But here, sometimes you need to get tissues to measure activities,

00:43:09.880 phosphorylation events. And also, you have the power of mouse genetics. You can knock genes in and out

00:43:15.960 of mice to really rigorously test hypothesis. I should tell you one experiment before I move to

00:43:22.040 this cartoon that we did in humans is we did biopsies in these humans when we raised fatty acids

00:43:29.080 and found this block in transport and asked the question, is a lipid intermediate, a fatty acid

00:43:36.120 metabolite interfering with insulin signaling cascade, which we just discussed, receptor and somewhere to

00:43:42.680 biopsies. And what we found was indeed in healthy individuals, just give glucose and insulin,

00:43:49.960 you get activation of PI3 kinase. This is the step you mentioned. This is the required step

00:43:55.480 for translocation. And in the follow-up study, same individuals, we raise glucose and insulin and also

00:44:02.840 raise fatty acids. And then we totally abrogate insulin activation of PI3 kinase. That study basically in

00:44:10.440 humans, in the model we care about is saying, yeah, somehow a fatty acid metabolite is leading to this

00:44:17.080 block in insulin action somewhere between PI3 kinase and the receptor. So we've narrowed it down to that.

00:44:24.360 I'll walk you through the steps that I think then are the biochemical metabolite that's mediating this,

00:44:31.400 the lipid fatty acid mediator that's leading to this, and then the biochemical mechanism. Does that sound

00:44:37.800 good? Yeah, that sounds fantastic. Here we have a cartoon of a muscle cell. And my view, again,

00:44:44.920 thinking about flux, it has to do with relative imbalance. So basically doing focused lipidomics,

00:44:51.560 we zeroed in on this metabolite, fatty acid metabolite called diacylglycerol. And yeah,

00:44:58.280 I heard you mention that before. It's the precursors, the penultimate step in triglyceride

00:45:03.000 synthesis, diacyl, two fatty acids on a glycerol backbone. This is a bioactive metabolite. It's

00:45:09.720 been known for years to activate novel PKCs. This is what we found tracked in our animal models with

00:45:17.400 lipid induced insulin resistance, do high fat feeding in a mouse or rat, get muscle insulin

00:45:22.200 resistance. And it was this metabolite that tracked with insulin resistance. And then we did the lipid,

00:45:29.160 same type of lipid infusion we did in humans, simply raise plasma fatty acids by giving triglyceride

00:45:35.080 and heparin. We saw acyl coase go up. We saw DAGs go up. Right when DAGs reached a peak, then we got

00:45:44.040 activation of novel PKCs, PKC theta and epsilon in the muscle. Then we linked to this block in insulin

00:45:52.280 action, which I'll show you in a second at the level of the receptor and one step downstream of the

00:45:57.560 receptor. The concept that I'd like to impart on you is it's this imbalance between fluxes. So fatty acids

00:46:06.360 are continuously being delivered to muscle cells. And we're going to do the same thing if we have time

00:46:12.680 to talk about the liver, because that's the other key insulin responsive organ. But we'll start with

00:46:16.280 muscle. Fatty acids are being delivered either through fatty acids or even hydrolysis of triglycerides

00:46:22.920 through LPL, endothelial bear, delivering more fatty acids to the muscle cell. When it's the flux of fatty

00:46:29.800 acids into the muscle cell that exceeds the ability of the mitochondria to oxidize the fat or store this

00:46:41.080 fatty acids, acyl coase as triglyceride, you get net accumulation of diacylglycerol. This is a very

00:46:48.280 important point. Triglycerides are neutral. So I want to emphasize this. So even though triglycerides

00:46:54.680 often track with insulin resistance, we've dissociated it inside the muscle cell and liver

00:47:00.440 cell from insulin resistance. It's a marker for DAGs, typically tracks very well, but it's an inert

00:47:07.000 storage form of lipid. So triglycerides are not the culprit. We've dissociated in liver and muscle,

00:47:13.400 but it's a pretty good marker if you can't measure the DAGs with mass spec.

00:47:18.120 Let's go back to that for a second. I want to make sure people understand what we're saying

00:47:21.560 here. So triacylglyceride or triglyceride, those two we use interchangeably, has this three carbon

00:47:27.480 glycerol backbone with three free fatty acids on it. That's the way that we very, very efficiently

00:47:33.720 store energy in the most energy dense hydrocarbon in our body. The DAG by extension has only two of those

00:47:41.560 free fatty acids. What typically sits on that third carbon? And what is it about that confirmation

00:47:49.080 that renders the DAG, in this case at least, seemingly much more of a problem than the TG or TAG?

00:47:57.080 Basically, it's a hydroxyl group, a simple hydroxyl group. It's the two fatty acids of the DAG

00:48:03.880 that sit into the bilayer, membrane bilayer. And then the hydrophilic hydroxyl group sits in the

00:48:11.560 cytoplasm. And that's what then will pull the novel PKCs to the plasma membrane. So that's the

00:48:18.760 troublemaker. That's the troublemaker. Basically, then when you get this imbalance between fatty acid

00:48:26.120 uptake versus oxidation in the mito versus storage as neutral lipid, you get activation of these two

00:48:33.480 novel PKCs in muscle, theta and epsilon. Theta blocks insulin action at the level somewhere between

00:48:42.600 the receptor and IRS-1 tyrosine phosphorylation. And epsilon, and we'll get into this for the liver,

00:48:49.560 directly binds to the insulin receptor, and then hits the receptor kinase. If we have a chance,

00:48:56.360 I'd love to share this with you and your listeners, because this, I think, has important evolutionary

00:49:02.040 mechanisms behind it. Why does this exist? And it's going to be very important for

00:49:06.280 survival during starvation. But nevertheless, when both of these NPKCs in muscle are activated,

00:49:14.200 you have reduced insulin tyrosine phosphorylation of IRS-1, less PR3 kinase activation. And as we

00:49:21.160 talked about, then less FUT4 translocation. So to me, the real culprit, and we've been able to

00:49:27.240 just quickly really test this rigorously, gene knockout, we've been able to inactivate isoforms,

00:49:33.400 NPKC theta, you get protection. We've been able to block mito-oxidation, and you make these animals 0.99

00:49:40.920 prone to fat buildup, insulin resistance. We block fat entry into the myocyte, inactivate FAT4,

00:49:48.440 they're protected. You overexpress lipoprotein lipase in the muscle, more fatty acid delivery,

00:49:54.120 muscle-specific insulin resistance. And then finally, if you rev up mitochondrial fat

00:49:59.000 oxidation, let's say through uncoupling, overexpress UCP3 in the muscle, you get protection

00:50:05.560 from insulin resistance. And all these track with DAGs going up or down with the insulin resistance,

00:50:10.760 or protection from insulin resistance. Let's talk a little bit about how we think

00:50:17.720 this is different in an active versus inactive person. Because at the outset, you said,

00:50:22.920 look, when we're trying to find this in the youngest cohort of patients, these 20-year-old,

00:50:27.560 basically undergrads at Yale that we're going to study, we screen on many things,

00:50:31.960 but an important thing we screen for is sedentary behavior. You mentioned that at the very outset,

00:50:36.920 which leads me to believe that if you did a sampling across the cross-country team,

00:50:41.720 the crew team, you wouldn't find this phenomenon. So what is it about activity or the lack thereof

00:50:49.400 that presumably points to this elevation of intracellular DAGs that kicks off this cascade?

00:50:56.520 Dr. Let me just show you. So this is where we talked about Riven and his hypothesis of insulin

00:51:02.840 resistance and how what we wanted was to build on it. Because I'm going to answer your question

00:51:08.040 about exercise. And I want to do two things that I want to show you how exercise reverses this muscle

00:51:13.640 insulin resistance. But I also want to show you and your listeners why exercise in muscle actually

00:51:20.200 will prevent fatty liver and liver insulin resistance. I think that this is a useful segue.

00:51:25.880 And so this is from Jerry Riven's Banting lecture in 1988. And at that time, people were still arguing

00:51:34.040 whether insulin resistance was driving all these other things we see around the circle,

00:51:38.440 atherosclerosis, hypertension, type 2 diabetes, polycystic ovarian disease, inflammation,

00:51:44.680 or are these just common things clustering together? So what we wanted to do was actually ask the

00:51:50.040 question. What we see in these young 20-year-olds, these volunteers, is the first thing we see is

00:51:54.600 muscle insulin resistance. And maybe that's driving atherogenic dyslipidemia, who is going to lead to

00:52:01.960 heart disease, high triglycerides, low HDL, and non-alcoholic fatty liver disease by changing the fate of

00:52:10.120 ingested carbohydrate from glycogen to fat. So this is the distribution I was telling you about. And

00:52:16.600 healthy, young, sedentary individuals. We're going to get into exercise in a second. We simply take

00:52:21.880 the bottom quartile, one and four, versus the top quartile, and we give them two high carbohydrate

00:52:27.960 meals. And we say, where's the energy going from that carbohydrate? How is it being stored? Getting

00:52:32.440 at the very first question you asked me. We can use our NMR to measure changes in fat storage in liver and

00:52:40.040 muscle, as well as glycogen in liver and muscle. And what we found then is you give them two high

00:52:47.800 carbohydrate milkshakes. And there's virtually no difference in the plasma glucose concentrations at

00:52:54.360 at this late breakfast and lunchtime high carbohydrate shake. But you can see it at the

00:52:59.240 expense of severe hyperinsulinemia is what we talked about. So the reason these young insulin resistant,

00:53:05.560 as well as every insulin resistant person is perfectly fine, is the beta cells are pumping

00:53:10.440 out two to three times the amount of insulin just to maintain new glycemia. So these beta cells are

00:53:15.320 just being whipped, working really hard. And that's why no one's diabetic. You're insulin resistant.

00:53:20.040 That's why virtually every obese insulin resistant person is normal glycemia, because the beta cells

00:53:26.040 are working so hard to maintain this. And you can see that here. The other thing I want to point out is

00:53:31.160 the insulin levels are, I'll give a number, you know, so normal, maybe 100 at the peak and maybe

00:53:36.840 180 at the peak in the resistant individuals. But this is in plasma, the portal vein with the liver

00:53:43.400 seeds is three times the liver seen huge amounts of insulin in these insulin resistant individuals just

00:53:49.320 to maintain normal glycemia. We use carbon NMR to look at changes in muscle glycogen and liver glycogen.

