In this episode of Sleep Review Conversations, Thomas S. Kilduff, PhD, who received this year’s Distinguished Scientist Award from the Sleep Research Society, talks about a career spent investigating sleep and the brain. He is particularly well known for his contributions to the field of narcolepsy research.
Hosted and produced by Rose Rimler, associate editor of Sleep Review
Run time: 40:19
Rose Rimler, Sleep Review associate editor (RR): Hello and welcome to Sleep Review Conversations, a podcast from Healthcare Media Company Allied 360. This episode is an expanded interview with Thomas S. Kilduff PhD, who is the director of the Center for Neuroscience at SRI International in Menlo Park, California. He received the Distinguished Scientist Award from the Sleep Research Society for original and sustained scientific contributions made over a career. I included an excerpt of my conversation with Dr. Kilduff in the Sleep 2017 Preview Podcast, which is also available on our site.
This is the interview in its entirety for those who’d like to hear more about Dr. Kilduff’s career.
RR: I was looking at some of your work. You’ve had a really interesting career. You started off researching hibernation in ground squirrels. Is that right?
Kilduff: That’s right.
RR: How did you transition to researching sleep in humans?
Kilduff: Well that’s not … Believe it or not it’s not that far a change. When animals enter hibernation in the fall … So, first of all, one of the misconceptions about hibernation is animals go down, lower their body temperature in the fall, and they don’t wake up again until the spring, but, in fact, that six to seven month period defines the hibernation season. And during that period of time animals will repeatedly lower their body temperature and then warm up again for about a 24 hour or less period and then reenter. So, over the course of the hibernation season, they may go through 20 of these cycles or so. And the lab in which I did my graduate work showed that, as animals are entering hibernation, they are selectively doing it during … by increasing the amount of slow wave sleep that they have and reducing the amount of REM, or rapid eye movement, sleep. So, that was among the lines of evidence that suggested hibernation is a regulated phenomenon and it may be, in essence, an extension of slow wave sleep.
RR: Is that true for all mammals that hibernate?
Kilduff: Well, I don’t think we’ve examined all mammals that hibernate to this point in terms of those type of electrophysiological studies. The ones that have been most tractable to study have been ground squirrels which are bit larger than a laboratory rat, which is the standard species for EEG, for sleep studies. So, it was easy to kind of transfer that technology from a rat to a squirrel. And then subsequently over the last, certainly somewhere in the last five to ten years, people have done a study on the black bear in Alaska. And those animals, of course, they are much larger than a ground squirrel. So, just to give some perspective, a ground squirrel in the fall – at least the species that I’ve studied and am most familiar with – may weigh 400 grams when they enter hibernation and they fast all winter, relying on the stored body fat. And when they wake up in the spring, they’re down to about 180 grams.
So, they go through a very distinct annual cycle of fattening and then fasting, resulting in massive weight loss and cycles of weight gain and weight loss. But because they’re relatively small, they can lose heat very efficiently compared to a bear, and so bears – being much larger in size – they’re not able to lose as much heat even in the winter when they may be subjected to temperatures, you know in Alaska, of -40. I think, to my knowledge, the lowest temperature that a bear has been recorded in terms of their body temperature is 29 degrees Celsius which – I don’t know, I don’t have the conversion off the top of my head – but it’s approximately 78 degrees Fahrenheit. So, still very warm. But they clearly undergo a period of hypometabolism where they’re lowering their metabolism and lowering their body temperature and those two, of course, are closely associated. And what happens in the bear is also that they do seem to show much more of the slow wave activity in the brain that is typically associated with restoration -restorative sleep in humans.
RR: So, does any of that information about how hibernation works in some mammal species, does that help you understand the nature of human sleep, or what the purpose of sleep is? Is that a connection that can be made?
