The Nocturnal Brain

by Guy Leschziner

  In narcolepsy, however, this is not the case. The neurological mechanisms preventing you from entering REM quickly after going to sleep go awry. People with narcolepsy can flick into REM very early in the night, sometimes directly from waking. Multiple episodes of REM sleep in the mean sleep latency test, with its brief, twenty-minute opportunities to nap, are characteristic of narcolepsy. And so it is easy to see why some of the symptoms of narcolepsy arise. If you go straight into dreaming sleep from being awake, then having dreaming experiences in wakefulness may manifest as seeing or hearing things in the room as you lie there. If the muscle weakness of REM sleep is switched on while you are awake, then you will feel paralysed. And if you are really unlucky, you may experience both together, resulting in the really unpleasant sensation of feeling pinned down to the bed while you hallucinate that someone or something is in the room with you.

  Imagine two seesaws in a children’s playground. When swinging free, balanced without any children on either end, a strong gust of wind may swing the beam one way or the other. Think of one seesaw controlling the switch between sleep and wake, and think of the other controlling the switch between REM and non-REM sleep. The hypocretin system is the equivalent of a 16-stone prop forward, perched delicately right at the far end of each seesaw, firmly planting the end that pushes wake into the ground, and when you are asleep, holding down the end that promotes non-REM sleep. Lose the prop forward, and the seesaws go back to swinging in the breeze. Without hypocretin, people with narcolepsy flick in and out of sleep uncontrollably, and switch in and out of REM. So the loss of hypocretin-producing neurones results in sleepiness during the day, sometimes irresistible sleep attacks, sleep paralysis and hallucinations on sleep onset, termed hypnagogic hallucinations.

  In addition, flicking in and out of REM sleep into wake overnight results in extremely vivid dreams. Most people only remember their dreams when they wake from REM sleep, but for people with narcolepsy, they are doing this constantly throughout the night. And actually, although we consider people with narcolepsy to be very sleepy, over a 24-hour period they don’t really sleep any more than anyone else. They sleep more during the day, but overnight their sleep is broken and of poor quality.

  When I look at the results of Adrian’s study, however, searching for the early onset of REM overnight, or the short time to drifting off to sleep and the presence of REM sleep in the daytime nap test, there is nothing. No hint of narcolepsy. His sleep study, apart from showing a bit of snoring and some leg movements, is totally normal. Cataplexy usually develops alongside sleepiness or up to several years afterwards. It is rare for cataplexy to start before any other symptoms of narcolepsy.

  When I see him next, we discuss these results. I tell him that a positive sleep study is often seen in people without narcolepsy, and can be caused by sleep deprivation, but a negative sleep study is rather against a diagnosis of narcolepsy. I have occasionally seen apparent cataplexy in people with psychological issues, a manifestation of psychological distress, but Adrian strikes me as totally stable, without any obvious psychological triggers. We decide to proceed to a spinal tap, to analyse his cerebrospinal fluid for hypocretin.

  Two weeks later I see him again in clinic with the results. Normal. I am slightly stunned. Pretty much everyone with cataplexy has low or deficient levels of hypocretin. We discuss his symptoms again. I remain totally convinced that what he is describing is cataplexy, and despite the nagging doubt in my mind raised by the negative results, I prescribe him medication specific for it.

  * * *

  While it is easy to understand how a deficiency of hypocretin can cause sleepiness, hypnagogic hallucinations and sleep paralysis, explaining cataplexy is a bit more difficult. Why this bizarre symptom occurring in the middle of the day should be related to a sleep disorder is incomprehensible at first glance. But there are a few clues. Monitoring muscles during cataplectic attacks demonstrates repeated brief silence in muscle activity, and looks very similar to the silencing of muscle tone seen in REM sleep. It is this repetitive nature of short bursts of weakness that people sometimes describe as a juddering to the ground. The weakness usually builds up gradually, and people can often manage to lower themselves to the ground to avoid injury. If seated, the muscles of the neck may briefly sag, causing some nodding as the person tries to combat the muscle weakness, and if the face is involved, there will be a slackening of the facial musculature that looks like twitching.

