In work conducted at Stanford University and published by Luo et al. (2018) in Proceedings of the National Academy of Sciences today, a major step has been made in our understanding of the sleep condition narcolepsy, the discovery of specific immune cells that have receptors recognizing both hypocretin and specific strain of flu.  This strengthens the hypothesis that in some people the Influenza A flu is creating an immune response that can go awry and starts to cross-react with a brain molecule called hypocretin. These immune cells, called CD4 T cells, confuse specific pieces of flu with a piece of hypocretin, orchestrating the destruction of approximately 20,000 cells in the brain that produce hypocretin.  Once these cells are gone, the affected person experiences narcolepsy because the hypocretin molecule is important to stay awake and control dreaming. The importance of this discovery deserves our attention, as it opens the possibility of preventing narcolepsy using modified flu vaccines. It is also the first time that such a direct relationship has been found between an infectious disease (the flu) and an autoimmune disease (narcolepsy).

Understanding the condition

Narcolepsy affects children (Waking up Mathilda by Claire Crisp) as well as adults (Wide Awake and Dreaming: A Memoir of Narcolepsy by Julie Flygare) and wrecks lives if not rapidly recognized and treated, as is often the case.  

Before puberty, narcolepsy often starts very abruptly.  Affected children suddenly gain an enormous amount of weight, sleep all the time, start to dream vividly, talk in their sleep, and have episodes of muscle weakness (called cataplexy) that can affect the entire body.  Some of the weakness also characteristically affects the face, forcing the child to often have the mouth half open and the tongue protruding. Rapid diagnosis and treatment is needed to avoid further complications, such as these children missing school, losing friends, and becoming ashamed and withdrawn.

In adolescents and adults, although the picture can be as dramatic as in younger children, it often starts more insidiously.  The weight gain is also present but is more progressive, sometimes over a year. The sleepiness is more fluctuating, the adolescent is often unable to sleep well because of dreams, nightmares, or simply waking up in the middle of the night is often an issue.  Sleepiness waxes and wanes, and culminates in irresistible sleep attacks, followed by naps, which leave the patient rejuvenated but only for a couple of hours. Muscle weakness or paralysis episodes most often only occur suddenly and briefly when the patient is happy, notably in the context of a funny joke or when having a good time with family and friends.  These episodes also affect the neck and face, but also the knees, sometimes forcing the patient to sit down or to crumple down slowly to the floor.

The autoimmune cause

For over 10 years, the cause of narcolepsy has been known to be the loss of about 20,000 so-called “hypocretin/orexin” neurons in the hypothalamus, a deep brain structure.  How this cell loss occurs has, however, remained a mystery.

The main theory, based on genetic studies, has long been a suspected autoimmune attack.  Specifically, more than 15 narcolepsy genetic-predisposing factors have been identified and all involve the immune system, with most also involved in genetic predisposition to other autoimmune diseases, most notably type 1 diabetes (see Ollila et al., 2018).  The genetic work, done in my laboratory but in collaboration with researchers and clinicians from all over the world, pinpoints very clearly the role of a specific immune protein subtype called DQ0602 reacting with immune cells called CD4+ T lymphocytes, and probably subsequently involving CD8+ T cell lymphocytes that destroy hypocretin/orexin cells.  

In autoimmune diseases, immune cells called T cells and antibodies that normally respond to bacteria or viruses, mistakingly attack part of our own body resulting in symptoms that can affect pretty much any organs:  the guts in Crohn’s disease or gluten intolerance, joint articulations in rheumatoid arthritis, or insulin-secreting cells of the pancreas in type 1 diabetes.

What has been challenging in narcolepsy is that whereas in almost all autoimmune disease autoantibodies can be detected in blood targeting a specific organ, nothing has ever been found in the blood of narcoleptic patients, which has made it impossible to develop a blood diagnostic test.

A main trigger of narcolepsy: the flu

Another big clue in understanding narcolepsy came simultaneously from China and Scandinavia in 2010.  In the spring of 2009, a new form of influenza A nicknamed the “H1N1 swine flu” appeared in Mexico and caused great alarm as high mortality was observed. Because it was a new flu most probably originating from swine, humans had limited immunity, and it was predicted to infect half of the world population the following season starting in September 2009.  

These new flus appear from time to time (the last one was in 1967 in Hong Kong called H3N1) and are usually associated with higher mortality, notably in younger individuals, although this is variable.  One of the deadliest known flu viruses in history has been the H1N1 Spanish flu of 1918, estimated to have killed 50 million people in 1918. Considering the risk involved, and the fact the vaccine that had been prepared for the 2009-2010 season did not contain this new strain, separate swine flu vaccines were prepared by multiple companies, including a specific brand called Pandemrix, produced by Glaxo Smith Kline (GSK).