Kilduff: So, people have been looking for that connection I must admit because as I mentioned when the animals warm up to … for this 24 hour period or so in the middle of the hibernation season, it doesn’t really make sense energetically why you would want to warm up. Because if the purpose of hibernation is to preserve your energy, why would you rewarm? Because that’s a very energetically expensive process. The way the animals do that is the squirrels, when their body temperature is two degrees, initially they’ll burn brown fat, which is this very specialized organ that’s found in small mammals and infant humans, babies, right? And then after, if they –let’s say raise their body temperature from two degrees up to say, 15 — at that point, they start shivering and they shiver very violently and it is a very energetically expensive process as they’re trying to raise their temperature by shivering.
So, it doesn’t really make much sense why animals were doing that, but what people found — now I’m going back 25 years ago or so — is that when the animals arouse from hibernation, they would have this big burst of slow wave activity and slow wave sleep. So, for a while, the hypothesis was that the reason the animals were waking up from hibernation is to sleep. That there must be some process that is associated with slow wave sleep that has to occur even when your body temperature is down at two degrees. But then subsequent studies found that if you deprive animals of sleep during that first six hours, for example, after they’ve awakened from hibernation, they don’t increase their slow wave sleep higher than you would otherwise expect. So, it still remains a mystery why the animals are waking up. But, nonetheless, what you see in a hibernating organism like this is you see the same in the non-hibernating season. You see the same types of sleep cycles that exist in humans and other mammals that have slow wave sleep and REM sleep, and that those are recurring with a particular cyclicity.
But in the fall, in association with hibernation, there seems to be a shut off, if you will, of this REM sleep process and whatever it is that underlies the drive for REM sleep to occur in our brain. So, for that reason, that’s one of the many mysteries that we’re interested in studying. A relationship between sleep and hibernation.
RR: So, it seems like you could have, early in your career, you could have chosen to become a mammalogist and to continue studying squirrels and maybe go on to study black bears and look at this phenomenon of hibernation. Did you always know that you wanted to put your research to effect in humans? Or is that something that kind of happened serendipitously when you began working in humans and looking at human sleep?
Kilduff: Well, I think all biologists working in whatever aspect of biology you’re working on –whether it’s at a cellular level, at a molecular level, or at an organismal level such as studying a phenomenon such as hibernation – we’re all of course interested in how what we’re studying could be translated in some way to understanding the human condition, and perhaps improving sleep and sleep disorders, in my case. So that’s one factor. Another factor is, quite frankly, the market, if you will. Market forces. It’s much— the sources of research funding for sleep are much broader than the sources of research funding for hibernation.
RR: The discovery of hypocretin or orexin, that’s probably your most widely known contribution to the field so far. Was it also the most exciting for you personally? Did you know that you were on to something really interesting when you did that research?
Kilduff: Well, not initially, but as time went on that was certainly the case. I’ve got to acknowledge at this point some of the … I was very fortunate to stumble into this project, I have to say. When I was on sabbatical at the Scripps Research Institute in the laboratory of Greg Sutcliffe — I was a researcher at Stanford University at the time — but you may know that there’s this process , if you have put in a period of time in academics, you’re allowed to take a sabbatical. So, I took a sabbatical year and I went down to Sutcliffe’s lab to learn a particular technique, and this technique was being applied even before I got to Sutcliffe’s lab to isolate new messenger RNAs that were expressed in the hypothalamus. So, I stepped into this project after some of the very difficult work had been done at the early stages, and they had a number of what we called “anonymous clones” to identify. And there was this one clone that had a very specific localization to a region of the hypothalamus that had been shown in the rat brain in the rat hypothalamus.
So, basically, I worked on that for a year with other folks in Greg Sutcliffe’s lab. And we certainly knew, based on the localization, that whatever this was going to be — that turned out be hypocretin — that it was going to be special. But we were uncertain what the function was. We thought because of its localization in the hypothalamus, it probably had more to do with feeding than anything else because it was found in an area of the hypothalamus that anyone who had taken Psych 101 courses knew that, when you lesion that area of the hypothalamus, rats in particular would become obese. But once we found out more about the structure of this messenger RNA and figured out what peptides, neuropeptides, it could encode, we then injected those neuropeptides into the brain of rats and never really saw an increase in feeding.
