Multiple sclerosis: risk factors, prodromes, and potential causal pathways

www.thelancet.com/neurology Vol 9 July 2010 727

Review

Lancet Neurol 2010; 9: 727–39

See In Context page 659

Blizard Institute of Cell and Molecular Science, Queen Mary University of London, Barts and The London School of Medicine and Dentistry, London, UK (S V Ramagopalan DPhil, R Dobson MRCP, U C Meier DPhil, Prof G Giovannoni FRCP); and Wellcome Trust Centre for Human Genetics, University of Oxford, and Department of Clinical Neurology, John Radcliff e Hospital, Oxford, UK (S V Ramagopalan)

Correspondence to: Prof Gavin Giovannoni, Blizard Institute of Cell and Molecular Science, Queen Mary University of London, Barts and The London School of Medicine and Dentistry, 4 Newark Street, London E1 2AT, [email protected]

Multiple sclerosis: risk factors, prodromes, and potential causal pathwaysSreeram V Ramagopalan, Ruth Dobson, Ute C Meier, Gavin Giovannoni

Multiple sclerosis (MS) is a common, complex neurological disease. The precise aetiology of MS is not yet known, although epidemiological data indicate that both genetic and environmental factors are important. The evidence that the environment acts long before MS becomes clinically evident is well established and suggests the existence of a prodromal phase for the disease. The increasing incidence of MS emphasises the need for strategies to prevent this chronic disorder, and the possibility of a prodrome indicates a window of opportunity to potentially reverse early disease processes before clinical disease becomes evident. Studying a prodrome requires techniques other than clinical observation such as monitoring endophenotypes that result from associated risk factors. However, our current knowledge of causal pathways and endophenotypes in MS is limited. Identifying and studying individuals with a high risk of developing the disease provides a powerful opportunity to understand the MS causal cascade and is highly relevant to strategies that are aimed at preventing this debilitating disease.

IntroductionMultiple sclerosis (MS) is the most common disease of the CNS to cause permanent disability in young adults.1 On the basis of strong circumstantial evidence, MS is thought to be an organ-specifi c autoimmune disorder,1,2 but much remains to be understood about the initiation of the disease. MS seems unlikely to result from a single causative event; instead, the disease seems to develop in genetically susceptible populations as a result of environmental exposures.

The concept of a prodrome is being studied intensively in neurological disease, and is defi ned as the time period between the onset of decline in a baseline level of functioning until criteria for disease diagnosis are met.3 The constellation of symptoms in a prodrome tends to be non-specifi c, particularly in the early stages. Thus, prodromal symptoms are not prospectively deterministic, and research is directed towards identifying which patients with prodromal symptoms will later develop disease.3 The fi ndings of these studies will help us to understand the trajectory of changes in the disease process from genetic risk factors to clinical diagnosis.3

The question of whether there is a prodrome in MS has so far not been extensively studied. A diagnosis of MS is made after a historical review of events in a patient’s life, fi ndings observed on neurological examination, data acquired from diagnostic tests, and after the exclusion of other diseases that could account for the clinical and paraclinical fi ndings.4 Patients with MS typically present with a clinically isolated syndrome (CIS), which is defi ned by a distinct fi rst neurological event with observed demyelination involving the optic nerve, cerebrum, cerebellum, brainstem, or spinal cord.5 CIS has, until recently, been thought to be the fi rst sign of MS. However, radiological abnormalities might be identifi ed in the absence of clinical symptoms, leading to use of the term “radiologically isolated syndrome” (RIS) to specifi cally describe individuals who have structural anomalies in the CNS that are highly suggestive of demyelination.6 Individuals with RIS are at

increased risk of developing MS.6 This observation, in conjunction with evidence for known MS risk factors that act many years before disease onset, renders unsurprising the notion that the disease develops subclinically. Thus, the existence of a prodrome in MS is compelling, but has so far received little attention. Endophenotypes, measurable components unseen by the unaided eye along the pathway of disease development, are also likely to exist in MS.7 In this Review, we summarise current understanding of disease susceptibility to elucidate potential causal pathways and endophenotypes, and suggest avenues for further study.

Genetic risk factorsThe development of MS must start in individuals who are genetically susceptible. The importance of genetic factors in susceptibility to MS has been shown by genetic epidemiological studies.8 Family studies assessing risks to relatives of MS probands have revealed a marked familial aggregation of the disease.9–12 First-degree relatives are generally at 10–25 times greater risk of developing MS than the general population. This risk correlates with degree of kinship (fi gure 1), with parent-of-origin eff ects and sex infl uencing the size of these risks.9–12

Genetic associationsHLA types exert the strongest genetic eff ect in MS, but the association is not straightforward. Whereas association with HLA-DR2 (HLA-DRB1*15) has long been known in northern Europe (heterozygosity conferring an odds ratio [OR] of 2·7, and homozygosity of 6·7),13,14 in other regions (eg, Sardinia), association is predominantly seen with HLA-DRB1*0301, HLA-DRB1*0405, and HLA-DRB1*1303.15 Revisiting the HLA association in northern European MS populations uncovered other haplotypes (HLA-DRB1*03, HLA-DRB1*01, HLA-DRB1*10, HLA-DRB1*11, HLA-DRB1*14, and HLA-DRB1*08) that are both positively and negatively associated with the disease, diff er in magnitude of eff ect, and either act on their own or have an eff ect in combination with another haplotype (fi gure 2).16–19

728 www.thelancet.com/neurology Vol 9 July 2010

Review

The relative frequencies of susceptibility and protective HLA haplotypes, which vary between countries, play a part in determining the distribution of the disease.19,20 Even more complexity is likely to exist at the HLA region than has been described to date, and more associations are likely to be found.21

Genome-wide association studies,22 and subsequent replication eff orts,23 have uncovered other genes with modest eff ects in MS, including interleukin-7 receptor α (IL7RA), interleukin-2 receptor α (IL2RA), C-type lectin-domain family 16 member A (CLEC16A), CD58 (formerly lymphocyte function-associated antigen 3; CD58), tumour-necrosis-factor receptor superfamily member 1A (TNFRSF1A), interferon regulatory factor 8 (IRF8), and CD6 (CD6).22

Ethnic originEthnic origin is also thought to play a part in the development of MS, with some groups being at higher risk than others. Studies that have controlled for confounding factors have shown lower prevalence of MS in African Americans (African American men had approximately 40% lower MS risk than white men), Native Americans, Mexicans, Puerto Ricans, and Japanese, and a virtual absence of the disease in Chinese and Filipinos.24 This eff ect is almost certainly genetically determined.

Parent-of-origin eff ectEpidemiological data strongly hint at a maternal parent-of-origin eff ect in MS.25–27 For example, maternal half-siblings of patients with MS have nearly double the risk of MS compared with paternal half-siblings (2·35% vs

1·31%, p=0·048).12 Risk of MS in maternal half-siblings compared with full siblings does not diff er signifi cantly (2·35% vs 3·11%, p=0·1), suggesting that this parent-of-origin eff ect is a major component of familial aggregation of the disease.12 The mechanism of the increased risk conferred maternally remains to be elucidated, but epigenetic mechanisms (DNA and chromatin alterations, including DNA methylation of C-G dinucleotides and histone modifi cations that regulate genomic function28) have been strongly implicated.29

SexMS is more common in women than in men.30–33 However, genome-wide studies have failed to provide any convincing support for any MS-associated genes on the X chromosome, so the increased incidence of MS in women might be related to female-specifi c physiology, and could thus be hormone related.34 Furthermore, the increase in female MS incidence over time suggests that this sex eff ect also makes women more susceptible to MS environmental risk factors.30

Environmental risk factorsAlthough genes are needed for MS to develop, genetic epidemiological studies clearly illustrate a prominent role for the environment in determining MS risk. Factors with the strongest evidence for involvement in MS are Epstein-Barr virus (EBV), smoking, and latitude/vitamin D. Reports on other factors such as geographical region and data from migration studies suggest that the timing of exposure is a crucial determinant of risk for MS, particularly for some factors that operate very early in life.

Figure 1: Age-adjusted percentage recurrence risks for relatives of multiple sclerosis probands Data are mean (standard error). Data from Willer et al.11

General population

ChildFirst cousinNiece/nephewAunt/uncle Parent Dizygotic twin

Sibling Monozygotic twin

Recu

rrenc

e ris

k (%

)

30

20

15

10

5

0

40

35

25

Relationship to index case

Relative of male probandRelative of female proband

www.thelancet.com/neurology Vol 9 July 2010 729

Review

Epstein-Barr virusNearly all individuals with MS (>99%) have been found to have been infected with EBV compared with approximately 94% of age-matched controls.35 Thus, MS is very rare in adults who have not been infected with EBV (OR of MS in EBV-negative vs EBV-positive individuals is 0·06 [95% CI 0·03–0·13]).35,36 People with high titres of anti-EBV antibodies have a higher risk of developing MS compared with those with low titres.37,38 The relationship seems to be temporal: plasma antibody titres against the EBV nuclear antigen 1 (EBNA1) increase several years before the onset of neurological symptoms of MS.37,38 In addition, individuals with a history of infectious mononucleosis have an increased risk of developing MS. A systematic review and meta-analysis of 14 case-control and cohort studies reported a combined conservative OR of MS after infectious mononucleosis of 2·3 (95% CI 1·7–3·0).36 This risk has subsequently been confi rmed in large population-based studies.39,40

In paediatric cases, the association between EBV infection and the development of MS is weaker. In a North American study, 108 (86%) children with MS were seropositive for remote EBV infection, compared with 61 (64%) matched controls.41 Another North American study on paediatric patients with MS reported greater than 91% seropositivity for EBV compared with 57% in controls.42 In a European seroprevalence study of 147 paediatric patients, 145 (99%) children with MS had detectable anti-body against EBV compared with 106 (72%) age-matched controls.43 MS diagnosis is more diffi cult in children, and some uncertainty remains about the nosology of MS in

this group. Therefore, long-term follow-up is crucial to establish whether EBV-seronegative children with an MS diagnosis subsequently have typical MS as adults. There is also an apparent paradox in that EBV infection in children increases the risk of paediatric MS, whereas delayed infection, as manifested by infectious mononucleosis, is associated with adult onset MS. However, the temporal sequence in both paediatric and adult onset MS supports causality, and given that high viral titres in both adults and children increases MS risk,38,42 infectious mono-nucleosis is likely to be a marker for an abnormal immune response, and it is this abnormal response that underlies MS risk rather than later infection per se.