In 2010, two strange phenomena were observed. In China, Dr. Fang Han had been recording the number of cases arising every year in children, and had noticed that the onset in these cases was most often in the spring or summer, suggesting that perhaps a winter infection (at the time, strep throats were suggested as the culprit) was responsible for triggering narcolepsy, and that it maybe took a few more months for hypocretin cells to be killed.  In 2010, however, a 3 times higher number of cases was reported in the spring and summer, suggesting that the swine flu had been a strong trigger.

In parallel with this, Scandinavia went through a successful vaccination campaign against the swine flu with almost half of the population being vaccinated with Pandemrix in addition to the regular seasonal vaccine.  Surprisingly, however, cases of abrupt, severe childhood narcolepsy started to appear a few months following the Pandemrix vaccination campaign. This raised alarm, and the few sleep scientists who raised the possibility of a vaccination side effect were vilified, not surprisingly considering the many false claims that have been made against vaccines.  Ten years, later, this effect is not in doubt, and although only 1/10,000 vaccinees ever developed narcolepsy (a rare occurrence), the risk of narcolepsy after Pandemrix was found to be increased approximately 10-fold: a clear connection. Pandemrix was not used in the United States, and no association was found with regular flu vaccines.

A digression on the vaccination debate

For full disclosure: The author of this piece is pro-vaccination.  Any effective treatment always carries the risk of side effects. Doctors must inform patients of risks but patients must also take responsibility for these small known and unknown risks.  Hundreds of people become deaf or partially destroy their kidneys when taking antibiotics, yet the net result is vastly better than the alternative. In medicine, all effective treatments can have side effects.  Trained as a pharmacologist, I like to say “no effect, no side effects”.

There is no doubt that protecting the entire population against deaths from flu, measles, or other illnesses using vaccination has been one of the most positive developments of medicine to date.  In fact, it has generally been shown to be one of the safest interventions ever developed. As with any medicine or procedure, however, it is difficult to predict the possibility of relatively rare side effects from a vaccine, especially in the context of a rapidly evolving epidemic.   

On the anti-vaccine side, frivolous claims have been made suggesting vaccines cause hundreds of problems without any proof.  These are frequently dismissed based on the argument that there is no direct statistical proof of an association. Although very often the science associated with these claims is non-demonstrative at best, it should also be acknowledged that in many cases, associations cannot be excluded in specific cases either, considering that no epidemiology study will ever have the power to demonstrate rare occurrences.  In this context, the only solution is for the patient to understand that anything we do always carries a risk whether it is to be—or to not be—vaccinated, and that no doctor or scientist will ever be in a position to deliver 100 percent safe therapies.

The case of vaccines is also somehow more complicated than for other medicines, however, as it is not just a matter of weighing your own benefit/risk ratio, but also raises civic duty issues.   Indeed, vaccinations are, by design, not only protecting yourself but also blocking transmission of the disease to other people through you, a phenomenon called herd immunity. In this context, when many people are not vaccinated, the virus can transmit more rapidly and it is also a societal issue.

A finding published in Nature

In a paper published by Latorre et al. 2018, investigators used ultrasensitive screens to see if a small population of blood CD4 cells of narcolepsy patients reacted to hypocretin/orexin itself, or to another human protein that has been implicated before called TRIB2.

The work is a tour de force, involving isolating many individual CD4 cells in each patient that react to hypocretin or TRIB2, and making them grow in the presence of these antigens until hundreds of single cell lines reacting to these are generally obtained (narcolepsy) or not obtained (controls).  Testing 19 patients, they found that all had CD4 T cells that reacted to hypocretin/orexin and many reacted to TRIB2. Twelve of 12 controls were negative, even so they all had the main narcolepsy genetic factor called DQ0602, found in 25 percent of the population and almost necessary to develop the disease.  These CD4 cells recognized various pieces of hypocretin and TRIB2, suggesting a strong but somewhat diffuse immune reactivity that could explain why hypocretin/orexin cells were killed. A few CD8 cells with the same reactivity were found. No correlation with flu reactivity was found.

The works showed that many more CD4 T cells in the blood of narcolepsy patients react to hypocretin/orexin and TRIB2 than controls, suggesting these cells are recognizing hypocretin cells as “foreign” and initiating hypocretin cell killing, resulting in narcolepsy.  No connection with the flu was found, however, so that why flu such as the 2009 H1N1 swine influenza A and some flu vaccinations could sometimes induce narcolepsy is a question that stays unanswered.

