The ASCI Interviews
Michael W. Schwartz: ‘Lots of angles that can be taken’

Published September 10, 2014

Dr. Michael W. Schwartz is the Robert H. Williams Endowed Chair in Medicine; Director of the Diabetes and Obesity Center of Excellence; Director of the Nutrition Obesity Research Center; and Professor in the Division of Metabolism and Endocrinology at the University of Washington (UW). Dr. Schwartz received his M.D. from Rush Medical College and completed his residency in medicine and his fellowship training in endocrinology and metabolism at UW. Dr. Schwartz’s research focuses on hypothalamic and neuroendocrine control of energy balance and glucose metabolism and on CNS mechanisms involved in obesity, insulin resistance, and diabetes. Elected to the ASCI in 1999, Dr. Schwartz is also a member of the AAP.

John Hawley is the ASCI’s Executive Director.

John Hawley: Tell me a little bit about your inspiration for a career in medicine.

Michael Schwartz: When I took my first college course in animal physiology, it felt like something that I could be excited about as a career. That I could spend my life trying to answer the questions raised in that type of science — it just clicked as something I wanted to do. Then it was a matter of how and what to do. My father, Theodore Schwartz, who passed away in 2008, was an academic physician and a chair of medicine at Rush Medical College for many years. That had an impact on the direction I chose. I decided that I could go to medical school, and that whatever sort of research career I might have, I would prefer to do it from the point of view of being a physician and understanding medicine.

JH: Your father was also a member of ASCI – always interesting to discover multigenerational ASCI members. What focused you as you got through medical school and started focusing on a career in research?

MS: Even back in college, I had gotten interested in the brain-behavior interface — the concept that you could understand emotion or even conscious thought from a biological perspective. When I went to medical school, I wondered, “What would be a good specialty to pursue that type of question within the context of how academic medicine is organized?” I thought about neurology, neurosurgery, psychiatry. But I decided that a neuroendocrine approach would be the best option. I enjoyed medicine the most during my training, and endocrinology is a subspecialty of medicine, and I thought the scientific approach used in classical endocrinology might make the most sense as I pursued my own interests. And those approaches are a little bit different from what you might see in psychiatry or neurosurgery. I also felt that if you’re going to pursue a science career, you never know if it will work. It’s a matter of keeping as many good options open as you can. I enjoyed the clinical practice of endocrinology, so if the research career didn’t work out, I would still end up doing something I wanted to do.

JH: Aside from your father, were there other mentors that kept you on the track, gave you the confidence to know that you could work through it?

MS: Aside from my father, who was a source of inspiration and advice and a sounding board for my work, the main person was my primary research mentor when I was an endocrine fellow, Dan Porte, who is also an ASCI member. He recruited me as a trainee during my residency here at the University of Washington. He provided a great environment for me to learn the craft of being a good scientist. He was one of the leaders in metabolism and type 2 diabetes in his era. He had won the Banting Medal of the American Diabetes Association, was the president of the American Diabetes Association, and was a world leader in understanding insulin secretion and how it becomes abnormal in type 2 diabetes. But he also had an interest in the central nervous system and was the first to document that insulin secretion from the pancreas could be regulated by neural inputs. He had this interest in the interface between brain and metabolism, which is what I became interested in as well. He’s a really critical thinker and a creative person.

JH: When you exited your training pathway and began establishing your own lab, what were the challenges you encountered?

MS: I think the biggest challenge in an academic medical environment — which is complicated, with many moving parts — is the realization that the buck stops with you. You no longer have a person you can turn to with problems. Whether you’re dealing with personnel issues in the lab, funding issues, administrative concerns, trying to get a paper published, whatever it is, it’s on you to get through that process and come out of it in a way that allows you to stay competitive and successful.

That’s the big transition: You might think you’re a good scientist when you’re still working under your mentor’s umbrella — and you may be — but you haven’t found out what it means to be an independent scientist until that safety net is removed. Related to that, in an academic medical center, I realized fairly early on that if you want to have control over your destiny as a scientist, you need to command resources. If you get one grant, you can’t say, “Okay, now I’m all set.” You need to have as much funding as you can commandeer. There’s a tension in a clinical department between how much clinical, teaching, and administrative work and how much research work is expected of you. The best way to leverage that situation so that you end up having the time and resources to do what you think is important is to be effective in bringing in the funds to run your research program.