Because EBV infection is nearly ubiquitous in adults with MS, inhibiting EBV infection would seem to be a potential option for preventing MS. However, the implications of a preventive EBV vaccination strategy at a population level are currently unknown and would need to be carefully explored.

SmokingA recent retrospective meta-analysis gave a pooled OR for developing MS of 1·51 (95% CI 1·24–1·83) for ever versus never smoking.44 Earlier studies showed a dose-dependent relation to MS risk.45 In a Swedish study, snuff use did not increase the risk of MS (OR 0·3 [95% CI 0·1–0·8]), suggesting that factors present in smoked tobacco or the route of administration are important.46 Promoting smoking cessation might be one of the most straight forward interventions available to reduce the incidence of MS.

Figure 2: Genotypic odds ratios for multiple sclerosis for combinations of alleles at the HLA-DRB1 locusX/X=individual with no disease-associated alleles with baseline odds ratio of 1·0 (dotted line). X=any non-disease-associated allele. Numbers (01, 08, 10, 11, 14, 15, 17) indicate HLA-DRB1 alleles associated with multiple sclerosis. Data have been log-transformed. Odds ratios from Ramagopalan et al.19

15/15–2·0

–1·5

–0·5

0·5

0

–1·0

1·0

2·0

2·5

1·5

15/17 X/15 11/15 17/17 X/17 14/15 X/X

HLA-DRB1 genotype

Log

(odd

s rat

io)

11/17 X/11 11/14 14/17 11/11 X/14

730 www.thelancet.com/neurology Vol 9 July 2010

Review

Latitude Within regions of temperate climate, MS incidence and prevalence increase with latitude.47 In Australia, the prevalence of MS in Hobart, southern Tasmania, is 75·6 per 100 000 compared with a prevalence of 11 per 100 000 in northern Queensland.48 However, more complex patterns of disease distribution do exist. In Norway, MS prevalence does not increase with latitude, but does negatively correlate with proximity to coastal fi shing areas and subsequent fi sh consumption.49 Some of the geographical distribution of MS can be explained on the basis of ethnicity and genetic factors,50 but latitude remains the strongest factor for risk after controlling for ethnic origin.31 Despite these fi ndings, the latitude eff ect has decreased over the past few decades (OR of MS, 2·02 for northern vs southern US states in Vietnam veterans, and 2·64 for the earlier-born World War II veterans).31

Vitamin DSunlight exposure and associated vitamin D status are potential explanations for the association between latitude and MS incidence.51 Past sunlight exposure has been found to be inversely related to MS susceptibility.52

Although questionnaire studies are prone to recall bias, the eff ect of sun exposure on MS has been confi rmed by the objective measure of actinic damage,52 with greater damage associated with decreased odds of MS (OR 0·32 [95% CI 0·11–0·88] for grade 4–6 damage). However, the timing of damage could not be determined in this retrospective study.52

Experimental and epidemiological data suggest that vitamin D is the mediator of the sunlight eff ect. Consumption of fatty seafood and cod liver oil in Norway, both rich sources of vitamin D, has been noted to provide protection against the risk of MS,49 although this outcome might also arise from the biological eff ects of omega-3 fatty acids.49 A prospective cohort study found that taking vitamin supplementation that included vitamin D was associated with about a 40% reduction in the risk of developing MS,53 although the amounts of vitamin D taken are thought to be insuffi cient to make much change in circulating vitamin D concentrations,54 and eff ects of multivitamin intake might be confounded by behavioural diff erences.55 The strongest evidence for a role for vitamin D comes from a prospective, nested case-control study in US military personnel, in which

Population studied

Findings

Alter et al62 Israel Prevalence of MS in immigrants from northern and central Europe was 30–51 per 100 000; from southern Europe, 9–18 per 100 000; from North Africa, 6 per 100 000 population. Prevalence of MS among native inhabitants of Israel was 4 per 100 000

Alter et al63 Israel Incidence of MS in European immigrants increased sharply if migration occurred after the age of 15 years

Dean64 South Africa UK immigrants to South Africa had an MS prevalence of 51 per 100 000 compared with a prevalence of 11 per 100 000 in white South-African-born individuals

Alter et al65 Hawaii Prevalence of MS in Japanese was 7 per 100 000 compared with 11 per 100 000 in white Hawaiian and 34 per 100 000 in white immigrants to Hawaii

Dean and Kurtzke66 South Africa Risk of developing MS in UK immigrants was reduced by nearly 70% among those who emigrated to South Africa under the age of 15 years

Dean et al67 UK In immigrants from Europe, Australia, North and South America, Egypt, Turkey, and Iran, incidence of admission to hospital for MS was high or moderately high; immigrants from India, Pakistan, Africa, and the West Indies had a low incidence of hospital admission for MS

Kurtzke et al68 USA White World War II veterans had signifi cant reductions in MS risk for moves southward from either the north or middle tiers of the USA and increases in risk for moves northward from the middle tier

Elian and Dean69 UK UK-born children of West Indian immigrants had an incidence and prevalence of MS similar to that of the UK average

Elian et al70 UK Children born in the UK of Asian, African, and West Indian immigrants had a high prevalence of MS similar to that occurring in the UK general population

Dean and Elian71 UK Indian and Pakistani immigrants who entered England before 15 years of age had a higher risk of developing MS than those who entered after this age; Caribbean immigrants did not show this diff erence

Kurtzke et al72 France Prevalence of MS in migrants from North Africa was higher for those who developed MS in France (77 per 100 000) than for those who developed it in North Africa (17 per 100 000)

Hammond et al73 Australia Prevalence of MS among those migrating before the age of 15 years from the UK to Australia was not signifi cantly diff erent to the prevalence among those migrating after 15 years

Karni et al74 Israel Prevalence of MS in immigrant Arab Jews was 22 per 100 000 and in Israeli-born Arab Jews was 52 per 100 000

Cabre et al75 French West Indies

Individuals returning to the French West Indies after a period of residence in France had a 1·7 times greater risk of MS (95% CI 1·19–2·38); this increased risk was more marked if the stay was made before the age of 15 years (4 times increase [95% CI 2·17–6·83])

Smestad et al76 Norway Prevalence of MS in patients of Norwegian origin was 170 per 100 000, 85 per 100 000 for immigrants from the middle east, 21 per 100 000 for immigrants from Asia, and 20 per 100 000 for immigrants from Africa

MS=multiple sclerosis.

Table 1: Migration studies in MS

www.thelancet.com/neurology Vol 9 July 2010 731

Review

high serum concentrations of 25-hydroxycholecalciferol were found to have a protective eff ect against MS.56 The relationship between MS risk and vitamin D metabolism has been reviewed in detail elsewhere.57

As 1 billion people worldwide have vitamin D defi ciency or insuffi ciency due to reduced sun exposure or in-adequate vitamin D intake,58 vitamin D supplementation might have the greatest eff ect on the prevention of MS.

Month of birth Willer and colleagues59 estimated that the risk of developing MS north of the equator for individuals born in May was 9·1% higher than expected, and was 8·5% lower than expected for those born in November. Variation in the deviation pattern of MS births by latitude would suggest an ultraviolet radiation eff ect, which is supported by studies of individuals born in the southern hemisphere where more patients with MS are born in November than in May.59 However, the month-of-birth eff ect in Sardinia was similar to that of northern Europeans.60 The month-of-birth eff ect is more pronounced in Scotland, with 31% more MS births in April and 20% fewer MS births in November (p=0·001).61

Timing of exposureBecause the average age of MS onset is approximately 30 years, there is a long period from birth to MS diagnosis for environmental factors to act. The eff ects of migration between high-risk and low-risk geographic regions have been examined in many populations (table 1), and provide clues about when some environmental factors are operative. Although there is potential for bias due to diff erential migration,77 these studies consistently show that MS risk is infl uenced at least to some extent by the migrant’s country of origin.71 Elian and colleagues70 reported that fi rst-generation Afro-Caribbean and Asian immigrants to Britain have a much lower incidence of MS than their second-generation counterparts born in the UK. Additional support comes from space-time cluster analysis on the location of a cohort of patients with MS in Sardinia, which showed that patients were signifi cantly more likely to live near to one another between the ages of 1 and 3 years.78 Despite the limits of small sample sizes, a critical age has been suggested: immigrants who migrate before adolescence acquire the risk of their new country, whereas those who migrate later retain the risk of their home country.64 However, data on a migrant population in Australia suggest that this critical age could extend into adulthood.73

The importance of adolescence in acquiring risk for MS, as suggested by migration data, is further supported by associations of lower age of menarche and adolescent obesity with increased risk of MS.79,80 Other epidemio-logical studies (eg, on occupation) suggest that such infl uences might extend into adult life.81 The risk of developing MS falls markedly after the age of 50 years,4 thus implying that environmental risk factors are unable

to induce the development of MS after a certain timepoint.

