Genetics of Human Circadian Rhythms and Sleep

Part 1: Connections between Clock and other Phenotypes

00:00:07.05 Hi, I'm Louis Ptacek, a neurologist and scientist at the Department of Neurology
00:00:13.07 at the University of California, San Francisco.
00:00:16.04 And I'll be talking about genetics of sleep and circadian behavior.
00:00:21.08 Dr. Fu and I have spent a large part of our careers studying human genetic diseases where
00:00:28.13 mutations in single genes can cause fragile X, myotonic dystrophy, epilepsy, muscle disease, etc.
00:00:36.23 And through study of these families, it's been possible to map the genes that cause
00:00:41.06 the phenotype, and thereby study the proteins in the laboratory in test tubes and physiological
00:00:49.04 experiments to understand normal function, normal biology, that's occurring in the human organism,
00:00:55.22 as well as pathophysiological processes that contribute to disease.
00:01:00.20 A change in direction for our research occurred when we decided to embark on studying
00:01:06.10 the genetics of human behavior.
00:01:09.19 Behavioral phenotypes are very challenging to characterize at a clinical level.
00:01:15.07 And this has made finding genes for behaviors very difficult.
00:01:18.26 But we felt that it was possible because of families that we identified that have
00:01:24.12 circadian sleep behaviors that are genetic, and passed from generation to generation.
00:01:31.24 Sleep behavior is very complicated.
00:01:35.05 There is a biological tendency for us to go to sleep and wake up, but to emphasize
00:01:40.21 the challenges of studying human behavior, we often override our biological tendencies,
00:01:48.05 because we have to stay up to meet a deadline, to study for an exam, or to spend time with
00:01:55.07 our families, even though we might be very tired and want to go to sleep.
00:01:59.10 We drink compounds like alcohol and coffee, which affect our sleep.
00:02:03.21 Sometimes we take medicines that affect our sleep.
00:02:06.14 And so there are lots of different things that we do that are not listening to our biological
00:02:12.07 tendency to sleep and wake, but rather overriding our biological systems.
00:02:17.11 But, again, the goal here is to identify genes and mutations that cause genetic behaviors
00:02:24.02 in humans, with the long term goal of better understanding biology and, in some cases, diseases.
00:02:31.10 Now, we've evolved on this beautiful planet, the third planet from the Sun.
00:02:38.15 And as a result of this, and the fact that this planet turns on its axis every 24 hours,
00:02:44.20 with corresponding changes in light and dark and also temperature, everything about us
00:02:50.14 has evolved so that we can function optimally in an environment where the light and dark cycles
00:02:56.02 are changing every 24 hours, along with the temperature cycles.
00:03:01.20 And it's not surprising that, through evolution, all living organisms on this planet have also
00:03:08.00 evolved in such a way that they can function maximally, or function optimally,
00:03:13.11 in this changing environment.
00:03:17.00 We know, for example, that all kinds of different things change and oscillate in our own bodies
00:03:23.26 and the bodies of other animals, for example.
00:03:27.28 We have a... our deepest sleep in the wee hours of the morning.
00:03:32.17 Our body temperature is going up and down and reaching it... reaches its low point
00:03:37.21 at about 4:30 in the morning, on average, that is, for humans.
00:03:42.27 And... and our pain thresholds are oscillating, our immune system function is oscillating,
00:03:48.01 etc, etc.
00:03:51.13 Now, the system that we have evolved we call the circadian system.
00:03:57.28 This is from the word circa dies, Latin words meaning "about a day" or "around a day".
00:04:06.27 So, that means that this refers to these oscillating systems that are cycling every 24 hours.
00:04:15.03 In the center of the... the circadian system for all living things is a clock which is
00:04:20.07 ticking inside the brain, in the case of humans.
00:04:24.11 And this is the rhythm generator that is oscillating on a 24-hour period.
00:04:31.02 However, because of changing day length and also changes in time zones, for example,
00:04:38.17 when we travel, we need a mechanism whereby we can reset the clock.
00:04:43.21 And that is called entrainment.
00:04:46.15 Entrainment is largely done through light coming in through the eyes, although
00:04:53.04 other things can entrain the clock, including changes in temperature, exercise, and also feeding.
00:04:59.26 Then, the things that the... the clock, while oscillating in our brain, is then telling
00:05:06.13 certain things in our body to oscillate.
00:05:10.11 And so these are called pacemaker outputs.
00:05:13.04 That is, our clock, then, is regulating our sleep/wake cycle.
00:05:17.25 Our clock is regulating our immune system function.
00:05:20.23 Our clock is regulating our body temperature rhythms.
00:05:24.04 And... and so forth.
00:05:25.24 And then, in the periphery, in our body, the rest of our body, we also have oscillators.
00:05:32.11 Individual cells that we could biopsy from... from your skin, for example, can oscillate
00:05:37.23 when we grow them in a culture system in the laboratory.
00:05:42.00 And using different paradigms, with altered feeding schedules, for example, we can dissociate
00:05:48.03 the rhythm of enzymes in the liver from the wake and sleep rhythms that are coming
00:05:54.25 out of the brain.
00:05:56.08 So, taken together, this... we call this the mammalian circadian system,
00:06:01.02 the circadian timing system that regulates all elements of our biology and metabolism.
00:06:08.26 Anatomically, we know that... shown here in this part of this section of a brain that
00:06:16.10 the nucleus right here, the suprachiasmatic nucleus, which contains about 10,000 neurons
00:06:22.05 on each side of the brain... that this is the master regulator of the clock.
00:06:27.12 This is the... the part of the brain that's regulating and coordinating all of the different
00:06:32.08 things that are going on, metabolically, in our bodies to help us survive and prosper.
00:06:38.18 Now, one way that we can look at circadian rhythms in mice is to put them into a cage
00:06:44.25 with a running wheel and record the activity of running on the wheel.
00:06:48.26 And when we do that, as you can see here, under conditions of light and dark, you can
00:06:53.22 see that this activity, shown by these black marks, begins at approximately
00:06:59.09 the same time every day.
00:07:00.26 And that the... the line, then, is a vertical line.
00:07:05.16 And that represents a circadian period of approximately 24 hours.
00:07:10.07 And when the lights are going on and off every 12 hours, these mice will live on a 24-hour day.
00:07:16.15 However, if we turn the lights off, as we do at... at this point, and now continue to record,
00:07:22.14 you can see that the mice begin getting on the wheel a little bit earlier each day.
00:07:28.23 And that signifies the fact that mice have a circadian period length that's slightly
00:07:35.07 less than 24 hours.
00:07:37.19 Shown here for an individual mouse is a... a period length of 23.7 hours.
00:07:43.11 And that's average for a mouse of this type, a C57 black 6 mouse.
00:07:49.08 Now, I've already mentioned, but I'd like to emphasize that the clock in our brain is
00:07:55.21 regulating not only our sleep/wake cycle, but it's affecting our learning and memory.
00:08:01.14 It's affecting our GI function, our gastrointestinal tract.
00:08:05.09 Our heart and lung function.
00:08:07.02 Immune function.
00:08:08.14 Cell cycle regulation.
00:08:10.05 And... and so the clock is tied very intimately to all elements of our normal biology.
00:08:17.10 And therefore, when it's not working properly it can also contribute to diseases of all
00:08:23.18 different systems in our body, as I'll mention later on.
