Antibodies to influenza nucleoprotein cross-react with human hypocretin receptor 2

In their recent publication in Science Translational Medicine, Ahmed et al. proposed that the increased risk of narcolepsy following vaccination with Pandemrix™ could potentially be explained by antibody cross-reactivity between the H1N1 influenza virus nucleoprotein (NP) and the human hypocretin receptor 2 (HCRT-R2) (1). Subsequent comments in Science (2) and Science Translational Medicine (3) discussed the relevance of these findings in explaining any potential links between vaccination, influenza and narcolepsy.

The starting point for Ahmed and colleagues was the observation that very few cases of narcolepsy were reported after vaccination with Focetria™ thus prompting a search for sequence differences between the re-assortant vaccine strains used to produce Pandemrix and Focetria (seed virus strains X-179A and X-181, respectively). Where differences between X-181 and X-179A sequences were found, they searched for sequence homologies between these influenza virus sequences, and the sequences of hypocretin and hypocretin receptors 1 and 2. The bio-informatic analysis was then based on the premise that a single amino acid difference existed at position 116 of the nucleoprotein (NP) sequences in the re-assortant strains for Pandemrix and Focetria, X-179A (I) and X-181 (M), respectively. This analysis resulted in the identification of amino acid similarity between the influenza A(H1N1)pdm09 virus nucleoprotein (NP) (YDKEEIRRIWR, defined as the mimicry sequence) and the human hypocretin receptor 2 (HCRT-R2) (YDDEEFLRYLWR). Next, it was shown that antibodies recognizing HCRT-R2 were cross-reactive with the influenza NP mimicry sequence. Cross-reactive antibodies were detected in sera from narcoleptic patients with a history of Pandemrix vaccination but also in control sera. The authors propose that the findings provide a potential link between Pandemrix, as well as influenza virus infection, and narcolepsy, based on the induction of (HCRT-R2) receptor-blocking cross-reactive antibodies, while acknowledging that the mechanism by which such cross-reactive antibodies induce narcolepsy symptoms remains elusive. The authors further conclude that the “findings shift the focus away from the adjuvant alone and instead suggest the H1N1 viral antigens, and particularly viral NP, as the prime suspect in the search for the trigger of vaccine-induced narcolepsy”. We would like to make several comments on these results and their interpretation.

1. Ahmed et al show that the M/I amino acid difference has no effect on the capacity of peptide to compete with sera for HCRT-R2 binding but that it does affect DQB1*0602-binding. It was shown that the I-containing peptide binds DQB1*0602 whereas the M-containing peptide does not (1). However, our sequence analysis of the seed lots for Pandemrix and Arepanrix™ revealed that both contain an M at position 116 (Figure 1), as does Focetria. This is consistent with Genbank database entry KJ942731. Thus, Pandemrix and Focetria contain the same peptide sequence that may mimic HCRT-R2 for antibody recognition and that does not bind DQB1*0602. It is not clear whether DQB1*0602-binding of the cross-reactive peptide is essential for induction of cross-reactive antibodies. The inability of a candidate mimicry peptide to bind DQB1*0602 may question a role for that peptide in the induction of narcolepsy.

2. The link between the detection of cross-reactive antibodies and narcolepsy is not fully clear. Cross-reactive antibodies were detected not only in sera from Finnish pediatric patients with history of narcolepsy but also in sera from Finnish children not known to have narcolepsy and individuals that had (H1N1)pdm09 infection. As stated by the authors, this could mean that cross-reactive antibodies are required but not sufficient to induce narcolepsy symptoms (1). Interpretation of the results is complicated by the fact that the narcolepsy patients are expected to be DQB1*0602+, whereas this is unlikely to be the case for the all Focetria control subjects (for which no HLA-typing data were generated). As shown previously by Vaarala et al (4), the DQB1*0602 allele itself is important for antibody responses against NP. The incidence of DQB1*0602 in Focetria vaccinated Italian subjects is ~1.5% (1), implying that differences in DQB1*0602 status among the subjects from whom samples were tested could be important.

3. The authors’ proposed mode of action for anti-HCRT-R2 antibodies involves receptor blockade, leading to impaired hypocretin signalling (1). Indeed, deficient HCRT receptor signalling can lead to induction of narcolepsy symptoms, as shown in narcoleptic dogs carrying a mutation in HCRT-R2 (5, 6). HCRT-R2-deficient narcoleptic dogs appear to have normal hypocretin levels (7) suggesting that a negative feedback loop between HCRT-R2 deficiency and hypocretin levels is not a default mechanism. However, and as pointed out by the authors, a mechanism of receptor blockade does not explain why narcolepsy patients, including pediatric patients in Finland with history of Pandemrix vaccination (8) have hypocretin deficiency.