If the administration sees that you’re doing that well, at least at the University of Washington, they will not try to get you to do things that you don’t want to do; they realize that if you’re taking on too many other responsibilities, you may not be able to continue to support your lab in that way. So investigators who prove that they can be highly productive and secure funding early in their careers are the ones who end up with the most protected time to do their work. It’s a balancing act, and you need to know that from the beginning. There are always plenty of things that the administration can have you do — extra clinical work, extra teaching in the medical school — and it’s hard for people to say no. The best way to protect your time and ensure you have the chance to be successful is to establish a large, well-funded research program, and that has to be your top priority.

JH: What was your first major project grant?

MS: The first one was an R01, and I had that when I was still under the mentorship of Dan Porte. Actually, by the time I really was independent, I had two R01s from the NIH. That was really what I used launch my independent career.

JH: You still working on one of those R01s, “Mechanisms of Diabetic Hyperphagia”?

MS: Yes. That was my second grant. My first grant did not get renewed the last time, I think that was in 2012, so that was kind of a heartbreak, but that’s the way it goes.

JH: Working on a grant that started 18 years ago, you have been successful focusing on a very particular topic. Could you talk about how you envisioned your work on hyperphagia and where it is now?

MS: It began with trying to understand how the brain controls food intake and body weight, how that is related to normal control of body weight and abnormal control, as in obesity. Then it expanded to include how the brain controls blood sugar, with the realization that it plays a more important role than previously thought, both under normal circumstances and in the setting of diabetes.

I had adopted a metabolism study paradigm approach, but I was applying it to a question about the brain, and there wasn’t really much of a precedent at the time. We know that, for example, if an animal is fasted for a day, they’ll eat more than they normally do to catch up. That is a form of hyperphagia, so-called re-feeding hyperphagia. A simple question you could ask is, “What drives them to eat more?” You can’t explain it just on the basis of behavioral analysis. It has to be tied into metabolism, because as soon as body weight catches up to normal, food intake becomes normal. So there is a connection between the change in body weight and the food intake. The approach we took was to ask, “What happens in the body during a fast that might be a signal to the brain to cause it to eat more?”

At that time, in the early 1990s, we knew that fasting lowers circulating levels of insulin quite a lot. There was evidence that if you put insulin in the brain of normal animals, they eat less. So we thought, maybe when the animals are fasted, the drop in insulin levels in the blood is a signal that stimulates feeding behavior. That was a fundamental premise, and it’s a classic metabolism paradigm: fasting lowers a hormone, it elicits a response, and the question is, Is the response related to the change in hormone? And can you get it back by giving the hormone back during the fast?

There was known to be a particular set of neurons in the hypothalamus called neuropeptide Y (NPY) neurons that are activated by fasting. And if you treat the brain of a rat with NPY, it will stimulate feeding. So now you have this idea that fasting lowers insulin levels and activates neurons that stimulate feeding. So can we connect the dots and say that the low insulin is a factor that stimulates the NPY neurons and therefore is connected to the food intake? That’s the way this type of science has worked: trying to build a chain, but in a way that can be tested. So in one of my earlier papers, we asked, “If you directly infuse the brain of a fasted rat with insulin, can you block the activation of the NPY neurons?” The answer was yes. That provided the first link between a change in a peripheral hormone and a change in a neural circuit that was related to feeding behavior.

This was fortuitous, because those NPY neurons are very important for feeding behavior — and not just because of NPY, but because of other things that they do and because it turns out they’re regulated not only by insulin but by leptin, which at that time hadn’t been discovered. But this paradigm for insulin had been established when leptin was discovered in 1994, and it took about six months after the identification of leptin to get it for our experiments. It was a no-brainer, because we knew that if you fast an animal, leptin levels drop too, just like insulin. The question was, “Are these NPY neurons also targets for the action of leptin?” We were then able to test both whether leptin administration to normal animals inhibits NPY neurons — and we showed that it does — and what happens to NPY if you give leptin to mice that are obese because they don’t produce leptin, called ob/ob mice. We showed that ob/ob mice have the same type of activation of the NPY neurons that you see with fasting. When you give them leptin, you shut the neurons off.

