Retraction of Paper “Confirming” Narcolepsy as an Autoimmune Disease

Today is a disappointing day in the narcolepsy community.The scientific
manuscript reported to have “confirmed” narcolepsy as an autoimmune disease has been formally retracted.

The retraction notice states, “The researchers report that they have been unable to reproduce the paper’s key findings.” (source


The paper, entitled  “CD4+ T cell autoimmunity to hypocretin/orexin and cross-reactivity to a 2009 H1N1 influenza A epitope in narcolepsy,” and published by Science Translational Medicine in December 2013, caused more than media ripple.  Submitted by co-senior authors Mellins and Mignot and collaborators, the original article was highlighted in more than 100 news and press releases, and received featured articles and posts on major news sites including Nature News, Scientific AmericanThe Huffington PostMedical News Today, National Geographic, Psychology Today — it even made waves in The New Reddit Journal of Science

The article has been cited 4 times since it’s publication in Dec. 2013, according to PubMedCentral.

At the time of publication, the paper was reported to demonstrate immune cross-reactivity between a flu antigen and the orexin protein — so-called molecular mimicry. In layman’s terms, if an immune response was generated recognizing flu antigen (such as during natural infection or in a vaccine), that immune response could also target neurons in the brain expressing orexin. Specifically reported was (1) characterization of narcolepsy-specific auto-immune CD4+ T cells, (2) their corresponding epitopes, and (3) evidence of a mimicry-based mechanism potentially explaining the association between narcolepsy and influenza infection.

Orexin is a hormone critical for maintaining wakefulness. Loss of orexin-producing neurons is believed to be the primary cause of symptoms in many individuals with narcolepsy. You can learn more about narcolepsy here.

These results were poised to explain the increases in narcolepsy following Pandemrix vaccinations in Europe during the 2009 H1N1 influenza pandemic.  Earlier studies suggested that the adjuvant used in the Pandemrix vaccine, AS03 (which was not approved for use in the United States) may have been to blame.

The findings of the paper in question caused a stir because for many years studies have failed to demonstrate orexin immunoreactivity or a cell-specific immune response to orexin producing neurons, despite the fact that narcolepsy contains other “autoimmune signatures” including a high (90%) association with a specific HLA genotype (HLA-DQB1*0602), and a strong association with the T-cell receptor alpha locus, among others.

A quote from the Nature News article summarizes the general sentiment well:

“Thomas Scammell, a neurologist at Harvard Medical School in Boston, Massachusetts, says that the results are welcome after “years of modest disappointment”, marked by many failures to find antibodies made by a person’s body against their own hypocretin. “It’s one of the biggest things to happen in the narcolepsy field for some time.”


Co-senior author of the study Mellins is quoted:

“Up till now, the idea that narcolepsy was an autoimmune disorder was a very compelling hypothesis, but this is the first direct evidence of autoimmunity. I think these cells are a smoking gun,”


With many basic science manuscripts which make significant discoveries, the excitement is often confined to the scientific community. What is different about this paper is that the enthusiasm bubbled over into patient circles and prominent outreach organizations. Narcolepsy has been long suspected to be of immune origin, and the fact that this theory hasn’t been yet confirmed has meant the often stimatized patient population continues to struggle with being incorrectly diagnosed with mental and behavioral disorders, often up to 15 years before receiving a correct diagnosis. One of the important advancements that might have been made from the discoveries of this paper was the development of a new diagnostic test for narcolepsy, which could improve diagnosis and time-to-diagnosis for patients.

Interestingly, a letter written by inventors of Pandemrix was submitted in February 2014 stating that  “CD4+ T cell cross reactivity” was part of a “research plan.(update 9-9/30/14: the letter has since been retracted). The research plan, authored by GlaxoSmithKline researchers is entitled “Narcolepsy and A(H1N1)pdm09 vaccination: Shaping the research on the obeserved signal” was published in December 2013.  In it, they listed key areas of research needed to fill the information gap about how the Pandemrix AS03-adjuvanted influenza vaccine could have caused increases in narcolepsy:

Therefore, the following key areas of research can be identified, (1) characterization of hypothetical narcolepsy-specific auto-immune CD4+ T cells, (2) mapping epitopes of such T cells, and (3) evaluating potential mechanisms that would enable such cells to gain access to the hypothalamus. Addressing these questions could further our understanding of the potential links between narcolepsy and A(H1N1)pdm09 vaccination and/or infection. Of particular interest is that any evidence of a mimicry-based mechanism could also explain the association between narcolepsy and A(H1N1)pdm09 influenza infection.