But shortly after we published this work, another group led by Masashi Yanagisawa at University of Texas Southwestern Medical Center described the exact same peptides and they, on the other hand, found that, when they injected their peptides in the brain, they saw an increase in feeding. The only difference was we were doing our experiments at a different time of day than they did. We did our injections just before the lights went out. So, that’s the animal’s normal active period, whereas Yanagisawa’s lab– really, at that time didn’t have a background in terms of sleep or circadian biology– so they would do their injections at the beginning of the researcher’s work day, which for a rat is the major sleep period.
RR: So, it would keep them up and they would be eating at strange times of day?
Kilduff: Exactly. So, the peptides we now know as the hypocretins (or the orexins) have this profound wake-promoting effect in rodents. And that’s most evident, of course, when you give them, the animals, the peptide early in the workday or the animal’s inactive period. So it was, in retrospect, it’s not surprising that they saw these animals … That they thought the peptide had an effect for increasing food intake because if you compare the amount of food that those animals, who were injected with the peptide, ate versus the control animals who were injected with a vehicle and went back to sleep, it’s obvious that the increased food intake is a secondary effect due to the fact the animals were just awake longer.
RR: What did you think when you heard about their research?
Kilduff: Well, we were incredulous that another group that was working completely in parallel (neither of our groups knew anyone from each other) and that we would both arrive at discovering the same peptides. I neglected to point out these two papers were published within a matter of six weeks of one another. And it wasn’t really apparent immediately that these were the same peptides. In fact, they published their paper in a journal called Cell, and there was a commentary associated with their paper pointing out what an exciting time this was for researchers who worked on the hypothalamus because, in the last six weeks, four new peptides had been discovered in the hypothalamus. And, in fact, there were only two new peptides that had been discovered, but they were discovered independently by two groups but they were the same peptides.
So, I guess the next major development in this story was about 18 months later when the group at UT Southwestern went on to make the knockout mouse. That is, they eliminated the orexin gene (or hypocretin gene) and found those mice had what we call a phenotype that looks very similar to narcolepsy, that is, the animals would scurry around their cage and all of a sudden they would fall over and be inert, essentially asleep, for a period close to a minute, maybe a little bit more, and then get up and resume their activities. And they would do this repeatedly over the course of the dark period. And so, just prior to that, my colleague at Stanford, Emmanuel Mignot, who had been working on narcoleptic dogs for many years before that … It had been shown that narcoleptic dogs, particularly in Doberman pinschers and Labrador retrievers, that those animals have what looks like narcolepsy because they certainly have this profound cataplexy which is essentially a loss of muscle tone and resulting … You’ll be playing with animal, and all of a sudden, they’ll fall onto the floor and they’ll go to what looks like sleep.
It turns out that cataplexy is a different state. It’s not quite wakefulness, it’s not quite sleep. I’ll explain that in a moment. But Emmanuel was trying for many years then to find out what the gene defect was in the dog because it had been known that this gene defect was a single gene. But at that time there were no… no genomes had been sequenced, and it was a very difficult process to hone in on what the gene defect was in a species like a dog which, of course, breeds relatively slowly compared to a mouse, for example. And it turned out that Emmanuel’s eventually … His group eventually found the defect, and that it was in the receptor for hypocretin … the hypocretin peptides. There’re two peptides, hypocretin-1, hypocretin-2, or orexin-A, or orexin-B, depending on which nomenclature you use. And it turns out there are two receptors for those peptides as well. And in the dogs, they had a mutation in the hypocretin receptor-2.