00:53:56.600 And you can see again, young insulin resistant, 20 year olds can't get glucose into muscle glycogen

00:54:03.080 due to a block in transport because they have increased ectopic fat in the myocyte. Dags are up,

00:54:09.160 no problem in liver. And then you look at the changes in fat and this carbohydrate, this is change

00:54:17.080 in liver triglyceride. It's up two and a half, 2.3 fold. You put some heavy water, stable heavy water

00:54:24.360 into the milkshake to track de novo lipogenesis. That's the conversion of glucose to fat. And that's

00:54:30.280 also up greater than two fold. Quick question there. There was a very famous experiment. It's

00:54:35.880 been so long since I've read it. I certainly know I spent many hours on it. It was by Mark Hellerstein,

00:54:41.160 circa 94-ish. And he looked at this question of how much carbohydrate could be converted to fat via

00:54:50.200 de novo lipogenesis. And if I recall correctly, the answer was, at least from that paper, was not

00:54:56.920 that much. But also I believe one of the criticisms of that was that he was looking at an insulin

00:55:02.520 sensitive population. Am I remembering that correctly? Because what you're showing here would

00:55:07.720 suggest the opposite, which suggests that an insulin resistant person is capable of significant de novo

00:55:13.880 lipogenesis. Everything you've said is correct. When you're thinking about de novo lipogenesis,

00:55:20.200 two things is, again, what conditions are you studying this? Is it after meal ingestion? Is it

00:55:25.720 in a fasting state when a lot of people have measured this in the past? It's minimal and it makes sense. It

00:55:30.760 only gears up what substrate is taking in. And then depending on the type of substrate, you can alter this

00:55:39.400 quite a bit. So it can be changed by simply putting more fructose, more glucose in the meal,

00:55:45.400 by increasing the meal size. Mark's done beautiful work in the past. It is what it is. Those studies

00:55:50.760 are what they are. But clearly what we're learning here is just, as you say, DNL is significant. It's not

00:55:56.440 the majority of the fat. I think most investigators would agree the majority of fat synthesis in liver is

00:56:03.240 occurring through esterification. That is fatty acids coming to the liver, getting corporate into

00:56:08.200 triglyceride. But there is a significant importance for DNL. And again, especially if you track it

00:56:15.880 chronically in patients who are continuously high carb feeding, especially high sucrose, high fructose

00:56:22.280 corn syrup, we want to get into that. But fructose basically gets funneled into the liver, into the DNL

00:56:26.920 pathway. It's ubiquitous. You can push DNL to be significant, and it is a significant contributor to

00:56:34.200 metabolic fatty liver disease. And it's upregulated with peripheral sensitivity. I think this is the

00:56:39.800 major message I want to give here, is when you have muscle insulin resistance, specifically, it will

00:56:46.040 drive the liver fat synthesis by DNL. When you have that, when your liver is making more fat through DNL, it

00:56:55.880 makes more BLDL exports. So plasma triglycerides go up and HDL goes down.

00:57:01.880 So what I find interesting about this, before you go further, Jerry, is this is all from the

00:57:07.000 2007 PNAS paper by your wife, actually, right? Kit Peterson.

00:57:10.280 Yeah, Kit Peterson.

00:57:11.560 So what I find interesting about these data is that these patients were euglycemic. I mean,

00:57:17.480 that to me is the staggering piece of this. These patients are still potentially a decade away from

00:57:26.040 seeing an interference in glucose homeostasis. They're a decade away from their doctor saying,

00:57:33.080 hey, your glucose is a little higher than it should be postprandially, never mind even at the fast.

00:57:39.640 And yet they're already seeing an 80% increase in triglyceride, which I just want to sort of

00:57:45.480 talk a little bit about this clinically. Most laboratory assays will say a triglyceride level

00:57:51.240 of 150 milligrams per deciliter is considered normal. Well, we don't say that in our practice,

00:57:57.560 we view anything over a hundred is abnormal. That's a red flag. And if your trigs are more than

00:58:03.560 two X, your HDL cholesterol, that's a very big red flag. Although most people would accept

00:58:09.400 triglycerides of three or four, if not five times above HDL cholesterol before the sirens would go off.

00:58:16.120 And yet when you look at these patients, again, euglycemic, you see a difference of approximately,

00:58:22.520 you know, a hundred to 105 of the trigs in the insulin resistant to 60 in the insulin sensitive.

00:58:29.080 So it's all kind of right here in front of you, sort of in a way that unfortunately just doesn't

00:58:34.760 get appreciated, but it's the more intense stuff that's mind boggling to me, which is the two

00:58:40.040 and three-fold difference we see in de novo lipogenesis, hepatic synthesis of fat, impaired

00:58:47.000 hepatic glucose sensitivity. And I guess it speaks to the point you made earlier, Jerry, which is

00:58:52.600 when the portal vein amplification of insulin differences is as big as it is, it becomes

00:58:57.800 basically a magnifier of everything we're seeing in the periphery.

00:59:01.480 Exactly. Yeah. And our normative data, in my view, we need to reset. What we consider normal is,

00:59:07.480 to me, this is when we look at our insulin sensitive, that's what our normal should be

00:59:11.800 and guiding us. You asked about exercise. That's something we're quite passionate about. And I want

00:59:16.600 to kind of see, tell you how that fits in here. So again, conceptually here, we have a normal person

00:59:22.760 ingesting carbohydrate. First question, how is this distributed? It's in glycogen is where you want

00:59:28.280 to store your ingested carbohydrate. It gets stored in glycogen and liver and muscle. And again,

00:59:33.800 this is one quarter of our young, lean, healthy volunteers are insulin resistant. And again,

00:59:39.400 if you're overweight or obese, you're there already because these are lean individuals. And that's

00:59:44.440 still one quarter of the population. You can't get that ingested glucose into glycogen due to this

00:59:50.760 block in transport, due to the block in DAG PKC inhibition of insulin signaling. It's diverted to liver.

00:59:58.760 You have that insulin in the portal vein. That's three times per if it's up to five,

01:00:03.240 600 micro units per mil. That turns on SRIBP1C, the master transcriptional regulator of triglyceride

01:00:11.640 synthesis gears up all the DNL enzymes. So you have increased DNL that leads to this increase. We just

01:00:18.680 reviewed plasma triglycerides, this reduction in HDL. This is going to set these healthy individuals up to

01:00:26.280 to atherogenic dyslipidemia, heart disease in their forties and fifties. With time, it's metabolic

01:00:33.160 associated fatty liver disease now. And again, most common cause of liver disease now in the world.

01:00:39.640 It's now leading cause of NASH, leading cause of liver fibrosis, cirrhosis and stage liver disease and

01:00:47.800 going to be liver cancer. So it's all going to be metabolic driven and from that hyperinsulinemia,

01:00:53.880 in my view. So exercise, can we do anything about this? This hypothesis is right. We can test it.

01:01:00.280 And so you asked about exercise. So this is a study we did some years ago, published in the New

01:01:04.920 England Journal, took these young insulin resistant offspring. And this is with parents with type 2

01:01:10.120 diabetes. And the Jocelyn group did a really nice study. They found that if you have two parents with

01:01:16.600 type 2 diabetes, and if you're insulin resistant, that single parameter is the best predictor whether or

01:01:21.800 not you would go on to develop type 2 diabetes. So we've tried to study these individuals with our

01:01:27.560 methodology extensively. And Gianluca Persagin, who did this study when he was a fellow with me,

01:01:34.280 took these and just studied them in the basal state, shown here that, you know, again, in these

01:01:38.680 young insulin resistant, again, everyone's here is lean, non-smoking, no medications. They're in their

01:01:44.120 20s and 30s, BMI 23, 24 to factor out obesity, confounding the factors of obesity, medications,

01:01:51.320 smoking, other things. So young, lean, healthy individuals, but just parents with diabetes,

01:01:55.400 insulin resistant, you study them. And in the basal state, take up less than half the amount of

01:02:00.920 glucose in muscle, and it's due to a block in transport. So same thing as I've gone on and on

01:02:06.760 before in the diabetics and the obese individuals, this block in transport.

01:02:10.040 And we asked the question, does exercise, can we bypass this abnormality? And the answer is yes. So

01:02:16.360 here you can see this was after six weeks of being on a StairMaster, three 15-minute bouts at about 65%

01:02:23.880 MVO2 max. And here we're normalizing insulin-stimulated muscle glycogen synthesis. And we

01:02:30.440 usually measuring glucose 6-phosphate, we've opened up that door of getting glucose into the myocyte. So,

01:02:36.840 and I think molecular explanation for this is this protein called AMPK, which we can talk about,

01:02:43.160 gets activated with exercise. And that has been shown to cause boric translocation independent,

01:02:49.880 independent of pi-3 kinase. And so we're kind of short-circuiting that block with exercise.

01:02:55.320 To test our overall hypothesis, does muscle insulin resistance drive fatty liver and VNL and high

01:03:03.000 triglycerides? We took these young insulin-resistant individuals and we showed,

01:03:07.640 John Lucas showed in that New England Journal study, even a single bout, 45-minute bout,

01:03:12.840 was sufficient to open up the door to glucose, cause that GLUT4 translocation. And Rasmus Raval,

01:03:19.960 when he was a clinical fellow with me, did one single bout in these same individuals I showed you

01:03:24.920 before, insulin resistance in muscle. The ones had high triglycerides, low HDL, and prone to

01:03:30.440 increased DNL. With a single bout, we were able to show that that same ingested glucose would lead to

01:03:36.520 more glucose deposition as muscle glycogen. And we got significant reductions in de novo lipogenesis,

01:03:44.520 significant reductions in liver triglyceride.

01:03:46.920 I just want to make sure I understand that. And it's relevant to another question I have about

01:03:50.840 the difference between insulin-dependent and independent glucose uptake. So do we know if

01:03:55.560 that single bout of exercise, which particular piece of the pathway got released? Did it have some

01:04:04.840 direct effect on the root cause, the DAG, or some of the kinases downstream? Was it even further

01:04:12.680 downstream at the very last step where the transporter gets released? Where was the actual

01:04:18.360 bottleneck alleviated with that single bout of exercise? I can speculate. In these were human

01:04:23.960 studies, I can tell you that we open up the door, we measure glucose 6-phosphate in them, and that

01:04:29.720 goes up. So we open the door for that defect in insulin-stimulating transport is now reversed. So

01:04:36.120 glucose transporters are in the membrane, glucose is coming in. What I can't tell you is whether or not

01:04:42.200 we've altered DAGs and we're getting improved insulin signaling at the level of the receptor and IRS-1,

01:04:49.320 and or is it just AMPK causing this GLUT4 translocation? If I had to speculate, I would

01:04:55.480 think most of it is through the latter. We were simply, with an acute bout, causing AMPK-induced GLUT4

01:05:03.720 translocation, which we know happens independent of pediatric kinase. That's established. So we're

01:05:09.000 short-circuit. We're just causing GLUT4 right at all the lower mechanisms to get to the membrane.