00:08:29.03 I'd like now to also emphasize the difference between the circadian clock, which I've described here,
00:08:36.27 and sleep, which is a separate process, but it's intimately related to the circadian clock.
00:08:43.18 The clock is something ticking in our brain that... even if we turn off all the lights
00:08:48.07 and have no time cues for the day, that will relentlessly oscillate day to day.
00:08:55.04 And keep our body temperature oscillating, our melatonin rhythms oscillating, and our
00:09:00.02 sleep/wake cycles oscillating, for example.
00:09:02.16 Sleep, however, is somewhat different.
00:09:06.00 We sleep approximately eight hours each day, but if we go for a period of time
00:09:11.26 without sleeping -- that is, sleep deprivation -- we can be very tired and fall asleep in the earl...
00:09:17.18 in the morning, when ordinarily we would... we would be waking up.
00:09:21.13 So, sleep deprivation can lead to overriding the biological tendency to wake, because we've
00:09:29.02 accumulated something that we... we don't understand at all but something that we
00:09:33.11 call sleepiness.
00:09:35.18 And so Borbely many years ago described this two process model, whereby we get the circadian time
00:09:44.24 oscillating like this, in a sinusoidal manner, making us sleepy at night,
00:09:55.20 waking in the morning, sleepy at night, waking in the morning, and then, on top of that,
00:09:59.19 there's something that we call process S. And this is something that accumulates as a function
00:10:04.15 of how long we've been awake.
00:10:06.21 And so, as we've been up longer and longer, we become more and more sleepy.
00:10:10.18 And then, as we sleep, this process S, or this sleepiness -- whatever that really is,
00:10:15.16 we don't understand -- dissipates as a function of the time that we are sleeping.
00:10:21.06 So, shown on the right side of this slide is increased sleep pressure, as the...
00:10:27.23 is the terminology we use to describe a sleepiness that comes from extended sleep deprivation,
00:10:35.01 despite the fact that the clock continues to... to oscillate relentlessly inside our brains.
00:10:41.15 And so this person could be sleepy this whole time, even though the clock is now saying
00:10:47.01 "be awake", for example.
00:10:49.09 So, for a long time, we've recognized that there are differences in the tendencies for
00:10:55.18 people to sleep and wake in the general population.
00:10:59.17 And for today's purposes, I'm going to refer to the "conventional" sleepers as people who
00:11:04.09 have their sleep onset -- they fall asleep -- around 11 or 12 at night
00:11:10.05 -- I think that's pretty average -- and who wake up perhaps at 7 or 8 in the morning.
00:11:16.02 We recognize that there are people in the general population who we refer to as "night owls"
00:11:21.06 or delayed sleep phase, because they can't fall asleep until much later at night.
00:11:27.26 And consequently, they want to stay asleep, they want to stay in bed, until much later
00:11:33.05 in the morning.
00:11:34.19 These people get in trouble sometimes, because they can't fall asleep early and yet
00:11:39.08 they often have to wake up at a proscribed time because of their job, because of school,
00:11:45.13 and other responsibilities.
00:11:46.20 Now, people who are truly retinally blind can actually drift in time.
00:11:52.24 And this is the result of the loss of any light cues into the clock that we...
00:12:00.01 that I mentioned earlier, that can help us to reset the clock each day.
00:12:04.26 And then, I'm going to focus most of... or, much of the attention today talking about
00:12:09.06 these "morning larks".
00:12:10.11 And these are the individuals who spontaneously wake up very early in the morning, and then
00:12:16.24 tend to go to sleep much earlier.
00:12:18.15 So, they sleep the same quantity, the same amount of time, as a conventional sleeper,
00:12:24.07 but their sleep bout is shifted forward.
00:12:26.18 So, we call this advanced sleep phase, in the same way that we would call
00:12:31.27 the night owl's delayed sleep phase.
00:12:34.17 The sleep bout is advanced, in the case of ASP, or delayed in the case of the night owls,
00:12:40.26 or DSP, relative to the solar day.
00:12:44.15 In the suprachiasmatic nucleus, the part of the brain that regulates... is the
00:12:48.01 master regulator of the clock, when we disrupt that surgically in... in rodents or in non-human
00:12:54.22 primates, we know that that leads to arrhythmia, sleeping on and off through different
00:13:00.15 times of the day without a consolidated bout, typically during the dark phase of the day.
00:13:07.26 Now, we also recognize that there are individuals who are... who can stay up to an... the time
00:13:15.07 that a conventional sleeper would stay up, but who wake very early in the morning.
00:13:20.10 And this is because there are some people who require a lot less sleep than the average person.
00:13:26.14 And we call these natural short sleepers.
00:13:29.08 There are also natural long sleepers, who are not lazy people, but rather they're people
00:13:34.04 who absolutely require more sleep than average to perform at a normal high level.
00:13:41.02 Now, a real breakthrough for us in this area was when we identified families that have
00:13:49.03 a genetic form of morning larkism, or advanced sleep phase.
00:13:53.22 We call it familial advanced sleep phase, FASP.
00:13:58.04 Delayed sleep phase are the night owls, or familial delayed sleep phase.
00:14:02.04 And we also have families who are familial natural short sleepers.
00:14:08.04 All of these families have a mutation in a gene that's traveling from generation to generation,
00:14:13.16 and leading them to be morning larks, night owls, or short sleepers.
00:14:18.15 So, our work in this area first began when a 69-year-old woman came to the attention
00:14:25.24 of our colleague, Chris Jones, a clinical sleep neurologist, because she had a lifelong
00:14:30.20 history of waking up extremely early in the morning and going to sleep very early at night.
00:14:36.26 When she was a young woman, she would wake up spontaneously at 3 or 4 in the morning
00:14:41.19 and would be very tired around dinner time, and struggled to stay up until 8, 9, 10 o'clock.
00:14:48.15 As a 69-year-old woman, she was already sleeping, often, at dinner time and waking up spontaneously
00:14:55.22 at midnight or 1 in the morning.
00:14:58.11 She had had this trait forever and she didn't like the trait.
00:15:02.11 But she came to see our colleague, Chris Jones, because she noticed the same trait not only
00:15:07.09 in her daughter but in her granddaughters, and wondered if there was anything that
00:15:10.27 could be done for them.
00:15:13.28 Study of this family and this woman, who's signified by the arrow, here, led to uncovering
00:15:21.15 of this very large pedigree.
00:15:23.24 As you'll see, the... there are many circles and squares.
00:15:26.08 The circles represent women.
00:15:27.27 The squares represent men.
00:15:30.00 The ones that are shaded in are men and women who have this trait, familial advanced sleep phase,
00:15:36.09 early morning wakening.
00:15:40.01 And there are some people who we are... who are clearly unaffected, and then there are
00:15:44.11 others who we were uncertain of, and we called these undetermined, as... as shown below.
00:15:50.12 So, as you can see, this is a very strong trait, segregating or traveling through many
00:15:55.05 generations of this very large Utah pedigree, where approximately 50% of children of
00:16:02.03 affected individuals are affected themselves by this morning lark trait.
00:16:08.03 We began by collecting a lot of clinical data, first of all, sleep logs.
00:16:13.01 We asked these individuals to keep track of their sleep times.
00:16:16.10 The Horne-Ostberg Questionnaire is a questionnaire that allows us to assess the tendency of someone
00:16:24.01 to be a morning lark versus a night owl.
00:16:27.23 We had all of these individuals...
