Article Text
Abstract
Objective To clarify whether genetic and common infectious backgrounds are distinct, according to anti-aquaporin 4 (AQP4) antibody status in Japanese patients with neuromyelitis optica (NMO).
Methods We analysed human leucocyte antigen (HLA)-DRB1 and HLA-DPB1 alleles, and IgG antibodies against Helicobacter pylori, Chlamydia pneumoniae, varicella zoster virus and Epstein–Barr virus nuclear antigen (EBNA) in 116 patients with NMO, including 39 patients with neuromyelitis optica spectrum disorder (NMOSD), 145 multiple sclerosis (MS) patients and 367 unrelated healthy controls. 77 NMO/NMOSD patients were seropositive for AQP4 antibody while 39 were seronegative.
Results Compared with healthy controls, NMO/NMOSD patients showed a significantly lower frequency of DRB1*0901 and significantly higher frequencies of DRB1*1602 and DPB1*0501, which conferred susceptibility to anti-AQP4 antibody positive NMO/NMOSD, but not antibody negative NMO/NMOSD. DRB1*0901 was a common protective allele, irrespective of the presence or absence of anti-AQP4 antibody. Anti-H pylori and anti-C pneumoniae antibodies were more commonly detected in anti-AQP4 antibody positive NMO/NMOSD patients than healthy controls. Antibody negative NMO/NMOSD patients did not differ from healthy controls regarding the presence of these antibodies. The presence or absence of antibodies against varicella zoster virus and EBNA did not vary among the groups. The frequencies of antibodies against these four pathogens were not significantly different between MS patients and healthy controls.
Conclusions Our results suggest that HLA-DRB1*1602 and DPB1*0501 alleles and H pylori and Chlamydia pneumonia infection are risk factors only for anti-AQP4 antibody positive NMO/NMOSD but not for anti-AQP4 antibody negative NMO/NMOSD.
- Multiple Sclerosis
- HLA
- Neurogenetics
Statistics from Altmetric.com
Introduction
Multiple sclerosis (MS) is an inflammatory demyelinating disease of the CNS. In contrast, neuromyelitis optica (NMO) is an inflammatory disease of the CNS selectively affecting the optic nerves and spinal cord. In this condition, longitudinally extensive spinal cord lesions (LESCLs) extending over three or more vertebral segments are regarded as characteristic. The nosological position of NMO has long been a matter of debate. However, the discovery of an IgG specific for NMO, initially designated NMO-IgG but now known to be anti-aquaporin 4 (AQP4) antibody, suggested that NMO is a distinct disease entity with a fundamentally different aetiology from MS.1 ,2 The classification of NMO has recently been expanded, and the limited form of NMO is now named neuromyelitis optica spectrum disorder (NMOSD).3
MS is rare in Asians, but selective and severe involvement of the optic nerve and spinal cord is characteristic when it does occur, and this form is termed opticospinal MS (OSMS).4 Because NMO-IgG/anti-AQP4 antibody is present in 30–60% of Japanese OSMS patients with LESCLs,5–8 OSMS in Asians is said to be the same entity as NMO.3
MS and allied disorders are thought to be caused by multiple contributing genetic and environmental factors. The largest genetic effect comes from the major histocompatibility complex class II region. In Caucasians, the allele human leucocyte antigen (HLA)-DRB1*1501 is associated with MS.9 In the Japanese population, several studies have reported that conventional MS is associated with HLA-DRB1*1501, while OSMS is associated with HLA-DPB1*0501.4 ,10 After the discovery of NMO-IgG/anti-AQP4 antibody, the DPB1*0501 allele was shown to be associated with NMO in Japanese11 and Southern Han Chinese.12 However, it is still unclear whether HLA class II association differs according to the presence or absence of anti-AQP4 antibody in NMO patients.
Regarding infectious factors in the environment, it has been repeatedly shown that Epstein–Barr virus (EBV) infection is more prevalent in Caucasian MS patients than in healthy controls (HCs); therefore, EBV infection has been assumed to increase susceptibility to MS.13 By contrast, in OSMS patients, we reported a higher frequency of Helicobacter pylori infection compared with that in HCs.14 H pylori infection was later shown to be associated with the presence of anti-AQP4 antibody.15 However, it remains to be elucidated whether NMO or the presence of anti-AQP4 antibody is associated specifically with H pylori infection or generally with some persistent bacterial or viral infections. Therefore, in the present study, we aimed to clarify whether immunogenetic background and common infection profiles are the same or differ between anti-AQP4 antibody positive and negative NMO patients. Thus, in addition to H pylori infection, we studied Chlamydia pneumoniae infection as a representative other type of chronic bacterial infection because we and others have shown that anti-AQP4 antibody positive OSMS and NMO patients develop a prominent Th17 and Th1 shift in CSF cytokine profiles16 ,17 and peripheral blood CD4 positive T cell cytokine profiles,18 ,19 and C pneumoniae can induce a predominantly Th1-type cytokine profile.20 EBV and varicella zoster virus infections were also investigated as representative types of chronic viral infection.