Symptoms and signs before MS diagnosisA substantial number of people who are diagnosed with MS recall earlier symptoms that could be attributable to a fi rst demyelinating event. Individuals who later develop MS might also experience fatigue, depression, and/or cognitive impairment before the diagnosis of MS.82,83 Many patients fi rst present with CIS, but recent work has highlighted that radiological (RIS) and immunological changes might occur earlier.

Radiologically isolated syndrome Some individuals with pathological features of MS at post mortem are never diagnosed as having MS in life, implying that they either had asymptomatic disease or that their symptoms were not severe enough to warrant an MS diagnosis.84,85 Some individuals have an MRI scan that is highly suggestive of MS, but have no signs or symptoms. Lebrun and colleagues86 identifi ed 30 patients fulfi lling MRI Barkhof-Tintoré criteria for MS,87,88 but with normal neurological examination. 11 patients subsequently converted to clinically apparent disease, with a mean time between fi rst brain MRI and CIS of 2·3 years.86 Okuda and colleagues6 studied a cohort of 44 individuals who met at least three of four of the Barkhof criteria.87 Among the 30 patients who were followed-up clinically, ten developed CIS or defi nite MS after a median of 5·4 years.6 Thus, patients who have RIS are at risk of developing CIS and MS, but a larger cohort of individuals with RIS who are followed prospectively is needed to determine the actual risk of conversion to defi nite MS.

Clinically isolated syndrome85% of people who develop MS are likely to present with CIS,5 although this has not been validated in a population-based study. Of those who present with CIS, only 63% overall will develop MS over a period of 20 years.89 MRI can help to predict who is more likely to develop MS: 82% of individuals with an abnormal baseline MRI will go on to develop MS compared with only 21% with a normal baseline MRI.89 The most readily visible abnormality is the presence of focal white matter lesions on T2-weighted MRI, which correlates with lesions seen on inspection of pathological samples. However, the number of T2 lesions is no better predictor of MS.89 More recently, grey matter abnormalities and progressive grey matter atrophy, cortical and thalamic reductions in N-acetyl-aspartate, magneti-sation transfer ratio changes, and evolving diff usion-tensor-imaging abnormalities have been observed in patients who convert from CIS to MS, although these were small studies and require validation.90–93

MS immunological signature Until recently, MS has been thought to be a CD4 T-cell-mediated disease, because experimental autoimmune

732 www.thelancet.com/neurology Vol 9 July 2010

Review

encephalomyelitis (EAE; the animal model of MS) can be adoptively transferred to naive animals by the injection of encephalitogenic myelin-specifi c CD4 T cells.94 However, clinical trial data suggest that, by contrast with EAE, targeting CD4 T-cell function in MS might not be of therapeutic benefi t as patients who received anti-CD4 antibody showed no clinical improvement.95

T-helper (Th) subpopulations of CD4 T cells (ie, Th-1, Th-2, and Th-17 subsets) have characteristic cytokine profi les.94 Th-1 cells were thought to be the prime drivers of the autoimmune process in MS; however, treatment with the monoclonal antibody ustekinumab, which blocks interleukin 12, a cytokine that is pivotal in Th-1 diff erentiation, produced no benefi t in phase 2 clinical trials.96 Another subpopulation of CD4 cells, CD4+CD25+ T regulatory (Treg) cells expressing Fork head box protein P3 (FOXP3), have been shown to play a part in immune homoeostasis. Impaired Treg-cell activity seems to be important in MS aetiology. Studies have shown that CD4+CD25+ Treg cells are present at the same frequency in patients with MS as in controls, although the eff ector function of CD4+CD25+ Treg cells is impaired in patients with MS.97,98 A particular subset (CD127+) of Treg cells might be aberrant.99 Studies are needed to determine whether this abnormality in Treg cells exists before the diagnosis of MS. A recently identifi ed lineage of CD4 T cells that produces interleukin 17 has been shown to be essential for inducing EAE.94 Th-17 cells were also found to be enriched in active MS lesions, implicating them in disease pathogenesis.100 The eff ects

of Th-17 and Treg-cell populations in MS thus require further study.

CD8 T cells are also thought to be involved in the development of MS.94 An early study reported numerical and functional abnormalities of CD8 cells in patients with MS.101 Because the prevailing view at the time was that these cells were suppressor T lymphocytes, the emphasis was on analysis of suppressive function.101 More recently, CD8 cells have been found to be potential active participants in the pathogenesis of MS. Within MS plaques, CD8 T cells outnumber CD4 T cells.94 T cells isolated from peripheral blood or CSF of patients with MS show a selective enrichment of memory CD8 T cells with oligoclonal expansion, suggesting that these cells are activated and antigen driven.102 In a study of 52 patients with suspected MS, the recruitment of highly diff erentiated (C-C chemokine receptor 7 negative) CD8 T cells in the CSF compared with peripheral blood was determined, and patients were followed for 12 months.103 A CD8/CD4 cell ratio of 0·94 or greater was found in patients with relapsing-remitting (RR) MS and possible MS compared with those with other neurological diseases (risk ratio 2·29 [95% CI 1·13–4·66]; p=0·006), suggesting that CD8 T cells play a part in the early stages of RRMS.103 In support of this fi nding, the frequency of interferon-γ-secreting EBV-specifi c CD8 T cells in patients with CIS was signifi cantly higher than in patients with MS or other neurological diseases, and healthy controls.104 Regulatory and eff ector CD8 T cells need to be studied in more detail in MS and people at risk of developing MS.

Although MS is widely thought to be caused by T-cell-mediated immunity, substantial evidence exists for humoral immunity in the disease process. The most con-sistent immunological abnormality described in patients with MS is the intrathecal synthesis of IgG, generating CSF oligoclonal bands (OCBs).105 The appearance of IgG in CSF indicates abnormal B-cell-related processes active within the brain parenchyma, although the specifi city of CSF IgG remains to be resolved. CSF OCBs have been found in more than 95% of patients with MS,106 but are not specifi c to MS. Indeed, some patients have CSF OCBs many years before a diagnosis of MS is made.107 CSF OCBs are both negative and positive predictors of risk for MS in people presenting with CIS.105,108,109

The immunology of MS is not just restricted to the adaptive immune system, and we are just beginning to understand the role of innate immune responses. Arguably, the changes in the natural-killer (NK) cell population are the most frequently reported peripheral blood irregularity in MS (table 2).119,121,123 This abnormality seems to exist before the diagnosis of MS because CD56bright NK-cell frequency seems to be reduced in CIS.123 NK cells showing a NK type 2 phenotype that can be induced under similar conditions as the Th-2 subset117 might also help to regulate autoreactive eff ector T cells.119

Insight into the immunopathogenesis of MS is also provided by current monoclonal antibody therapies that

Finding

Benczur et al110 NK activity was found to be signifi cantly lower in patients with MS

Uchida et al111 NK activity was found to be signifi cantly lower in patients with MS

Braakman et al112 NK cells from patients with MS showed a signifi cantly lower increase in activity in response to IL2 than those from healthy controls

Vranes et al113 Patients with MS had a lower number and activity of NK cells than controls

Kreuzfelder et al114 Patients with MS had a lower number of NK cells than controls

Munschauer et al115 Patients with MS had a lower number of NK cells than controls

Kastrukoff et al116 Functional activity of NK cells correlated signifi cantly with the development of MRI lesions and clinical relapses

Takahashi et al117 NK cells in the remission of MS were characterised by an increase in IL5 mRNA and a decreased expression of IL12Rβ2 mRNA, as well as a higher expression of CD95

Kastrukoff et al118 Clinical relapses were associated with troughs in NK-cell functional activity

Takahashi et al119 In CD95+ NKhigh patients, NK cells could actively suppress potentially pathogenic autoimmune T cells that could mediate infl ammatory responses in the CNS

Bielekova et al120 In a phase II trial of daclizumab (an anti-CD25 monoclonal antibody) in RRMS, suppression of contrast-enhancing lesions on MRI was signifi cantly associated with the expansion of circulating CD56bright NK cells

Rinaldi et al121 Disease activity as measured by MRI was characterised by perturbations in NK cell subsets

Saraste et al122 Changes in the CD56bright population tended to normalise on IFNβ therapy

De Jager et al123 CD8lowCD4− cells were reduced in frequency in patients with untreated RRMS and in patients with CIS; these diff erences were due to a reduction in CD8lowCD56+CD3−CD4− NK cells

CIS=clinically isolated syndrome. IFNβ=interferon β. IL2=interleukin 2. IL5=interleukin 5. IL12Rβ2=interleukin 12 receptor β2. NK cell=natural killer cell. MS=multiple sclerosis. RRMS=relapsing-remitting MS.