00:16:29.27 I... interviewed by a psychologist with expertise in sleep.
00:16:34.10 And we also performed a Beck Depression Inventory, since major depression can also affect our sleep,
00:16:40.22 and we wanted to make sure that we were not misdiagnosing people because
00:16:46.10 they had some other problem, like depression, which can lead to very early morning awakening.
00:16:52.05 And then all of these individuals were evaluated by myself or, mainly, by Chris Jones,
00:16:59.04 our colleague at the University of Utah.
00:17:02.11 So, what we found is that to be affected we... we... we set up this classification criteria,
00:17:09.18 where we require that people can fall asleep spontaneously at 8:30 or before.
00:17:16.18 Now, frequently, these individuals will struggle to stay up later than that, but they can easily
00:17:22.10 fall asleep by 8:30 in the evening, which is the time of day, dinnertime through 8 or 8:30,
00:17:28.15 when we're least likely to sleep.
00:17:31.06 We required that they only had one major sleep episode per day.
00:17:36.05 That is, they didn't have some funny sleep pattern.
00:17:39.00 We required that it began before the age of 40 years, because as we get older it is normal
00:17:45.11 for many people to become more early bird in their phenotype.
00:17:50.26 And we also required that this was not a phenotype that was supported or brought out by taking
00:17:57.23 a lot of ligh... having a lot of bright light in the morning or drinking a lot of coffee
00:18:02.02 early in the morning.
00:18:03.02 This is a phenotype that's being driven by the... the innate biological tendency of these individuals.
00:18:10.18 And so, then we brought a subset of the affected and unaffected individuals from this large
00:18:15.15 family into the clinical research center, where we studied them on a weekend where
00:18:20.11 we could control the lights and the telephone and everything else.
00:18:24.22 And we allowed them to go to sleep when... and... and asked them to sleep when they
00:18:28.12 really felt like it.
00:18:30.02 And to wake up, obviously, when they felt like it in the morning.
00:18:34.13 We required that the lights be dim after 4pm.
00:18:38.14 We measured melatonin, core body temperature, actigraphy... we measured activity as a surrogate
00:18:45.07 for wakefulness, with inactivity being a surrogate for sleep.
00:18:50.02 But we also performed polysomnography.
00:18:52.17 This is EEG tests of the brain, measuring brainwaves, where we can tell when a person
00:18:59.12 actually falls asleep and when they technically wake up.
00:19:02.08 And then we did a test called a multiple sleep latency test, which is a test of how well
00:19:07.09 rested these individuals are after they've woken up, regardless of whether they're morning...
00:19:14.10 early morning risers or more conventional sleepers.
00:19:18.05 And here's what we found.
00:19:20.13 We measured total sleep time.
00:19:22.26 We measured the pattern of the sleep as assessed by the brainwaves.
00:19:27.19 We measured sleepiness, the MSLT, or multiple sleep latent latency test, for the next day
00:19:34.04 following sleep in the clinical research center.
00:19:38.12 What we found was that individuals, regardless of whether they were affected or unaffected,
00:19:45.10 had essentially normal sleep quality.
00:19:48.21 And they had equal amounts of sleep quantity, or length of total sleep.
00:19:54.03 So, the actual sleep itself was not changed.
00:19:57.09 What was changed, as I show in the next slide, was the onset of their sleep and the offset
00:20:04.12 of their sleep, when they awoke early in the morning.
00:20:08.00 The affected individuals, shown in the second column here, were going to sleep at approximately 7:25,
00:20:15.04 waking at 4 o'clock in the morning, and their sleep architecture is all shown here.
00:20:22.01 What we showed was that the onset of sleep in the evening was much earlier for individuals
00:20:27.23 who are affected by FASP.
00:20:30.02 And that they woke spontaneously much earlier in the morning, despite having completely
00:20:34.19 normal sleep quality and quantity when compared to controls.
00:20:40.25 And in one individual, the first 69-year-old woman who I mentioned, we measured her sleep
00:20:48.11 and her temperature rhythms, melatonin rhythms, while she was living in an apartment with
00:20:54.24 no cues to the time of day -- this is a modern day... one of the modern day cave experiments
00:21:00.02 that you may have heard about -- and show here the temperature rhythm, which is shifting
00:21:07.04 a little earlier each day.
00:21:08.27 That is to say, her period is a little less than 24 hours.
00:21:12.22 It's about 23.3 hours.
00:21:14.23 Now, normal humans, conventional sleeping humans, have a period of slightly longer than
00:21:20.26 24 hours, 24.3 hours.
00:21:24.02 And so this woman has a period that's a full hour shorter than normal conventional sleepers.
00:21:31.25 So, in summary, by studying this family, we showed that sleep quality and quantity were normal;
00:21:38.27 that the sleep wake times, melatonin rhythms, temperature rhythms, and
00:21:42.26 everything else that we measured were all advanced earlier, by 4 to 6 hours;
00:21:47.28 and, finally, that in this one subject, that her circadian period was
00:21:54.23 a full hour shorter than an age- and gender-matched control.
00:21:59.15 And this was reported in 1999 and, again, I find this rather remarkable.
00:22:05.11 Despite that these individuals live among us, it had not been recognized nor reported
00:22:11.25 until 1999, when Chris Jones and we reported this family and two others like it.
00:22:19.27 And this set us on our way to... to begin characterizing the genetics of this phenotype
00:22:26.05 in these human families.
00:22:28.12 Now, why is this important?
00:22:31.01 It's interesting to understand why and how we sleep, of course, but it also has real
00:22:36.08 implications for human health.
00:22:38.13 We know that when there's dysregulation of the clock, when the clock is not working properly
00:22:43.13 or synchronized properly, that this can contribute to sleep disorders, to depression and other
00:22:50.15 affective disorders, to pain syndromes, to stroke and heart attacks, and... as well as
00:22:56.18 other phenotypes including cancer.
00:22:59.06 So, sleep... chronic sleep deprivation, as well as desynchrony of the clock, or being
00:23:04.22 out of sync with our clock, as happens often, for example, with shift workers who chronically
00:23:10.20 work the night shift, for example, as a nurse who's constantly working the night shift,
00:23:19.14 she will be at increased risk for having breast cancer compared to a nurse who's working the
00:23:26.05 day shift chronically.
00:23:28.00 And so, to date, we've collected a large number of families.
00:23:31.03 We now have 90 families with familial advanced sleep phase.
00:23:36.02 And let me note also that we've shown in a population-based study that FASP, which had
00:23:42.09 never been recognized prior to 1999, when we published the first paper, that this affects
00:23:48.20 0.5% of the population.
00:23:51.02 That is to say, 1 in 200 people have a familial form of ASP.
00:23:56.22 And because, for many of these people, it's not a problem at all -- many of them feel virtuous
00:24:01.12 for being the morning lark or for being the early bird that gets the worm --
00:24:08.04 they never see physicians.
00:24:09.14 It's only when people don't like this phenotype that they seek out help from physicians,
00:24:15.21 and why the first woman who came to our attention came to see our colleague, Chris Jones.
00:24:20.16 So, this is a pretty common phenotype.
00:24:23.11 And we've got lots of families.
00:24:25.25 We also have approximately 30 families with a strong night owl phenotype.
00:24:31.19 And over 50 families who have a genetic form of this natural short phenotype.
00:24:39.20 And to date, through a lot of hard work and genetics, we've found 15 to 20 putative mutations