Methods
Subjects
All patients enrolled in the present study were thoroughly examined in the neurology departments of the university hospitals of the South Japan MS Genetic Consortium (see appendix for list of co-investigators) between 1987 and 2010. All NMO patients fulfilled the 2006 revised criteria for NMO.21 We defined NMOSD patients as those meeting either two absolute criteria plus at least one supportive criterion, or one absolute criterion plus more than one supportive criterion from the 2006 NMO criteria. MS patients were defined using the 2005 revised McDonald criteria for MS22 and included those who did not meet the above mentioned NMO/NMOSD criteria. Patients with primary progressive MS were excluded. In the present study, 116 patients who were diagnosed with NMO or NMOSD, 145 patients with MS and 367 unrelated HCs were enrolled. Among the 116 NMO/NMOSD patients, 77 were seropositive for anti-AQP4 antibody while 39 were seronegative.8 Among the same group, 77 patients were diagnosed with NMO and 39 with NMOSD. There were 65 patients with NMO and 12 patients with NMOSD among the anti-AQP4 antibody positive NMO/NMOSD subgroup while there were 12 patients with NMO and 27 patients with NMOSD among the antibody negative NMO/NMOSD subgroup. Among the 116 NMO/NMOSD patients in the present study, there were five cases (4.3%) showing seroconversion from negative to positive (4.3%) and five cases showing seroconversion from positive to negative (4.3%). We classified these seroconvertants as seropositive. Among the 39 seronegative NMO patients, anti-AQP4 antibody was assayed in the relapse phase (within 1 month after the onset of acute exacerbation) in 22 cases (56%) and in the remission phase (either the stable stage or more than 1 month after exacerbation) in 17 cases (44%). Among the 17 seronegative NMO patients whose anti-AQP4 antibody was assayed in the remission phase, four cases (24%) were administered prednisolone (5–20 mg/day), one case (6%) was administered both prednisolone (5 mg/day) and azathioprine (50 mg/day), and two cases (12%) were administered both prednisolone (5–15 mg/day) and interferon β-1b (8 million international units every other day). Finally, among the 22 seronegative NMO patients whose anti-AQP4 antibody was assayed in the relapse phase, four cases (18%) were receiving prednisolone (5–40 mg/day), one case (5%) was receiving azathioprine (50 mg/day), one case (5%) was receiving interferon β-1b (8 million international units every other day) and three cases (14%) were receiving both prednisolone (5–20 mg/day) and interferon β-1b (8 million international units every other day). We collected demographic data from the patients by retrospective review of their medical records. These data included gender, age of onset, disease duration, expanded disability status scale (EDSS) of Kurtzke scores,23 annualised relapse rate, progression index,24 CSF oligoclonal IgG bands (OB) and IgG index, brain MRI lesions that met the Barkhof criteria for MS25 and LESCLs extending over three or more than three vertebral segments. This study was approved by each of the institutional ethics committees.
Magnetic resonance imaging
All MRI studies were performed using 1.5 T units (Magnetom Vision and Symphony, Siemens Medical Systems, Erlangen, Germany), as previously described.6 ,8 Brain MRI lesions were evaluated according to the Barkhof criteria for MS.25 Spinal cord lesions extending over three or more vertebral segments in length were considered to be LESCLs.
HLA-DRB1 and HLA-DPB1 genotyping
The genotypes of the HLA-DRB1 and -DPB1 alleles of the subjects were determined by hybridisation between the products of PCR amplification of the HLA-DRB1 and HLA-DPB1 genes and sequence specific oligonucleotide probes, as described previously.11
Anti-AQP4 antibody assay
The presence of anti-AQP4 antibody was assayed, as described previously,26 using green fluorescent protein-AQP4 (M1 isoform) fusion protein transfected human embryonic kidney (HEK) cells. Serum samples diluted 1 : 4 were assayed for anti-AQP4 antibody, at least twice using identical samples, with the examiners blinded to the origin of the specimens. Samples that gave a positive result twice were deemed positive. When the judgment was equivocal, we measured the anti-AQP4 antibody using M23 isoform (green fluorescent protein-AQP4) transfected HEK cells.
