Correspondence
Does autoreactivity have a role in narcolepsy? In our recent Review of the sleep disorder narcolepsy,1 we included the hypothesis that type 1 narcolepsy is an autoimmune disease. As part of this discussion, we identified hypocretin as a possible target of an autoimmune attack leading to narcolepsy. At the time of writing, this idea was supported by data from a report by De la Herrán-Arita and colleagues,2 which has since been retracted. The retraction was based on an inability of the researchers to reproduce their finding of T-cell reactivity towards hypocretin and H1N1 antigenic epitopes.2 In view of this retraction, no direct evidence exists of autoreactive T cells playing a part in the pathophysiology of narcolepsy. The autoimmune hypothesis of narcolepsy, however, is based on many other observations described in our Review. Briefly, these observations include the following: a strong association with HLA class II DQB1*06:02 and genetic associations with other genes involved in the regulation of host immune responses (such as TCRα, OX40L, CTSH, and P2RY11); and environmental associations with upper airway infections, Streptococcus pyogenes, and, importantly, influenza A virus H1N1 vaccination and infection. Furthermore, several articles have reported the presence of autoimmune antibodies in serum samples from patients with narcolepsy. A common theme in these studies is that only a minority of patients is seropositive for the autoantibody under investigation, which is the case for hypocretin antibodies and tribbles 2 antibodies, and some patients were recently shown to have other brain-specific autoantibodies.3 Three immunohistochemical patterns were shown in rat brain. For one of these antibody staining patterns, the target epitope was the common 1072
C-terminal epitope of neuropeptide glutamic acid-isoleucine/α-melanocyte-stimulating hormone (NEI/αMSH); for the others, the target epitope is unknown. None of the autoantibodies mentioned targeted hypocretin neurons and, curiously, these autoantibodies were detected in the serum of patients with other sleep disorders such as idiopathic hypersomnia, KleineLevin syndrome, obstructive sleep apnea, delayed sleep phase, restless legs syndrome and periodic limb movements and parasomnias, and even in some healthy controls. The results argue against these autoantibodies being related directly to the pathogenesis of narcolepsy, but NEI/αMSH IgG antibodies from a patient with narcolepsy affected sleep patterns when injected intracerebroventricularly in mice, which suggests some functional relevance. 3 Intracerebroventricular injection of immunoglobulins from patients with anti-tribbles 2 antibodies also affected sleep patterns in mice.4 These results support the idea that autoantibodies might be involved in the symptomatology of narcolepsy. Autoantibodies might be produced as part of a secondary immune response to the initial neuronal damage in narcolepsy, or they might contribute to the development of narcolepsy by an as yet unknown mechanism. Reports that the H1N1 Pandemrix vaccinations and H1N1 infection have been associated with an increased incidence of narcolepsy strongly point to an immune-mediated pathogenesis but since the retraction of the report by De la Herrán-Arita and colleagues,2 direct evidence of autoimmunity (table 2 in our Review) is weaker than the evidence seemed to be at the time of writing.1 The retraction of the article2 means that we do not have any clear evidence for molecular mimicry between hypocretin and H1N1 virus. A paucity of evidence, however, does not exclude
the possibility that molecular mimicry might be at play in the pathogenesis of H1N1 virus infection-induced narcolepsy. Other non-proteinaceous components of the vaccine might also have an important role. The notion that autoimmune antibodies have a role in the pathogenesis of narcolepsy is supported by many studies. However, final proof of autoimmunity directed towards hypocretin neurons is still missing, because autoreactive T cells against hypothalamic neurons or neuronal molecules have not been shown. Additionally, autoantibodies against neuronal or other host factors have been shown only in a subgroup of patients. No convincing animal model of narcolepsy-type disease exists to show that the disorder can develop from the transfer of autoreactive T cells or antibodies to host factors. Future research should aim to address these gaps in our knowledge. We declare no competing interests.