00:24:46.20 or mutations that we've proven to be causative.
00:24:50.01 15 to 20.
00:24:51.11 So, a small minority of these 90 families have mutations in the kinds of genes that
00:24:57.20 we know are important for regulating the circadian clock.
00:25:01.02 And about which Dr. Fu will talk in the... in the next section.
00:25:07.04 At the same time, we've found 3 genes and 3 mutations that regulate human sleep and
00:25:14.08 sleep length.
00:25:15.08 So, these 3 mutations lead to individuals who sleep 4 to 6 hours per night, and feel
00:25:20.11 well rested and perform at a very high level.
00:25:24.02 And so what this tells us is that despite a tremendous amount of wonderful work in the
00:25:28.24 genetics of circadian clocks in other organisms, like fruit fly or like mice, that there's
00:25:36.15 still a huge amount for us to learn from human genetics, in the many... not only from the
00:25:42.25 families that we've studied and where we've cloned genes but also, more excitingly,
00:25:47.20 from the ones where we haven't found mutations in any of the usual suspects, where there's...
00:25:53.28 there are new genes to discover and new pathways in the biology of circadian regulation and sleep
00:26:00.16 for us to learn about.
00:26:02.24 And so it's not surprising that some things are conserved all the way from fruit flies
00:26:06.26 to humans.
00:26:08.00 But we are also unique as humans and have different kinds of biology that... that have
00:26:14.05 evolved in our species that we can only begin to study through the study of human families.
00:26:21.16 Now, let me just talk briefly about some other phenotypes.
00:26:28.13 We know that sleep and circadian function are linked to all kinds of normal biology.
00:26:33.21 And that disruption can lead to increased risks for many diseases.
00:26:38.15 But I'd like to give a couple of examples of really exciting families that that have
00:26:42.21 not only familial advanced sleep or other phenotypes but... but those kinds of phenotypes
00:26:49.13 linked to non-circadian, non-sleep phenotypes.
00:26:53.16 This was the second family that we studied that... and led to the identification of the
00:26:58.03 second human circadian rhythm gene.
00:27:00.17 This is a family collected by our colleague, Bob Shapiro in Vermont, that has
00:27:05.15 familial advanced sleep phase.
00:27:07.13 They are very early sleepers and they wake up very early in the morning, compared to
00:27:13.06 the unaffected individuals in this family, who go to sleep at 11... approximately 11:30
00:27:19.04 and wake up at 8 in the morning.
00:27:21.07 So, we've cloned the gene, and Dr. Fu will talk about this, and identified the mutation
00:27:27.06 that causes FASP in this family.
00:27:30.20 But in addition, all of the affected individuals in this family, who all carry a mutation in
00:27:37.10 a gene called casein kinase 1 delta, all of these individuals also have asthma
00:27:43.02 and migraine with aura.
00:27:45.13 And so a very interesting question is, are these phenotypes, asthma and migraine with aura,
00:27:50.17 which are very common in the general population, are they coincidentally occurring
00:27:55.24 in the six individuals here who have FASP, or is this mutation actually causing those
00:28:02.15 phenotypes in these people?
00:28:05.14 And we have gone on to show that the mutation in this gene is actually causing the migraine
00:28:11.15 with aura phenotype.
00:28:13.25 And we still don't know whether or not this mutation causes the asthma phenotype.
00:28:20.05 Here's another interesting family.
00:28:22.04 These individuals were first characterized as FASP because they're early morning risers.
00:28:29.17 You can see that they're waking up at 3 or 4 in the morning, even on vacation, when they
00:28:34.25 have an opportunity to sleep in.
00:28:37.03 But we went on to show that they actually are able to stay up pretty late too.
00:28:41.10 And so they have a component of short sleep as well.
00:28:45.03 And the affected individuals in this family have very severe depression and anxiety
00:28:51.01 which is helped by Lexapro in... in these individuals.
00:28:55.00 This is an antidepression drug that benefits many patients with... with depression.
00:29:00.17 And so, let me just tell you that once we identified the gene that causes the circadian
00:29:05.17 phenotype in this family, we then began to study that mutation in the mice that carry
00:29:09.26 this mutation, and have accumulated some evidence suggesting that the depression phenotype in
00:29:16.06 this family might actually be caused by the mutation in this circadian gene.
00:29:20.25 So, there's still a lot more work to do, but this is what everybody expects, that there
00:29:25.05 will be very intimate connections between the clock and sleep, and other phenotypes,
00:29:32.11 in this case, like depression or migraine with aura, perhaps asthma, etc, etc.
00:29:41.00 And so, I made reference early to the... verbally, to the two process model, where the circadian
00:29:46.20 oscillator, that is going up and down and making a sleep and wake, is interacting with
00:29:52.13 another process of sleepiness, that we don't understand but sleepiness that accumulates
00:29:59.21 as a function of how long we've been up.
00:30:02.23 But I'd like to add another element to that model and... and perhaps you could think about it
00:30:08.12 as a three process model of sleep.
00:30:12.13 Through evolution, it makes sense that Mother Nature would reuse different pieces of biology
00:30:21.00 to regulate related and interconnected processes like sleep and wakefulness.
00:30:28.24 We know, and from studies in flies and mice and our own studies in humans, a lot about
00:30:34.23 the genetics of the circadian clock and why it oscillates in a day-and-night kind of rhythm.
00:30:41.21 We also are beginning to learn something from studies in human families about what regulates
00:30:46.12 human sleep.
00:30:47.12 We still understand very little about sleep, despite the fact that it is something that
00:30:54.02 most of us do for approximately a third of our lives.
00:30:58.04 But through evolution, other behaviors, like... like our energy level, have to oscillate in
00:31:06.18 a circadian way as well.
00:31:08.13 When we wake up in the morning, to be successful we have to go out and find mates and
00:31:14.04 make our families, begin our families.
00:31:15.28 We need to hunt and gather and find food.
00:31:19.17 We need to protect ourselves from dangers and... and survive through the day and night.
00:31:28.17 But to get restful sleep, that helps us to perform well again tomorrow, three things
00:31:34.15 have to happen.
00:31:35.15 We have to have our clock telling us that we're sleepy.
00:31:39.20 We have to accumulate something during the day, because we've been up, that we call sleepiness,
00:31:45.11 that helps reinforce that sleep phenotype.
00:31:49.05 And finally, we have to let go of our concerns about hunting and gathering, finding mates,
00:31:56.13 and everything else, and let ourself... selves go that... so that we can actually fall asleep.
00:32:02.14 And so we think that this makes a strong argument, a strong teliologic argument, that some of
00:32:10.06 the machinery that's important for the circadian clock and some of the machinery that's important
00:32:15.00 for the sleep homeostatic process in our brain are also going to be the kinds of machineries
00:32:20.28 in biology that regulate our behavior.
00:32:24.07 Our behavioral activation, our affect, and so forth.
00:32:27.23 And that, through evolution, these pathways have been retooled in different ways,
00:32:35.02 not only to regulate the clock, sleep, but also behaviors like... like our affect.
00:32:46.05 Finally, I'd like to acknowledge my longtime collaborator, Ying-Hui Fu and her laboratory;
00:32:51.19 Chris Jones, our sleep colleague at the University of Utah; and sources of funding,
00:32:56.15 which have made all of this work possible.