Table 2: Evidence for a role for NK cells in MS

www.thelancet.com/neurology Vol 9 July 2010 733

Review

target molecules of the immune system, which seem to be highly eff ective. Natalizumab targets an adhesion molecule, α4 integrin, that blocks entry of immune cells into the CNS.124 α4 integrin is expressed on all T, B, and NK cells and most macrophages, monocytes, and granulocytes. Daclizumab targets interleukin 2 receptor α, which is expressed on all activated T cells, some regulatory T cells (FOXP3-positive cells), the CD56bright NK population, and activated macrophages and dendritic cells.120 Alemtuzumab is directed against CD52, which is also expressed on all T, B, and NK cells and most macrophages, dendritic cells, and granulocytes.125 By contrast, rituximab is B-cell specifi c; it is directed against CD20 and is likely to aff ect antibody-mediated immunity.126 The common link between all of these treatments is their eff ect on B cells, but it is clear that many cell types play a part in the pathogenesis of MS. Of interest will be how rituximab treatment aff ects viral titres of EBV, which predominantly resides in B cells. More details on the immunology of MS can be found elsewhere.127–129

Potential causal pathwaysOur current understanding of MS development is that RIS leads to CIS and then eventually to MS. The associated risk factors for MS also seem to delineate a putative causal cascade (fi gure 3).130 Factors largely defi ned from birth (ie, sex, HLA status, place of birth) require the incitement of environmental factors (vitamin D defi ciency, late EBV exposure) to develop the abnormalities required that subsequently lead to MS. The latitude eff ect (early life) and infectious-mononucleosis associations (adolescence) would support the notion that vitamin D defi ciency precedes EBV infection,130 but the Australian migration data and the evidence for vitamin-D-related infl uences on risk during adult life (eg, outdoor occupations decrease MS risk131) suggest that vitamin D has the potential to play a part over a longer time period. These two environmental candidates might push those people at risk into developing asymptomatic MS or CIS.

Smoking might be involved in the initial stages of MS development, but there is now strong evidence for an eff ect of smoking on disease progression.132 The

Figure 3: Prodromal and potential causal pathway for MSMS risk is partly defi ned by factors acting at or around the time of birth. During childhood and adolescence, additional dynamic risk factors alter MS risk. Dynamic protective factors might also act over the same period. As an individual is exposed to dynamic risk factors over time, immunological changes become apparent in the periphery. The probability of developing MS increases until a biological threshold is reached and immunological changes occur in the CNS. Demyelination becomes inevitable. Disease-modifying factors then act and interact with other factors in driving disease progression through to clinically defi nite MS. If we can act to modify some of the dynamic risk factors at an early stage (ie, before stage 2), before biological disease becomes inevitable, then MS might be preventable. The current challenge is how to identify individuals while they are still asymptomatic. Further work is needed to defi ne the order and time at which dynamic risk factors act, and any interactions that might take place between them, to develop MS preventative strategies. CIS=clinically isolated syndrome. EBV=Epstein-Barr virus. MS=multiple sclerosis. NK cells=natural killer cells. OCB=oligoclonal bands. RIS=radiologically isolated syndrome. Treg cells=T-regulatory cells.

RIS CIS MS

Dynamic factorsStatic factors

1 Declining physiology: peripheral immunological endophenotype2 Biological disease threshold: CNS endophenotype3 Asymptomatic disease: RIS (abnormal MRI and/or evoked potentials)4 Clinical disease a CIS b Relapsing MS c Relapsing secondary-progressive MS d Non-relapsing secondary-progressive MS

Disease-modifying factors

Protective factors:Protective HLA haplotypes

4b 4c 4d

4a

1

2

3

In utero Childhood Adolescence/early adulthood Adulthood

Risk factors:Family history/genetic risk factorsFemale sexBorn in MayBorn in high latitude regions

Risk factors:Exposure to EBVSmokingVitamin D deficiencyResident in high latitude regions

Protective factors:Outdoor activity/sun exposureVitamin D supplementsDiet high in fish oils

Favourable factors:None identified

Low risk Very low risk

Unfavourable factors:Exposure to EBVSmoking

At risk High risk

Peripheral immunological changesTreg cells (CD4+CD25+), NK cells, CD8+

CNS changes(OCBs and microscopic pathology)

MRI/evoked potentials changes

Clinical disease

734 www.thelancet.com/neurology Vol 9 July 2010

Review

continuum of CIS to MS is also aff ected by the environment, and smoking increases the risk of converting to MS after CIS. In a study of 129 patients with CIS, 75% of smokers versus 51% of non-smokers developed MS after 3 years of follow-up.133 CIS patients who converted to MS within 5 years had higher concentrations of EBNA1 IgG compared with those who did not convert.134,135 However, whether this is an indication of the eff ect of EBV on susceptibility or an

eff ect on disease progression remains to be determined.

Whether MS risk factors act in sequence and depend on each other or whether they act independently and in an additive or multiplicative fashion is not known. That no factor has yet been shown to be present in all patients with MS, with the possible exception of EBV in adult-onset MS, suggests that causal pathways are likely to diff er between individuals and would support an additive eff ect.

Study OR (95% CI)

Family history

Non-twin fi rst-degree relative aff ected59 19 615 fi rst-degree relatives of 8205 Danish patients with MS 7·1 (5·8–8·8)

Ethnic origin

White male compared with black male31 US cohort study of 4951 patients with MS and 9378 controls 1·49 (1·09–2·27)

HLA haplotype

HLA-DRB1*15 homozygote16 Canadian cohort study of 2454 patients with MS and 4639 unaff ected fi rst-degree relatives

5·42 (4·12–7·16)

HLA-DRB1*15 heterozygote16 Canadian cohort study of 2454 patients with MS and 4639 unaff ected fi rst-degree relatives

2·91 (2·42–3·51)

HLA-DRB1*15/HLA-DRB1*14 heterozygote16 Canadian cohort study of 2454 patients with MS and 4639 unaff ected fi rst-degree relatives

1·06 (0·56–2·03)

Immune marker genes

Interleukin 2 receptor α (IL2RA)22 Case-control study of 4839 patients with MS and 9336 controls 1·15 (1·04–1·27)

Interleukin 7 receptor α (IL7RA)22 Case-control study of 4839 patients with MS and 9336 controls 1·13 (1·02–1·23)

C-type lectin domain family 16 A (CLEC16A)22 Case-control study of 4839 patients with MS and 9336 controls 1·15 (1·04–1·25)

CD58 (CD58)22 Case-control study of 4839 patients with MS and 9336 controls 1·30 (1·14–1·47)

Tumour necrosis factor receptor superfamily, member 1A (TNFRSF1A)22 Case-control study of 4839 patients with MS and 9336 controls 1·20 (1·10–1·31)

Interferon regulatory factor 8 (IRF8)22 Case-control study of 4839 patients with MS and 9336 controls 1·25 (1·12–1·39)

CD6 (CD6)22 Case-control study of 4839 patients with MS and 9336 controls 1·18 (1·07–1·30)

Place of birth

Migration before vs after age 15 years71 Cohort study of 76 immigrant patients in the UK 2·07 (1·13–3·77)

Age

Incidence at age 30 years vs age 55 years4 Cohort study of 1099 Canadian patients 4·5 (1·52–13·3)

Clinically isolated syndrome

Abnormal MRI vs normal89 UK cohort study of 107 CIS patients 3·99 (1·65–9·65)

Presence of oligoclonal bands, independent of MRI109 Spanish cohort study of 415 CIS patients 1·7 (1·1–2·7)

Sex

Female30 Population-based study of 27 074 Canadian patients with MS 6·62 (6·21–7·13)

EBV infection

Infectious mononucleosis36 Meta-analysis of 11 case-control and 3 cohort studies totalling 1667 patients with MS and 3606 controls

2·30 (1·70–3·01)

Anti-EBNA1 antibody geometric mean titre >320 vs <80141 Nested US case-control study of 148 women with MS and 296 healthy female controls

1·66 (1·32–2·08)

Smoking

Ever vs never44 Meta-analysis of case-control studies totalling 1155 patients with MS and 153 182 controls

1·51 (1·22–1·87)

Month of birth

May59 17 874 Canadian, 11 502 British, 6276 Danish, and 6393 Swedish patients with MS compared with population controls

1·10 (1·07–1·13)

Vitamin D

Serum 25-hydroxycholecalciferol increased in the lower quintile (<63·3 nmol/L) vs the upper quintile (>99·1 nmol/L)56

Nested US case-control study of 148 white patients with MS and 296 matched healthy white controls

1·69 (1·03–2·78)

Meta-analyses were included if available, and if not, the largest (or, in some cases, the only) study available was included. EBV=Epstein-Barr virus. EBNA1=EBV nuclear antigen 1. MS=multiple sclerosis. OR=odds ratio.

Table 3: Risk factors for MS

www.thelancet.com/neurology Vol 9 July 2010 735

Review

Endophenotypes An endophenotype is an intermediate phenotype that fi lls the gap in the causal chain between genes and clinical disease,136 and is typically defi ned by several traits and/or characteristics that in isolation might not be considered pathological. Endophenotypes are most commonly applied to psychiatric diseases, but can be applied to any complex disorder.7 The disease markers that are used to defi ne the endophenotype are usually associated with disease in the population, might be heritable, manifest in an individual whether or not the disease in question is active, are found in non-aff ected family members at a higher rate than in the general population, and co-segregate with the disease.136

The existence of endophenotypes in MS has been poorly studied. However, some studies have shown that people at risk of MS are more likely to have the physiological changes seen in MS.137–139 The potential endophenotypes that are currently available are either immunological or radiological, but require further investigation to determine their temporal relationship.

Several potential immunological endophenotypes for MS risk have been identifi ed. Viglietta and colleagues97 found that patients with MS had impaired eff ector function of CD4+CD25+ Treg cells. Another study found that EBV-specifi c interferon-γ-secreting CD8 T-cell response increases signifi cantly close to disease onset.104 Transcriptional profi ling of naive CD4 T cells in patients with CIS at diagnosis identifi ed an expression profi le that characterised patients at high risk of conversion to MS (34 [92%] of 37 patients converted within 9 months). Consistent downregulation of TOB1, a gene thought to repress T cell proliferation, in these patients was also observed.140 All of these changes might be useful in predicting an individual’s risk of developing MS.

