Neuronal Control of Appetite, Metabolism and Weight

Last week, I attended a Keystone conference, "Neuronal Control of Appetite, Metabolism and Weight", in Banff.  Keystone conferences are small, focused meetings that tend to attract high quality science.  This particular conference centered around my own professional research interests, and it was incredibly informative.  This post is a summary of some of the most salient points.

Rapid Pace of Scientific Progress


The meeting really underscored the incredible pace of scientific progress on feeding, body weight and metabolism, and the key role the brain plays in these processes.  This is largely due to remarkable technical breakthroughs being applied by people who know how to use them to answer important questions.  Conditional knockout and transgenic mice, in which genes can be manipulated in specific cell types, continue to play an important role in neuroscience research.  However, the more recent introduction of optogenetic and DREADD technology has been a quantum leap.  Optogenetics (light-activated ion channels) allows researchers to manipulate specific neuron populations using light, which offers exquisite spatial and temporal control.  This is important because  1) you can fire (or inhibit) neurons in a way that roughly approximates how they would fire naturally, and 2) even within tiny sub-nuclei of your brain, there can be a number of different cell types performing different functions, so having a technique with cell type specificity is useful (in contrast, imagine trying to determine the function of a nucleus by lesioning it, which might destroy 10 different intermingled cell populations all doing different things).  One early example of the application of optogenetics to feeding showed that activating AgRP neurons in the arcuate nucleus "evokes voracious feeding within minutes" (1).  AgRP neurons were already known to be an important neuron population for hunger and feeding, but this experiment demonstrated it in the most convincing manner to date.

DREADD channels are used to activate or inhibit neuron activity using injections of the synthetic ligand CNO.  This is a nice technique because it's non-invasive (as opposed to optogenetics).  DREADD technology has bolstered the finding that activating AgRP neurons drives feeding and rapid fat gain (2).

There are a number of other interesting emerging technologies, but those two stand out.

Palmiter Lab

Richard Palmiter (University of Washington) gave the opening keynote address, describing his research on a feeding circuit from AgRP neurons in the arcuate nucleus to the parabrachial nucleus and then the central nucleus of the amygdala.  This circuit seems to shut off food intake in response to aversive stimuli (sickness, pain, fear), but may not play much of a role in feeding under normal circumstances.

Dr. Palmiter also presented the remarkable finding that ablating AgRP neurons reverses the obesity and infertility phenotype of ob/ob (leptin-deficient and genetically obese) mice.

As the inventor of the transgenic mouse, Dr. Palmiter has played a major role in advancing scientific progress in the biological sciences.

Ozcan Lab

Umut Ozcan (Harvard University) presented data from his lab suggesting that leptin resistance in the hypothalamus is caused by endoplasmic reticulum stress (ER stress; a form of cellular stress that can be caused by misfolded proteins).  He has applied a compound (SR-01) in obese mice that reduces ER stress, restores leptin sensitivity and causes remarkable fat loss.  This compound has no effect in genetically obese mice that lack leptin receptors, and it has no effect on lean mice, both of which support his hypothesis that it specifically increases leptin sensitivity in diet-induced obese mice.  He has been presenting these data at meetings for going on two years and they haven't been published yet, which I find puzzling.  But they certainly do look promising.

Lam Lab

Tony Lam (University of Toronto) presented his data on the role of glucagon in blood glucose control via the brain.  Glucagon is an important player in blood glucose control via the liver, but its effects were generally assumed to result from direct actions on the liver until Dr. Lam's findings.  He showed that glucagon also acts in the brain to regulate blood glucose, but in the opposite direction to its actions in the liver (suppresses blood glucose), and that diet-induced obese animals lose the ability to activate this mechanism, potentially contributing to elevated blood glucose in obesity.

Lowell Lab

Brad Lowell gave a remarkable talk.  His lab is focused on understanding how AgRP and POMC neurons are regulated by other neurons.  As stated above, this small population of neurons in the arcuate nucleus is the most potent feeding circuit known in the brain.  POMC neurons are a well-characterized neuron population in the arcuate nucleus that suppress feeding.  The Lowell lab uses a combination of neuron tracing techniques (tracing the brain's wiring diagrams), optogenetics, and electrophysiology (measuring the 'electrical' activity of neurons), to great effect.