Part 2: Understanding our Sleep Behaviors

00:00:07.18 Hello.
00:00:08.18 I'm Ying-Hui Fu from the Department of Neurology at University of California, San Francisco.
00:00:17.20 Today, I'm going to tell you two examples of the mutations that we have found for humans
00:00:25.03 who have advanced sleep phase syndrome.
00:00:29.18 I will also show you how, by studying these mutations, we can actually learn a great deal
00:00:35.15 about the molecular mechanisms of human sleep behavioral regulation.
00:00:41.02 Before I start, I want to first briefly remind you of the core molecular clock.
00:00:47.15 For the mammalian system, there are two major transcription factors, CLOCK and BMAL.
00:00:53.01 They form dimers and bind to the promoter of clock control genes.
00:00:57.22 And PERs and CRYs are the most well studied clock control genes.
00:01:03.11 The protein levels of PERs and CRYs are tightly regulated.
00:01:08.00 And the PER is regulated by casein kinase 1 delta.
00:01:13.14 Another important kinase for circadian regulation is the GSK3 beta.
00:01:19.02 PERs and CRYs can form dimers and then come back to suppress the transcriptional activation
00:01:25.07 of CLOCK and BMAL, therefore forming a feedback... transcriptional feedback loop.
00:01:33.24 The first example I want to tell you is a mutation we found for this family that Louis
00:01:40.16 told you in the first part of this talk.
00:01:44.12 The mutation carriers in this family, they all showed about four hours phase advance.
00:01:50.09 The mutation was found in the gene encoding for casein kinase 1 delta.
00:01:55.24 And the mutation changed the amino acid at 44 from a threonine to alanine.
00:02:03.03 And this mutation makes... the casein kinase 1 has a reduced enzyme activity.
00:02:08.18 Here, as you can see, where the gray bars represent normal casein kinase 1 activity,
00:02:14.22 and the white bars represent the mutant casein kinase 1 activity.
00:02:19.16 You see, with five different substrates, the mutant enzymes showed reduced activity compared
00:02:25.22 to normal enzymes.
00:02:28.06 Now, for the mice that we generated to carry this specific human mutation, these mice
00:02:37.14 have a period length shorter than the control mice.
00:02:41.21 And this is as we expected, because we expect the phase-advanced animals should have
00:02:47.11 a shorter period length.
00:02:49.00 Now, if we remove the endogenous CK1 delta gene from the mutant transgene, then the period
00:02:56.12 is even shorter.
00:02:58.21 Now, interestingly, when we made Drosophila with this specific human mutation, we found that when
00:03:07.02 the flies carry mutant human CK1 delta their period is longer.
00:03:13.18 And it's longer than not only in the wild-type flies but also for the flies that carry wild-type
00:03:21.10 human CK1 delta gene.
00:03:23.13 So, what this told us is that, despite the fact that individual components of the molecular
00:03:30.25 clock are highly conserved between the mammalian system and Drosophila, it is likely that there
00:03:38.02 will be differences in the regulatory mechanism between these two different systems.
00:03:45.05 Now, one interesting feature for this particular family, as Louis mentioned, is that all the
00:03:53.18 mutation carriers for this family not only have a sleep behavior trait, they also all
00:04:00.08 have asthma and migraine headaches.
00:04:03.08 So, our hypothesis for this is that... we know that kinases can have up to
00:04:10.04 hundreds of different substrates.
00:04:12.09 So, if a mutation has occurred in a kinase, it is possible that the mutation can affect
00:04:20.11 more than one substrate, therefore lead to multiple phenotypes.
00:04:26.02 Now, because CK1 delta is very important for circadian regulation... so, we decided to
00:04:34.01 carry out the proteomic study for CK1 delta and also for CK1 epsilon.
00:04:39.28 And while we were doing this proteomic study, we decided to also address another question.
00:04:45.27 From previous studies, people have shown that casein kinase 1's... their activities remain
00:04:53.09 the same throughout the day.
00:04:56.00 So we wondered, can these kinases then phosphorylate different substrates at different times of
00:05:02.18 the day, therefore serving as part of their regular... regulatory function?
00:05:08.14 And indeed, our results show that CK1 delta and CK1 epsilon can phosphorylate
00:05:15.20 different substrates at different times of the day.
00:05:21.04 Now, from the proteomic study, we identified a novel gene called PHB2, here.
00:05:30.00 And PHB2 also plays an important role in circadian clock regulation.
00:05:35.26 PHB2 can suppress clock control genes' expression in a manner that is independent from BMAL
00:05:44.19 and CLOCK.
00:05:46.09 Interestingly, PHB2's protein level is regulated by CK1 epsilon.
00:05:53.15 But messenger RNA level of PHB2 is regulated by CK1 delta.
00:06:01.12 Now, the second mutation I want to tell you is for this particular family that Louis also
00:06:09.05 showed you.
00:06:10.05 We called it kindred 2174.
00:06:12.23 And this is the very first family that we collected.
00:06:16.12 The mutation carriers in this family also have four hours phase advance behavior trait.
00:06:22.23 Now, this is a very large family, so we were able to use this family and used a traditional
00:06:29.14 human genetics method to map the gene to the telomere region on the long arm of chromosome 2.
00:06:38.28 We then used a standard positional cloning method to find the mutation in the
00:06:44.23 Period 2 gene.
00:06:47.11 And it turns out the mutation was found in the amino acid position 662, that changes
00:06:54.23 from serine to glycine.
00:06:59.24 Now, when we examine the amino acid sequence in this serine 662 region, we found that the
00:07:07.25 amino acid sequence in this region is highly conserved among different PERs
00:07:13.25 from human and mice.
00:07:16.20 In addition, we found that there are four serine residues immediately C-terminal to
00:07:24.14 this serine 662, the first serine, here.
00:07:29.09 And interestingly, they all follow this exact serine-X-X-serine motif.
00:07:37.20 And serine-X-X-serine motif is a consensus sequence for casein kinase 1 activity.
00:07:45.26 So, we carried out a series of biochemical studies and what we found is that the