  Another feature of cataplexy that gives us some insight into what causes it is that, occasionally, particularly with prolonged attacks, people begin to hallucinate or dream. It therefore appears that cataplexy is closely linked with REM sleep. In the same way that REM sleep, or at least aspects of it, are switched on inappropriately early in the night in people with narcolepsy, it seems that the absence of hypocretin during the day predisposes to the sudden onset of aspects of REM sleep in wakefulness. The paralysis of REM sleep is switched on while you are standing or sitting. In many ways, this is the mirror image of John’s acting out of his dreams at night, his REM sleep behaviour disorder. Whereas John’s problem is down to a failure to generate paralysis in REM sleep at night, Adrian’s is due to the uncontrolled switching on of paralysis in the day. In Adrian’s case, the cataplexy affects his whole body, but for some people it can be limited to only one part – the face, neck, arms or legs, for example.

  So, cataplexy results from aspects of REM sleep, and the paralysis that accompanies it, being switched on at inappropriate times during wakefulness. But what all of this fails to explain is why cataplexy is triggered by laughter. In fact, it is not just laughter that triggers it. Sometimes it arises spontaneously, without any prompts at all. For other people, surprise, anger, sorrow or anxiety spark it off. I have had patients collapse in the street when honked at by a car; others who have attacks when they argue with their family. But certainly laughter, or an ‘internal feeling of mirth’, as Adrian puts it, is the most common generator.

  Actually, mild muscle weakness with laughter is a normal phenomenon, hence the expression ‘weak with laughter’. Monitoring muscle electrical activity in normal individuals demonstrates that laughter suppresses what’s called the H-reflex, essentially a laboratory version of the reflexes elicited by neurologists when we tap the knee with a hammer. Clearly, in cataplexy, this suppression is dramatically amplified. In fact, during a cataplectic event, the knee jerk and other reflexes simply disappear. On the rare occasions I see a cataplectic event in a patient in clinic, if a tendon hammer is within reach I will check for these reflexes to confirm that this is true cataplexy.

  Experiments have shown that the hypocretin-producing neurones in the hypothalamus are very active when we experience strong emotions. So it seems that somehow hypocretin puts the brake on this normal phenomenon of weakness generated by strong emotions, dampening it down. The absence of these neurones somehow destabilises the brainstem’s regulatory system for muscle tone.

  Though that is not the whole story either. In addition, another part of the brain called the amygdala is implicated. This almond-shaped structure sits on either side of the brain, deep within the temporal lobe, and has an important function in the processing of emotional stimuli. Epileptic seizures arising in the amygdala often precipitate the experience of sudden strong emotion, such as overwhelming fear. Studies in patients with narcolepsy have shown alterations in activity in the amygdala while looking at funny pictures, and in narcoleptic dogs there appear to be changes in electrical activity in the amygdala with cataplectic attacks. The theory is that circuits from the amygdala project to the areas of the brainstem involved in maintaining muscle activity. In wakefulness, these connections are inhibited by hypocretin, but in narcolepsy, the lack of hypocretin takes the brake off this circuit, resulting in loss of muscle strength with high amygdalar activity.

  But all of this is very odd. What reason is there for any connection between strong emotion and muscle weakness? Why would our brains be designed to have these con
nections between the amygdala, the neurological cornerstone of emotions, and brainstem nuclei causing muscle weakness? It seems nonsensical from an evolutionary perspective to go weak at times of heightened emotion. The last thing you would want is for your legs to collapse from under you when you are terrified by a predator – and what would be the benefit of weakness with laughter?

  A recent intriguing but unproven hypothesis is that cataplexy is related to a phenomenon of ‘tonic immobility’. This state represents the feigning of death that many animals perform when under attack. The opossum that rolls over and plays dead when threatened is the classic example, but tonic immobility has been described in a wide variety of animals, including sharks, chickens, pigs and snakes. Although some animals remain in a particular posture, the muscles often go very slack, and other animals go floppy. While tonic immobility has not been described in humans, the phrase ‘to be paralysed with fear’ suggests that something similar may also occur in us.