01:05:13.960 So we fixed the block in insulin action. I think, though, with chronic exercise therapy,

01:05:19.560 we're going to be doing both, where we get melt-away, the lipid and DAGs go down. So we have improved

01:05:25.240 insulin signaling as well as more AMPK-induced GLUT4 translocation.

01:05:29.560 Yeah, I'll tell you just, I think I've even discussed this on a previous podcast. I've had a couple of

01:05:34.760 patients with type 1 diabetes that I've taken care of, not many, but in the phenotype of patients

01:05:40.120 with type 1 diabetes where there is a significant amount of exercise, specifically sort of modest

01:05:46.440 intensity aerobic exercise. So a person who is, for example, doing brisk walking, very brisk walking,

01:05:53.480 sort of to the tune of four miles an hour, an hour to two hours a day. These patients with type

01:05:59.720 1 diabetes can be virtually free of insulin and maintain reasonable glycemic control. So they

01:06:06.440 could walk around with a hemoglobin A1C of 6% using maybe 12 units of insulin a day and obviously

01:06:13.880 restricting carbohydrates. But again, it suggests, I say this having watched them change the intensity

01:06:20.920 duration of the exercise, that it seems that that exercise becomes a spigot to how much glucose they

01:06:27.000 can dispose in their muscle seemingly without insulin. It's almost like a total bypass of the

01:06:32.360 system, which again, I think to your point is chronic. I don't think this is something we see

01:06:37.240 acutely. I obviously can't comment on it. The first time I saw it, which was probably about six years

01:06:42.600 ago, it really sent a light bulb off, which is imagine now being able to maximize both insulin

01:06:49.080 dependent and insulin independent glucose uptake into a muscle that really becomes a powerful tool to combat

01:06:56.680 all of this sort of metabolic dysregulation. That's what AMPK does is insulin dependent glucose

01:07:01.800 uptake. And I can see in combination with reduced carbohydrate consumption, less coming into the

01:07:08.520 circulation and whatever little comes in is taken care of through AMPK, insulin independent

01:07:14.280 glucose translocation. So that fits. Before we go to the liver, and I do want to actually talk about

01:07:20.920 how all of this works in the liver. I want to go back to one other thing that you very briefly touched

01:07:26.840 on, which is the evolutionary explanation for some of this. That would be best done, if I might say, with

01:07:34.680 the liver. Okay, great. Let's do it. Because I want to understand this, yeah. That's kind of fun. So let's

01:07:40.040 now turn. So I kind of walked you through at least my thinking about insulin resistance, why it's so important

01:07:46.120 for not only diabetes, but so many diseases. I've shown you the physiological cause for

01:07:52.600 insulin resistance in muscle, can't get glucose in the glycogen. I've shown you that block is a

01:07:56.680 transport, and then I've given you a molecular understanding of how that insulin resistance

01:08:02.040 in muscle happens. My view is lipid diacylglycerol is block, leading to activation of a novel protein

01:08:08.920 kinase C, epsilon theta, blocking insulin signaling. Okay. So let's now, and then I've shown you how muscle

01:08:14.840 insulin resistance can lead to fat accumulation in liver, atherogenic dyslipidemia, and fatty liver.

01:08:21.960 Now we know fatty liver is what then leads to insulin resistance in the liver. And so I want to

01:08:27.800 take you through the molecular basis for how fat in liver causes insulin resistance. And it's pretty much,

01:08:32.920 what's nice now that you understand muscle, lipid induced muscle insulin resistance, it's pretty

01:08:37.480 close to the same story in liver. So here's a cartoon of the liver cell.

01:08:41.960 But is the direction of causation, Jerry, in the order in which you're telling the story? In other

01:08:47.400 words, is the hyperinsulinemia as a result of muscle insulin resistance? Let me clarify that.

01:08:55.560 Muscle insulin resistance, which leads to peripheral hyperinsulinemia, which is accompanied by portal

01:09:02.040 vein hyperinsulinemia, which leads to what you're about to tell us. Is that the order in which you think

01:09:07.800 this occurs? I do. As I say, this is what we see in our volunteers as we march through

01:09:13.240 the progression in different stages. We don't see liver abnormalities in these young 20 year olds.

01:09:19.080 It's all muscle and maybe a little bit of the fat cell, which we'll come to at the end, but it's the

01:09:23.960 muscle. There's no alterations in the liver until they get fatty liver. Once they get fatty liver, then we

01:09:30.280 see both insulin resistance in liver and insulin resistance in muscle. A very important distinction

01:09:36.840 between humans and rodents. We've studied both models quite extensively. Rodents develop insulin

01:09:42.440 resistance in the reverse direction. They get liver fat first, liver insulin resistance, and then muscle.

01:09:48.200 Most of the studies are done in rodents. It's a very important distinction in terms of

01:09:52.920 the progression and very different humans versus rodents. And we can talk about similarities and

01:09:57.800 differences if you want, but we're going to focus mostly on humans for this talk.

01:10:02.760 And that makes total sense. So it is, again, it's peripheral IR, hepatic IR,

01:10:09.720 hepatic consequences, which then basically amplify it.

01:10:13.080 That's my belief. Yeah. And again, leading to this beta cell compensation, compensation, and then

01:10:18.840 again, something when you get both muscle and liver, insulin resistance, and increased glucose

01:10:24.040 production by liver, then something happens to the beta cell. And that's when things really start

01:10:28.440 to spiral where you have very profound hyperglycemia, fasting and postprandial.

01:10:33.320 Here's the cartoon of the liver cell. And again, glucose transport is not rate controlling,

01:10:38.440 as you know, in the liver cell. Glucose just diffuses in through GLUT2 transporters.

01:10:42.680 And the insulin, again, binds to the receptor. Same thing, autophosphorylation. The key

01:10:48.840 intermediate there in liver is IRS-2, undergoes tyrosine phosphorylation. You use PI3 kinase,

01:10:54.440 just as you did in muscle AKT2. And in liver, what happens is you have a few things. One not shown

01:11:03.160 here is glucokinase translocation, and that we've recently shown is probably very important for rate

01:11:08.200 control getting glucose into the hepatocyte. You also get activation of glycogen synthase and more

01:11:14.120 glycogen synthesis. And then you have this phosphorylation of FOXO, which is a transcriptional

01:11:21.400 regulator. And that then is excluded from the nucleus and then down-regulates then gluconeogenesis

01:11:28.840 through a transcriptional mechanism. And if we have a chance, I'd like to come back to this because we

01:11:33.240 have some interesting data that speaks to really how insulin's inhibiting this key process.

01:11:39.560 So let's now just focus on how lipid causes insulin resistance in liver. Same metabolite,

01:11:45.800 it's the diisoglycerols. They go to activate epsilon. That's really the major isoform of PKC,

01:11:54.520 novel PKCs in liver. And work by Varmin Samuel, when he was doing his PhD with me in a series of studies,

01:12:02.040 Varmin showed that epsilon binds to the insulin receptor and directly inhibits the receptor

01:12:08.360 kinase itself. And that then leads to downstream abnormalities. What I want to share with you

01:12:14.120 now, which I think, and again, it gets into this evolutionary basis for insulin resistance,

01:12:18.120 which I think your listeners might find interesting, is how is epsilon inhibiting the

01:12:24.120 receptor kinase? We worked on this, Jesse Reinhardt and Max Peterson. He was an MD-PhD student with me.

01:12:31.400 We did untargeted phosphoproteomics. And what I'm showing here is the catalytic domain of the insulin

01:12:38.280 receptor. Yeah, I can just describe it for the listeners. It's a loop. You can picture it as a door

01:12:45.000 over the pocket for the catalytic domain of the insulin receptor. And this door is closed.

01:12:50.680 IRS-1, IRS-2 can't go into the pocket for tyrosine phosphorylation.

01:12:56.600 When insulin binds the receptor, these three tyrosines, the 1158, the 1162, and the 1163,

01:13:03.880 become phosphorylated. That opens the door. That loop flips out. And then IRS-1, IRS-2 go into the

01:13:09.880 pocket and undergo tyrosine phosphorylation to get the rest of the cascade going.

01:13:15.000 Using untargeted phosphoproteomics, we were able to show Jesse Reinhardt, who is our collaborator in

01:13:20.760 MassSpec Maven, identified using purified receptor, purified PKC epsilon, that when you add activated

01:13:28.920 epsilon to the receptor, you phosphorylate this threonine. And that got us very excited because,

01:13:35.640 golly, that's one amino acid away from these two tyrosines that are required for activation.

01:13:41.640 The receptor is maybe doing something important. And so the other thing that got us excited about,

01:13:47.160 and here's getting into evolution, is the sequence of the catalytic domain for the receptor. And it's

01:13:54.680 been conserved all the way from humans down to fruit flies. Those three tyrosines, same position.

01:14:02.040 And that threonine that sits right between the two tyrosines, 1158 and 1162,

01:14:08.280 has been conserved all the way, again, from homo sapiens down to drosophila through evolution of

01:14:14.120 something that's important that usually hangs around. That's a long time. So to prove this,

01:14:19.320 we very simply, we did some genetics. Again, that's what you can do is you can knock a glutamic acid,

01:14:25.800 replace that threonine with glutamic acid, mimic a phosphorylation event, and that kills the kinase

01:14:31.640 activity. You can mutate the threonine to an alanine so it can't get phosphorylated. And then you have

01:14:36.840 protection in vitro from epsilon-induced reduction in IRK activity. And then you can make the mouse.

01:14:43.160 And so here in this paper, we made mice where we replaced the threonine in that key position.

01:14:50.120 The 11, that's the mouse homolog. The 1150 is the homolog for the 1160 in humans. So all the threonines

01:14:56.520 are instead alanines. And I won't get into the data other than say the mice are perfectly normal,

01:15:01.960 normal chow, normal insulin sensitivity, nothing that, you know, normal size, normal growth.