I won't go through the entire talk, which was incredibly information-dense, but I will mention two key findings:

1. What neurons activate AgRP neurons and cause hunger and feeding?  PACAP and TRH eurons in the paraventricular nucleus send excitatory projections to AgRP neurons, and stimulating these neurons greatly increases feeding.  The DMH also sends excitatory projections to AgRP neurons, but these are weaker.  
2. The arcuate nucleus mediates a large portion of leptin action on appetite and body weight, yet deleting leptin receptors from arcuate neurons doesn't cause much obesity, implying that they are receiving leptin-responsive signals from upstream neurons.  What are these neurons?  Dr. Lowell showed that leptin receptor expressing neurons in the dorsomedial hypothalamus send dense inhibitory projections to AgRP neurons.  Activating these DMH neurons potently suppresses feeding.

Sternson Lab

Scott Sternson (NIH's prestigious Janelia farm campus) is interested in the downstream targets of AgRP neurons, and he also makes extensive use of optogenetics, DREADDs and related approaches.  Dr. Sternson argued for a paradigm shift in understanding feeding-related motivation.  Activating AgRP neurons increases motivation for food, probably reflecting hunger.  But what is hunger, and how does it shape future food selection behaviors (via reward processes)?  He presented data suggesting that hunger drives food intake behavior and learning through a process involving negative reinforcement.

He showed that activating AgRP neurons in mice (presumably inducing a hunger-like state) is aversive (unpleasant).  Also, experimentally suppressing the activation of AgRP neurons caused by hunger is reinforcing/rewarding (increases the likelihood of seeking situations/flavors associated with the cessation of AgRP neuron activity), basically suggesting that the satiety state is reinforcing because it relieves the unpleasant state of hunger.  So animals learn to prefer situations associated with the relieving of the aversive 'hunger' state, and this is part of how the reward system learns how to guide food-seeking behavior

Zeltser Lab

Lori Zeltser (Columbia University) had a fascinating talk on anorexia that also contained a very interesting tibit on insulin and obesity.  I won't get into all the details of her talk, but she is researching the intersection between genes, calorie restriction, and psychological stress in promoting anorexia.

The Val66Met allele of the BDNF gene is strongly associated with anorexia in at least some human studies.  BDNF is part of the leptin signaling pathway and therefore plays an important role in the control of food intake and body fatness.  BDNF Val66Met mice are relatively normal, but when stressed (by social isolation) or temporarily calorie-restricted, mice will go through spontaneous periods of aphagia (no eating).  I find it remarkable that it's possible to model anorexia in mice, but her data were fairly convincing that this human genetic susceptibility factor increases the risk of anorexia under conditions of psychological or metabolic stress.  We don't know if the same is true in humans, but it seems plausible based on her data.

She presented another interesting tidbit that's relevant to the insulin-obesity idea: mice with a hypothalamus-specific deletion (Nkx2.1) of the insulin receptor have normal body weight, body fatness, and leptin levels.  The implication is that (complete) insulin resistance in the hypothalamus does not lead to leptin resistance or obesity, which bears on one of the more popular versions of the insulin-obesity hypothesis (e.g., the idea that hyperinsulinemia causes hypothalamic insulin resistance, which causes hypothalamic leptin resistance, which leads to obesity).

Cowley Lab

Michael Cowley (Monash University) gave one of the most fascinating talks, partially because I wasn't aware of this line of investigation prior to the conference.  In short, he has shown that elevated leptin is largely responsible for obesity-induced hypertension (high blood pressure) in rodents, and the evidence suggests the same is true for humans, who also tend to develop hypertension with obesity.  He (with Stephanie Simonds) recently wrote an excellent review article in Trends in Neurosciences that covers most of the evidence (3).  The illustrations of the sympathetic (SNS) and parasympathetic (PNS) nervous systems, and the brain regions involved in SNS and PNS outflow are really outstanding in that paper.  I'll be hanging on to it as a reference.