  So it has been suggested that there are similarities between tonic immobility and cataplexy, and that the connections between the amygdala and brainstem are an evolutionary hangover, but this still does not explain why positive emotions such as laughter or joy could trigger this pathway. Nor does it explain why cataplexy is more likely to occur in the presence of friends and when relaxed – it is rare to see cataplexy in the clinic, because people are usually anxious about coming into hospital to see a doctor – but it may at least provide an explanation for the existence of this circuitry, the link between emotions and paralysis.

  * * *

  I touch base with Adrian a few weeks after starting him on his treatment. The response is startling and rewarding. His collapses have almost completely stopped on a small dose of medication. ‘[The medication has] almost been like a light switch,’ he tells me.

  I don’t think I’ve had any full collapses since we started that medication. I’ve had what I would call a kind of near-collapse – this sort of weakness that I used to feel before a full-on collapse. I’ve felt at different times that I was right on the very edge of being controlled. I’ve had these kind of situations where I’m trying to be a little bit humorous and just kind of felt . . . It’s a very difficult feeling to describe. It just feels like a sort of a swirling feeling in my lower back. But just as it has started to appear, then it’s gone away.

  Over the next couple of years, Adrian and I see each other regularly. His cataplexy remains well controlled, but it becomes obvious that he is becoming a little more sleepy. Initially he puts this down to his stressful job, his limited sleep, his long commute, but eventually we realise that this is not the only reason. I bring him back in for a sleep study, to see if this time we can demonstrate with more certainty that he has narcolepsy. To my surprise, once again his sleep study is totally unremarkable.

  Driven by his and my curiosity to understand his condition, however, we decide to repeat the spinal tap to assess his hypocretin levels. When I see him again to review the results, his hypocretin is the same as it was two years previously. In fact, exactly the same! On closer inspection, it becomes clear that the lab has actually sent me the results from two years previously. I call through to hear that, actually, the new tests show that Adrian’s hypocretin is almost completely undetectable, some three years after the onset of his cataplexy . . .

  * * *

  While the loss of hypocretin explains the features of narcolepsy with cataplexy, there remains the question as to why these hypocretin-producing neurones disappear in the first place. What has caused the vanishing into thin air of these neurones in Adrian’s lateral hypothalamus? Clearly it is not as straightforward as in the Stanford dogs, where a single genetic mutation results in a failure to detect hypocretin. No single genetic abnormality has been found in humans, with the exception of one person, mentioned earlier.

  It was in the 1980s, during research into a variety of diseases of unknown cause, that scientists found that most patients with narcolepsy carried a particular genetic marker. Narcoleptics carried a particular type of genetic variant of the human lymphocyte antigen (HLA), a complex of proteins involved in the regulation of the immune system, first identified in the search for an explanation to tissue rejection following organ transplants. Most patients with narcolepsy with cataplexy were found to be positive for a variant called HLA DR2, a protein complex responsible for presenting fragments of infective agents – antigens – to the white blood cells that combat infection.

  This was the first indication that the immune system might play a role in the development of narcolepsy; many other conditions associated with particular HLA types, like lupus and rheumatoid arthritis, have a very clear autoimmune basis. Subsequent studies have confirmed an even stronger association with a related HLA type, called DQB1*0602, present in almost all people, up to 98 per cent, with cataplexy. But this HLA type is also not the complete explanation. About one in four of the population also carries this HLA type, while narcolepsy is much rarer, affecting about one in 2,000 people. Additionally, more recent genetic studies have also demonstrated that other genes increase the risk of narcolepsy, such as those that encode receptors on T-cells, the white blood cells that are the mainstay of our immune system.