01:15:06.920 But when Max fed these mice a high fat diet, the wild type mice get profound hepatic insulin

01:15:13.160 resistance. And this we see, and everyone else on the planet sees, you feed mice high fat diet,

01:15:18.920 even for three days, they get fat accumulation, DAG accumulation, hepatic insulin resistance.

01:15:24.760 Does it have to have sucrose in it as well, or just fat?

01:15:28.520 It doesn't need to be. You can make it worse if you add a little sucrose. They like that in the

01:15:32.520 drinking water and they have even more fatty liver if you put sucrose in the drinking water. But

01:15:37.320 this is just with fat alone, but it's even more greater when you put sucrose or fructose or whatever

01:15:42.680 sugar you want in the drinking water. And here then you can see when you simply mutate that threonine

01:15:49.160 tonaline, now you have perfectly normal hepatic insulin sensitivity as reflected by insulin's

01:15:56.280 ability to suppress hepatic glucose production. And this is despite the same amount of liver fat,

01:16:01.960 same amount of liver DAGs in the liver. This tells us that that single amino acid is doing something

01:16:08.680 very important in terms of mediating lipid induced insulin resistance. And this actually just came

01:16:13.480 out this last week, this paper now, just to summarize where we've now shown that there's

01:16:19.720 different isoforms. We didn't get into this, into the, of diacylglycerol. And it really matters which

01:16:25.960 isoform it is and what compartment it is. Just to summarize this paper that just came out in cell metabolism,

01:16:32.920 we were able to show by measuring the three different stereoisomers of diacylglycerol.

01:16:38.280 It's really the SN12 isoform and measuring these different isoforms in five different

01:16:45.960 intracellular compartments, the plasma membrane, the cytosol, lipid droplet, ER, and the mitochondria.

01:16:52.200 It's really specifically the SN12 isoform in the plasma membrane that's important.

01:16:59.000 If you just measure total DAGs, you may easily miss this. We learned that this recent study and that we,

01:17:05.160 we showed both that PKC epsilon is both necessary and sufficient for this process by doing the knock

01:17:10.840 in and overexpression. But I just want to basically touch on the question you asked me about why do we

01:17:17.880 have insulin resistance? Why should it exist? And the reason I think it exists is it's protective

01:17:24.760 for us during starvation. When you starve, this is true pretty much in all mammals, mice, rats,

01:17:31.720 and humans. When we starve, we get fatty liver. Here in this study, this is Rachel Perry's paper in cell

01:17:39.160 from a couple of years ago. Take rats, just starve them for 48 hours. You have increased lipolysis,

01:17:45.560 more fatty acids delivered to the liver, hepatic fat accumulation, DAGs we show go up, SN12,

01:17:53.480 PKC epsilon translocation, and insulin resistance in liver. And the main thing that insulin does in the

01:18:00.760 liver is it promotes glucose uptake and storage as glycogen. When you think about it, that's what you

01:18:07.880 want turned off during starvation. Because during starvation, glucose is a very precious molecule,

01:18:14.360 and you want to preserve this in circulation for the CNS, which is critically in need. It's really

01:18:20.840 the major source of energy for the CNS. And so by promoting hepatic insulin resistance,

01:18:28.440 we're promoting glucose in circulation for basically the CNS to operate. And so that to me is why that

01:18:35.800 threonine is preserved all the way from humans to fruit flies. And I just wanted to show you this cover

01:18:42.520 of nature, this Mexican cave fish. It's a fun story because after our paper came out,

01:18:49.000 this little fish made the cover of nature. And what was so fascinating about it, so these little

01:18:55.080 fish, they live inside caves. They spend most of their life starving. The only time they are able to

01:19:00.760 eat is when something smaller than them swims in front of the cave, and then they can reach out and

01:19:06.920 grab it and pull it back into the cave and gobble it up. And these workers who studied this Mexican 1.00

01:19:13.320 cave fish found this cave fish had a mutation in the insulin receptor, had profound hepatic insulin

01:19:19.560 resistance. And they also went on to say this was important to allow them to survive.

01:19:25.400 In my view, insulin resistance was a protective mechanism throughout evolution that allowed us to

01:19:31.960 survive all species during starvation, which was probably the predominant environmental exposure

01:19:38.120 we've had for the last many, many millennia. And it's only in recent years, recent decades, that now

01:19:45.160 we're in this toxic environment of overnutrition. And it's when these same pathways now are going the

01:19:52.440 opposite direction, promoting disease by doing what they were at one time was protective. And now they're

01:19:59.320 actually being told metabolic disease that we just discussed. So I want to make sure I can unpack this

01:20:04.920 a little bit. So I want to start in the muscle because I think it's easier. And again, we'll even

01:20:09.320 talk about it in humans, which means we can do it on a sort of different timescale, because obviously

01:20:13.960 48 hours of fasting in a mouse is a seismic fast, a near fatal fast. But let's say 48 to 72 hours of

01:20:21.640 fasting in a human, we still would expect to see significant muscle insulin resistance. And there would be a

01:20:28.680 great reason for that evolutionarily because you would want to make sure that as much glucose as

01:20:35.400 possible in that circulation, which by this point is all coming through hepatic glucose output is not

01:20:41.160 being, quote unquote, wasted in muscle glycogen synthesis. To your point, every gram of gluconeogenic

01:20:49.320 substrate that's going through the liver and then coming out the liver should be preserved for the brain.

01:20:54.040 Because even Cahill's studies showed that after 40 days of starvation, humans were still getting

01:21:00.520 about 40% of CNS energy from glucose, the remainder from ketones. So glucose never went away as a substrate

01:21:09.400 for the brain. So I think I have a handle on the muscle side of things. I'm still struggling a little

01:21:15.480 bit to understand the physiologic consequence of hepatic insulin resistance and how that feeds into

01:21:25.480 what I think should be an environment that says, figure out a way to make as much glucose for the CNS

01:21:33.160 as possible. Why does more fat accumulation in the liver make it better served to protect the brain?

01:21:41.400 So first of all, let me step back. So both organs during starvation, both liver, even though I focus

01:21:47.320 here on liver, muscle will become insulin resistant also through increased circulating fatty acids through

01:21:52.680 the mechanisms. We talked about DAGs building up, PKC-theta. So insulin resistance in all organs

01:21:59.240 are going to preserve glucose for the CNS. I was just focusing on the threonine here in liver because

01:22:06.200 that's where epsilon was taking us. To understand the liver, I want to just take you to another

01:22:12.760 cartoon because you're asking a very important question about processes, about regulation, how

01:22:19.720 insulin works in liver. And I think to do this, let me just step back. The conceptual view, again,

01:22:26.440 this is a cartoon I always like to show. How does insulin work? This was from 20 years ago when I was

01:22:31.400 first studying it maybe 30 years ago. Insulin binds the receptor, magic happens, something happens and

01:22:36.680 you have an effect. And so even though insulin's been, since it's discovered, we're still trying to

01:22:42.200 really understand what's happening in different tissues, how it works and getting surprises. So this

01:22:48.200 is the canonical view we just went through of how insulin works on liver, it binds the receptor,

01:22:54.760 it activates the cascade to promote glycogen synthesis and turn off gluconeogenesis.

01:23:00.520 And what we're finding is this simple view doesn't explain many things and I think needs modification,

01:23:07.960 especially in terms of insulin regulating gluconeogenesis, this process that is required

01:23:14.040 to keep us alive during starvation. Without gluconeogenesis, we're not going to wake up in the

01:23:18.680 morning because it's gluconeogenesis that supplies glucose for the CNS while we're sleeping. And

01:23:23.720 certainly during starvation, without this process, we're in trouble.

01:23:27.160 I don't think that can be overstated by the way. Let's go back to what you just said. We couldn't

01:23:32.440 survive, by my calculation, Jerry, we'd have a hard time surviving 10 minutes without gluconeogenesis as

01:23:39.000 a species. Well, I'll modify that a little bit. As passionate, I'd love to hear you state the importance

01:23:46.280 of gluconeogenesis. No, we know clinically you can. And again, from the lessons learned from

01:23:52.520 gene knockout. Unfortunately, there are patients with inherited disease, Von Gerke's disease.

01:23:57.880 As you know, patients who don't have glucose 6-phosphatase, the last key step, getting glucose

01:24:02.200 6-phosphate out. We do know that can be compatible with life. We have patients with glucose 6-phosphatase,

01:24:08.840 and the way we keep them alive is just continuously to feed them.

01:24:12.840 Yeah, that's my point. Without continuous glucose feeding, your lifespan would be measured in minutes

01:24:18.840 to hours without gluconeogenesis to regulate glucose homeostasis.

01:24:23.160 It's critical for life function. We're on the same page. So let's just talk about then how it's

01:24:29.640 thought to operate and regulate it. It's also important to be able to modulate it. So we eat

01:24:35.320 a meal, and we have to suppress gluconeogenesis. Otherwise, glucose would go up to 400 or 500 after

01:24:41.480 eating a carbohydrate meal. So it has to be a process that's turned on, turned off,

01:24:45.960 and not turned on too much, you know, in terms of diabetes, because that's what drives fast and

01:24:49.800 hyperglycemia. Traditionally, pretty much the major textbooks, physiology, biochemistry,

01:24:56.600 insulin is thought to be turned off gluconeogenesis through transcriptional mechanisms. And again,

01:25:03.960 this is this FOXO phosphorylation by AKT, exclusion from the nucleus. Then you get downregulation of

01:25:11.880 PEPCK, excuse me, and 6-phosphatase. FOXO is the transcription regulator for these downregulation.

01:25:18.680 The problem with this view, and again, there's some beautiful molecular biology, and I don't want to

01:25:24.280 deny this doesn't happen, but the problem with this being the predominant regulating mechanism is

01:25:30.200 threefold. One is you can knock out AKT or FOXO and give insulin to the mouse, and you can still turn

01:25:39.480 off gluconeogenesis in a fasted mouse, which is totally dependent on gluconeogenesis. That speaks

01:25:44.760 to the fact you don't need these key insulin signaling pathways to regulate gluconeogenesis.

01:25:50.280 The second thing in terms of its role in mediating fast and hyperglycemia and diabetes is

01:25:56.840 we got liver from patients with poorly controlled diabetes. So when patients go in for Roux-en-Y

01:26:05.080 or bariatric surgery, the surgeon can take a little piece of liver under direct visualization,

01:26:10.200 so it's very safe, and give us enough liver, we can do actually protein measurement and enzyme

01:26:16.920 measurements of PEPCK, 6-phosphatase, not just message, but actually the proteins themselves.