00:07:53.11 first serine of this five-serine region can be phosphorylated by a priming kinase.
00:08:01.08 And when this occurred then the following four serines can be phosphorylated by
00:08:07.22 casein kinase 1 delta.
00:08:11.03 Now, in the mouse model that we made to carry this specific PER2 serine 662 to glycine mutation,
00:08:21.27 the mutant transgenic mice have a period shorter than the wild-type mice, as we expected.
00:08:28.00 And this is also in agreement with the human subjects as a shorter period length.
00:08:33.03 Now, if we remove the endogenous PER2 allele from these mice, then the period became
00:08:40.01 even shorter.
00:08:41.21 Now, interestingly, when we generated the mouse model that changes the serine to an
00:08:49.13 aspartic acid to mimic the constitutive phosphoserine condition at this specific site,
00:08:58.14 then the transgenic mice have a longer period than the wild-type mice.
00:09:03.07 Now, again, if we remove the endogenous PER2 allele from serine to aspartic acid mutant
00:09:11.03 transgenic mice, their period became longer.
00:09:16.27 So, what we have here is that with this five-serine region the first serine is phosphorylated
00:09:24.20 by a priming kinase.
00:09:26.14 And then CK1 delta then phosphorylates the four additional serines.
00:09:31.00 And when this happens, then the animals will have a longer period, which we can think of it
00:09:36.06 as a slower clock, because it takes longer to finish the cycle.
00:09:41.09 Now, if the first serine is blocked, it cannot be phosphorylated, therefore none of them
00:09:48.09 are phosphorylated, then the animal will have a shorter period.
00:09:53.15 And we can think of this as a faster clock.
00:09:56.10 Now, when we look at the activity recording for our serine to glycine mutant transgenic mice,
00:10:04.09 we found that four hours before we turn the light off the animal starts to become
00:10:11.17 active.
00:10:12.17 The activity turned on.
00:10:15.16 And four hours before the light on, the activity starts to wind down.
00:10:21.08 And this is very similar to human subjects with four hours phase advance.
00:10:25.15 So, this mouse model actually recapitulates the human phenotype very well.
00:10:32.20 So, we showed that this serine to glycine mutation will make the PER2 protein
00:10:41.06 become hypophosphorylated.
00:10:43.20 But how does a hypophosphorylated PER2 then lead to an advanced sleep phase behavior trait?
00:10:50.24 Well, there were three obvious possibilities.
00:10:54.22 First is that the mutation could affect PER's protein stability.
00:10:59.21 Second is that the mutation will affect PER2's nuclear translocation timing or mechanism.
00:11:06.11 Third is that a mutation could affect PER2's repressor activity,
00:11:12.09 therefore change the transcriptional regulation.
00:11:15.17 And we examined each of these possibilities.
00:11:19.05 And our results show that the major effect for this mutation is changing PER2's
00:11:25.23 repressor activity.
00:11:27.10 Now, here you can see that the wild-type PER2 messenger RNA, peaked as... here, this peak.
00:11:38.07 And the SG PER2 messenger RNA actually peaked earlier than the wild-type PER2 messenger RNA.
00:11:46.15 And the SD PER2 messenger RNA peaked later than the wild-type, agreeing with... that
00:11:52.28 SG has a shorter period length and advanced sleep behavior trait.
00:11:58.14 Now, not only that the SG PER2 messenger RNA peaked lower than the wild-type PER2 messenger
00:12:06.26 and the peak for SD PER2 is higher than the wild-type PER2, and this is true not only
00:12:14.24 for the mouse endogenous PER2 messenger RNA, also for human PER2 transgene messenger RNA.
00:12:22.11 So, what this result told us is that the SG PER2 is a stronger repressor than the wild-type protein,
00:12:30.10 and the SD PER2 is a weaker repressor than the wild-type PER2 protein.
00:12:38.14 So, we knew that PER2 protein is regulated by casein kinase 1 delta, so we wondered whether
00:12:47.23 we could use our mouse model to study genetic interaction between CK1 delta and PER2.
00:12:55.18 And this slide is just to show you that the copy number of CK1 delta actually does not
00:13:02.10 affect mouse period length.
00:13:05.23 With the heterozygous CK1 delta knockout mice with only one copy of the normal gene and
00:13:15.00 CK1 delta wild-type transgene with 4-5 copies of the normal gene, their period remains the
00:13:21.25 same with wild-type mice.
00:13:24.09 Again for the serine to glycine PER2 mutant transgene, the period is shorter than the
00:13:31.05 wild-type mice.
00:13:33.08 But if we cross these SG mutant transgenic mice with the CK1 delta heterozygous knockout mice,
00:13:41.14 their period became longer.
00:13:43.28 Now, if we cross the SG mutant transgenic mice with the CK1 delta wild-type transgene,
00:13:51.19 then the period became shorter.
00:13:54.03 So, what this result told us is that on PER2 protein, in addition to this
00:14:03.05 phosphorylation site by CK1 delta around 662, there are other CK1 delta phosphorylation sites for PER2.
00:14:12.03 And these other sites, when they are phosphorylated, they can lead to a shorter period length.
00:14:18.09 So, we proposed this model to explain all of our data together.
00:14:24.15 We think that PER2 has multiple phosphorylation sites by casein kinase 1 delta.
00:14:30.14 And one of these sites is around serine 662.
00:14:33.25 And then... when this site is phosphorylated, it can lead to increased PER2 messenger RNA
00:14:41.06 and protein, and lead to a longer period length.
00:14:45.08 Now, when these other sites are phosphorylated, it can lead to increased PER2 degradation,
00:14:52.22 therefore lower PER2 protein level.
00:14:55.15 This then leads to a shorter period length.
00:14:58.03 And under normal conditions, these different pathways have to maintain a delicate balance
00:15:03.15 in order for the animals to have a stable, normal period length.
00:15:08.25 So, all these results really point out the importance of PER2 phosphorylation in setting
00:15:15.28 the speed of the clock.
00:15:18.14 In fact, this is true not only for PER2 protein but also for other clock proteins as well.
00:15:26.19 So, then this, then, raised another question, which is, are other post-translational modifications
00:15:34.15 also important for regulating the circadian clock?