Detection of anti-H pylori, anti-C pneumoniae, anti-varicella zoster virus and anti-EBV nuclear antigen IgG antibodies
Serum anti-H pylori (H pylori), anti-C pneumoniae (C pneumoniae), anti-varicella zoster virus and anti-EBV nuclear antigen IgG antibodies were measured using commercial ELISA kits according to the manufacturer's instructions (Vircell, Spain), as described previously.15 The antibody index was determined by dividing the optical density values for target samples by the optical density values for cut-off control samples and then multiplying by 10. As recommended by the manufacturer, an ELISA test index value was considered positive if >11, equivocal if between 9 and 11 and negative if <9. According to the manufacturer's instructions, the ELISA assay used in the present study has 100% sensitivity and 83% specificity for H pylori infection and 100% sensitivity and 93% specificity for C pneumoniae infection. Samples with equivocal results were retested for confirmation, and if samples were equivocal twice they were considered negative.
Statistical analyses
The phenotype frequencies of the HLA-DRB1 and HLA-DPB1 were compared using either the χ2 test or Fisher's exact probability test when the criteria for the χ2 test were not fulfilled. Uncorrelated p values (puncorr) were corrected by multiplying them by the number of comparisons indicated in the footnote of each table (Bonferroni–Dunn's correction) to calculate corrected p values (pcorr). Fisher's exact probability test was used to compare gender, the presence or absence of CSF IgG abnormalities and the frequencies of antibodies against common infectious agents among subgroups. Other demographic features were analysed using the Wilcoxon rank sum test. Brain MRI lesions that met the Barkhof criteria25 and LESCLs were not compared among NMO/NMOSD subgroups and the MS group because these items constitute one of the criteria for either NMO21 or MS.22 All analyses were performed using JMP 8.0.3 (SAS Institute, Cary, North Carolina, USA). In all assays, statistical significance was determined by p<0.05.
Results
HLA-DRB1 and HLA-DPB1 alleles in all NMO/NMOSD patients
Compared with HCs, NMO/NMOSD patients showed significantly higher frequencies of DRB1*1602 and DPB1*0501 (pcorr=0.0223, OR=8.988, 95% CI 2.344 to 34.468, and pcorr=0.0124, OR=2.394, 95% CI 1.424 to 4.023, respectively) and a significantly lower frequency of DRB1*0901 (pcorr<0.0001, OR=0.169, 95% CI 0.076 to 0.376) (tables 1 and 2).
HLA-DRB1 and HLA-DPB1 alleles in NMO/NMOSD patients according to the presence or absence of anti-AQP4 antibody
Compared with HCs, anti-AQP4 antibody positive NMO/NMOSD patients demonstrated significantly higher frequencies of DRB1*1602 and DPB1*0501 (pcorr=0.0080, OR=12.133, 95% CI 3.063 to 48.059, and pcorr=0.0074, OR=3.175, 95% CI 1.619 to 6.227, respectively) and a significantly lower frequency of DRB1*0901 (pcorr=0.0023, OR=0.183, 95% CI 0.072 to 0.466) (tables 1 and 2). However, anti-AQP4 antibody negative NMO/NMOSD patients only showed a significantly lower frequency of DRB1*0901 (pcorr=0.0406, OR=0.142, 95% CI 0.034 to 0.602).
HLA-DRB1 and HLA-DPB1 alleles in NMO/NMOSD patients according to NMO and NMOSD classification
Compared with HCs, NMO patients showed significantly higher frequencies of DRB1*1602 and DPB1*0501 (pcorr=0.0080, OR=12.133, 95% CI 3.063 to 48.059, and pcorr=0.0368, OR=2.605, 95% CI 1.382 to 4.910, respectively) and a significantly lower frequency of DRB1*0901 (pcorr=0.0016 OR=0.144, 95% CI 0.051 to 0.405) (see supplementary tables S1 and S2, available online only). NMOSD patients also showed similar trends, with a higher frequency of DPB1*0501 and lower frequency of DRB1*0901, but the differences did not reach statistical significance owing to the small sample size.