01:26:21.560 And to my surprise, I thought all these enzymes from everything I was thinking about biochemistry,

01:26:29.480 and at least what I learned when I was a lecture in medical students, I expected PEPCK and 6-phosphatase,

01:26:35.400 and fructose 1, 6-biphosphatase, all to be upregulated two to threefold in the poorly controlled diabetic that

01:26:42.920 was undergoing through and bypass surgery compared to the non-diabetic. And we found no relationship between

01:26:50.840 protein expression of these enzymes, gluconeogenic enzymes, and at least fasting glucose and insulin

01:26:56.520 and history of diabetes. Finally, when you develop methods, the flux methods we won't get into to

01:27:03.160 actually quantify this flux of gluconeogenesis, which has not been easy to measure by the way, but

01:27:09.400 we have methods now, they're very good to measure this flux. We can turn off gluconeogenesis within

01:27:15.080 five minutes, and that's much faster than you'd expect in transcriptional, translational mechanisms.

01:27:21.720 Just to kind of talk about how gluconeogenesis, this is the gluconeogenic pathway, lactate to glucose,

01:27:28.280 you can have transcriptional regulation, you can have substrate regulation. So

01:27:32.440 glycerol, we've shown from lipolysis, there is no rate control. The more glycerol that comes from fat

01:27:39.480 breakdown in the fat cell, that fluxes the liver comes right out of glucose. There's no rate control,

01:27:44.760 it's just all substrate driven. Redox we've shown in the liver cell regulates gluconeogenesis. And this,

01:27:51.480 in a series of studies that Anila has done, that's how I think metformin works. And we can talk about

01:27:57.400 that if you're interested. But finally, I want to emphasize is this allosteric regulation of gluconeogenesis

01:28:04.200 by acetyl-CoA. This had been known for decades to be a regulator of pyruvic carboxylase and had kind

01:28:11.400 of been forgotten because it was very hard to measure and no one looked at it in vivo because

01:28:16.520 it's hard to measure in vivo or especially in the diabetic situation. We said, well, wait a minute,

01:28:22.280 let's go back and look at acetyl-CoA. We developed the methods, tandem mass spec methods, very sensitive,

01:28:27.640 very specific, to do this in freeze clamp tissues from animals with varying degrees of diabetes

01:28:33.800 hyperglycemia. The bottom line is found a very robust relationship between acetyl-CoA, which is,

01:28:40.520 as you know, the end product of beta-oxidation. Take fatty acids and break them down to beta-oxidation,

01:28:46.360 the end product is before it enters the TCA cycle. And there's this very robust relationship, just all

01:28:53.480 these different studies. But basically every study we do, we give insulin, we get suppression of

01:28:59.160 acetyl-CoA. This explains how insulin acutely suppresses gluconeogenesis. When diabetic models,

01:29:07.000 when you have increased gluconeogenesis, it's two-fold increases in acetyl-CoA. But it perfectly

01:29:12.520 follows rates of gluconeogenesis, which we quantify, track perfectly with concentrations of hepatic

01:29:19.480 acetyl-CoA content. I just want to take you how insulin normally works in the liver cell,

01:29:25.800 and then how it becomes dysregulated in diabetes. And this is going to answer your question about

01:29:31.000 how do we distinguish insulin promoting storage as glycogen, yet keeping gluconeogenesis going for

01:29:36.520 the brain. So this is very important to answer that question. So in my view, insulin binds the

01:29:44.040 receptor and it has direct effects through the receptor. That is mostly to promote glucose uptake

01:29:51.560 and storage as glycogen. The effects on gluconeogenesis, the process that keeps us going during starvation,

01:29:59.240 is really mostly regulated not through the receptor in liver, but it's through its effect on the fat cell

01:30:06.680 in the periphery. In studies we've done in awake rats, and we're translating this to humans,

01:30:12.680 it's really insulin putting the brake on peripheral lipolysis, less fatty acid delivery to liver,

01:30:21.480 less generation of acetyl-CoA. And we've shown this, the more fatty acids that flux the liver

01:30:26.680 track almost perfectly with acetyl-CoA content, less pyruvate carboxylase activity. And again,

01:30:33.160 there's about 10-15% of this gluconeogenesis is simply coming from less glycerol from lipolysis to

01:30:38.680 liver through substrate push. So you have two very different processes here. One is glycogen synthesis.

01:30:45.320 That's what the receptor is doing in liver. Gluconeogenesis is mostly 90%, I would say,

01:30:51.400 there may be a little bit of intrapatic lipolysis regulation, but mostly through its effect to put

01:30:56.680 the brake on peripheral lipolysis. And this model, by the way, will explain, in my view, the explanation

01:31:03.000 for all the controversies of insulin action that have been described through the last decades in mice,

01:31:09.080 where you knock out AKT in the mouse, insulin still works. You do things to the periphery,

01:31:15.480 fat cell, and you affect glucose metabolism, gluconeogenesis. All these studies that appear

01:31:21.000 to be conflicting can be explained if you use this model as a template to understand insulin action.

01:31:27.240 And again, I have short-term fasts and long-term fasts. This is important species differentiation.

01:31:33.240 Mice, and as you pointed this out, Peter, even after an overnight fast, boom, all the glycogen is

01:31:39.160 gone. Very different from humans. Humans hold onto their glycogen like dogs, probably for two days.

01:31:45.720 We've done these measurements with starvation in humans. We've shown that it takes about two days to

01:31:50.360 deplete liver glycogen. When you have glycogen in liver, it's really these direct effects of insulin

01:31:57.000 on liver will predominate. But as you move to the fasting state, so in a mouse, after a 12-hour fast

01:32:03.480 or longer, and in a human, probably have to go 24 or longer fast, then it's really insulin,

01:32:09.320 these indirect effects will predominate. And this will also explain all the controversies in

01:32:15.080 dogs, Sherrington, Bergman, in terms of direct and indirect. They've each published a dozen papers

01:32:20.440 on going back and forth, which predominates. This mechanism would explain, I believe, all of those

01:32:25.480 findings. And then I just want to now show you how I view the dysregulation and diabetes. So now,

01:32:32.440 typically on the background of obesity, which is what happens in most of our diabetics,

01:32:37.720 although you have lean individuals who also have this, you have expanded fat stores in the periphery,

01:32:44.600 but now you have insulin resistance in the fat. So insulin can't put the brake on lipolysis. And

01:32:51.000 we can talk about that mechanism, which we're now working on, but it's going to be very similar in

01:32:55.080 terms of liver and muscle. But you also have this component of inflammation. It's been described by

01:33:00.600 many, many individuals. You get crown-like structures, macrophages move in, they release TNF-alpha-IL-6.

01:33:08.680 And what we were able to discern, a lot of people would argue it was inflammation. If you go back to the

01:33:14.280 insulin resistance literature 10, 20 years ago, everyone was discussing inflammation,

01:33:19.640 circulating cytokines, TNF-alpha-IL-6, resistant RBP, circulating factors that were released from

01:33:27.400 inflammation, driving insulin resistance. What we found is, again, you can dissociate inflammation

01:33:33.880 from insulin resistance. That's what I spent the first three decades of my life doing, showing that

01:33:38.280 just ectopic lipid DAGs would drive insulin resistance independent of inflammation.

01:33:43.880 But the transition from just insulin resistance in liver and muscle to fasting hyperglycemia

01:33:51.640 depends on inflammation. And it's through this mechanism where now you have localized inflammation

01:33:57.880 in the fat cell. TNF-alpha-IL-6, I'm sure there's other things, will promote increased lipolysis in the

01:34:06.520 fat cell. More lipolysis, more fatty acid delivery to liver. DAGs go up. Epsilon gets activated. You block

01:34:17.240 insulin action, so you have less glucose being taken up into glycogen. This is what happens in virtually most

01:34:23.480 patients with fatty liver disease. But again, what takes you to fasting hyperglycemia is this.

01:34:31.000 And that's where acetyl-CoA goes up. And again, now your rates of lipolysis, when you measure

01:34:36.840 turnover, not just fatty acid concentrations, but turnover, palmitate turnover production and glycerol

01:34:42.760 turnover, it's up twofold. This increases acetyl-CoA concentrations twofold. This activates pyruvic

01:34:50.440 caboxylase activity and flux twofold. And then in addition, your glycerol delivery to liver is up

01:34:57.080 twofold. And now your rates of gluconeogenesis are increased twofold. And this is now what's driving

01:35:04.680 fasting hyperglycemia in every poorly controlled type two diabetes. It's this gluconeogenic process

01:35:10.920 that we've shown using many, many methods and others have shown this too. This is what now is driving

01:35:16.840 hyperglycemia in type two diabetics. Okay. I have several questions, Jerry. First,

01:35:24.040 these adipocytes that are undergoing lipolysis, these are peripheral adipocytes. Is that correct?

01:35:30.120 Yes. You can have situations where even fat in the liver is probably contributing to this,

01:35:35.560 especially in the lipodystrophic individual who has no peripheral fat cells. So under conditions,

01:35:41.000 the liver fat is playing a role, but most of it in most of, you know, I would say garden variety,

01:35:46.280 what I see is going to be peripheral lipolysis. So when we think about an insulin resistant,

01:35:54.120 obese person with metabolic syndrome, so this is what, 20% of the US population, maybe even more.

01:36:01.880 We've clearly established they are insulin resistant in the muscle. We've established that

01:36:06.440 they are insulin resistant in the hepatocyte. They are obese. So would we still say they are

01:36:12.920 insulin resistant at the fat cell or would we say they are insulin sensitive at the fat cell because

01:36:17.480 they are correctly undergoing lipogenesis in the fat cell? They're at least taking up a

01:36:23.640 sterified fat and they're presumably impairing lipolysis, which is why they retain adipocel mass.

01:36:30.120 In other words, there's a, the flux through the fat cell is negative. They're holding on to fat,

01:36:35.080 correct? Yeah. But I think, and this is a question, a very important question we're going to next. I

01:36:41.320 would still predict if you do careful studies of measuring rates of lipolysis, my definition,

01:36:48.280 they will have insulin resistance in the fat cell. And that's because the reason they're doing

01:36:54.200 everything you just said, they're holding on to fat. They're not happy about it. The doctor's not

01:36:58.440 happy about it is because it's at hyperinsulinemia. So their insulin concentrations are two to threefold.