It all fit together that NPY neurons are inhibited by both insulin and leptin, and that insulin and leptin are two hormonal signals that the brain uses to adjust behavior to the metabolic needs of the animal. And that explained where the diabetic hyperphagia comes from: animals that have lost their insulin-secreting cells, which causes diabetes, show markedly elevated blood sugars, but also markedly increased food intake. But since leptin is made by fat cells, and in the absence of insulin, fat cells can’t store fat and all the fat in the body starts dissolving, leptin levels also plummet. This situation is like fasting, with very low levels of both insulin and leptin. The difference is that when you’re fasted, you’re not eating, whereas in uncontrolled diabetes, you’re eating continuously but you can never raise the insulin or leptin level, no matter how much you eat.

So the animals are in a perpetual state of being driven to eat, and there’s nothing they can do about it unless you give them back leptin or insulin. We showed that you can reduce food intake if you shut off the NPY neurons, for example, by giving leptin to a diabetic animal. And that took us in a different direction over the last five years, because we realized if you treat the brain of the uncontrolled diabetic rat that has low leptin and low insulin with a small amount of leptin, you normalize not only food intake, but also blood sugar, even though you’re not giving them insulin.

This was among the first observations that the brain can lower blood sugar without insulin. And you can fix diabetes by targeting the brain independent of insulin, and that’s a major focus for us now — that’s where FGF19 comes in. So ob/ob mice, the leptin-deficient obese mice, have elevated blood sugar, like a mild diabetes and severe glucose intolerance; this means that if you give them a bolus of glucose, the blood glucose levels will go much higher than normal, and it will take longer to clear the sugar out of the body. But part of what’s causing the glucose intolerance of ob/ob mice is that they are — as you would expect, because they’re obese — severely insulin resistant. But they also have a major impairment of the component of glucose disposal that doesn’t involve insulin at all, which in normal mice mediates about half of glucose disposal: insulin-independent glucose disposal is reduced by 70% in ob/ob mice.

We figured that you couldn’t fix their glucose intolerance unless you increased the insulin-independent and insulin-dependent components. But there was no precedent for rapid regulation of insulin-independent glucose disposal. It was thought to be something that could not be regulated, much less by the brain. In a study we published in the JCI last fall, we selected FGF19, a hormone member of the FGF family that’s made in the gut in response to a meal; as early as 2004 investigators at Lilly had shown that if you inject FGF19 into an ob/ob mouse, you could dramatically improve its glucose tolerance, and they saw the effect even when they put FGF19 into the brain. So we thought, “If you give FGF19 into the brain and improve glucose tolerance, there’s a good chance that this involves an improvement of insulin-independent glucose disposal.” We gave FGF19 or vehicle into the brain of ob/ob mice and showed that within 90 minutes, it had a dramatic effect to improve glucose tolerance and that the effect was exclusively due to stimulation of insulin-independent glucose disposal. This showed, for the first time — basically very similar to what we found with giving leptin into the brain of diabetic animals — that you can dramatically improve glucose homeostasis through the brain and through mechanisms that don’t have anything to do with insulin.

This has raised the possibility that there’s a side of glucose metabolism that’s been neglected, because most of the focus since the discovery of insulin has been how insulin is secreted, how and where it acts, what it does to the liver, what is the signal transduction mechanism, and so on. Thousands and thousands of papers have been published on these questions over the last 90 years since insulin was discovered. There’s been very little published on the insulin-independent component, and yet we’re finding that if you activate it through the brain, it can have effects just as powerful as insulin does. That has now become a major interest, and we have two R01s now that are focused on different aspects of this brain control of blood sugar.

JH: This sounds like a tremendous area of potential for therapeutics.

MS: There are lots of angles that can be taken here. Beginning with leptin: obviously it’s not a panacea — it failed for drug treatment of obesity. But recently it was approved by the FDA for humans with a disease called lipodystrophy, a genetic defect that prevents you from making fat cells properly. And because those individuals can’t make fat cells properly, they don’t make leptin properly either. There’s lots of evidence now that if you give leptin to those patients, they get better metabolically. The patients have diabetes despite having insulin levels that are sky high. The problem is, number one, since they don’t have the fat cells that they need to store calories, the brain is always getting the message that there isn’t enough body fat, just as if they were starving to death. When the brain gets that message, it activates responses that drive up food intake and also drive up blood sugar because the brain doesn’t like to be in a situation where its going to run out of fuel. So the two most dramatic things that it does are driving up appetite and driving up blood sugar. And one of the ways that the loss of fat storage capacity is communicated to the brain is by the low leptin signal. So if you give leptin back to those patients, you shut down the brain’s signal, food intake comes back to normal, and the blood sugar drops. That’s the only treatment that has been found to be effective for those rare cases with congenital lipodystrophy. You can give them as much insulin as you want, and it’s not going to fix their diabetes; but if you give them leptin, there’s a good chance that their diabetes will go away.