Comparison of demographic features among patients with anti-AQP4 antibody positive and negative NMO/NMOSD and MS
We compared demographic features among patients with anti-AQP4 antibody positive and negative NMO/NMOSD and MS (table 3). The proportion of women was significantly higher among patients with anti-AQP4 antibody positive NMO/NMOSD than among patients with MS (pcorr<0.0001). Age at onset was significantly higher in patients with anti-AQP4 antibody positive NMO/NMOSD than in patients with MS (pcorr=0.0001). EDSS scores were significantly higher in patients with NMO/NMOSD, regardless of anti-AQP4 antibody status, than in patients with MS (pcorr<0.0001, both groups). The annualised relapse rate was significantly higher in patients with anti-AQP4 antibody positive NMO/NMOSD than in patients with MS (pcorr<0.0001), and tended to be higher in patients with anti-AQP4 antibody positive NMO/NMOSD than in patients with anti-AQP4 antibody negative NMO/NMOSD (pcorr=0.0597). The progression index was significantly higher in patients with NMO/NMOSD, regardless of anti-AQP4 antibody status, than in patients with MS (pcorr<0.0001, pcorr=0.0237 for antibody positive and negative patients, respectively). The frequency of OB/increased IgG index was lower in patients with anti-AQP4 antibody positive NMO/NMOSD than in patients with MS (pcorr=0.0012).
Comparison of demographic features among patients with NMO, NMOSD and MS
We compared demographic features among patients with NMO, NMOSD and MS (see supplementary table S3, available online only). The proportion of women was significantly higher among patients with NMO than among patients with MS (pcorr=0.0002). Age at onset was also significantly higher in patients with NMO and NMOSD than in patients with MS (pcorr=0.0063, pcorr=0.0105, respectively). EDSS scores were significantly higher in patients with NMO and NMOSD than in patients with MS (pcorr<0.0001, both). The annualised relapse rate was significantly higher in patients with NMO than in patients with MS (pcorr=0.0002). The progression index was significantly higher in patients with NMO and NMOSD than in patients with MS (pcorr<0.0001, pcorr=0.0029, respectively). The frequency of OB/increased IgG index was lower in patients with NMO than in patients with MS (pcorr=0.0009).
Positivity rates for anti-H pylori, anti-C pneumoniae, anti-varicella zoster virus and anti-EBV nuclear antigen IgG antibodies
Finally, we measured levels of antibodies against common infectious agents present in anti-AQP4 antibody positive NMO/NMOSD patients, anti-AQP4 antibody negative NMO/NMOSD patients, MS patients and HCs (table 4). Anti-H pylori IgG antibodies were detected significantly more frequently in patients with anti-AQP4 antibody positive NMO/NMOSD than in patients with antibody negative NMO/NMOSD or MS, and HCs (pcorr=0.0264, pcorr=0.0008, pcorr=0.0088, respectively); however, there was no difference in the frequency of anti-H pylori IgG antibodies between anti-AQP4 antibody negative NMO/NMOSD patients and HCs. In addition, anti-C pneumoniae IgG antibodies were detected significantly more frequently in anti-AQP4 antibody positive NMO/NMOSD patients than in HCs (pcorr=0.0284), and tended to be more frequently detected in anti-AQP4 antibody positive NMO/NMOSD patients than in patients with MS (pcorr=0.0624) while, again, such a trend was not seen in anti-AQP4 antibody negative NMO/NMOSD patients. Neither EBV nuclear antigen nor anti-varicella zoster virus antibody positivity rates differed significantly among the four subgroups.
Discussion
The main new findings in the present study are: (1) DRB1*1602 and DPB1*0501 confer susceptibility to anti-AQP4 antibody positive NMO/NMOSD but not to anti-AQP4 antibody negative NMO/NMOSD; (2) DRB1*0901 is a common protective allele, irrespective of the presence or absence of anti-AQP4 antibody; and (3) compared with HCs, anti-H pylori and anti-C pneumoniae antibodies are more commonly seen in anti-AQP4 antibody positive NMO/NMOSD patients, but they are no more commonly seen in anti-AQP4 antibody negative NMO/NMOSD patients.