01:37:04.360 So again, their curve is right shifted. Insulin is doing the thing, but if you brought them down

01:37:09.480 to normal levels of insulin, then you might see more lipolysis and other things. So I think if you

01:37:14.360 were to do those studies and they've been done, there is peripheral insulin resistance,

01:37:18.520 but then you superimpose in addition. And I'll just say, I'll share with your listeners,

01:37:24.440 we're finding actually the same mechanism that we have in liver and muscle. And we're seeing this in many

01:37:30.360 other tissues too. In the fat cell, the diacylglycerol epsilon pathway is also accounting for

01:37:37.320 this defect in insulin action in the fat cell. So it's going to actually be a common mediator. And

01:37:43.000 again, most of the fat, of course, in the fat cells in the lipid droplet. So again, the plasma membrane

01:37:49.240 diacylglycerols that lead to epsilon activation in the membrane of the fat cells. And we're seeing the

01:37:54.360 same thing. And we see those same mice that I showed you before, the IRK knockin mice are protected from

01:38:01.240 lipid induced fat insulin resistance. On the fat topic, we've talked a lot about the

01:38:06.360 intra myocellular lipid. You've distinguished it from say marbling or fat between cells. One thing we

01:38:14.280 haven't spoken about that clinically gets a lot of attention is visceral fat. So you alluded to doing

01:38:20.040 an MRI. So we do a T1 weighted image of a person on an MRI gives us a beautiful resolution anatomically

01:38:27.240 of what's happening. And you can see the difference between a healthy person and an unhealthy person.

01:38:32.040 And one of the most glaring differences between people on that type of proton imaging is the amount

01:38:39.560 of fat that is inside the fascia. So you have subcutaneous fat that may not be aesthetically pleasing,

01:38:47.480 but more importantly, when you go inside the corset of fascia, you have some people that will have a

01:38:53.240 heavy ring of fat around their kidneys, their spleen, their liver. We call this visceral fat and the

01:38:58.520 association between this amount of visceral fat and poor health is very well understood. Whereas there

01:39:05.080 seems to be very little association between subcutaneous fat and poor health. How does that visceral fat

01:39:11.720 identification square with the intralipid myocellular component that you've described so elegantly at a

01:39:19.000 cellular level? In my view, and everything you said is correct, subq, if you're going to store fat

01:39:24.920 somewhere, that's the best place to store it. You certainly don't want to keep it inside liver and

01:39:29.960 muscle cells. In my view, and again, studies have been done to look at the visceral fat, and it's very

01:39:36.200 clear. It is, again, a very apple-shaped people have visceral fat. It's a very good predictor of insulin

01:39:41.480 resistance. It's really more of a marker for intrapatic fat. So anytime when you're doing your imaging, if

01:39:48.760 you just look at the liver too, they're going to correlate one to one 99 out of 100 times. So what you're

01:39:55.480 really doing there is a marker. Now, the visceral fat will also pour fatty acids into the portal vein, presumably. And 0.50

01:40:04.040 again, fatty acid delivery portal vein is probably going to lead to increased acetyl-CoA. You know,

01:40:09.720 again, will contribute some degree. To me, the major abnormality is really the fat inside the hepatocyte,

01:40:16.200 and more importantly, this acetyl-CoA within the hepatocyte. I want to give one example that makes

01:40:22.120 this point clearly, at least to me, the lesson I learned, and that's lipodystrophy. And as you know,

01:40:28.600 that's a situation where there is no fat, no sub-Q fat or visceral fat. These patients have no visceral

01:40:36.520 fat, huge livers, hepatomegaly, chock full of fat and liver, and again, diabetes through these mechanisms,

01:40:43.160 acetyl-CoA driving gluconeogenesis. And that's independent of the visceral fat. So that shows you

01:40:48.360 you don't need the visceral fat at all to drive this. It's fat in the hepatocyte. If I had to pick two

01:40:54.280 molecules that are driving metabolic disease, it's acetyl-CoA driving perfect carboxylase. And again,

01:41:02.680 the diacylglycerol is activating epsilon. And again, it's the epsilon that drives insulin resistance,

01:41:09.240 no diabetes, no hyperglycemia. Then it's this accelerated gluconeogenesis through this mechanism

01:41:16.200 that's taking you from just pure insulin resistance to fats and hyperglycemia and diabetes.

01:41:20.920 So let's again, pause there for a moment and unpack something very profound. If we've just

01:41:26.920 established that the accumulation of liver fat is effectively the hallmark of death to come,

01:41:35.080 and you just said acetyl-CoA and DAGs are two of the biggest culprits. Well, acetyl-CoA of course

01:41:42.040 is abundance of nutrient on some level, which speaks to something you said earlier. You take a patient with

01:41:48.520 type 2 diabetes, put them on 1,200 calories a day. By definition, that has to lower acetyl-CoA.

01:41:55.720 That immediately is going to improve things, which it does, whether that's sustainable indefinitely,

01:42:00.360 we can discuss. And of course, we've already established where these DAGs are coming from.

01:42:05.480 Again, I want to pause for a moment on that because I think a listener of this right now is going to say,

01:42:09.680 guys, you've lost me, okay? They don't know the difference between PEPCK, GSK3, AKT2, PI3 kinase.

01:42:17.200 I don't think you have to know those things. I think what you have to understand is that abundance

01:42:22.960 of nutrient is a relative term. It's not an absolute term. An athlete versus a sedentary person has a

01:42:29.600 very different amount of what that abundance looks like. I think we've also discussed that not all

01:42:34.800 nutrients are created equal. You've alluded to it already that sucrose and fructose disproportionately

01:42:40.000 prime the liver for this. And then of course, we're dealing with carbohydrate metabolism.

01:42:45.600 This is perhaps an interesting time to also start talking about both the modifications that we can

01:42:52.820 make. Because again, when we start to think about, you've talked about Western diet and sedentary

01:42:57.440 behavior a lot. So there's no doubt that there is an R environmental triggers contributing to these 0.75

01:43:03.560 epidemics, which largely began here in the United States, but we have fabulously spread to the west

01:43:09.460 of the world. And then of course, there's a whole pharmacologic side of this. I would like to revisit

01:43:14.720 the metformin question. I think it's a very interesting question. Metformin works presumably by

01:43:19.680 sort of weakly poisoning the mitochondria at complex one that would lead to a redox change of NAD and NADH,

01:43:26.340 which goes back to something you talked about. But as of this time, at least we don't really have many

01:43:31.620 exciting compounds in the pipeline for NAFLD, which as you also alluded to in about 10 years

01:43:37.700 is going to, through NASH and cirrhosis, be the leading indication for liver transplant in the

01:43:44.000 United States. Something that when I was in medical school accounted for less than 2% of liver

01:43:48.440 transplants just 20 years ago. In 30 years, admittedly with the advent of a cure for hep C,

01:43:55.020 it's now leapfrogged into the lead candidate for liver transplant. And yet what are we doing for it?

01:44:00.680 Not a lot. That's a lot I want to unpack. And as much as you still have time to discuss it,

01:44:05.240 let's proceed in any order you see fit. To add on to that, I just did a Zoom conference for

01:44:09.980 University of Pittsburgh and they're a big liver center. And one of their big problems with

01:44:14.860 transplanting livers is living donors. They're limited by donors because they all have fatty

01:44:21.120 liver, which they will not transplant because they don't do well. So not only is it the problem

01:44:25.820 in treating it in terms of at least this most commonly, that's the most common thing that they do,

01:44:29.940 but that's an aside. So what can we do about this if we can get our patients to lose weight? This,

01:44:36.400 of course, is the best diet and exercise, of course, is the best thing. And that's the first thing

01:44:41.460 I tell my patient. We really drill into them how we can really fix everything that's wrong with them

01:44:47.320 through this process. And unfortunately, as you know, and I know, it just doesn't work in the vast

01:44:52.240 majority of our patients. So in terms of pharmacology, my view, and here, again, it's the

01:44:58.420 liver. If I had to pick one organ to target, it's the liver. As important as muscle insulin resistance

01:45:04.160 is at the very beginning, if we actually want to reverse the disease and make the biggest impact,

01:45:09.240 if I had to pick one organ, it's the liver. If you're going to target, probably the easiest organ

01:45:14.520 to target. The way I think about the liver is in terms of thermodynamics. It's a thermodynamic

01:45:22.220 problem. It goes back to my physics training. And it's really energy in and energy out. The whole

01:45:28.180 metabolic problem with the liver is this imbalance of energy. Too much energy in relative to the ability

01:45:35.300 of the hepatocyte, the liver, to oxidize the energy and convert it to CO2 or export it. The one thing

01:45:42.800 the liver is also able to do is export energy as a form of the LDL triglyceride. If it's energy,

01:45:49.060 how do we fix it? Well, one way, again, we said diet and exercise, limit energy in. That works. And that

01:45:54.660 we talked about, Kit Peterson did this 20 years ago and it's shown many, many times. To get the patient

01:45:59.560 to stay on this is challenging. Bariatric surgery works, again, limiting energy in. We just saw a nice

01:46:05.760 study in the New England Journal. There's no magic to ruin why. It's simply if you pair feed individuals,

01:46:11.820 lose same amount of weight, same effect. Everything the bariatric surgery is doing,

01:46:16.560 at least for one why, is really through reducing through the weight loss. How can we do this

01:46:21.480 pharmacologically? Well, GLP-1 agonists are out there now. They're becoming very popular.

01:46:27.240 Their major effect is energy intake. Our patients eat less. Because they eat less, they lose weight,

01:46:34.680 induces nausea, mild nausea. Some people get into issues with vomiting. Nausea, mommy,

01:46:38.820 have to cut back on the dose. But this is how the GLP-1 agonists, I believe, are having its major

01:46:43.560 effect is weight loss. And they are what they are. They do accomplish reversal fatty liver to some

01:46:49.500 degree. They don't normalize, but it does come down in the right direction. Why do you think the GLP-1

01:46:54.840 agonists lead to reduced appetite? I just think through working through a central mechanism. All these gut

01:47:01.240 peptides lead to nausea, vomiting. Glucagon will do it. Sematostatin will do it. All these things,

01:47:07.880 if you give them a high enough concentrations, lead to some degree of nausea and vomiting. To me,

01:47:13.480 it's part of a spectrum. And if you just get it right, you just get people less interested in food

01:47:18.760 and they eat less. Metformin, that's the one agent we have that lowers gluconegenesis. I would just

01:47:24.060 come back. It's not complex one. I want to challenge you on that. We can talk about that. But

01:47:29.060 to me, it's complex one inhibition happens at millimolar concentrations, clinically not relevant.