One of the major clinic treatments for type 2 diabetes is a class of drugs based on GLP1, a gut hormone that’s a so-called incretin hormone believed to work primarily by increasing insulin secretion from the pancreas. If you have type 2 diabetes, GLP1 can boost the amount of insulin that you make. Exenatide, a long-acting GLP1 receptor agonist, and liraglutide are widely used, successful drugs in type 2 diabetes. Actually, liraglutide is now under review with the FDA for treatment of obesity, distinct from diabetes. This raises a separate issue: If the only thing that liraglutide did was to increase insulin secretion, it wouldn’t be useful for the treatment of obesity. But it also reduces food intake. How does it do that? It acts in the brain. It also reduces secretion of the hormone glucagon, and glucagon tends to drive up blood sugar levels.

The point is that GLP1 and GLP1 agonists represent a class of drugs that are already in use for the treatment of diabetes. Although they were marketed as being useful primarily because they acted to stimulate insulin secretion, it turns out that they also have beneficial effects in the central nervous system. We have unpublished work where it looks like these GLP1 analogs can improve blood sugar control not only by increasing insulin secretion but also by increasing the insulin-independent component of glucose disposal like FGF19. This might explain why these drugs are now being used in type 1 diabetes. Stimulation of insulin secretion in type 1 diabetes doesn’t get you very far, because there is no insulin secretion, so it’s interesting that these drugs are still useful in lowering blood sugar. I believe that that probably is mediated through the central nervous system.

It’s a funny evolution, because people have shied away from using drugs to treat obesity and diabetes that act in the brain in the fallout of Sanofi’s attempts to get a drug called Rimonobant approved for obesity treatment, and it turned out that Rimonobant increased the risk of suicide in humans in large drug trials. The drug companies just said, “We’re dropping our obesity program,” because you can’t target the brain without having unexpected side effects. Everyone moved away from the central nervous system as a target.

After GLP1 was introduced — and I know this because at the time I was an advisor to companies that were making GLP1 analogs before they were approved by the FDA — the focus was on accumulating insulin secretion. And it was beneficial and did increase insulin secretion, and it ultimately became part of mainstream treatment for type 2 diabetes. Once it was approved, these drugs were eventually found to be helpful in type 1 diabetes as well, but no one stopped to say, “Wait a minute, doesn’t that mean that it’s working in the brain to lower blood sugar?” The point is, through this backdoor pathway, I believe we’re going to come to realize that GLP1 analogs have a beneficial effect on blood sugar that involves the brain as well as the pancreas. It’s a precedent that I believe will be seen retrospectively by the pharmaceutical industry as having given a green light to consider the brain a legitimate target for metabolic diseases, even though we have come to that conclusion by a pathway never intended to even ask the question.

JH: That’s very interesting — the unexpected intersection of the course of research on one stream and the course of direct development on an entirely different stream. To go back to the training and mentoring angle, what is the size of your lab?

MS: When I say how big the lab is, that actually incorporates one of my trainees, Greg Morton, who came to me in the early 2000s as a post doc. Now he’s an associate professor, but we decided to run the lab together; we’re like a team. Greg Morton was the first author on the JCI paper from last year about FGF19. Another trainee is Josh Thaler, and he was the first author on our 2012 JCI paper that demonstrated hypothalamic injury in gliosis in association with diet-induced obesity. Josh is now an assistant professor. With him, his fellow, and Greg and my group, we’re around 20 people.

JH: That’s a fairly significant micro-enterprise. In looking for people to join your lab, you’ve had success, and you now have a team approach to your work. Are there aspects that are particularly important to you when you interview people for positions in your lab?