Memory CD8+ T cells are selectively enriched in patients with MS compared with controls.102 However, this diff erence does not exist in monozygotic twins who are discordant for MS,137 suggesting that individuals at risk of MS might have skewed CD8 T-cell distributions. Furthermore, the ratio of highly diff erentiated CD8 and

Risk factors Study details Findings

Islam et al142 Latitude of birth, familial risk

400 monozygotic twin pairs Concordance was 1·9 times (95% CI 1·2–3·2) greater among northern-born twins

Haghighi et al143 OCBs, familial risk Case-control study of 47 healthy siblings of patients with MS vs 50 unrelated healthy controls

9 (19%) healthy siblings of patients with MS had OCBs vs 2 (4%) controls

Soderstrom et al108 HLA, OCBs, MRI Population-based study of 147 consecutive patients with acute monosymptomatic optic neuritis

Abnormal MRI and presence of OCBs were strongly associated with MS; of 25 individuals with a normal MRI and no OCBs, none developed MS; presence of HLA-DR2 was related to MS, but did not add to the PPV of MRI and OCBs

Celius et al144 HLA, sex, age Cohort of 286 Norwegian patients with MS

HLA-DR2, DQ6 was signifi cantly more frequent among women than men (p=0·025), and was negatively correlated with age at diagnosis (p=0·0254)

Hensiek et al145 HLA, sex Cohort of 729 patients with MS HLA-DR15 was associated with younger age at diagnosis and female sex

De Jager et al141 HLA, EBV Nested case-control study of 148 women with MS and 296 age-matched healthy women

MS among HLA-DRB1*15-positive women with increased anti-EBNA1 titres (>1:320) was 9·7-fold (95% CI 3·2–29·2) higher than that of HLA-DRB1*15-negative women with low anti-EBNA-1 titres (<1:80)

Sundstrom et al146 HLA, EBV Case-control study of 109 individuals with MS and 212 age-matched and sex-matched controls

OR of developing MS in individuals with HLA-DRB1*15 and high EBNA1 reactivity was 16·0 (95% CI 5·1–50·0) compared with HLA-DRB1*15-negative subjects with low titres of EBNA1

Nielsen et al147 HLA, EBV Case-control study of 76 MS patients with IM, 1836 MS patients without IM, and 62 blood donors with history of IM and 484 without a history of IM

HLA-DRB1*15-positive individuals with a history of IM had a 10 times (95% CI 6·0–17·9) greater risk of MS than HLA-DRB1*15-negative individuals without a history of IM

De Jager et al148 HLA, non-HLA genes, EBV, smoking, sex

Case-control validation populations: 1340 patients with either MS or CIS matched to 1109 controls; 143 patients with MS and 281 controls from the Nurses’ Health Study

A weighted genetic risk score produced for 16 susceptibility loci including HLA-DRB1*15 gave modest prediction of MS, which was slightly enhanced by addition of sex, EBV, and smoking

Kelly et al149 HLA, MRI Cohort of 70 patients with CIS MS developed in 86% of MRI-positive, HLA-DRB1*1501-positive patients vs 55% of MRI-positive, HLA-DRB1*1501-negative patients (p<0·025)

Hauser et al150 HLA, MRI Cohort of 178 patients with optic neuritis

HLA-DR2 was present in 85 (48%) patients, and was associated with increased odds of probable or defi nite MS at 5 years (OR 1·92 [95% CI 1·01–3·67]; p=0·04); the association was most apparent among patients with signal abnormalities on baseline brain MRI

CIS=clinically isolated syndrome. EBV=Epstein-Barr virus. EBNA1=EBV nuclear antigen 1. IM=infectious mononucleosis. MS=multiple sclerosis. OCBs=oligoclonal bands. OR=odds ratio. PPV=positive predictive value.

Table 4: Studies of combined risk factors for MS

736 www.thelancet.com/neurology Vol 9 July 2010

Review

CD4 cells in the CSF seems to be a strong predictor for the development of MS.103

Potential endophenotypes have also been identifi ed in CNS studies of MS risk. Evoked potentials in relatives of patients with MS have been shown to have a latency that was signifi cantly longer than that for unrelated individuals controlled for age and sex.138 Asymmetries in evoked potentials were seen more often in HLA-identical siblings than in siblings with diff erent HLA genotypes.138 This suggests that subclinical electrophysiological changes do occur in individuals at risk of developing MS. A Sardinian study using a mobile MRI scanner to study 240 asymptomatic relatives of patients with MS found that 4% of relatives from families with sporadic MS and 10% from families with multiple aff ected members had focal MRI changes that met validated MRI criteria for MS compared with none of 56 controls.139

Conclusions and future perspectives Modelling all known risk factors for MS (table 3) might enhance our ability to predict who will develop the disease and to select individuals in whom to intensively study the MS prodrome. This could help us to elucidate the steps leading to the development of MS and to identify robust and easily measurable biomarkers and endophenotypes.

Some attempts at combining risk factors have been made, leading to substantial improvements in the potential prediction of MS (table 4). The most comprehensive study to date was that by De Jager and colleagues148 who attempted to combine 16 genetic risk loci, sex, and EBV titres into a prediction model. However, the model fell short of being clinically useful, and more detailed studies of these and additional risk factors in large cohorts are needed to better understand their predictive power in combination. The extent to which the known risk factors for MS explain the development of MS and the potential sensitivity and specifi city of a hypothetical test that would combine them and how much better at predicting MS it would be than just taking a family history are not yet known. The timing of risk factors (eg, EBV/vitamin D) that seem to require exposure many years before the development of MS means that longitudinal, prospective cohorts are required, that will need to be followed for many years.

Although some progress has been made, our under-standing of the stages involved in the development of MS remains limited. Further study is required to investigate the changes involved in the MS prodrome. There could be subtle changes in physiology that have so far been missed due to the current resolution of medical technology or simply by virtue of not being looked for. The pre-MS stage (fi gure 3), when at-risk individuals have had the relevant environmental exposures, deserves particular attention because immunological alterations (eg, changes in Treg, CD8, and NK-cell frequency and function), and likely biomarkers, will undoubtedly manifest at this point. On the basis of migration studies and data on infectious mononucleosis, the likely time period for this pre-MS stage is between adolescence and the peak onset age of MS.

CIS is generally viewed as the earliest accessible timepoint for MS studies. However, although people with CIS do not have MS as defi ned by contemporary diagnostic criteria, some already have evidence of grey matter atrophy and substantial cognitive impairment.89 Entirely asymptomatic individuals with a high risk of developing MS as defi ned by endophenotypes are thus of great interest. These individuals could be studied longitudinally for genetic, immunological, and epidemio-logical factors to investigate the initiation of the disease processes and to fully understand the trajectory of disease processes underlying MS causation. Ultimately, MS endophenotypes could be used to identify high risk individuals to enrol in interventional studies that target environmental risk factors (eg, smoking or vitamin D defi ciency) to prevent this often devastating disease.

Contributors GG conceived the idea of this Review. SVR, RD, UCM, and GG

undertook the literature search and wrote the paper.

Confl icts of interest SVR, RD, and UCM have no confl icts of interest. GG has received

consulting fees from Bayer-Schering Healthcare, Biogen-Idec,

GlaxoSmithKline, Merck-Serono, Novartis, Protein Discovery

Laboratories, Teva-Aventis, and UCB Pharma, lecture fees from Bayer-

Schering Healthcare, Biogen Idec, and Teva-Aventis, and grant support

from Bayer-Schering Healthcare, Biogen-Idec, Merck-Serono, Merz,

Novartis, Teva-Aventis, and UCB Pharma.

AcknowledgmentsWe thank our colleagues and collaborators. GG is generously funded by

grants from the MRC, National MS Society, the MS Society of Great Britain

and Northern Ireland, AIMS2CURE, and the Roan Charitable Trust. SVR

is funded by the MRC and is a Goodger Scholar at the University of

Oxford. UCM is funded by AIMS2CURE and the Roan Charitable Trust.

RD is funded by a fellowship from the Guarantors of Brain.

References1 Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshenker BG.

Multiple sclerosis. N Engl J Med 2000; 343: 938–52.

2 Giovannoni G, Ebers G. Multiple sclerosis: the environment and causation. Curr Opin Neurol 2007; 20: 261–68.

3 White T, Anjum A, Schulz SC. The schizophrenia prodrome. Am J Psychiatry 2006; 163: 376–80.

4 Paty DW, Ebers GC. Multiple sclerosis. Philadelphia: F A Davis Company, 1998.

5 Miller D, Barkhof F, Montalban X, Thompson A, Filippi M. Clinically isolated syndromes suggestive of multiple sclerosis, part I: natural history, pathogenesis, diagnosis, and prognosis. Lancet Neurol 2005; 4: 281–88.

Search strategy and selection criteria

A comprehensive search of the PubMed/Medline database was undertaken for all publications in English up to February, 2010. The following search terms were used in combination with “multiple sclerosis OR MS”: “vitamin D OR 25-hydroxyvitamin D”, “Epstein-Barr virus OR EBV”, “smoking”, “genet*”, “environment*”, “timing”, “exposure”, “clinically isolated syndrome OR CIS”, “radiologically isolated syndrome OR RIS”, “immunology”, and “migration”. Abstracts were manually searched for relevance, and selected publications were searched for further relevant references.

www.thelancet.com/neurology Vol 9 July 2010 737

Review

6 Okuda DT, Mowry EM, Beheshtian A, et al. Incidental MRI anomalies suggestive of multiple sclerosis: the radiologically isolated syndrome. Neurology 2009; 72: 800–05.