01:47:35.460 Our concentrations of metformin in humans, metformin are about 50 micromolar, 40 to 50 micromolar,

01:47:42.040 not millimolar, which is what inhibits complex one. And I think it's downstream. It does affect the

01:47:46.740 mitochondria. It does lead to the redox, but it's not through the complex one. It's probably

01:47:51.160 indirectly inhibiting mitochondrial glycerophosphate, the hydrogenase. That's what

01:47:56.200 leads to the redox. But we can come back to that if you want. I'd love to. That's very interesting.

01:48:01.600 To focus then on other mechanisms. So GLP-1, limit food intake, energy expenditure, SGLT-2 inhibitors

01:48:09.560 cause glucose loss in the urine, 400 calories a day loss. So they lose weight. Unfortunately,

01:48:16.300 it seems to plateau after several weeks. And you get very mild reductions in liver fat,

01:48:22.220 unfortunately, not as much, but maybe in combination with other things that might be

01:48:26.360 certainly helping the right direction. My favorite target is to promote mitochondrial

01:48:31.880 efficiency. And so one of the things we're working on now is to mitochondria is where you burn the fat.

01:48:38.680 That's the organelle that burns the fat through oxidation. If you can promote, then the mitochondria

01:48:44.320 would be a little bit less efficient. So they have to burn more fat to generate the same amount

01:48:48.220 of ATP. This we've shown in various forms, preclinical models, mice, rats with fatty liver,

01:48:55.680 NASH, liver fibrosis. It reverses fatty liver through these mechanisms, reverses NASH, reverses the

01:49:02.160 insulin resistance through reductions in DAGs, acetyl-CoA, reverses diabetes and ZDF models.

01:49:07.340 For the NASH world, it reverses the inflammation and will reverse liver fibrosis. And so I'm very

01:49:14.000 excited about this because I think it can be done safely. More recently, we've done this in

01:49:18.680 non-human primates and showed safety and efficacy of this approach in non-human primates. So based on

01:49:24.440 the mechanisms I've described, I think it fits. And not only what I'm very gratified by is it

01:49:32.200 actually reinforces the mechanisms I've described here by reversing diabetes, insulin resistance by

01:49:37.940 lowering DAGs and acetyl-CoA. But it's also going to be heart healthy. And I want to emphasize this

01:49:42.860 point because many drugs we have now for NAFYL and NASH reduce liver fat, maybe reverse the fibrosis or

01:49:50.780 slow down the fibrosis, but they may lead to alterations of cholesterol in the wrong direction.

01:49:56.600 Cholesterol goes up. And again, I have to come back to a nice point you made is it's heart disease that

01:50:02.440 is killing not only our diabetic, but also fatty liver patients. It's the heart disease. So whatever

01:50:07.960 we're doing to reverse fix NAFYL, NASH, liver fibrosis, it has to be heart healthy. And so when you burn fat

01:50:15.600 in liver through this mechanism, you decrease VLDL export, you lower triglycerides, you raise HDL,

01:50:22.740 and you actually have secondary beneficial effects on the periphery. So you actually will

01:50:27.180 secondarily improve muscle fat, reduce muscle fat and muscle insulin resistance. So this again fits

01:50:33.840 into my conceptual view of insulin resistance and would be, I think, a nice therapeutic approach that

01:50:39.340 we're going after. Now, does the uncoupling lead to excess ROS creation or anything else? Anytime I hear

01:50:47.620 of uncoupling in the mitochondria, which is a deliberately induced form of inefficiency,

01:50:52.740 you wonder, is this an unintended consequence potentially?

01:50:56.800 So uncoupling by definition, the biophysics of uncoupling, the energy has to go somewhere.

01:51:01.340 It's dissipated as heat. You're burning more fat and changes in the energy is going to lead to a

01:51:06.580 little bit of heat production. You will get energy production in the form of heat, but because it's

01:51:12.380 liver targeted, has no effect on body temperature, will not affect whole body weight. It's interesting.

01:51:17.300 I can just tell the story of uncouplers. Your listeners might be interested in this. So

01:51:21.600 they were first discovered actually in the early 1900s in the munitions factories. Europe was getting

01:51:29.800 ready. They knew a world war was coming. The munition factories were all getting geared up.

01:51:35.080 Some of the workers in the munition factories were getting this dust, yellow dust on their hands and

01:51:41.500 actually losing weight. They were just going home and despite eating the usual amount, they're finding

01:51:46.940 their weight was going down and maybe they were sweating a little bit more, a little diaphoresis.

01:51:51.220 And they went to their doctors and told them about the weight loss despite eating the same and

01:51:56.500 it's a little bit more diaphoresis, more sweating. And the doctors just said, what is this yellow dust

01:52:01.820 on your skin? And why don't you just wear gloves, wash your hands and wear gloves? And they got better.

01:52:07.300 This was dinitrophenol. This was a substance that was used in the munition factories to make TNT.

01:52:13.840 So dinitrile TNT. A physician, Tainter in the 1930s, basically said, maybe this is good for weight

01:52:21.880 loss. Actually introduced dinitrophenol as a weight loss drug. It was available over the counter. It

01:52:29.920 wasn't a prescription. So anyone could go like buying vitamins, get some BNP for weight loss. It

01:52:36.520 actually worked. So a lot of people, hundreds of thousands of people took dinitrophenol for weight

01:52:42.780 loss. And it worked. The paper is published in very good journals, JAMA by Tainter and others,

01:52:47.520 really described its beneficial effects. Unfortunately, and a very big unfortunately is again

01:52:53.200 one of the on-target effects that we just talked about. When you uncouple, you promote heat generation.

01:52:59.560 And this is in the whole body. DNP is going everywhere and promoting heat generation.

01:53:05.080 Unfortunately, a handful of these people took too much. They got into problems with hyperthermia,

01:53:09.760 increased body temperature, and got very sick from that. And some died. With the very first thing,

01:53:15.180 a newly created FDA, 1937, the first act they did was actually to pull DNP from the counters as an

01:53:24.140 over-the-counter kind of drug or medication. And the second act they had actually was thalidomide,

01:53:30.080 which they pulled and now is actually back in the clinic. That was always the problem with DNP. Why,

01:53:35.560 again, we say this is not a good thing. This is a toxic drug and everything else. And as it is,

01:53:41.400 it occurred to us that the reason it's generating the heat is you're uncoupling all the organs in the

01:53:47.820 body. And what if we just picked one organ, i.e. the liver, where the fat is accumulating? This is where

01:53:53.440 the organ that's driving lipidemia, hyperlipidemia, and diabetes. And if we could just melt the fat away

01:54:00.020 within a liver-specific manner, maybe we can have that beneficial effect without the toxicity.

01:54:06.120 And so in a series of studies, we were able to show proof of concept that by simply uncoupling the

01:54:12.700 liver, you could avoid hyperthermia and all the toxicities that have typically been associated

01:54:18.420 with the parent compound, DNP, and increase the therapeutic window. Every drug has a therapeutic

01:54:24.820 window, even aspirin and Tylenol, by a hundredfold. Based on this thinking, I think it can be done

01:54:30.460 very safely and be a treatment for very important metabolic diseases like Nafil and Nash.

01:54:36.660 So the IND has already been filed for this. Is it in phase one human yet?

01:54:40.380 No, no. We're still exploring preclinical models, thinking potentially about first starting out where

01:54:47.160 there are no indications for things like lipodystrophy, where leptin is not working. So I

01:54:52.800 think my thinking is I'd like to go slowly here. Hopefully within the next year or two, we may be

01:54:57.680 in humans. I think initially going after orphan diseases where there simply is no other treatment,

01:55:03.500 and that would be certain forms of lipodystrophy where they get very bad diabetes, Nafil, Nash,

01:55:09.540 and especially in conditions where leptin is not working.

01:55:12.860 Jerry, this has been obviously, as I said, a pretty technical discussion, even by the standards

01:55:17.440 of our podcast. I think the show notes are going to be integral because your figures, I think,

01:55:24.040 frankly, are very helpful. As I said, I understand this content probably better than most, and yet I

01:55:28.280 still find it very helpful to be able to kind of go through schematics. So I'm going to encourage the

01:55:32.720 listeners to do that. You also have some fantastic lectures online. I think for the people who really want

01:55:39.500 to go deep into this stuff, I think, frankly, some of your review articles and some of your recent

01:55:43.700 publications are just a great place to go. As I said at the outset, I just think that this

01:55:49.040 is the nexus from which all diseases stand. And therefore, we are really making a mistake

01:55:56.660 if we want to treat chronic diseases in their silos and just think about atherosclerosis and just think

01:56:03.100 about cancer and just think about Alzheimer's disease without understanding how these diseases are fed.

01:56:09.500 And unfortunately, that means rolling up our sleeves and understanding insulin resistance.

01:56:13.960 There's simply no getting around this. If this topic were easy, you would have presented it in

01:56:19.160 an easier manner. It's not easy. If I were to just kind of leave you with sort of, we've talked about

01:56:24.600 exercise, we've talked about nutrition. Do you feel strongly about any form of dietary thinking? So for

01:56:32.540 example, I have found clinically that carbohydrate restriction is a very effective way for patients

01:56:39.400 with insulin resistance to lose weight, not uniformly, but it's quite effective. It also seems to be easier

01:56:45.960 to adhere to than outright caloric restriction, though periodic fasting also seems to do a good job.

01:56:53.040 But have you observed anything similarly from a clinical perspective that fructose restriction

01:57:00.200 specifically or sugar restriction specifically as a vehicle to weight loss becomes a more effective

01:57:08.620 tool to ultimately produce what's understood to be efficacious, which is some reduction of weight

01:57:15.300 either as the cause or effect of the improvement? My thinking here is what I tell my patients is

01:57:20.220 whatever works to everyone is so different, different likes, different dislikes. I say, look at the scale,

01:57:26.520 whatever works for you to lose weight, because I know if you lose the weight, your diabetes is going

01:57:32.360 to get better. So I say you find something, whatever works for you, stick with it. That's the challenge

01:57:38.900 because we're very successful in the short term getting patients to lose weight. The unfortunate part

01:57:44.760 is they're able to get the weight off and then three months later, six months later, they come back to

01:57:49.440 the office and they're right back where they started. So it's a matter of, I tell them, you have to find

01:57:54.500 something that works for you, get the weight off, but then you have to be able to stick to it.