00:15:39.26 So, as I showed you earlier, we did a proteomic study for CK1 delta.
00:15:48.21 And GSK3 beta is another important kinase for circadian regulation.
00:15:54.24 So, we also did a proteomic study for GSK3 beta.
00:16:00.04 In this proteomic study, we found more than 400 potential GSK3 beta substrates.
00:16:06.24 And these more than 400 proteins can be mapped onto many biological pathways, including some
00:16:13.07 of them previously shown and others are novel biological pathways for GSK3 beta.
00:16:22.25 Out of these more than 400 proteins, we decided to follow up on one particular protein, which
00:16:29.18 is O-GlcNAc transferase, or OGT in short.
00:16:35.22 OGT and OGA, which is O-GlcNAcase, they are responsible for the O-GlcNAc post-translational
00:16:44.06 modification for proteins.
00:16:49.17 Interestingly, the GSK3 beta phosphorylation can significantly increase OGT's activity,
00:16:58.01 here, where the phosphorylation by GSK3 beta significantly increases the activity.
00:17:06.00 Now, O-GlcNAc modification also occurs on serine and threonine.
00:17:11.10 So, the first question we asked was, does O-GlcNAcylation also play a role in regulating
00:17:18.17 the circadian clock?
00:17:21.23 We first used an in vitro system to see if O-GlcNAcylation can affect circadian period length.
00:17:31.12 When we add OGA inhibitor to the cells, we found that the period length becomes longer.
00:17:38.21 Now, if we add OGT inhibitor to the cells, then the period became shorter.
00:17:46.26 We also confirmed this in vivo.
00:17:50.00 With OGT conditional knockout mice, we found that their period is shorter than the control mice.
00:17:58.02 For the Drosophila, the OGT knockdown, the period is shorter.
00:18:05.13 The OGA knockdown, the period is longer than the wild-type fly.
00:18:10.04 So, all these different systems suggest that under higher O-GlcNAcylation conditions,
00:18:17.15 it can lead to a longer period length.
00:18:20.10 And the lower O-GlcNAcylation level leads to a shorter period length.
00:18:28.23 We then found that at least two of the core clock components are modified by O-GlcNAc.
00:18:35.25 CLOCK protein, when it's modified by the O-GlcNAc, it actually reduced its transcription activity.
00:18:43.22 But the PER2 protein, when it's modified by O-GlcNAc, it actually further enhanced its
00:18:51.19 repressor activity.
00:18:53.05 Now, with PER2 luciferase activity as an index, we see that CLOCK and BMAL can turn on PER2 promoter.
00:19:04.04 In the presence of OGT, the transcription activity is significantly suppressed.
00:19:11.07 Now, wild-type PER2 protein is a repressor for CLOCK and BMAL activity.
00:19:18.16 In the presence of OGT, the PER2's repressor activity is further enhanced.
00:19:24.07 Now, as I told you earlier, the PER2 S to G mutant is a stronger repressor than the
00:19:33.26 wild-type protein.
00:19:35.19 However, when we add OGT to the reaction with mutant PER2 protein, it did not affect
00:19:43.21 the transcription activity.
00:19:46.06 This then gave us an idea that maybe O-GlcNAc modification plays a role on this serine 662 site.
00:19:55.09 We used a culture system, together with western blot, to show that when PER2 protein
00:20:02.13 is O-GlcNAc modified it actually blocks the phosphorylation at serine 662 site.
00:20:09.16 And also, if the serine 662 of PER2 is phosphorylated, the PER2 protein does not get O-GlcNAc modified
00:20:18.15 as well as the PER2 protein that is not phosphorylated at serine 662.
00:20:24.00 So, together, these results suggest that there is an interplay between O-GlcNAcylation and
00:20:31.10 phosphorylation at PER2 serine 662.
00:20:37.17 We also used mass spectrometry to map all the O-GlcNAc modification sites,
00:20:44.16 focusing on the serine 662 region.
00:20:47.05 And we found many O-GlcNAc modification sites, marked in red here, including the serine 662,
00:20:55.21 serine 668, and serine 671.
00:21:01.17 Now, OGT and OGA are responsible for O-GlcNAc modification of the protein.
00:21:09.03 And O-GlcNAcylation is dependent on its substrate, UDP-GlcNAc.
00:21:15.22 UDP-GlcNAc is derived from nutrients, including glucose, amino acids, and nucleic acids, through
00:21:24.26 hexosamine biosynthetic pathways.
00:21:27.24 Now, O-GlcNAcylation can regulate circadian rhythm.
00:21:32.13 So, we then wondered, well, can glucose level modulate the circadian clock?
00:21:41.00 So, we did this simple experiment.
00:21:44.11 Here, under low glucose concentration, we found that the serine 662 region of the PER2
00:21:52.25 can be phosphorylated as long as CK1 delta is present, regardless of if OGT or OGA
00:22:00.12 is there or not.
00:22:02.04 However, under high glucose concentration, the phosphorylation in this region is blocked
00:22:10.00 as long as OGT and OGA is there, regardless of if CK1 delta is there or not.
00:22:17.03 In summary, we found that in this serine 662 region there are five serines in a row.
00:22:23.13 When the first serine is phosphorylated by a priming kinase, therefore the four additional
00:22:29.26 serines are phosphorylated by CK1 delta.
00:22:33.07 This then leads to a longer period for the animal.
00:22:38.06 Under high glucose concentration, the O-GlcNAcylation can take over and block the phosphorylation
00:22:44.24 for this region, therefore leading to a shorter period and a faster clock.
00:22:49.28 And we call this "clock sugar rush".
00:22:54.20 So, we have so far collected more than 90 advanced sleep phase families.
00:23:01.13 And we are continuing to use these families to identify genes and mutations responsible
00:23:07.16 for this behavior trait.
00:23:10.02 When we find genes and mutations, we study their protein functions.
00:23:13.27 At the same time, we also generate mouse and fly models
00:23:18.12 to help us understand the molecular mechanism.
00:23:21.25 And with this parallel approach, we hope to gain a better understanding of human circadian
00:23:28.08 and sleep mechanisms.
00:23:30.12 With a better understanding of these regulatory mechanisms, we'll be able to come up with
00:23:35.25 a better therapeutic intervention for sleep-related disorders.
00:23:43.02 I want to acknowledge my long-term collaborators, Louis Ptacek and Chris Jones.
00:23:48.00 And also all the people who contributed to this work.