The finding published in Proceedings of National Academy of Science in 2018

In parallel with this exciting research, our laboratory also followed this route of investigation, and came to a complementary conclusion.  As these studies have not yet gone through the scrutiny of peer review, they will only be discussed briefly, as they complement the Latorre et al. 2018 finding.  Our laboratory started with the influenza connection.  A few years ago, it was noticed that a piece of the swine H1N1 virus “resembled” a piece of hypocretin/orexin, and that these bound DQ0602, the genetic factor that could link these findings to CD4 T cells.  Pursuing this hypothesis, one of my students, Guo Luo, examined T cell reactivity to hundreds of pieces of the swine H1N1 (notably those contained in Pandemrix) when complexed with DQ0602, looking for a difference in reactivity between narcolepsy cases and controls.  

Luo also looked if these T cells reacted to hypocretin itself.  After an exhaustive search, he found that patients with narcolepsy had increased reaction to two pieces of flu virus, including the one resembling hypocretin/orexin.  He also found that CD4 cells of patients with narcolepsy also have increased reaction to the corresponding piece of hypocretin. Another important finding was that the piece of hypocretin had to be modified to an activated form for the immune system to recognize it as foreign, which may explain why the body is vulnerable to narcolepsy.  Finally, the CD4 cells recognizing these pieces of virus and hypocretin were isolated, and the T cell receptors recognizing these pieces of virus and hypocretin in conjunction with DQ0602 sequenced.

Another surprise was made when Aditya Ambati, another student in my lab, analyzed the CD4 receptor sequence data.  Indeed, we found that the receptors on CD4 T cells recognizing hypocretin and H1N1 were not only similar, explaining the confusion, but were also of a very specific subtype that was directly implicated as a genetic risk factor for narcolepsy.  This is a very important finding, as it showed that the CD4 recognition of hypocretin and H1N1 was not only a by-product of narcolepsy and hypocretin cell loss, but somehow causing narcolepsy.

The conclusion and what it all means

We are witnessing a revolution of our understanding of narcolepsy.  Although many details need to be filled in, it is likely that, as shown by Luo et al. 2018 narcolepsy is indeed caused by the immune system confusing specific pieces of flu, notably a particular segment of the swine flu that resembles hypocretin when attached to the genetic factor DQ0602.  The immune system then starts to attack hypocretin cells and the process becomes even more activated, recognizing many pieces of hypocretin/orexin, not just associated with DQ0602, but also in the context of a broader and stronger response, as detected by Latorre et al. 2018.  This all converges at destroying all hypocretin/orexin cells, resulting in narcolepsy.

Although many small pieces of the puzzle remain to be put in place, the work raises the possibility that blood tests for the diagnosis of narcolepsy will be developed.  In addition, it may be possible to modify the flu vaccine to prevent the development of narcolepsy in subjects who are susceptible genetically and carry DQ0602. This, together with the fact researchers are now on their way to developing hypocretin/orexin agonists that will replace the missing molecules in affected patients, promises big change is ahead for all narcolepsy patients.  

The work also has indirect implications for our understanding of many other autoimmune diseases, diseases that affect approximately 5 percent of individuals in the world.  It indeed offers a roadmap to understanding how abnormal immune response develop in these other diseases.

To read more:

Waking up Mathilda, by Claire Crisp,

Wide Awake and Dreaming: A Memoir of Narcolepsy by Julie Flygare,

“T cells in narcolepsy patients target self-antigens of hypocretin neurons.” Daniela Latorre et al., Nature

“Narcolepsy risk loci are enriched in immune cells and suggest autoimmune modulation of the T cell receptor repertoire.” Ollila et al., (non peer-reviewed)

“Narcolepsy onset is seasonal and increased following the 2009 H1N1 pandemic in China.” Han et al. Ann Neurol. 2011 Sep;70(3):410-7.

“Autoimmunity to hypocretin and molecular mimicry to flu antigens in Type 1 narcolepsy.”

Luo et al., Proceedings of the National Academy of Science, 2018 Dec 12.

Follow us here and subscribe here for all the latest news on how you can keep Thriving.

Stay up to date or catch-up on all our podcasts with Arianna Huffington here.


  • Dr. Emmanuel Mignot, M.D., Ph.D., is the Director of the Center for Sleep Sciences and Medicine at Stanford University.