ARTICLE IN BRIEF
Researchers found an increased frequency of antibodies to hypocretin receptor 2 in narcoleptic patients with a history of Pandemrix vaccination that were cross-reactive with influenza nucleoprotein.
Why did Pandemrix, a vaccine used during the 2009-2010 H1N1 influenza pandemic, seem to trigger a puzzling increase of nearly 1,300 reported cases of narcolepsy in Europe, while another H1N1 vaccine, Focetria, did not? Was the reaction related to the vaccine, or to the immune response to the virus itself?
Setting out to answer these questions, an international team of researchers analyzed sera from individuals who had been vaccinated with either one of the vaccines, some of whom later developed narcolepsy. They concluded in a July 1 report in Science Translational Medicine that differences in the viral proteins used to make the vaccines could explain the association between narcolepsy and the Pandemrix vaccine.
The researchers identified antibodies to a particular protein fragment contained in a piece of the virus used to make the Pandemrix vaccine. That fragment, part of the influenza nucleoprotein (NP), cross-reacted with human hypocretin receptor 2 (HCRT R2), which is involved in the sleep dysregulation observed in narcolepsy, they reported.
The antibodies were seen at much higher frequency in the sera of Finnish patients who had been vaccinated with Pandemrix before they were diagnosed with narcolepsy compared with sera from normal Italian individuals who had received Focetria in 2009 (p<0.001) or Finnish individuals who had the 2009 H1N1 infection (p<0.001).
“The findings suggest that natural infections can trigger an autoimmune response that involves molecular mimicry” — a similarity between a vaccine antigen and a hypocretin-related structure in the brain, said S. Sohail Ahmed, MD, a rheumatologist and co-author of the study with Lawrence Steinman, MD, a professor of pediatrics and neurology and neurological sciences at Stanford University School of Medicine, and their colleagues.
“The benefits of the influenza vaccine still far outweigh the risks,” said Dr. Ahmed, who was global head of clinical sciences at Novartis Vaccines in Italy up to the time of the study completion. He is now global head of clinical sciences at GSK Vaccines, the company that recently acquired the non-influenza business from Novartis Vaccines. “This is a very rare signal that was picked up in the backdrop of millions of individuals who were protected by and benefited from vaccination and did not develop narcolepsy.”
That said, Dr. Steinman added, “If this mechanism is confirmed, it suggests that natural influenza infection may pose a much greater risk for people at risk for narcolepsy, since we demonstrated that the immune response to influenza infection generated very high amounts of antibodies to influenza nucleoprotein compared to the adjuvanted influenza vaccine tested.”
Dr. Ahmed pointed out that scientists have long considered that narcolepsy might be an autoimmune disease because most patients have one particular histocompatibility type, known as HLA DQB1*0602. Almost all diseases associated with HLA types are autoimmune, he explained.
STUDY METHODOLOGY, FINDINGS
As a first step in their analysis, the researchers compared protein sequences in the viral proteins used to make the 2009 H1N1 influenza pandemic vaccines. They identified differences in three protein fragments from the influenza virus and compared these for similarity to fragments of the human HCRT ligand. However, unlike previous studies looking for mimicry, they also included in their analyses the two receptors normally bound by HCRT — HCRT receptor 1 and HCRT receptor 2. Using this approach, they identified a high similarity between the influenza virus NP and the hypocretin receptor 2.
“This was a key finding,” said Dr. Ahmed, “since the influenza virus nucleoprotein is highly conserved in many influenza strains and there were reports previously of narcolepsy following influenza infections from other years. Furthermore, another human sleep disorder called excessive daytime sleepiness has been linked to changes in the HCRT receptor 2 that impair hypocretin binding-related signaling through this receptor.”
Dr. Steinman added, “The similarity to hypocretin receptor 2 was intriguing, as mutations in the same region of the receptor of narcoleptic dogs have been demonstrated to block signaling by the HCRT ligand, and mice with loss of function of the HCRT receptor 2 also develop narcolepsy.”
The researchers next wanted to determine whether these findings translated into differences related to the pandemic influenza vaccines (Pandemrix or Focetria) or the immune response elicited in individuals vaccinated with them. Working with researchers at the National Institute for Health and Welfare in Finland, they accessed previously collected sera from narcolepsy patients who had the disease-associated HLA marker (HLA DQB1*0602) and had received the Pandemrix vaccine; sera from healthy-appearing individuals without this sleeping disorder, which were collected in 2004/2005, thus predating the circulation of the 2009 H1N1 pandemic virus; and from individuals who had the 2009 H1N1 influenza infection.
Finland had the highest rates of pediatric and adolescent influenza vaccinations in the world during the 2009 to 2010 H1N1 pandemic, providing the researchers with a rich set of clinical data and samples.
Similarly, collaborators at the University of Siena in Italy had collected sera from individuals who received Focetria, which was not associated with narcolepsy.
The researchers conducted laboratory tests on the influenza vaccines and on the sera obtained from Finland and Italy for detection of antibodies to influenza NP, antibodies to influenza hemagglutinin, assays for antibody binding to human HCRT receptors expressed on cell lines, microscopy, and antibody absorption studies for determining cross-reactivity.