As a reminder, the reported findings of the now retracted “CD4+ T cell cross reactivity…” paper were: (1) characterization of narcolepsy-specific auto-immune CD4+ T cells, (2) their corresponding epitopes, and (3) evidence of a mimicry-based mechanism potentially explaining the association between narcolepsy and influenza infection.

A timeline of events is shown below:

retraction timeline crop

It should be noted that the publications of GlaxoSmithKline researchers did not reportedly contribute to the rectraction of “CD4+ T cell Autoimmunity…”

According to, Mignot has said regarding the retraction:

Mignot tells us:

We were just continuing our work based on the finding, trying to establish it as a diagnostic test, but could not replicate it.  No other work is affected, and in fact the DQ binding studies of that article are perfectly fine.  Only the [Enzyme-Linked ImmunoSpot (ELISPOT)] results are in question.

The retraction notice first appeared on the mobile site of Science Translational Medicine on July 23, 2014, but linked to a “Content Not Found” page. The “Retraction” section was missing from the non-mobile version of the website, and no reply was received when the Editors of STM were contacted to clarify whether the article had been retracted or not.

In spite of this disappointing turn of events, the community is hopeful that we will continue to move forward in our pursuit of understanding the molecular mechanisms of narcolepsy and that these developments will lead to improved diagnostic tools and treatments.


A. K. De la Herrán-Arita, B. R. Kornum, J. Mahlios, W. Jiang, L. Lin, T. Hou, C. Macaubas, M. Einen, G. Plazzi, C. Crowe, E. W. Newell, M. M. Davis, E. D. Mellins, E. Mignot, CD4+ T Cell Autoimmunity to Hypocretin/Orexin and Cross-Reactivity to a 2009 H1N1 Influenza A Epitope in Narcolepsy. Sci. Transl. Med. 5216ra176 (2013).

M. Brandt. “Narcolepsy is an Autoimmune Disorder, Researcher Says.”  Stanford Medicine News Center. (2009.)

K. Conger. “H1N1-triggered narcolepsy may stem from ‘molecular mimicry,’ study finds.” Standford Medicine News Center. (2013)

E. Yong. “Narcolepsy confirmed as autoimmune disease.” Nature News. (2013).

S. Harris. Notes on NarcolepsyPsychology Today. (2012).

J. Hallmayer et al. Narcolepsy is strongly associated with the T-cell receptor alpha locus. Nature Genetics. 41, 708 – 711 (2009)

Trouble with Tribbles?

Currently, the theory that narcolepsy is of a neurodegenerative autoimmune cause is the most widely supported among narcolepsy researchers, but the data to support this theory has been somewhat lacking (admittedly in part because there are simply fewer studies on narcolepsy than other autoimmune diseases).

In the first case, narcolepsy has a high HLA association, which is seen with other autoimmune diseases. In fact, narcoleptics with cataplexy have a 90% chance of being a carrier of the gene HLA DQB1*0602, an association among the highest of all autoimmune diseases.

Secondly, there is a reduction of hypocretin (a.k.a. orexin) producing neurons in the hypothalamus of narcoleptic patients, which leads to a reduction of hypocretin in the cerebrospinal fluid (CSF).  While it is still possible that the neurons are there and just not producing hypocretin, current data suggests there is a total loss of the neurons according to some imaging data via Mignot et al, which is covered in a separate post. In spite of this lack of hypocretin, no antibody to neurons or hypocretin have been found in narcoleptic patients.

So, if narcolepsy is indeed autoimmune and neurodegenerative, where is the proof?

Back in 2010, a paper came out in the Journal of Clinical Investigation suggesting that there was an association with narcolepsy and antibodies to a protein called Trib2 (tribbles homolog 2).  This work was redone by Mignot et al. and also replicated by a Japanese group, demonstrating that the antibodies seen in the study were found in multiple populations, including those of different ethnic origin. While tribbles is made by many cells in the brain, hypocretin-producting neurons make larger amounts of it, and it is theorized that these neurons could be preferentially attacked by the immune system.

In short, they found that 25% of individuals with the so-called narcolepsy gene (HLA DQB1*0602) with cataplexy had antibodies to Trib2, while antibodies to Trib2 were rare in individuals without cataplexy or controls. In addition, anti-trib2 antibody positivity correlated with disease progression, with antibody being higher in those individuals who were more recently diagnosed.