MS: I’m looking for several things. One, are they really inquisitive and do they have the fire in the belly — not necessarily to be promoted or to get recognition, but to answer questions scientifically? If you have that, it will get you through a lot of other things. Two, can they assimilate and process information in a logical way? If we’re having a conversation about something scientific, you can identify those who can think about what we’re talking about in an analytical way that allows them to see the connections that you’re going to need to be successful, at least in an integrative biology perspective, which is what our lab focuses on. Third is a team approach. Is this person going to fit in well? Are they able to look at the needs of others as being not so different from their own? And understand the need to help other people? We have a number of technical lab support people, but none of them is assigned to any one individual. If Greg Morton or Josh Thaler said to me, “I want to have my own technician to do this one assay exclusively in support of my research,” I would say, “Sorry, we tried that and it doesn’t work.” Nobody has control over any other person in the lab. Instead, each week at the lab meeting, we go through everything that every investigator in the lab wants to do that week, and we assign the personnel necessary to do it. When everyone feels their needs will be taken care of as long as they make sure everyone else’s needs are taken care of, that has proven to be the best formula for us. We’re looking for people who can buy into that. And that’s not everybody. Last is a fount of knowledge and basic intelligence, and finding out whether someone can really speak the language and think critically.

JH: Funding is a major issue, trainee pathways are longer. What would you tell someone considering a career either in academic medicine or life sciences, given that and also your significant accomplishments in your research area?

MS: We’re in a transitional period. It used to be that when there was a reduction of NIH funding, which has happened once a decade or so, the percentage of grants funded would drop, and then it would come back up again. I’m not sure that’s true now. This failure to increase NIH funding despite the fact that the cost of doing research continues to increase may be the new normal. Certainly looking politically, it’s hard to imagine the scenario where we would seeing increases that would make up for the deficit we’ve experienced over the last decade; it used to be that you could be in the top 18%–20% of applicants and get funded, and now you have to be in the top 10%–12% typically. And that’s a somewhat misleading number: If you’re requesting five years of support and you get funded, you often only get four years with 20% cut off the top of the number you requested. The amount of federal funding available to support a research program like mine has dropped a lot over the last decade, in particular since about 2011. The question is, what is the future going to look like? I’m not sure I know the answer, but the programs that are successful are those that have merged the ability to compete for federal funding with successful philanthropy in support of their programs.

In the context of the income inequality that’s now being talked about as a national issue, revenue is disproportionately being earned by a limited number of wealthy individuals, and those individuals are increasingly part of the future solution to these problems. A billionaire whose daughter has diabetes, for example, might say, “Not only do I want to get the best care for my daughter, I want to help figure out how to solve this problem.” If you try to build a program based on investigators who are funded by the NIH, and that’s where all your money comes from, that doesn’t work any more.

National Institute of Diabetes and Digestive and Kidney Diseases. DK090320. Hypothalamic Inflammation and Energy Homeostasis.

National Institute of Diabetes and Digestive and Kidney Diseases. DK101997. Novel Anti-Diabetic Actions of Hypothalamic FGF19-FGFR1 Signaling.

National Institute of Diabetes and Digestive and Kidney Diseases. DK083042. Novel Brain Mechanisms Controlling Glucose Homeostasis.

National Institute of Diabetes and Digestive and Kidney Diseases. DK035816. Nutrition Obesity Research Center.

Schwartz MW, Seeley RJ, Campfield LA, Burn P, Baskin DG. Identification of targets of leptin action in rat hypothalamus. J Clin Invest. 1996;98(5):1101–1106.

Seeley RJ, Yagaloff KA, Fisher SL, Burn P, Thiele TE, van Dijk G, Baskin DG, Schwartz MW. Melanocortin receptors in leptin effects. Nature. 1997;390(6658):349.

Thaler JP, Yi C-X, Schur EA, Guyenet SJ, Hwang BH, Dietrich MO, Zhao X, Sarruf SA, Izgur V, Maravilla KR, Nguyen HT, Fischer JD, Matsen ME, Wisse BE, Morton GJ, Horvath TL, Baskin DG, Tschöp MH, Schwartz MW. Obesity is associated with hypothalamic injury in rodent models and humans. J Clin Invest. 2012;122(1):152–162. PMCID: PMC3248304.

Morton G, Matsen M, Bracy D, Meek T, Nguyen H, Stefanovski D, Bergman R, Wasserman DH, Schwartz MW. FGF19 action in the brain induces insulin-independent glucose lowering. J Clin Invest. 2013;123(11):4799–4808. PMCID: PMC3809800.

Schwartz MW, Seeley RJ, Tschöp MH, Woods SC, Morton GJ, Myers MG, D’Alessio D. Cooperation between brain and islet in glucose homeostasis and diabetes. Nature. 2013;503(7474):59–66. PMCID: PMC3983910.