  So, narcolepsy may be a type of autoimmune disease. Perhaps it is Adrian’s own immune system that somehow mounted an attack on his hypocretin-producing neurones, causing this devastating loss. But, if that is the case, and this destruction is so strongly linked to the genes that influence how our immune system functions, what triggers this abnormal immune response? And why does it only occur in a tiny fraction of people with the right HLA type? A possible explanation may be forthcoming. Observations have long shown seasonal fluctuations in the onset of narcolepsy throughout the year, suggesting a link with winter infections, like influenza or streptococcal throat infections. Indeed, narcoleptics often report influenza in the year prior to onset of narcolepsy.

  It was the H1N1 swine flu global epidemic in 2009–10 that gave us the strongest indication of a link between infection and narcolepsy, however. I remember that winter well. The media was full of stories regarding the aggressive nature of this new strain of influenza sweeping across the world. The NHS had stockpiled large quantities of Tamiflu, a treatment said to lessen the effects of influenza; our intensive-care unit had beds set aside for swine flu patients; and there was a general feeling of mild panic in the air. No one knew what to expect, but the omens were bad. Within a couple of months, thirty countries around the world had already reported cases of H1N1. A concerted global campaign of vaccination for the H1N1 influenza strain was set in motion, and I recall waiting in line at the hospital for my jab. Ultimately, that year’s flu season passed without it becoming the public health disaster people feared, but we certainly had some very sick patients in the hospital, and some deaths.

  It was within a year of the vaccination campaign, however, that researchers noticed something rather unusual. Several cases of narcolepsy appeared to be related to one of the H1N1 vaccines used in Europe, called Pandemrix. The numbers of cases of narcolepsy rose dramatically in countries where this vaccine was used, particularly in children, but a similar increase in numbers was not seen in the USA, which used a different vaccine. In later years, the numbers of new cases dropped down to levels prior to this vaccination programme. It was not only the vaccination that was associated with onset of narcolepsy, though. In China, cases of narcolepsy were found to be related to the H1N1 influenza virus itself. Interestingly, in mice bred without intact immune systems, the H1N1 virus has been shown to migrate to the hypothalamus and brainstem, causing sleep disruption directly. Expression of high levels of viral components in these areas of the brain may make them particularly vulnerable to autoimmune damage, if the immune system is intact. So, it is not just the Pandemrix vaccine that appears to trigger narcolepsy, but the flu virus strain itself.

  Subsequent studies, including our own, have demonstrated this strong association between Pandemrix and narcolepsy. Th
e vaccine has been associated with a two- to twentyfold increased risk of developing narcolepsy. But this marked increase in cases has not been seen with the other commonly used H1N1 vaccine. The reason for this is not entirely understood, although there are subtle differences between the two vaccines, in the nature and quantities of the viral fragments contained. Studies have shown that the chemical structure of fragments of the H1N1 virus are very similar to bits of the hypocretin receptor, implying that the immune system, in particular circumstances, may mount an immune response against the virus or vaccine that also erroneously targets aspects of the hypocretin system. The hypocretin-producing cells are ‘collateral damage’.

  This concept of collateral damage of the immune system is not a new one in the world of neurology. There are many neurological conditions that have a similar basis. Guillain–Barre syndrome, a devastating and life-threatening destruction of peripheral nerves, has long been known to be triggered by Campylobacter, a bacterium that causes food poisoning, and post-infection and post-vaccination damage to the brain or peripheral nerves is well recognised. Infectious agents will often use ‘molecular mimicry’ – appearing structurally similar to the body’s own molecules – to outsmart the immune system. But if your immune system is configured in a particular way, it is liable to recognise these ‘mimics’, destroying the infectious agent but also aspects of your body that look chemically identical.

  For narcolepsy, it appears that Pandemrix is somehow better at triggering this response than natural infection or the other vaccine widely used, if you carry the HLA DQB1*0602 type. Perhaps this is due to the process of production of the vaccine. While the precise nature of this immune process remains to be unravelled, class actions are currently being prepared against the makers of the Pandemrix vaccine.