So, what that meant was when the neuropeptide hypocretin was being released by a pre-synaptic cell, even if that peptide reached the receptor and bound to the post-synaptic cell which is expressing that receptor, it was a non-functional receptor. So, it couldn’t transduce that information to the next cell. So, this, of course, is a very highly technical discussion but, between the mouse studies and the dog studies, the very exciting thing was that we knew right away that this system must have something to do with narcolepsy in humans because, in the mouse, you had a defect on the presynaptic side because the way the knockout had been made the animals weren’t making the peptide. In the case of the dogs, they were making the peptide, but the receptor was non-functional. So, whether you interrupted neurotransmission either pre-synaptically in the case of the mouse, or post-synaptically in the case of the dog, you end up with what looked like the same phenotype, that is, the cataplexy that’s characteristic of human narcolepsy.
RR: And have you found that to be the case in humans? Has that been the same mechanism that shows to be maybe the culprit in human narcolepsy?
Kilduff: So, it turns out that what human narcoleptics have is indeed a defect of the hypocretin or orexin system, but they have yet a different problem than either the mouse or the dog because, in humans, the hypocretin or orexin cells degenerate for some reason we don’t fully understand yet. So, in the case of the mouse, the gene had been knocked out so the cells that are normally there in the hypothalamus and are making hypocretin and other things are still present, but it’s only that one hypocretin gene that’s missing. But, in humans, the entire cell is dying in human narcoleptics. And we don’t understand why these cells are dying, but we think that there’s a link with the immune system. And it’s been speculated that there may be an autoimmune attack that is specifically targeting the hypocretin cells for some reason.
RR: That seems like a very specific type of autoimmune attack. Are there narcoleptics … Are they frequently … Do they show co-morbidities with other autoimmune diseases like lupus, or MS, or eczema or anything like that?
Kilduff: Right. So, first of all, let me mention something about the prevalence of narcolepsy in humans. We understand — at least in North America I think there’s the best data — the frequency is about 1 in 2,000 individuals. It is comorbid with some other diseases and, in all the ones that you mentioned, there are at least some cases of some narcoleptics who have both multiple sclerosis and narcolepsy, or may have depression and narcolepsy– well, it’s a complicated issue. But there’s no one disease association that exists in every human narcoleptic. But from a genetic point of view, there are a number of markers in the immune system, in particular the HLA system, as it’s called. So, there has been identification of probably the tightest gene-disease association, or one of the tightest gene-disease associations, between the immune system and narcolepsy of any disease that’s known to this point.
So, for that reason, people have been thinking there must be some autoimmune attack on these cells that results in narcolepsy. So, most recently there have been, in the last few years, people doing retrospective studies– that is, a narcoleptic patient will come in or a series of patients come in to the clinic, they get diagnosed and then they retrospectively analyze what other health issues the patient has had over the last year. And there have been a number of examples of strep infections in the year prior to the onset of narcolepsy. It’s thought that perhaps their immune system, under conditions of a strep infection and perhaps other types of infections, starts making an immune response activating particular T-cells and for some reason under … Normally the brain is protected from immune attack, but the blood brain barrier is compromised in some way, those T-cells can get though the blood brain barrier, get into the CNS and why the hypocretin cells, in particular, are so sensitive or attacked in these cases is, as I said, not fully understood.
But the hypothesis would be that whatever the viral or other infectious agent is when the immune system is activated that there may just happen to be some epitope on the surface of the virus that resembles an epitope that is on the hypocretin cells.
RR: It’s like it’s smuggled into the brain.
RR: One thing that’s really interesting about your career is that the discovery of hypocretin, or orexin, eventually down the road that led to some pharmaceutical research and even a drug that’s currently on the market, Belsomra or suvorexant, and that drug has gotten a lot of media attention. And then even the discovery of, or the process of getting that drug FDA approved and all that, that was the subject of a really long piece in the New Yorker. So, I wonder what it’s like as a scientist whose work, or the fruits of that labor down the road have gotten a lot of attention. What has that been like for you? Not every scientist gets … Sees their work resounding over the media or the real world quite as much as you have.