*Markku Partinen, Birgitte Rahbek Kornum, Giuseppe Plazzi, Poul Jennum, Ilkka Julkunen, Outi Vaarala markku.partinen@helsinki.fi Helsinki Sleep Clinic, Vitalmed Research Centre, Helsinki, Finland (MP); Department of Clinical Neurosciences (MP) and Institute of Clinical Medicine (OV), University of Helsinki, Helsinki, Finland; Molecular Sleep Laboratory, Department of Diagnostics (BRK) and Danish Center for Sleep Medicine, Department of Clinical Neurophysiology, University of Copenhagen (BRK, PJ), Glostrup Hospital, Glostrup, Denmark; Department of Biomedical and Neuromotor Sciences, University of Bologna, Italy (GP); IRCCS Istituto delle Scienze Neurologiche, Bologna, Italy (GP); Department of Virology, University of Turku, Turku, Finland (IJ); Department of Infectious Disease Surveillance and Control, National Institute for Health and Welfare, Helsinki, Finland (IJ); and Respiratory, Inflammatory and Autoimmune Diseases, Innovative Medicine, AstraZeneca, Mölndal, Sweden (OV) 1
2
Partinen M, Kornum BR, Plazzi G, Jennum P, Julkunen I, Vaarala O. Narcolepsy as an autoimmune disease: the role of H1N1 infection and vaccination. Lancet Neurol 2014; 13: 600–13. De la Herrán-Arita AK, Kornum BR, Mahlios J, et al. Retraction of the Research Article: “CD4+ T cell autoimmunity to hypocretin/orexin and cross-reactivity to a 2009 H1N1 influenza A epitope in narcolepsy”. Sci Transl Med 2014; 6: 247rt1.
www.thelancet.com/neurology Vol 13 November 2014
Correspondence
3
4
Bergman P, Adori C, Vas S, et al. Narcolepsy patients have antibodies that stain distinct cell populations in rat brain and influence sleep patterns. Proc Natl Acad Sci USA 2014; 111: E3735–44. Katzav A, Arango MT, Kivity S, et al. Passive transfer of narcolepsy: anti-TRIB2 autoantibody positive patient IgG causes hypothalamic orexin neuron loss and sleep attacks in mice. J Autoimmun 2013; 45: 24–30.
Simvastatin in subarachnoid haemorrhage: beyond the short-term We are grateful to the STASH trial1 investigators for their efforts to advance the use of simvastatin in subarachnoid haemorrhage from bench to bedside. The results of the STASH trial do not support the use of 40 mg simvastatin daily to decrease delayed cerebral infarction and to improve long-term clinical outcome. The investigators, therefore, concluded that patients with subarachnoid haemorrhage should not be treated routinely with simvastatin. Unfortunately, the STASH trial and a similar study by Wong and colleagues2 underway to evaluate a dose of simvastatin higher than that used in STASH were not designed to evaluate neurocognitive outcomes. Cognitive dysfunction, such as memory deficits and executive dysfunction, is increasingly recognised as an important, and perhaps even the most prevalent, long-term complication of subarachnoid haemorrhage.3,4 We are also concerned that the length of the treatment period—up to 3 weeks—was not sufficient to investigate potential benefits of simvastatin on outcomes after subarachnoid haemorrhage. The pathophysiology of subarachnoid haemorrhage is poorly understood, as is our understanding of the duration of action of underlying mechanisms. Ischaemic stroke (eg, delayed cerebral infarction after subarachnoid haemorrhage) results in microglial activation that might persist for several
months, potentially increasing brain injury;5 and endothelial injury can be observed throughout the arterial tree for months after subarachnoid haemorrhage.6 In our opinion, it would be reasonable to continue treatment with simvastatin until the damaged thrombogenic and dysfunctional cerebrovascular endothelium has recovered fully and the neuroinflammatory responses have subsided. In fact, experimental studies that replicate cognitive decline in subarachnoid haemorrhage show that the beneficial effects of simvastatin on regional cerebral blood flow, motor function, cognitive function, and neuroprotection are forfeited when simvastatin is stopped after 2 weeks of treatment. A sustained improvement in cognitive function and neuronal survival was shown only when simvastatin administration was extended for the duration of the study7 (6 weeks). Although the STASH trial investigators report that simvastatin at a dose of 40 mg for up to 3 weeks did not prevent delayed cerebral infarction, it is unclear whether longer use could improve long-term outcomes. Statins are well tolerated with long-term use. Perhaps it is time to depart from the 2–3 week cerebral vasospasm treatment paradigm, and extend the treatment period in clinical trials of interventions for subarachnoid haemorrhage. Otherwise, we risk false negatives and might prematurely discard potentially useful treatments. We declare no competing interests.
*Frederick Lombard, Gavin W Britz, David S Warner [email protected] Duke University, NC 27705, USA 1
2
www.thelancet.com/neurology Vol 13 November 2014
Kirkpatrick PJ, MSc CLT, Turner CL, Hutchinson PJ, Murray GD, STASH Collaborators FTS. Simvastatin in aneurysmal subarachnoid haemorrhage (STASH): a multicentre randomised phase 3 trial. Lancet Neurol 2014; 13: 666–75. Wong GKC, Liang M, Tan H, et al. High-dose simvastatin for aneurysmal subarachnoid hemorrhage: a multicenter, randomized, controlled, double-blind clinical trial protocol. Neurosurgery 2013; 72: 840–44.