7 Flint J, Munafo MR. The endophenotype concept in psychiatric genetics. Psychol Med 2007; 37: 163–80.

8 Dyment DA, Ebers GC, Sadovnick AD. Genetics of multiple sclerosis. Lancet Neurol 2004; 3: 104–10.

9 Robertson NP, Fraser M, Deans J, Clayton D, Walker N, Compston DA. Age-adjusted recurrence risks for relatives of patients with multiple sclerosis. Brain 1996; 119: 449–55.

10 Sadovnick AD, Baird PA, Ward RH. Multiple sclerosis: updated risks for relatives. Am J Med Genet 1988; 29: 533–41.

11 Willer CJ, Dyment DA, Risch NJ, Sadovnick AD, Ebers GC. Twin concordance and sibling recurrence rates in multiple sclerosis. Proc Natl Acad Sci USA 2003; 100: 12877–82.

12 Ebers GC, Sadovnick AD, Dyment DA, Yee IM, Willer CJ, Risch N. Parent-of-origin eff ect in multiple sclerosis: observations in half-siblings. Lancet 2004; 363: 1773–74.

13 Jersild C, Fog T, Hansen GS, Thomsen M, Svejgaard A, Dupont B. Histocompatibility determinants in multiple sclerosis, with special reference to clinical course. Lancet 1973; 2: 1221–25.

14 Barcellos LF, Oksenberg JR, Begovich AB, et al. HLA-DR2 dose eff ect on susceptibility to multiple sclerosis and infl uence on disease course. Am J Hum Genet 2003; 72: 710–16.

15 Marrosu MG, Murru R, Murru MR, et al. Dissection of the HLA association with multiple sclerosis in the founder isolated population of Sardinia. Hum Mol Genet 2001; 10: 2907–16.

16 Dyment DA, Herrera BM, Cader MZ, et al. Complex interactions among MHC haplotypes in multiple sclerosis: susceptibility and resistance. Hum Mol Genet 2005; 14: 2019–26.

17 Masterman T, Ligers A, Olsson T, Andersson M, Olerup O, Hillert J. HLA-DR15 is associated with lower age at onset in multiple sclerosis. Ann Neurol 2000; 48: 211–19.

18 Barcellos LF, Sawcer S, Ramsay PP, et al. Heterogeneity at the HLA-DRB1 locus and risk for multiple sclerosis. Hum Mol Genet 2006; 15: 2813–24.

19 Ramagopalan SV, Morris AP, Dyment DA, et al. The inheritance of resistance alleles in multiple sclerosis. PLoS Genet 2007; 3: e150.

20 Dean G, Yeo TW, Goris A, et al. HLA-DRB1 and multiple sclerosis in Malta. Neurology 2008; 70: 101–05.

21 Lincoln MR, Ramagopalan SV, Chao MJ, et al. Epistasis among HLA-DRB1, HLA-DQA1, and HLA-DQB1 loci determines multiple sclerosis susceptibility. Proc Natl Acad Sci USA 2009; 106: 7542–47.

22 De Jager PL, Jia X, Wang J, et al. Meta-analysis of genome scans and replication identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis susceptibility loci. Nat Genet 2009; 41: 776–82.

23 Hafl er DA, Compston A, Sawcer S, et al. Risk alleles for multiple sclerosis identifi ed by a genomewide study. N Engl J Med 2007; 357: 851–62.

24 Kurtzke JF, Beebe GW, Norman JE Jr. Epidemiology of multiple sclerosis in U.S. veterans: 1. Race, sex, and geographic distribution. Neurology 1979; 29: 1228–35.

25 Herrera BM, Ramagopalan SV, Lincoln MR, et al. Parent-of-origin eff ects in MS. Observations from avuncular pairs. Neurology 2008; 71: 799–803.

26 Hoppenbrouwers IA, Liu F, Aulchenko YS, et al. Maternal transmission of multiple sclerosis in a Dutch population. Arch Neurol 2008; 65: 345–48.

27 Ramagopalan SV, Yee IM, Dyment DA, et al. Parent-of-origin eff ect in multiple sclerosis. Observations from interracial matings. Neurology 2009; 73: 602–05.

28 Handel AE, Ebers GC, Ramagopalan SV. Epigenetics: molecular mechanisms and implications for disease. Trends Mol Med; 16: 7–16.

29 Chao MJ, Ramagopalan SV, Herrera BM, et al. Epigenetics in multiple sclerosis susceptibility: diff erence in transgenerational risk localizes to the major histocompatibility complex. Hum Mol Genet 2009; 18: 261–66.

30 Orton SM, Herrera BM, Yee IM, et al. Sex ratio of multiple sclerosis in Canada: a longitudinal study. Lancet Neurol 2006; 5: 932–36.

31 Wallin MT, Page WF, Kurtzke JF. Multiple sclerosis in US veterans of the Vietnam era and later military service: race, sex, and geography. Ann Neurol 2004; 55: 65–71.

32 Hirst C, Ingram G, Pickersgill T, Swingler R, Compston DA, Robertson NP. Increasing prevalence and incidence of multiple sclerosis in South East Wales. J Neurol Neurosurg Psychiatry 2009; 80: 386–91.

33 Alonso A, Hernan MA. Temporal trends in the incidence of multiple sclerosis: a systematic review. Neurology 2008; 71: 129–35.

34 Whitacre CC. Sex diff erences in autoimmune disease. Nat Immunol 2001; 2: 777–80.

35 Ascherio A, Munger KL. Environmental risk factors for multiple sclerosis. Part I: the role of infection. Ann Neurol 2007; 61: 288–99.

36 Thacker EL, Mirzaei F, Ascherio A. Infectious mononucleosis and risk for multiple sclerosis: a meta-analysis. Ann Neurol 2006; 59: 499–503.

37 Sundstrom P, Juto P, Wadell G, et al. An altered immune response to Epstein-Barr virus in multiple sclerosis: a prospective study. Neurology 2004; 62: 2277–82.

38 Levin LI, Munger KL, Rubertone MV, et al. Temporal relationship between elevation of Epstein-Barr virus antibody titers and initial onset of neurological symptoms in multiple sclerosis. JAMA 2005; 293: 2496–500.

39 Nielsen TR, Rostgaard K, Nielsen NM, et al. Multiple sclerosis after infectious mononucleosis. Arch Neurol 2007; 64: 72–75.

40 Ramagopalan SV, Valdar W, Dyment DA, et al. Association of infectious mononucleosis with multiple sclerosis. A population-based study. Neuroepidemiology 2009; 32: 257–62.

41 Banwell B, Krupp L, Kennedy J, et al. Clinical features and viral serologies in children with multiple sclerosis: a multinational observational study. Lancet Neurol 2007; 6: 773–81.

42 Yeh EA, Chitnis T, Krupp L, et al. Pediatric multiple sclerosis. Nat Rev Neurol 2009; 5: 621–31.

43 Pohl D, Krone B, Rostasy K, et al. High seroprevalence of Epstein-Barr virus in children with multiple sclerosis. Neurology 2006; 67: 2063–65.

44 Hawkes CH. Smoking is a risk factor for multiple sclerosis: a metanalysis [sic]. Mult Scler 2007; 13: 610–15.

45 Hernan MA, Olek MJ, Ascherio A. Cigarette smoking and incidence of multiple sclerosis. Am J Epidemiol 2001; 154: 69–74.

46 Hedstrom AK, Baarnhielm M, Olsson T, Alfredsson L. Tobacco smoking, but not Swedish snuff use, increases the risk of multiple sclerosis. Neurology 2009; 73: 696–701.

47 Kurtzke JF. Geographic distribution of multiple sclerosis: an update with special reference to Europe and the Mediterranean region. Acta Neurol Scand 1980; 62: 65–80.

48 Hammond SR, McLeod JG, Millingen KS, et al. The epidemiology of multiple sclerosis in three Australian cities: Perth, Newcastle and Hobart. Brain 1988; 111: 1–25.

49 Kampman MT, Brustad M. Vitamin D: a candidate for the environmental eff ect in multiple sclerosis—observations from Norway. Neuroepidemiology 2008; 30: 140–46.

50 Weinstock-Guttman B, Jacobs LD, Brownscheidle CM, et al. Multiple sclerosis characteristics in African American patients in the New York State Multiple Sclerosis Consortium. Mult Scler 2003; 9: 293–98.

51 Acheson ED, Bachrach CA, Wright FM. Some comments on the relationship of the distribution of multiple sclerosis to latitude, solar radiation, and other variables. Acta Psychiatr Scand Suppl 1960; 35: 132–47.

52 van der Mei IA, Ponsonby AL, Dwyer T, et al. Past exposure to sun, skin phenotype, and risk of multiple sclerosis: case-control study. BMJ 2003; 327: 316.

53 Munger KL, Zhang SM, O’Reilly E, et al. Vitamin D intake and incidence of multiple sclerosis. Neurology 2004; 62: 60–65.

54 Vieth R, Cole DE, Hawker GA, Trang HM, Rubin LA. Wintertime vitamin D insuffi ciency is common in young Canadian women, and their vitamin D intake does not prevent it. Eur J Clin Nutr 2001; 55: 1091–97.

55 van Dam RM, et al. Potentially modifi able determinants of vitamin D status in an older population in the Netherlands: the Hoorn Study. Am J Clin Nutr 2007; 85: 755–61.

56 Munger KL, Levin LI, Hollis BW, Howard NS, Ascherio A. Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA 2006; 296: 2832–38.