We used a cell based assay using the M1 isoform of AQP4 under the condition of a 1 : 4 serum dilution to increase sensitivity without affecting specificity.26 Moreover, we measured anti-AQP4 antibody using HEK cells transfected with the M23 isoform of AQP4 when the judgment was equivocal in an assay using M1 isoform. The positivity rate (66%) for anti-AQP4 antibody among NMO/NMOSD patients in this study was comparable with those (61.5%) recently reported from other institutions in Japan.7 Therefore, we believe that the anti-AQP4 antibody positivity rate determined by our cell based assay using either M1 or M23 isoform did not severely distort the results. Although Jarius et al27 reported that anti-AQP4 antibody titres increased during relapse relative to levels during the remission phase, we found no correlation of anti-AQP4 antibody titres with any clinical parameters, nor did we find any significant difference in titres between the relapse and remission phases in our series.26 Several reports from other institutions have also found no correlation of anti-AQP4 antibody titres with any clinical parameters.28 Because positive to negative seroconvertants comprised only a small proportion (<5%) of the patients in our series and because we included seroconvertants in the seropositive group where possible, we believe that the disease stages and immunomodulatory medications in our patients did not significantly affect the results of the present study.
In the anti-AQP4 antibody negative subgroup, the proportion of patients with NMOSD was larger than that in the antibody positive subgroup. According to the revised NMO criteria,21 patients who have both LESCLs and MS-like brain lesions on MRI are classified as having NMO if they have anti-AQP4 antibody while patients negative for anti-AQP4 antibody are classified as having NMOSD. This surely led to an increased proportion of NMOSD patients in the anti-AQP4 antibody negative subgroup. In the present study, we compared the demographic features and frequencies of HLA-DRB1 and HLA-DPB1 alleles between NMO and NMOSD patients and found that both groups showed similar trends with the exception of a higher frequency of HLA-DRB1*1602 in NMO patients. This supports the notion that the observed differences in HLA-DRB1 and HLA-DPB1 associations other than HLA-DRB1*1602 are attributable to anti-AQP4 antibody status rather than NMO/NMOSD classification.
In the present study, we clearly showed that the DPB1*0501 allele, a well known risk allele for OSMS and NMO in Japanese10 ,11 and Southern Han Chinese,12 increased susceptibility to anti-AQP4 antibody positive NMO/NMOSD, but not to anti-AQP4 antibody negative NMO/NMOSD. The same was true for the DRB1*1602 allele, which was also reported to be a risk factor for NMO in Southern Han Chinese.12 We found that anti-AQP4 antibody positive NMO/NMOSD patients tended to have higher relapse rates. These observations are partially consistent with the previous finding that anti-AQP4 antibody is associated with frequent relapses.8 Therefore, it is conceivable that the pathogenic mechanisms underlying anti-AQP4 antibody positive NMO/NMOSD may differ from those underlying the anti-AQP4 antibody negative type, and that development of AQP4 autoimmunity is in part based on genetic background (in Asians by DPB1*0501 and DRB1*1602 and in Caucasians by other HLA class II genes, such as DRB1*03).29 ,30
In the current study, it is noteworthy that anti-H pylori and anti-C pneumoniae IgG antibodies were significantly more prevalent only in anti-AQP4 antibody positive NMO/NMOSD patients, but not in the anti-AQP4 antibody negative group, compared with HCs. The ELISA assays used in the present study have reasonably high sensitivity and specificity,31 ,32 although H pylori and C pneumoniae infection should be confirmed by methods other than ELISA in future studies. The significance of C pneumoniae infection in MS has long been a matter of debate. However, a meta-analysis reported that C pneumoniae infection did not show a strong association with the risk of MS.33 In contrast, the increased EBV infection rate reported in MS13 was not observed in NMO/NMOSD patients, regardless of anti-AQP4 antibody status. These observations collectively suggest that the signature of common infections may be distinct between MS and NMO/NMOSD patients, and especially in anti-AQP4 antibody positive NMO/NMOSD patients.