3
4
5 6
7
Anderson SW, Todd MM, Hindman BJ, et al. Effects of intraoperative hypothermia on neuropsychological outcomes after intracranial aneurysm surgery. Ann Neurol 2006; 60: 518–27. Al-Khindi T, Macdonald RL, Schweizer TA. Cognitive and functional outcome after aneurysmal subarachnoid hemorrhage. Stroke 2010; 41: e519–36. Thiel A, Heiss WD. Imaging of microglia activation in stroke. Stroke 2011; 42: 507–12. Hughes JT, Schianchi PM. Cerebral artery spasm. A histological study at necropsy of the blood vessels in cases of subarachnoid hemorrhage. J Neurosurg 1978; 48: 515–25. Takata K, Lombard FW, Sheng H, Laskowitz D, Borel CO, Warner DS. A rat model of cerebral vasospasm resulting in chronic motor and neurocognitive deficits following experimental SAH. J Neurosurg Anesthesiol 2006; 18: 295–301.
Authors’ reply We welcome the comments by Frederick Lombard and colleagues in relation to the publication of results from the STASH trial.1 An important issue is raised of neurocognitive deficits after subarachnoid haemorrhage, which are often apparent despite a seemingly good physical outcome. The Rankin Disability Scale does not evaluate cognitive outcome but it is unusual in our experience for patients with severe cognitive impairment to fall into the category of a good outcome in terms of their mRS score. Although we did not do cognitive assessments in the STASH trial, we used the well characterised SF-36 questionnaire, which does substantially explore non-physical outcomes. Data for 656 patients did not show any significant differences in the mental, physical, or overall scores of those receiving statin therapy compared with patients receiving placebo. The details of these assessments are the subject of further post-hoc analyses to be published . Lombard and colleagues also invite commentary with respect to the duration of drug administration in trials of subarachnoid haemorrhage. Although concern exists that the processes involved in tissue damage after subarachnoid haemorrhage might be ongoing for some weeks after the ictus, most patients who 1073
Does autoreactivity have a role in narcolepsy? In our recent Review of the sleep disorder narcolepsy,1 we included the hypothesis that type 1 narcolepsy is an autoimmune disease. As part of this discussion, we identified hypocretin as a possible target of an autoimmune attack leading to narcolepsy. At the time of writing, this idea was supported by data from a report by De la Herrán-Arita and colleagues,2 which has since been retracted. The retraction was based on an inability of the researchers to reproduce their finding of T-cell reactivity towards hypocretin and H1N1 antigenic epitopes.2 In view of this retraction, no direct evidence exists of autoreactive T cells playing a part in the pathophysiology of narcolepsy. The autoimmune hypothesis of narcolepsy, however, is based on many other observations described in our Review. Briefly, these observations include the following: a strong association with HLA class II DQB1*06:02 and genetic associations with other genes involved in the regulation of host immune responses (such as TCRα, OX40L, CTSH, and P2RY11); and environmental associations with upper airway infections, Streptococcus pyogenes, and, importantly, influenza A virus H1N1 vaccination and infection. Furthermore, several articles have reported the presence of autoimmune antibodies in serum samples from patients with narcolepsy. A common theme in these studies is that only a minority of patients is seropositive for the autoantibody under investigation, which is the case for hypocretin antibodies and tribbles 2 antibodies, and some patients were recently shown to have other brain-specific autoantibodies.3 Three immunohistochemical patterns were shown in rat brain. For one of these antibody staining patterns, the target epitope was the common 1072