57 Ascherio A, Munger KL, Simon KC. Vitamin D and multiple sclerosis. Lancet Neurol 2010; 9: 599–612.

738 www.thelancet.com/neurology Vol 9 July 2010

Review

58 Holick MF. Vitamin D defi ciency. N Engl J Med 2007; 357: 266–81.

59 Willer CJ, Dyment DA, Sadovnick AD, Rothwell PM, Murray TJ, Ebers GC. Timing of birth and risk of multiple sclerosis: population based study. BMJ 2005; 330: 120.

60 Sotgiu S, Pugliatti M, Sotgiu MA, et al. Seasonal fl uctuation of multiple sclerosis births in Sardinia. J Neurol 2006; 253: 38–44.

61 Bayes HK, Weir CJ, O’Leary C. Timing of birth and risk of multiple sclerosis in the Scottish population. Eur Neurol 2009; 63: 36–40.

62 Alter M, Halpern L, Kurland LT, Bornstein B, Leibowitz U, Silberstein J. Multiple sclerosis in Israel. Prevalence among immigrants and native inhabitants. Arch Neurol 1962; 7: 253–63.

63 Alter M, Leibowitz U, Speer J. Risk of multiple sclerosis related to age at immigration to Israel. Arch Neurol 1966; 15: 234–37.

64 Dean G. Annual incidence, prevalence, and mortality of multiple sclerosis in white South-African-born and in white immigrants to South Africa. BMJ 1967; 2: 724–30.

65 Alter M, Okihiro M, Rowley W, Morris T. Multiple sclerosis among Orientals and Caucasians in Hawaii. Neurology 1971; 21: 122–30.

66 Dean G, Kurtzke JF. On the risk of multiple sclerosis according to age at immigration to South Africa. BMJ 1971; 3: 725–29.

67 Dean G, McLoughlin H, Brady R, Adelstein AM, Tallett-Williams J. Multiple sclerosis among immigrants in Greater London. BMJ 1976; 1: 861–64.

68 Kurtzke JF, Beebe GW, Norman JE Jr. Epidemiology of multiple sclerosis in US veterans: III. Migration and the risk of MS. Neurology 1985; 35: 672–78.

69 Elian M, Dean G. Multiple sclerosis among the United Kingdom-born children of immigrants from the West Indies. J Neurol Neurosurg Psychiatry 1987; 50: 327–32.

70 Elian M, Nightingale S, Dean G. Multiple sclerosis among United Kingdom-born children of immigrants from the Indian subcontinent, Africa and the West Indies. J Neurol Neurosurg Psychiatry 1990; 53: 906–11.

71 Dean G, Elian M. Age at immigration to England of Asian and Caribbean immigrants and the risk of developing multiple sclerosis. J Neurol Neurosurg Psychiatry 1997; 63: 565–68.

72 Kurtzke JF, Delasnerie-Laupretre N, Wallin MT. Multiple sclerosis in North African migrants to France. Acta Neurol Scand 1998; 98: 302–09.

73 Hammond SR, English DR, McLeod JG. The age-range of risk of developing multiple sclerosis: evidence from a migrant population in Australia. Brain 2000; 123: 968–74.

74 Karni A, Kahana E, Zilber N, Abramsky O, Alter M, Karussis D. The frequency of multiple sclerosis in Jewish and Arab populations in greater Jerusalem. Neuroepidemiology 2003; 22: 82–86.

75 Cabre P, Signate A, Olindo S, et al. Role of return migration in the emergence of multiple sclerosis in the French West Indies. Brain 2005; 128: 2899–910.

76 Smestad C, Sandvik L, Holmoy T, Harbo HF, Celius EG. Marked diff erences in prevalence of multiple sclerosis between ethnic groups in Oslo, Norway. J Neurol 2008; 255: 49–55.

77 Gale CR, Martyn CN. Migrant studies in multiple sclerosis. Prog Neurobiol 1995; 47: 425–48.

78 Pugliatti M, Riise T, Sotgiu MA, et al. Evidence of early childhood as the susceptibility period in multiple sclerosis: space-time cluster analysis in a Sardinian population. Am J Epidemiol 2006; 164: 326–33.

79 Ramagopalan SV, Valdar W, Criscuoli M, et al. Age of puberty and the risk of multiple sclerosis: a population based study. Eur J Neurol 2009; 16: 342–47.

80 Munger KL, Chitnis T, Ascherio A. Body size and risk of MS in two cohorts of US women. Neurology 2009; 73: 1543–50.

81 Freedman DM, Dosemeci M, Alavanja MC. Mortality from multiple sclerosis and exposure to residential and occupational solar radiation: a case-control study based on death certifi cates. Occup Environ Med 2000; 57: 418–21.

82 Lyon-Caen O, Jouvent R, Hauser S, et al. Cognitive function in recent-onset demyelinating diseases. Arch Neurol 1986; 43: 1138–41.

83 Young AC, Saunders J, Ponsford JR. Mental change as an early feature of multiple sclerosis. J Neurol Neurosurg Psychiatry 1976; 39: 1008–13.

84 Engell T. A clinical patho-anatomical study of clinically silent multiple sclerosis. Acta Neurol Scand 1989; 79: 428–30.

85 Gilbert JJ, Sadler M. Unsuspected multiple sclerosis. Arch Neurol 1983; 40: 533–36.

86 Lebrun C, Bensa C, Debouverie M, et al. Unexpected multiple sclerosis: follow-up of 30 patients with magnetic resonance imaging and clinical conversion profi le. J Neurol Neurosurg Psychiatry 2008; 79: 195–98.

87 Barkhof F, Filippi M, Miller DH, et al. Comparison of MRI criteria at fi rst presentation to predict conversion to clinically defi nite multiple sclerosis. Brain 1997; 120 (part 11): 2059–69.

88 Tintore M, Rovira A, Martinez MJ, et al. Isolated demyelinating syndromes: comparison of diff erent MR imaging criteria to predict conversion to clinically defi nite multiple sclerosis. Am J Neuroradiol 2000; 21: 702–06.

89 Fisniku LK, Brex PA, Altmann DR, et al. Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain 2008; 131: 808–17.

90 Dalton CM, Chard DT, Davies GR, et al. Early development of multiple sclerosis is associated with progressive grey matter atrophy in patients presenting with clinically isolated syndromes. Brain 2004; 127: 1101–07.

91 Van Au Duong M, Audoin B, Le Fur Y, et al. Relationships between gray matter metabolic abnormalities and white matter infl ammation in patients at the very early stage of MS: a MRSI study. J Neurol 2007; 254: 914–23.

92 Fernando KT, Tozer DJ, Miszkiel KA, et al. Magnetization transfer histograms in clinically isolated syndromes suggestive of multiple sclerosis. Brain 2005; 128: 2911–25.

93 Rovaris M, Judica E, Ceccarelli A, et al. A 3-year diff usion tensor MRI study of grey matter damage progression during the earliest clinical stage of MS. J Neurol 2008; 255: 1209–14.

94 Kasper LH, Shoemaker J. Multiple sclerosis immunology: the healthy immune system vs the MS immune system. Neurology 2010; 74 (suppl 1): S2–8.

95 van Oosten BW, Lai M, Hodgkinson S, et al. Treatment of multiple sclerosis with the monoclonal anti-CD4 antibody cM-T412: results of a randomized, double-blind, placebo-controlled, MR-monitored phase II trial. Neurology 1997; 49: 351–57.

96 Segal BM, Constantinescu CS, Raychaudhuri A, Kim L, Fidelus-Gort R, Kasper LH. Repeated subcutaneous injections of IL12/23 p40 neutralising antibody, ustekinumab, in patients with relapsing-remitting multiple sclerosis: a phase II, double-blind, placebo-controlled, randomised, dose-ranging study. Lancet Neurol 2008; 7: 796–804.

97 Viglietta V, Baecher-Allan C, Weiner HL, Hafl er DA. Loss of functional suppression by CD4+CD25+ regulatory T cells in patients with multiple sclerosis. J Exp Med 2004; 199: 971–79.

98 Haas J, Hug A, Viehover A, et al. Reduced suppressive eff ect of CD4+CD25high regulatory T cells on the T cell immune response against myelin oligodendrocyte glycoprotein in patients with multiple sclerosis. Eur J Immunol 2005; 35: 3343–52.

99 Michel L, Berthelot L, Pettre S, et al. Patients with relapsing-remitting multiple sclerosis have normal Treg function when cells expressing IL-7 receptor alpha-chain are excluded from the analysis. J Clin Invest 2008; 118: 3411–19.

100 Tzartos JS, Friese MA, Craner MJ, et al. Interleukin-17 production in central nervous system-infi ltrating T cells and glial cells is associated with active disease in multiple sclerosis. Am J Pathol 2008; 172: 146–55.

101 Antel JP, Rosenkoetter M, Reder A, Oger JJ, Arnason BG. Multiple sclerosis: relation of in vitro IgG secretion to T suppressor cell number and function. Neurology 1984; 34: 1155–60.

102 Jacobsen M, Cepok S, Quak E, et al. Oligoclonal expansion of memory CD8+ T cells in cerebrospinal fl uid from multiple sclerosis patients. Brain 2002; 125: 538–50.

103 Jilek S, Schluep M, Rossetti AO, et al. CSF enrichment of highly diff erentiated CD8+ T cells in early multiple sclerosis. Clin Immunol 2007; 123: 105–13.

104 Jilek S, Schluep M, Meylan P, et al. Strong EBV-specifi c CD8+ T-cell response in patients with early multiple sclerosis. Brain 2008; 131: 1712–21.