Videos in this Talk

Part 1: Connections between Clock and other Phenotypes
Audience:
- Researcher
Part 2: Understanding our Sleep Behaviors
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad

Speakers: Louis Ptacek, Ying-Hui Fu

Total Duration: 0:57:52

Recorded: April 2014

All Talks in Human Disease

Talk Overview

By studying families with sleep/wake disorders, Fu and Ptáček have shown that mutations that cause changes in the phosphorylation or acylation of the PER2 protein are responsible for regulating circadian rhythms.

Ptáček introduces the circadian clock and its relationship to sleep. He describes different sleep-wake behaviors including people who go to sleep and awaken exceptionally early or late. By studying families with an advanced sleep phase (ASP) phenotype, he and his colleagues showed that these individuals had a shortened circadian period. Further studies of families with similar sleep-wake behaviors have identified a number of mutations responsible for circadian rhythm regulation.

In Part 2 of the talk, Dr. Fu explains that studies of families with sleep disorders have shown that post-translational modifications of the PER2 protein are involved in regulating circadian rhythms. Casein kinase 1 phosphorylates PER2 and mutations in either CK1 or specific serine residues in PER2 result in an advanced sleep phase phenotype. O-GlcNAcylation of PER2 also participates in regulating circadian rhythms because O-GlcNAcylation blocks the sites usually phosphorylated by CK1.

Speaker Bio

Louis Ptacek

Louis Ptáček leads the Division of Neurogenetics, in the Department of Neurology, at the University of California, San Francisco. He is also an Investigator of the Howard Hughes Medical Institute.Ptáček’s interest is in inherited disorders of the nervous system in human families and his early research identified mutations in several ion channels that caused episodic… Continue Reading

Ying-Hui Fu

Ying-Hui Fu is a Professor of Neurology at the University of California, San Francisco where she collaborates with Dr. Ptáček in the Division of Neurogenetics. Research in her lab uses genetic tools to study neurodegenerative diseases, such as multiple sclerosis, and sleep/wake behaviors. Dr. Fu received her PhD from Ohio State University before moving to… Continue Reading

Comments

Kade Merritt says

June 30, 2021 at 4:55 pm

A guy told me at work of this DEC2 because he knows how I tell people how I sleep. And its very interesting, I can go to bed at almost anytime and sleep 4-5 hours and be ready to go in the morning. If I sleep 6-8 hours I have a headache.

Reply

Part 1: Connections between Clock and other Phenotypes

Part 2: Understanding our Sleep Behaviors

Talk Overview