The brain controls the SNS and PNS, which regulates many 'autonomic' (unconscious) functions like heart rate, lipolysis, glucose control, adrenaline release, heat production, digestion, and blood pressure.  Normally, it does so in response to leptin and other peripheral signals.  During the development of obesity, the brain becomes resistant to many of leptin's actions, such as its ability to suppress food intake and reduce body fatness.  However, it doesn't become resistant to leptin's ability to stimulate the SNS, so the SNS goes into overdrive in obese rodents and humans.  Dr. Cowley's data suggest that this is due in large part to increased leptin acting in the dorsomedial hypothalamus.

Insulin also activates the SNS, and obesity is also associated with elevated insulin.  Elevated insulin acting on the brain is also probably a factor that contributes to hypertension through similar pathways.  Interestingly, low-carbohydrate diets often have a dramatic effect to rapidly reduce blood pressure.  Dr. Cowley's results suggest that this is probably due to a reduction of insulin action in the brain.

Tschop Lab

Matthias Tschop (Institute of Diabetes and Obesity Research, Munich) was one of the organizers of the conference.  He runs a mega-lab in Germany and is the director of a research institute.  We have collaborated with him on understanding the role of hypothalamic inflammatory signaling in obesity.  The most interesting part of his talk was his description of obesity drugs he has created by fusing gastrointestinal satiety hormones (including glucagon, GLP-1 and PYY) to one another in a single molecule.  He is basically trying to create a drug that mimics the hormonal effects of gastric bypass, so that people can get the same weight loss and glucose control benefits without surgery.  His results in rodents are impressive, and there are some preliminary data suggesting a similar effect in humans.

Schwartz Lab

Mike Schwartz (my current mentor; University of Washington) gave a great closing keynote address.  He argued for a new framework for thinking about blood glucose regulation.

Researchers have long known that there are apparently two ways that blood glucose is regulated following a meal: 1) insulin increases tissue glucose uptake and reduces glucose production by the liver, and 2) another unknown mechanism of glucose disposal that does not rely on insulin.  The latter is referred to as 'glucose effectiveness' and accounts for ~50% of the body's ability to regulate blood glucose following a meal!  Originally, glucose effectiveness was thought to be a passive, unregulated process.

One of the most interesting recent discoveries in diabetes research is that under certain conditions, the brain can completely regulate blood glucose without any help from insulin.  This has been shown multiple times in type 1 diabetic rats that completely lack insulin.  Infusing leptin into the brain of these animals almost completely normalizes their blood glucose control, both by reducing liver glucose production and increasing tissue glucose uptake.  My colleague Greg Morton has done a lot of the work on this.  This suggests that the brain can have a major influence on blood glucose control, and that this system is probably partially redundant with the insulin system of blood glucose regulation (which is obviously very important under normal circumstances).

Dr. Schwartz presented data suggesting that glucose effectiveness may in fact be a regulated (not passive) process that results at least in part from gastrointestinal signals acting on the brain, and that it is impaired in obesity.  This has huge implications for diabetes treatment, because if we can figure out how this process works, it may pave the way for diabetes treatments that don't rely on insulin at all!  Most current drug treatments rely on insulin, insulin sensitizers, and/or insulin secretogogues.  Insulin therapy can have serious unwanted side effects such as hypoglycemia.

My Poster

I presented a poster on our finding that when rats are fed a fattening diet, they experience a rapid increase of activated caspase-3 in astrocytes of the arcuate nucleus and other brain regions (within 3 days).  Although rats gained fat rapidly on this diet, fat gain did not appear to explain the effect.  Caspase-3 is normally associated with a type of cell death called apoptosis, however in this case there was no apoptosis to be found.  We think that caspase-3 is part of the rapid cellular injury response to this diet that we reported previously (4).  This suggests a novel non-apoptotic role for caspase-3 in the brain's response to injury, and suggests that astrocytes may be trying to contain damage to important feeding centers early during exposure to a fattening diet.  The poster was well received.  These data were accepted for publication in the journal Brain Research two weeks ago.

Conclusion

The conference was excellent and left my mind extremely full.  There were many informative talks and posters that I left out of this summary.  The pace of scientific discovery right now is mind-boggling; it would be a full-time job just to keep up with current findings in my field alone.  I'm confident that this new research will pay dividends in understanding and improving human health, and particularly preventing and treating obesity and diabetes.  In the meantime, we already know how to prevent (and to a lesser extent, treat) most obesity and diabetes using diet and lifestyle.