01:57:58.820 And that's where the challenge is. A lot of diets, people are able to get on, get the weight off,

01:58:02.740 and they just can't adhere to it for the long term. And so it's a marathon. You have to find something

01:58:08.080 you like, like it enough to be able to stick with that. That's the most important thing because

01:58:12.240 we've all seen that where people lose the weight and then a few weeks, months later, right back to

01:58:18.140 where they started. So everyone has to find what works for them.

01:58:21.280 I guess I want to come back to the metformin thing because it's so interesting. So you mentioned

01:58:26.220 that the inhibition of complex one actually is probably not taking place because you actually

01:58:36.400 mentioned basically a thousand fold difference in concentration. Say a little bit more about that and

01:58:41.320 why you're then imputing that it's the impact of metformin presumably on NAD and NADH, which you could

01:58:49.640 also get out of an inhibition of complex one, but via some other mechanism, it sounds like.

01:58:53.740 Studies that we've done, and we're still working on this, clearly most of the literature, if you read

01:58:59.240 on metformin, let's talk about the big picture. So metformin lowers glucose in patients with poorly

01:59:06.240 controlled diabetes, mostly through inhibition of gluconeogenesis. I think most clinical physiologists

01:59:12.760 would agree with that. And so we've done studies quantifying gluconeogenesis, both by NMR, heavy

01:59:18.900 water, multiple methods, same individuals. And that's its major effect, not through inhibition

01:59:24.180 of glycogenolysis, not through gut biome. It's gluconeogenesis. And the other thing clinically

01:59:29.960 is the more poorly controlled diabetes, the greater the effect. You're not going to see much effect.

01:59:35.720 There's very confusing studies that have been published in non-diabetic individuals that find

01:59:40.280 all kinds of other things going on. I don't think that's clinically relevant. It's gluconeogenesis.

01:59:44.620 So then how does it do gluconeogenesis? So most of the literature, if you read it,

01:59:49.360 virtually all in animals that study mechanism, have implicated complex one. And we've known

01:59:54.040 about guanide inhibition. Metformin is a guanide, biguanide. Even before metformin,

02:00:01.100 we had fenformin and other guanides that have been studied. And they will inhibit complex one,

02:00:06.620 no doubt about it. And most have focused on complex one inhibition leading to either AMPK activation

02:00:14.180 or buildup of a metabolite that inhibits gluconeogenesis or something. 99% of the mechanisms

02:00:22.680 have talked about complex one inhibition. My issue with that is, again, not very many studies have done

02:00:29.600 careful measurements of this most commonly used drug on the planet. For your readership,

02:00:36.760 guanides have been used for diabetes for hundreds of years. The French lilac extracts have been used

02:00:43.040 300 years ago in description. They didn't know what diabetes was at that time. It wasn't defined,

02:00:48.400 but patients with polyuria, polydipsia, were overweight, treated with the extract,

02:00:54.160 the French lilac, and their symptoms improved. Most studies, if you look at, were used at millimolar

02:00:59.080 concentrations. And again, when they look at complex one inhibition, which has been implicated to then

02:01:04.720 lower ATP, raise ADP, and activate AMPK, it requires millimolar concentrations. And so when you

02:01:12.140 actually measure metformin in the patient who's taking one gram twice a day, which is your maximal

02:01:18.480 dose, pretty much the best efficacious dose, your levels in plasma are about 30 to 50 micromolar.

02:01:25.420 So you could say even, you know, in portal vein, it's pills are taken orally, give it two to three

02:01:30.440 times that. You're still talking about maybe 100 micromolar, tenfold less than what all of these

02:01:36.740 studies have been doing, even both the in vitro studies in the literature and well, the in vivo

02:01:40.980 studies, giving levels that achieve millimolar concentrations. So yes, you see things. Complex one's

02:01:48.340 an important, it's an electron transporter, it's important for function and health. And you're going to

02:01:52.760 see effects when you inhibit complex one at those high concentrations. In my view, they're not

02:01:58.460 clinically relevant. So the effects that I do think are clinically relevant that we have observed

02:02:03.680 at 50 and 100 micromolar of metformin are really on the enzyme glycerol-3-phosphate dehydrogenase,

02:02:11.660 the mitochondrial isoform that is required to move the protons from outside to inside the mitochondria.

02:02:18.880 And when you inhibit this enzyme, NADH goes up, NAD goes down. When you have this increase in the

02:02:25.680 cytosolic redox, you can't get lactate to pyruvate and you can't get glycerol to DHAP.

02:02:33.720 So if I'm right, it's going to be substrate dependent inhibition of gluconeogenesis. Whereas if you inhibit

02:02:39.760 complex one and AMPK or whatever mechanism downstream, it should be gluconeogenesis independent of substrate.

02:02:46.580 And what we've shown both in vitro and in vivo, most importantly in vivo, in two or three different

02:02:52.880 models, metformin at these clinically relevant doses and concentrations only inhibit gluconeogenesis from

02:03:00.720 glycerol and lactate. It doesn't inhibit it from alanine or DHAP or anything that does not depend on

02:03:09.460 the cytosolic redox state. This also explains why we rarely see clinically hypoglycemia on patients

02:03:17.300 treated with metformin because there's these alternative gluconeogenic substrates that can

02:03:22.040 come in, alanine can keep coming out. So you never see, rarely, unless they have another agent on top

02:03:27.380 of metformin like insulin or SU, you rarely see it if ever. And that's why also you see the lactic

02:03:32.680 acidosis, which is a fortunate toxicity of metformin, where again, it's specifically getting

02:03:39.420 that lactate to pyruvate conversion, which is dependent on the redox state. So that's the

02:03:43.600 mechanism I believe is clinically relevant. And now we're at last step is how is it inhibiting this

02:03:48.740 enzyme? And I believe it's actually through an indirect effect on this enzyme that we'll

02:03:52.720 hopefully have ready for prime time in the year.

02:03:55.880 And do you think that in a healthy individual who's eating well, is of normal weight, is insulin

02:04:02.580 sensitive, and is exercising robustly, metformin could actually be counteractive to benefit?

02:04:10.820 That's a profound question. I don't know the answer to that. And it gets into, I don't know if you're

02:04:16.660 going to take me there in terms of the use of metformin for aging. Healthy people are taking it for

02:04:21.740 aging now. I think that's why it's so important to understand this mechanism, then understand the

02:04:27.100 implications of it. It is redox. Is that a good thing or not for longevity and health? That's a

02:04:33.100 question that remains to be answered. I find myself very much on the fence with that question,

02:04:38.260 while in the insulin resistant patient, even without diabetes, feeling that this is a very net

02:04:43.700 positive agent. But my personal views on it, just based on clinical observation, is that in the

02:04:50.440 person I described earlier, the lean insulin sensitive, vigorously exercising individual,

02:04:55.840 it may actually not provide benefit. But again, there are studies in the works that are going to

02:05:01.640 hopefully be able to provide some fidelity to understanding that. It sounds like you're equally

02:05:06.680 kind of undecided on that as well. Yes. Well, Jerry, I can't thank you enough. Again, I say this

02:05:12.000 to many people I interview, but I really mean it here. It's not just for this discussion and the time

02:05:15.340 you put into it, but obviously much more importantly for the career and for this incredible body of work

02:05:21.580 that you've amassed through your pursuit and obviously remarkable collaborations with so many

02:05:27.500 people. I've enjoyed this discussion immensely. It's actually one of the discussions I'm going to

02:05:32.600 have to probably go back and listen to again. So I hope that a listener isn't hearing this and isn't

02:05:37.740 discouraged by the fact that you're at this point in the discussion and you're thinking,

02:05:40.840 oh my God, I might've only retained half of that. That's okay. I'm going to be listening to this one

02:05:45.400 and I just finished listening to it now and I'm going to listen to it again. So thank you very much,

02:05:50.300 Jerry, for that. Thank you, Peter. It's been a pleasure. Thank you for listening to this week's

02:05:54.620 episode of The Drive. If you're interested in diving deeper into any topics we discuss,

02:05:58.980 we've created a membership program that allows us to bring you more in-depth exclusive content

02:06:03.640 without relying on paid ads. It's our goal to ensure members get back much more than the price

02:06:08.820 of the subscription. Now to that end, membership benefits include a bunch of things. One,

02:06:13.840 totally kick-ass comprehensive podcast show notes that detail every topic, paper, person,

02:06:18.980 thing we discuss on each episode. The word on the street is nobody's show notes rival these.

02:06:24.460 Monthly AMA episodes or ask me anything episodes, hearing these episodes completely.

02:06:29.560 Access to our private podcast feed that allows you to hear everything without having to listen to

02:06:34.840 spiels like this. The Qualies, which are a super short podcast that we release every Tuesday

02:06:40.160 through Friday, highlighting the best questions, topics, and tactics discussed on previous episodes

02:06:44.900 of The Drive. This is a great way to catch up on previous episodes without having to go back and

02:06:49.960 necessarily listen to everyone. Steep discounts on products that I believe in, but for which I'm not

02:06:55.700 getting paid to endorse. And a whole bunch of other benefits that we continue to trickle in

02:07:00.200 as time goes on. If you want to learn more and access these member-only benefits,

02:07:04.180 you can head over to peteratiamd.com forward slash subscribe. You can find me on Twitter,

02:07:10.360 Instagram, and Facebook, all with the ID, peteratiamd. You can also leave us a review on

02:07:16.100 Apple Podcasts or whatever podcast player you listen on. This podcast is for general informational

02:07:22.040 purposes only and does not constitute the practice of medicine, nursing, or other professional

02:07:26.480 healthcare services, including the giving of medical advice. No doctor-patient relationship

02:07:32.240 is formed. The use of this information and the materials linked to this podcast is at the user's

02:07:37.920 own risk. The content on this podcast is not intended to be a substitute for professional medical advice,

02:07:44.260 diagnosis, or treatment. Users should not disregard or delay in obtaining medical advice

02:07:50.360 from any medical condition they have, and they should seek the assistance of their healthcare

02:07:54.960 professionals for any such conditions. Finally, I take conflicts of interest very seriously. For all

02:08:01.540 of my disclosures and the companies I invest in or advise, please visit peteratiamd.com forward slash

02:08:08.960 about where I keep an up-to-date and active list of such companies.

02:08:24.960 Thank you.

The Peter Attia Drive - December 07, 2020

#140 - Gerald Shulman, M.D., Ph.D.： A masterclass on insulin resistance—molecular mechanisms and clinical implications

Episode Stats

Summary

Transcript