Although frequent H pylori infection has previously been reported in anti-AQP4 antibody positive NMO/NMOSD patients,14 ,15 the results of the present study indicate that anti-AQP4 antibody positive NMO/NMOSD patients may have experienced frequent common bacterial infections. However, no specific infection has yet been clearly associated with the induction of anti-AQP4 antibody positive NMO/NMOSD, suggesting that antigen independent mechanisms may be involved. Microbial infection stimulates the innate immune system, resulting in enhanced expression of costimulatory molecules and proinflammatory cytokines.34 H pylori infection has been shown to induce Th1- and Th17-type cytokine production,35 while C pneumoniae infection produces a predominantly Th1-type cytokine profile.20 We and others have shown that anti-AQP4 antibody positive OSMS and NMO patients develop a prominent Th17 and Th1 shift in CSF cytokine profiles16 ,17 and peripheral blood CD4 positive T cell cytokine profiles.18 ,19 Therefore, chronic bacterial infections may enhance development of NMO/NMOSD through activation of Th17/Th1 cells, which may contribute to the induction of CNS inflammation in NMO patients.16 ,17 In particular, H pylori infection may contribute to the pathology of anti-AQP4 antibody related neural damage by inducing systemic inflammatory stimulators such as interleukin 17, which would activate neutrophils in NMO patients in whom systemic increases in the levels of myeloperoxidase36 and complement factors37 are also observed. Alternatively, in some bacterial infection associated autoimmune diseases, molecular mimicry between self and bacterial antigens has been implicated.34 In addition, regression of idiopathic thrombocytopenic purpura after eradication of H pylori may be explained by molecular mimicry between platelet and H pylori antigens.38 Although certain bacteria contain bacterial aquaporins, neither H pylori nor C pneumoniae do.39 However, it remains possible that proteins other than bacterial aquaporins from H pylori and C pneumoniae may have cross epitope mimicry with human AQP4.
In the present study, we did not observe any significant association of the genetic and infectious factors examined with anti-AQP4 antibody negative NMO/NMOSD, except for DRB1*0901 as a protective factor. Our study found that DRB1*0901, one of most prevalent DRB1 alleles in the Japanese population, had a strong protective effect against NMO/NMOSD, regardless of anti-AQP4 antibody status. It was previously reported that the DRB1*0901 allele is protective against MS in Japanese patients.11 Furthermore, a recent meta-analysis in Chinese patients disclosed that the DRB1*0901 allele is protective against MS in Chinese patients.40 The observation that NMO/NMOSD patients have at least one protective factor in common with MS patients, regardless of anti-AQP4 antibody status, may suggest the existence of a common mechanism(s) between the two conditions in some disease cascades. To date, no specific contributing factors have been identified for anti-AQP4 antibody negative NMO/NMOSD. Thus the further elucidation of pathogenic factors in this type of NMO/NMOSD is warranted. Alternatively, anti-AQP4 antibody negative NMO/NMOSD may represent a heterogeneous mixture of several disease phenotypes.
Acknowledgments
We thank Dr Junji Kishimoto (Centre for Clinical and Translational Research, Kyushu University Hospital, Kyushu University) for his assistance with the statistical analyses.
Appendix
Co-investigators
The following are members of the South Japan Multiple Sclerosis Genetic Consortium: Drs Katsuichi Miyamoto (Kinki University, Site Investigator), Susumu Kusunoki (Kinki University, Chairman), Yuji Nakatsuji (Osaka University, Site Investigator), Hideki Mochizuki (Osaka University, Chairman), Kazuhide Ochi (Hiroshima University, Site Investigator), Masayasu Matsumoto (Hiroshima University, Chairman), Takeshi Kanda (Yamaguchi University, Chairman), Hirofumi Ochi (Ehime University, Site Investigator), Tetsuro Miki (Ehime University, Chairman), Kazumasa Okada (University of Occupational and Environmental Health, Site Investigator) and Sadatoshi Tsuji (University of Occupational and Environmental Health, Chairman).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online tables
Footnotes
SY and NI contributed equally to this work.
-
Contributors SY, NI, TM and J-IK conceived and designed the study. SY, NI, TM and SS performed the experiments. SY and NI analysed and interpreted the data. All authors contributed reagents/materials/analysis tools. SY, NI and J-IK wrote the paper. All authors reviewed, amended and agreed on the final version of the manuscript.
-
Funding This work was supported in part by a Health and Labour Sciences Research Grant on Intractable Diseases (H20-Nanchi-Ippan-016) from the Ministry of Health, Labour and Welfare, Japan, and a grant-in-aid (B; No. 22390178) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
-
Competing interests J-IK is an advisory board member for Merck Serono and a consultant for Biogen Idec Japan. He has received payment for lectures from Bayer Schering Pharma, Cosmic Cooperation and Biogen Idec Japan. TM received a grant and payment for manuscript preparation and development of educational presentations from Bayer Schering Pharma, and also received a payment for the development of educational presentations from Mitsubishi Tanabe Pharma.
-
Ethics approval This study was approved by each of the institutional ethics committees.
-
Provenance and peer review Not commissioned; externally peer reviewed.