C-terminal epitope of neuropeptide glutamic acid-isoleucine/α-melanocyte-stimulating hormone (NEI/αMSH); for the others, the target epitope is unknown. None of the autoantibodies mentioned targeted hypocretin neurons and, curiously, these autoantibodies were detected in the serum of patients with other sleep disorders such as idiopathic hypersomnia, KleineLevin syndrome, obstructive sleep apnea, delayed sleep phase, restless legs syndrome and periodic limb movements and parasomnias, and even in some healthy controls. The results argue against these autoantibodies being related directly to the pathogenesis of narcolepsy, but NEI/αMSH IgG antibodies from a patient with narcolepsy affected sleep patterns when injected intracerebroventricularly in mice, which suggests some functional relevance. 3 Intracerebroventricular injection of immunoglobulins from patients with anti-tribbles 2 antibodies also affected sleep patterns in mice.4 These results support the idea that autoantibodies might be involved in the symptomatology of narcolepsy. Autoantibodies might be produced as part of a secondary immune response to the initial neuronal damage in narcolepsy, or they might contribute to the development of narcolepsy by an as yet unknown mechanism. Reports that the H1N1 Pandemrix vaccinations and H1N1 infection have been associated with an increased incidence of narcolepsy strongly point to an immune-mediated pathogenesis but since the retraction of the report by De la Herrán-Arita and colleagues,2 direct evidence of autoimmunity (table 2 in our Review) is weaker than the evidence seemed to be at the time of writing.1 The retraction of the article2 means that we do not have any clear evidence for molecular mimicry between hypocretin and H1N1 virus. A paucity of evidence, however, does not exclude
the possibility that molecular mimicry might be at play in the pathogenesis of H1N1 virus infection-induced narcolepsy. Other non-proteinaceous components of the vaccine might also have an important role. The notion that autoimmune antibodies have a role in the pathogenesis of narcolepsy is supported by many studies. However, final proof of autoimmunity directed towards hypocretin neurons is still missing, because autoreactive T cells against hypothalamic neurons or neuronal molecules have not been shown. Additionally, autoantibodies against neuronal or other host factors have been shown only in a subgroup of patients. No convincing animal model of narcolepsy-type disease exists to show that the disorder can develop from the transfer of autoreactive T cells or antibodies to host factors. Future research should aim to address these gaps in our knowledge. We declare no competing interests.
*Markku Partinen, Birgitte Rahbek Kornum, Giuseppe Plazzi, Poul Jennum, Ilkka Julkunen, Outi Vaarala markku.partinen@helsinki.fi Helsinki Sleep Clinic, Vitalmed Research Centre, Helsinki, Finland (MP); Department of Clinical Neurosciences (MP) and Institute of Clinical Medicine (OV), University of Helsinki, Helsinki, Finland; Molecular Sleep Laboratory, Department of Diagnostics (BRK) and Danish Center for Sleep Medicine, Department of Clinical Neurophysiology, University of Copenhagen (BRK, PJ), Glostrup Hospital, Glostrup, Denmark; Department of Biomedical and Neuromotor Sciences, University of Bologna, Italy (GP); IRCCS Istituto delle Scienze Neurologiche, Bologna, Italy (GP); Department of Virology, University of Turku, Turku, Finland (IJ); Department of Infectious Disease Surveillance and Control, National Institute for Health and Welfare, Helsinki, Finland (IJ); and Respiratory, Inflammatory and Autoimmune Diseases, Innovative Medicine, AstraZeneca, Mölndal, Sweden (OV) 1
2
Partinen M, Kornum BR, Plazzi G, Jennum P, Julkunen I, Vaarala O. Narcolepsy as an autoimmune disease: the role of H1N1 infection and vaccination. Lancet Neurol 2014; 13: 600–13. De la Herrán-Arita AK, Kornum BR, Mahlios J, et al. Retraction of the Research Article: “CD4+ T cell autoimmunity to hypocretin/orexin and cross-reactivity to a 2009 H1N1 influenza A epitope in narcolepsy”. Sci Transl Med 2014; 6: 247rt1.
www.thelancet.com/neurology Vol 13 November 2014
Correspondence
3
4
Bergman P, Adori C, Vas S, et al. Narcolepsy patients have antibodies that stain distinct cell populations in rat brain and influence sleep patterns. Proc Natl Acad Sci USA 2014; 111: E3735–44. Katzav A, Arango MT, Kivity S, et al. Passive transfer of narcolepsy: anti-TRIB2 autoantibody positive patient IgG causes hypothalamic orexin neuron loss and sleep attacks in mice. J Autoimmun 2013; 45: 24–30.