While the findings themselves are not disputable, it does beg the question of where the antibodies are coming from in the first place. Antibodies do not arise de novo to do the body harm, but rather are products of a sustained encounter by the body with antigen under the context of inflammation.  Antibodies are products of adaptive immunity; that is, B cells which have been activated through their B cell receptor in an antigen-specific manner, and have had the antigen presented to them by an antigen presenting cell. So the fact that antibodies against Trib2 are found does not necessarily mean that Trib2 is the primary target of the immune response, although it could be.  In reality, antibodies to Trib2 in narcoleptic patients could arise by several mechanisms, and may be secondary to neuronal cell death.  To me, this seems the most plausible explanation, as Trib2 is normally found intracellularly, and it is widely expressed in the brain.

In the first place, it is clear that an individual (probably) must be genetically susceptible to developing narcolepsy at some point in their lives. These individuals generally harbor the HLA DQB1*0602 gene, or a TCRalpha chain polymorphism.  On this background of genetic predisposition, environmental factors begin to play a role, as they do in other autoimmune diseases.  Whatever the inciting factor may be, the role of Trib2 antibodies is not clear, despite their association in narcoleptic patients. It is entirely possible that already damaged neurons are engulfed by the innate immune systems phagocytic cells, and their contents presented to adaptive immune systems B cells and T cells, which then coordinate to produce antibody to multiple components of the damaged neurons, one target of which may be Trib2.

While the finding of a self-protein or “autoantigen” response in narcolepsy is compelling, and serves an important piece in bolstering the autoimmune theory of disease progression in narcolepsy, it is by no means necessarily the causative agent or end of the story, particularly for individuals who do not have cataplexy or the classical HLA DQB1*0602 gene.  In this way, although an autoantibody has been found in some patients, a Trib2 antibody response is clearly not necessary or sufficient to cause narcolepsy in all cases.

In future studies, it will be interesting to determine if other autoantibodies will be found, or if there are multiple self-targets that may give rise to narcolepsy symptoms.


Narcolepsy gene in a celiac patient.

As many of you know, autoimmune diseases are generally complex diseases, where people on a certain genetic background (i.e. those whore are “predisposed”) come into contact with environmental triggers and eventually develop autoimmune disease.

Particularly important for autoimmune diseases, the genetic predisposition seems to lie within a region of our genes called Human Leukocyte Antigen or HLA for short. Narcolepsy has one of the highest genetic (HLA) associations, with more than 90% of individuals with narcolepsy possessing the MHC clas II gene HLA DQB1*0602.  

Celiac disease is also shown to be associated with certain HLA genes, in particular with MHC class II HLA-DQ2.  In this recent study, a patient with confirmed celiac disease was demonstrated to have the narcolepsy gene DQB1*0602.

Schizophrenia, autoantibodies and responses to gluten


Schizophrenia is a complex disease that many of us are peripherally familiar with, at least in part. For many of us, the term “schizophrenia” conjures up images of people talking to themselves, ranting and running naked in the park.

So why are we talking about it on Autoimmune Patient?  

Schizophrenia is not definitively recognized as an autoimmune disease, but it warrants attention here because there is increasing evidence that it may share some autoimmune etiology. Like all autoimmune diseases, there seems to be a genetic component to schizophrenia, although one responsible gene has not been found, indicating that environmental factors play a role in the development of the disease.  Celiac disease has long been suspected a being connected with schizophrenia, but a common link has yet to be found.

In the first place, at least for men, schizophrenia tends to appear in late adolescence/early adulthood, a feature that is also shared by many autoimmune diseases including multiple sclerosis, lupus, autoimmune thyroid disease, and narcolepsy, just to name a few.

Secondly, autoantibodies to various receptors (muscarinic, nicotinic, dopaminergic, and NMDA) are found in schizophrenia, and there is a strong association with schizophrenia and the risk of developing other autoimmune disease.   In a recent study, individuals with schizophrenia were reported to have an increased risk of Graves disease, psoriasis, autoimmune anemia, Sjogren’s syndrome, myasthenia gravis, and celiac disease.

More recently, it was discovered that individuals with schizophrenia elicit distinct immunological reactions to gluten that are clinically distinct form those reactions seen in celiac disease, and not controlled by the canonical “celiac disease genes” HLA DQ2/DQ8.