Kilduff: Yes, I would respond to that two ways. First of all, it wasn’t just my work. There were a number of other people who’ve … in multiple different laboratories that have contributed to this story. I’ve already mentioned the Yanagisawa lab, and the Mignot lab. So, there are many cooks in this broth. But to answer your question more specifically, yes I saw the article in the New Yorker and I watched the story evolve, and know people from Merck that have been involved in generating the compounds that ultimately led to Belsomra. And it’s one thing when you’re a researcher and academic and seeing science progress, but the thing that really blew me away was, one day I was walking through the living room while at home and the television was on, and there was a commercial for Belsomra on TV. And that just blew me away. So there’s … You wonder about other paths your career could have taken too because there were opportunities around that time to enter the pharmaceutical industry as opposed to doing basic research, but I chose the path that I felt most comfortable with and was enjoying most at the time.
RR: Well speaking of that, you’re working on a lot of projects now. What are you most excited about?
Kilduff: Well, there’s a couple of new things that, in addition to our hypocretin work that we’re very excited about. One I told you a while ago when we were talking about hibernation, I just happened to mention that slow wave sleep and slow wave activity, in particular, is thought to be associated with restoration. And the reason for that, and by restoration I mean … It’s still a mystery of why we sleep obviously, and what is happening at a biochemical or a molecular level that refreshes us. Right? I mean we all know what that feeling of sleepiness is and how it’s impossible to overcome sometime. But then if you lay down, depending on the time of day, whether it’s a 20 minute nap or a good eight hours sleep, you — hopefully, anyway — awaken refreshed both cognitively and physically, you’re “stronger” than you were earlier before you took that nap or sleep.
So, we don’t really understand what is happening at a cellular or molecular or biochemical level that underlies that. So, closely related to the restorative process is the circadian process, that is, a biological clock in our brains that determines, in large part, when we’re active and when we’re inactive. And we’ve been able to map out from a neural– when I say we, I mean neuroscientists, not my lab in particular, but neuroscientists as a whole– have been able to identify the locus of the biological clock in the mammalian brain. But, in contrast, we have not been able to figure out where in the brain this molecular or cellular level of this restorative process is, how that is manifest. Most of the work is really focused on the hypothalamus. But a few years ago, Dmitry Gerashchenko, when he was working in our group here found a particularly interesting link in the cortex. And that is that we and, of course, the cortex is the cerebral cortex is the locus of cognitive function) and what we found, in particular, was that there was a particular cell type that becomes activated when you are asleep, which is just the opposite of the large majority of cells in the cortex.
Most cells are activated when you’re awake, and become less active when you sleep but these cells were doing the opposite. They were quiescent during wakefulness, and were only active during sleep. And we showed that was true in three different species: mouse, rat and hamster. And subsequently we showed that the longer we kept animals awake when we did short-term sleep deprivation studies, that is, keeping the animal awake for two hours, or four hours, or six hours longer than it normally would, that when the animals finally did go to sleep, the higher the proportion of these cells would be activated. So, what became very interesting is that this marker for these cells is something called a neuronal nitric oxide synthase, nNOS. It’s an enzyme that results in production of nitric oxide, which you may know is a gas. Neurotransmitters … Nerve cells use three ways to communicate– at least three ways that we know of at the moment– to communicate with one another.
One way is they release classical neurotransmitters like, well amino acid neurotransmitters include glutamate, and GABA. Other classical transmitters are biogenic amines like serotonin, norephinephrine, dopamine. So, that’s one way. A second way is that cells use neuropeptides like hypocretin and other neuropeptides to communicate. And those are slower in their communication; they are neuromodulators. But then the third way that neurons communicate with one another is using gases or “gasomitters” as they’re called. And less is understood about those, but nitric oxide was among the first that was discovered. So, in contrast to the fast neurotransmitters, which typically are released at synapses, gases are released in a nonspecific way. And they have the ability to spread much more widely than a neurotransmitter. So, the fact that these cells became active during sleep and made nitric oxide was quite interesting because that meant that they were in a … Oh, there was one other fact I should mention. And that is that these cells also projected widely in the cortex.