4. Although HCRT-R2-specific antibodies are detected in narcolepsy patients as well as in healthy subjects, the link between the presence of these antibodies and their induction is not yet clear. Putative cross-reactivity with a peptide in NP is demonstrated by competition experiments, but it has not been demonstrated that these antibodies are actually induced or boosted by H1N1 influenza virus infection or vaccination with an inactivated influenza vaccine.

5. NP contents of various pandemic and seasonal influenza vaccines were compared by gel electrophoresis and western blotting using NP-specific monoclonal antibodies. Pandemrix was shown to contain the highest amount of NP. Relative to Pandemrix, Arepanrix and Focetria were reported to contain 28% and 73% less NP respectively, which is likely to be attributable to different processes for Pandemrix and Focetria. However, these values were only based on NP/HA ratio and not on absolute amount of NP in the vaccines. Of note, Focetria appears to contain twice as much HA as Pandemrix and Arepanrix, leading to an absolute decrease of NP in Focetria of 36.5% and in Arepanrix of 14% as compared to Pandemrix. H1N1 infection is also likely to trigger high levels of NP antibodies.

We conclude that the paper of Ahmed et al. identifies for the first time cross-reactive antibodies binding influenza nucleoprotein and HCRT-R2. However, several questions remain to be answered before conclusions on potential links between influenza vaccines and/or infection and increased risk of narcolepsy can be drawn. The authors clearly acknowledge this in their Discussion and we agree with that view.

Footnote: Pandemrix and Arepanrix are trademarks of the GSK group of companies. Focetria is a trademark of Novartis AG.

REFERENCES

(1) S.S. Ahmed, W. Volkmuth, J. Duca, L. Corti, M. Pallaoro, A. Pezzicoli, A. Karle, F. Rigat, R. Rappuoli, V. Narasimhan, I. Julkunen, A. Vuorela, O. Vaarala, H. Nohynek, F. L. Pasini, E. Montomoli, C. Trombetta, C. M. Adams, J. Rothbard, L. Steinman, Antibodies to influenza nucleoprotein cross-react with human hypocretin receptor 2. Sci. Transl. Med. 7, 294ra105 (2015).
(2) G. Vogel, Narcolepsy link to pandemic flu vaccine becomes clearer. Science 349, 17 (2015).
(3) H. Wekerle, Vaccination and narcolepsy: Immune link found? Sci. Transl. Med. 7, 294fs27 (2015).
(4) O. Vaarala, A. Vuorela, M. Partinen, M. Baumann, T. L. Freitag, S. Meri, P. Saavalainen, M. Jauhiainen, R. Soliymani, T. Kirjavainen, P. Olsen, O. Saarenpää-Heikkilä, J. Rouvinen, M. Roivainen, H. Nohynek, J. Jokinen, I. Julkunen, T. Kilpi, Antigenic differences between AS03 adjuvanted influenza A (H1N1) pandemic vaccines: implications for Pandemrix-associated narcolepsy risk. PLoS. One. 9, e114361 (2014).
(5) L. Lin, J. Faraco, R. Li, H. Kadotani, W. Rogers, X. Lin, X. Qiu, P. J. de Jong, S. Nishino, E. Mignot, The sleep disorder canine narcolepsy is caused by a mutation in the hypocretin (orexin) receptor 2 gene. Cell 98, 365-376 (1999).
(6) B. Ripley, N. Fujiki, M. Okura, E. Mignot, S. Nishino, Hypocretin levels in sporadic and familial cases of canine narcolepsy. Neurobiol. Dis. 8, 525-534 (2001).
(7) M.F. Wu, R. Nienhuis, N. Maidment, H. A. Lam, J. M. Siegel, Role of the hypocretin (orexin) receptor 2 (Hcrt-r2) in the regulation of hypocretin level and cataplexy. J. Neurosci. 31, 6305-6310 (2011).
(8) F. Pizza, H. Peltola, T. Sarkanen, K. K. Moghadam, G. Plazzi, M. Partinen, Childhood narcolepsy with cataplexy: comparison between post-H1N1 vaccination and sporadic cases. Sleep Med. 15, 262-265 (2014).