Antibodies to the HCRT receptor 2 cross-reacted with influenza NP peptide in 17 of the 20 sera samples from people with narcolepsy who had received the Pandemrix vaccine in 2009 before onset of disease. By comparison, none of the six normal subjects who received Focetria in 2009 and five of the 20 samples from normal people with H1N1 infection in 2009 had antibodies against HCRT receptor 2.
“The studies on the antigen content of influenza vaccines revealed that the Pandemrix vaccine had significant levels of influenza nucleoprotein compared to the other vaccine tested,” Dr. Ahmed said. Focetria had 72 percent less nucleoprotein than Pandemrix.
“These findings suggested that one of the differences between the two vaccines was the quantity of an influenza virus protein that resembled a piece of the protein involved in sleep. This is molecular mimicry,” Dr. Ahmed explained. “And this made it less likely that the adjuvant by itself triggered narcolepsy.”
Other studies conducted in patients with autoimmune diseases receiving vaccines contained these and other adjuvants, with no increase in autoimmune disease activity, he noted.
The investigators hypothesized that antibodies to influenza nucleoprotein that enter the brain could occupy the hypocretin receptor 2 site and impair binding and signaling by HCRT, which is necessary to regulate normal sleep patterns. “Our proposed mechanism could end up having the same functional effect as is observed in animal models, where mutations or loss of function in HCRT receptor 2 results in narcolepsy. This will need to be investigated in future studies,” said Dr. Steinman.
“We need to follow up and test more numbers and types of samples,” Dr. Ahmed added. “We did the best we could with the serum samples that we had access to, which enabled us to investigate only the antibody-related immune response.”
No one knows what causes narcolepsy, and the authors of this latest paper say that there is not enough evidence to point a finger only at H1N1. In their study, five of the 20 cases that had no narcolepsy also had antibodies to NP. “That means something else must be needed to trigger narcolepsy,” said Dr. Steinman.
“It is complex science,” said Dr. Ahmed. “We needed expertise from rheumatology, neurology, immunology, and vaccinology to connect the dots.”
“We have not proven that narcolepsy is an autoimmune disease,” Dr. Steinman stressed. “This is an intriguing possibility, but there is a lot more work that needs to be done to make that connection.”
To prove that narcolepsy is autoimmune and linked to an immune response to the hypocretin receptor 2, it would be helpful to show that such antibodies cause experimental narcolepsy in an animal model, Dr. Steinman said. So far, it has been shown that knocking out the hypocretin receptor 2 is associated with narcolepsy in mice. And mutations in the hypocretin receptor 2 have been linked to narcolepsy in dogs.
“If an antibody to hypocretin receptor 2 triggered narcolepsy in an experimental model, it would finally prove the autoimmune hypothesis of narcolepsy,” he said.
In 2013, another team of investigators at Stanford, led by Emmanuel Mignot, MD, PhD, published a finding suggesting that T cells of the immune system, primed by H1N1 influenza hemagglutinin (HA), cross-reacted with HCRT and destroyed the hypocretin-producing neurons in the hypothalamus. Approximately seven months later, the Stanford investigators retracted the study after they found that they could not replicate their earlier results.
Commenting on the study, Clifford B. Saper, MD, PhD, the James Jackson Putnam professor of neurology and neuroscience at Harvard Medical School and chairman of neurology at Beth Israel Deaconess Medical Center, said he wasn't convinced that narcolepsy can be brought on by an abnormal antibody response triggered by influenza.
“The authors have discovered an antibody against the orexin 2 receptor [the hypocretin receptor 2], not orexin itself,” Dr. Saper said. “In narcolepsy there is selective loss of the hypothalamic orexin neurons but not their targets. so the findings may not be an explanation for narcolepsy occurring after H1N1 infection or immunization.”
But Thomas Scammell, MD, a professor of neurology at Harvard Medical School and Beth Israel Deaconess Medical Center who specializes in sleep medicine, said the authors' hypothesis is fundamentally a good one, with some caveats.
“They identified antibodies that cross-react with the orexin 2 receptor, but we don't know whether this causes narcolepsy,” Dr. Scammell said. “That they found antibodies in people without narcolepsy suggests the possibility that people can have antibodies but they don't mean anything clinically. I don't think they overstate their claim.”
He said he encourages his narcolepsy patients to get the influenza vaccine, and he will continue to do that.
Hartmut Wekerle, MD, director of the Max Planck Institute of Neurobiology and head of the department of neuro-immunology, and the author of an accompanying editorial on the current study, said one of the weaknesses of the paper is that “the authors were forced to rely on archival serum samples taken from narcolepsy patients several years ago for unrelated purposes. Thus, it was not feasible to compare precisely matched patient and control cohorts, in particular with regard to the time after receipt of the H1N1 vaccine or exposure to the H1N1 influenza virus.”
He added that the implications of the study are important. “First of all, the findings show that in rare cases, vaccines can accidentally trigger autoimmune-like brain disease. The anti-H1N1 vaccine Pandemrix appears to make the case, stimulating in some individuals the production of hypocretin receptor cross-binding antibodies as a result of its high content of influenza NP. The data draw attention to the protein composition of a given vaccine, but they do not permit prediction of autoimmune side effects.
“The work elegantly provides insight into vaccine-related narcolepsy, but does not directly disclose whether similar mechanisms (autoimmune or parainfectious, for example) contribute to sporadic narcolepsy,” he added.
He agreed with other independent commentators that more work is needed to explore that association.