Simvastatin in subarachnoid haemorrhage: beyond the short-term We are grateful to the STASH trial1 investigators for their efforts to advance the use of simvastatin in subarachnoid haemorrhage from bench to bedside. The results of the STASH trial do not support the use of 40 mg simvastatin daily to decrease delayed cerebral infarction and to improve long-term clinical outcome. The investigators, therefore, concluded that patients with subarachnoid haemorrhage should not be treated routinely with simvastatin. Unfortunately, the STASH trial and a similar study by Wong and colleagues2 underway to evaluate a dose of simvastatin higher than that used in STASH were not designed to evaluate neurocognitive outcomes. Cognitive dysfunction, such as memory deficits and executive dysfunction, is increasingly recognised as an important, and perhaps even the most prevalent, long-term complication of subarachnoid haemorrhage.3,4 We are also concerned that the length of the treatment period—up to 3 weeks—was not sufficient to investigate potential benefits of simvastatin on outcomes after subarachnoid haemorrhage. The pathophysiology of subarachnoid haemorrhage is poorly understood, as is our understanding of the duration of action of underlying mechanisms. Ischaemic stroke (eg, delayed cerebral infarction after subarachnoid haemorrhage) results in microglial activation that might persist for several
months, potentially increasing brain injury;5 and endothelial injury can be observed throughout the arterial tree for months after subarachnoid haemorrhage.6 In our opinion, it would be reasonable to continue treatment with simvastatin until the damaged thrombogenic and dysfunctional cerebrovascular endothelium has recovered fully and the neuroinflammatory responses have subsided. In fact, experimental studies that replicate cognitive decline in subarachnoid haemorrhage show that the beneficial effects of simvastatin on regional cerebral blood flow, motor function, cognitive function, and neuroprotection are forfeited when simvastatin is stopped after 2 weeks of treatment. A sustained improvement in cognitive function and neuronal survival was shown only when simvastatin administration was extended for the duration of the study7 (6 weeks). Although the STASH trial investigators report that simvastatin at a dose of 40 mg for up to 3 weeks did not prevent delayed cerebral infarction, it is unclear whether longer use could improve long-term outcomes. Statins are well tolerated with long-term use. Perhaps it is time to depart from the 2–3 week cerebral vasospasm treatment paradigm, and extend the treatment period in clinical trials of interventions for subarachnoid haemorrhage. Otherwise, we risk false negatives and might prematurely discard potentially useful treatments. We declare no competing interests.
*Frederick Lombard, Gavin W Britz, David S Warner [email protected] Duke University, NC 27705, USA 1
2
www.thelancet.com/neurology Vol 13 November 2014
Kirkpatrick PJ, MSc CLT, Turner CL, Hutchinson PJ, Murray GD, STASH Collaborators FTS. Simvastatin in aneurysmal subarachnoid haemorrhage (STASH): a multicentre randomised phase 3 trial. Lancet Neurol 2014; 13: 666–75. Wong GKC, Liang M, Tan H, et al. High-dose simvastatin for aneurysmal subarachnoid hemorrhage: a multicenter, randomized, controlled, double-blind clinical trial protocol. Neurosurgery 2013; 72: 840–44.
3
4
5 6
7
Anderson SW, Todd MM, Hindman BJ, et al. Effects of intraoperative hypothermia on neuropsychological outcomes after intracranial aneurysm surgery. Ann Neurol 2006; 60: 518–27. Al-Khindi T, Macdonald RL, Schweizer TA. Cognitive and functional outcome after aneurysmal subarachnoid hemorrhage. Stroke 2010; 41: e519–36. Thiel A, Heiss WD. Imaging of microglia activation in stroke. Stroke 2011; 42: 507–12. Hughes JT, Schianchi PM. Cerebral artery spasm. A histological study at necropsy of the blood vessels in cases of subarachnoid hemorrhage. J Neurosurg 1978; 48: 515–25. Takata K, Lombard FW, Sheng H, Laskowitz D, Borel CO, Warner DS. A rat model of cerebral vasospasm resulting in chronic motor and neurocognitive deficits following experimental SAH. J Neurosurg Anesthesiol 2006; 18: 295–301.
Authors’ reply We welcome the comments by Frederick Lombard and colleagues in relation to the publication of results from the STASH trial.1 An important issue is raised of neurocognitive deficits after subarachnoid haemorrhage, which are often apparent despite a seemingly good physical outcome. The Rankin Disability Scale does not evaluate cognitive outcome but it is unusual in our experience for patients with severe cognitive impairment to fall into the category of a good outcome in terms of their mRS score. Although we did not do cognitive assessments in the STASH trial, we used the well characterised SF-36 questionnaire, which does substantially explore non-physical outcomes. Data for 656 patients did not show any significant differences in the mental, physical, or overall scores of those receiving statin therapy compared with patients receiving placebo. The details of these assessments are the subject of further post-hoc analyses to be published . Lombard and colleagues also invite commentary with respect to the duration of drug administration in trials of subarachnoid haemorrhage. Although concern exists that the processes involved in tissue damage after subarachnoid haemorrhage might be ongoing for some weeks after the ictus, most patients who 1073