105 Hintzen RQ, Giovannoni G. CSF analysis in suspected MS: do bands aid? Neurology 2008; 70: 1059–60.

106 Freedman MS, Thompson EJ, Deisenhammer F, et al. Recommended standard of cerebrospinal fl uid analysis in the diagnosis of multiple sclerosis: a consensus statement. Arch Neurol 2005; 62: 865–70.

www.thelancet.com/neurology Vol 9 July 2010 739

Review

107 Nilsson P, Larsson EM, Maly-Sundgren P, Perfekt R, Sandberg-Wollheim M. Predicting the outcome of optic neuritis: evaluation of risk factors after 30 years of follow-up. J Neurol 2005; 252: 396–402.

108 Soderstrom M, Ya-Ping J, Hillert J, Link H. Optic neuritis: prognosis for multiple sclerosis from MRI, CSF, and HLA fi ndings. Neurology 1998; 50: 708–14.

109 Tintore M, Rovira A, Rio J, et al. Do oligoclonal bands add information to MRI in fi rst attacks of multiple sclerosis? Neurology 2008; 70: 1079–83.

110 Benczur M, Petranyl GG, Palff y G, et al. Dysfunction of natural killer cells in multiple sclerosis: a possible pathogenetic factor. Clin Exp Immunol 1980; 39: 657–62.

111 Uchida A, Maida EM, Lenzhofer R, Micksche M. Natural killer cell activity in patients with multiple sclerosis: interferon and plasmapheresis. Immunobiology 1982; 160: 392–402.

112 Braakman E, van Tunen A, Meager A, Lucas CJ. Natural cytotoxic activity in multiple sclerosis patients: defects in IL-2/interferon gamma-regulatory circuit. Clin Exp Immunol 1986; 66: 285–94.

113 Vranes Z, Poljakovic Z, Marusic M. Natural killer cell number and activity in multiple sclerosis. J Neurol Sci 1989; 94: 115–23.

114 Kreuzfelder E, Shen G, Bittorf M, et al. Enumeration of T, B and natural killer peripheral blood cells of patients with multiple sclerosis and controls. Eur Neurol 1992; 32: 190–94.

115 Munschauer FE, Hartrich LA, Stewart CC, Jacobs L. Circulating natural killer cells but not cytotoxic T lymphocytes are reduced in patients with active relapsing multiple sclerosis and little clinical disability as compared to controls. J Neuroimmunol 1995; 62: 177–81.

116 Kastrukoff LF, Morgan NG, Zecchini D, et al. A role for natural killer cells in the immunopathogenesis of multiple sclerosis. J Neuroimmunol 1998; 86: 123–33.

117 Takahashi K, Miyake S, Kondo T, et al. Natural killer type 2 bias in remission of multiple sclerosis. J Clin Invest 2001; 107: R23–29.

118 Kastrukoff LF, Lau A, Wee R, Zecchini D, White R, Paty DW. Clinical relapses of multiple sclerosis are associated with ‘novel’ valleys in natural killer cell functional activity. J Neuroimmunol 2003; 145: 103–14.

119 Takahashi K, Aranami T, Endoh M, Miyake S, Yamamura T. The regulatory role of natural killer cells in multiple sclerosis. Brain 2004; 127: 1917–27.

120 Bielekova B, Catalfamo M, Reichert-Scrivner S, et al. Regulatory CD56(bright) natural killer cells mediate immunomodulatory eff ects of IL-2Ralpha-targeted therapy (daclizumab) in multiple sclerosis. Proc Natl Acad Sci USA 2006; 103: 5941–46.

121 Rinaldi L, Gallo P, Calabrese M, et al. Longitudinal analysis of immune cell phenotypes in early stage multiple sclerosis: distinctive patterns characterize MRI-active patients. Brain 2006; 129: 1993–2007.

122 Saraste M, Irjala H, Airas L. Expansion of CD56Bright natural killer cells in the peripheral blood of multiple sclerosis patients treated with interferon-beta. Neurol Sci 2007; 28: 121–26.

123 De Jager PL, Rossin E, Pyne S, et al. Cytometric profi ling in multiple sclerosis uncovers patient population structure and a reduction of CD8low cells. Brain 2008; 131: 1701–11.

124 Polman CH, O’Connor PW, Havrdova E, et al, AFFIRM Investigators. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N Engl J Med 2006; 354: 899–910.

125 Coles AJ, Compston DA, Selmaj KW, et al, CAMMS223 Trial Investigators. Alemtuzumab vs. interferon beta-1a in early multiple sclerosis. N Engl J Med 2008; 359: 1786–801.

126 Hauser SL, Waubant E, Arnold DL, et al, HERMES Trial Group. B-cell depletion with rituximab in relapsing-remitting multiple sclerosis. N Engl J Med 2008; 358: 676–88.

127 Bielekova B, Becker BL. Monoclonal antibodies in MS: mechanisms of action. Neurology 2010; 74 (suppl 1): S31–40.

128 McFarland HF, Martin R. Multiple sclerosis: a complicated picture of autoimmunity. Nat Immunol 2007; 8: 913–19.

129 Friese MA, Fugger L. Pathogenic CD8(+) T cells in multiple sclerosis. Ann Neurol 2009; 66: 132–41.

130 Goodin DS. The causal cascade to multiple sclerosis: a model for MS pathogenesis. PLoS One 2009; 4: e4565.

131 Ascherio A, Munger KL. Environmental risk factors for multiple sclerosis. Part II: noninfectious factors. Ann Neurol 2007; 61: 504–13.

132 Healy BC, Ali EN, Guttmann CR, et al. Smoking and disease progression in multiple sclerosis. Arch Neurol 2009; 66: 858–64.

133 Di Pauli F, Reindl M, Ehling R, et al. Smoking is a risk factor for early conversion to clinically defi nite multiple sclerosis. Mult Scler 2008; 14: 1026–30.

134 Farrell RA, Antony D, Wall GR, et al. Humoral immune response to EBV in multiple sclerosis is associated with disease activity on MRI. Neurology 2009; 73: 32–38.

135 Lunemann JD, Tintore M, Messmer B, et al. Elevated Epstein-Barr virus-encoded nuclear antigen-1 immune responses predict conversion to multiple sclerosis. Ann Neurol 2009; 67: 159–69.

136 Hasler G, Drevets WC, Gould TD, Gottesman, II, Manji HK. Toward constructing an endophenotype strategy for bipolar disorders. Biol Psychiatry 2006; 60: 93–105.

137 Somma P, Ristori G, Battistini L, et al. Characterization of CD8+ T cell repertoire in identical twins discordant and concordant for multiple sclerosis. J Leukoc Biol 2007; 81: 696–710.

138 Nuwer MR, Visscher BR, Packwood JW, Namerow NS. Evoked potential testing in relatives of multiple sclerosis patients. Ann Neurol 1985; 18: 30–34.

139 De Stefano N, Cocco E, Lai M, et al. Imaging brain damage in fi rst-degree relatives of sporadic and familial multiple sclerosis. Ann Neurol 2006; 59: 634–39.

140 Corvol JC, Pelletier D, Henry RG, et al. Abrogation of T cell quiescence characterizes patients at high risk for multiple sclerosis after the initial neurological event. Proc Natl Acad Sci USA 2008; 105: 11839–44.

141 De Jager PL, Simon KC, Munger KL, Rioux JD, Hafl er DA, Ascherio A. Integrating risk factors: HLA-DRB1*1501 and Epstein-Barr virus in multiple sclerosis. Neurology 2008; 70: 1113–18.

142 Islam T, Gauderman WJ, Cozen W, Hamilton AS, Burnett ME, Mack TM. Diff erential twin concordance for multiple sclerosis by latitude of birthplace. Ann Neurol 2006; 60: 56–64.

143 Haghighi S, Andersen O, Rosengren L, Bergstrom T, Wahlstrom J, Nilsson S. Incidence of CSF abnormalities in siblings of multiple sclerosis patients and unrelated controls. J Neurol 2000; 247: 616–22.

144 Celius EG, Harbo HF, Egeland T, Vartdal F, Vandvik B, Spurkiand A. Sex and age at diagnosis are correlated with the HLA-DR2, DQ6 haplotype in multiple sclerosis. J Neurol Sci 2000; 178: 132–35.

145 Hensiek AE, Sawcer SJ, Feakes R, et al. HLA-DR 15 is associated with female sex and younger age at diagnosis in multiple sclerosis. J Neurol Neurosurg Psychiatry 2002; 72: 184–87.

146 Sundstrom P, Nystrom L, Jidell E, Hallmans G. EBNA-1 reactivity and HLA DRB1*1501 as statistically independent risk factors for multiple sclerosis: a case-control study. Mult Scler 2008; 14: 1120–22.

147 Nielsen TR, Rostgaard K, Askling J, et al. Eff ects of infectious mononucleosis and HLA-DRB1*15 in multiple sclerosis. Mult Scler 2009; 15: 431–36.

148 De Jager PL, Chibnik LB, Cui J, et al. Integration of genetic risk factors into a clinical algorithm for multiple sclerosis susceptibility: a weighted genetic risk score. Lancet Neurol 2009; 8: 1111–19.

149 Kelly MA, Cavan DA, Penny MA, et al. The infl uence of HLA-DR and -DQ alleles on progression to multiple sclerosis following a clinically isolated syndrome. Hum Immunol 1993; 37: 185–91.

150 Hauser SL, Oksenberg JR, Lincoln R, et al. Interaction between HLA-DR2 and abnormal brain MRI in optic neuritis and early MS. Neurology 2000; 54: 1859–61.