So, they have the … Because they project widely in the cortex, they have the ability to influence large regions of the cortex by release of all three of these different types of neurotransmitters I mentioned: the classical type, the peptide because these cells make another peptide called somatostatin and one called Neuropeptide Y, but they also release this gas. So, we think these cells may be critical in understanding something about the restorative process because these cells are in a position to — because of their wiring and because of the neurotransmitters they use — they’re in a position to influence wide regions of the cortex and orchestrate the slow waves that are characteristic of slow wave sleep.
That’s one project we’ve worked on.
RR: That sounds like a big project. Is that something you’re looking at animal models for? Are you at that stage?
Kilduff: Yes. So, we have multiple animal models that we are using to understand that … these cells in more detail. One involves the standard knockout mouse that as you eliminate the gene that produces nNOS, or neuronal nitric oxide synthase, and we’ve found that animals that lack that gene have very abnormal sleep. And then there’s another set of tools that have become available to neuroscientists over the last decade called optogenetics, or chemogenetics (these are two different tools.) And we are using animal models– mouse models–to employ those tools in the nNOS cells and see if we can activate or inhibit those cells, and see what effect it has on the slow wave activity.
RR: So, there are a lot of topics that you could be discussing when you present your lecture at SLEEP. Do you know what you’re going to be focusing on?
Kilduff: Well there is a symposium … So, in addition to the lecture I’m giving there, I have a symposium the prior day on another project we worked on called trace amine-associated receptor-1 or TAAR-1, and this is a receptor that very, very little is known about. But the reason it’s relevant to the rest of our discussion, is that with my colleagues here at SRI International, we found that using drugs that are produced by a pharmaceutical company that activate this receptor may be a new pathway for treatment of narcolepsy. So, we have multiple mouse models now of narcolepsy. Some that involve degeneration of the hypocretin neurons. And, of course, those animals have cataplexy just as the gene knockout mouse that I mentioned earlier have.
And so we, over the years, have tested many different pharmaceutical agents on cataplexy in these mice, and to see which ones would mitigate cataplexy and which ones exacerbate it. This is very similar to work that was done in the narcoleptic dogs years ago by the Mignot group, but we have identified that this new receptor that I mentioned, TAAR-1, seems to hold promise as a new therapeutic pathway for treatment of narcolepsy. So, because I have that symposium the day before my lecture, the program committee has asked me not to talk about TAAR-1, or to de-emphasize when I talk about TAAR-1, and instead I’ll primarily talk about our recent work on hypocretin system and this nNOS story, and a little bit less on TAAR-1.
RR: So, have you been to this meeting in prior years?
Kilduff: Oh yeah. So, this is … I joined the Sleep Research Society when I was a student in the 1980’s, and it’ll be like … in addition to the thrill of going to a scientific meeting and learning new information, I’ll also be seeing a lot of old friends. So, it’s always a good meeting to go to.
RR: Well great. Is there anything else that you wanted to add, either about your work or about the meeting in Boston, or anything that you’d like the audience to know?
Kilduff: Well, I think I should say that I’m very fortunate over the years to have good funding from the National Institute of Health which, of course, we’re all grateful to the American taxpayers who are supporting the National Institute of Health which, as you may know from the early days of the Trump administration, the proposal was to reduce the funding for NIH by 20%. But, fortunately, Congress in its wisdom saw to increase the NIH budget and the budget was passed last year, er, last week and I’m certainly hopeful, as are, I’m sure, many other scientists, that that continues in the negotiation, that an increase in fondness for NIH that– demonstrated by Congress, will continue into the negotiations for FY-18 which are beginning shortly.
RR: Well great. Well thank you so much and congratulations on your award.
Kilduff: Thank you.
RR: Thanks for listening to Sleep Review Conversations. Go to sleepreviewmag.com and click on resources to find this and other podcasts and their transcripts.