Multiple Sclerosis Resource Centre

Welcome to the Multiple Sclerosis Resource Centre. This website is intended for international healthcare professionals with an interest in Multiple Sclerosis. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Peptide motif analysis predicts lymphocytic choriomeningitis virus as trigger for multiple sclerosis

Molecular Immunology, Volume 67, Issue 2, Part B, October 2015, Pages 625–635

Highlights

  • LCMV peptide matches MBP more closely than peptides shown to cross-react with MBP.
  • LCMV is concentrated in geographical regions where MS incidence is highest.
  • LCMV inhibits production of type I interferons.
  • Both immune dysregulation and molecular mimicry may contribute to initiate MS.

Abstract

The etiology of multiple sclerosis (MS) involves both genetic and environmental factors. Genetically, the strongest link is with HLA DRB1*1501, but the environmental trigger, probably a virus, remains uncertain. This investigation scans a panel of proteins from encephalitogenic viruses for peptides homologous to the primary autoantigen from myelin basic protein (MBP), then evaluates candidate peptides against a motif required for T cell cross-reactivity and compares viral prevalence patterns to epidemiological characteristics of MS. The only peptide meeting criteria for cross-reactivity with MBP was one from lymphocytic choriomeningitis virus (LCMV), a zoonotic agent. In contrast to current candidates such as Epstein–Barr virus, the distribution of LCMV is consistent with epidemiological features of MS, including concentration in the temperate zone, higher prevalence farther from the equator, and increased prevalence in proximity to regions of peak MS incidence, while lack of person-to-person transmission is consistent with low MS concordance across monozygotic twins. Further, LCMV blocks induction of type I interferon (IFN). Hypothetically this would dysregulate immune processes in favor of proinflammatory pathways as well as upregulating HLA class II and providing more binding sites for autoantigen. The combination of molecular mimicry with virally-induced immune dysregulation has the potential to explain aspects of autoimmunity not addressed by either mechanism alone.

Keywords: Multiple sclerosis, DRB1*1501, Lymphocytic choriomeningitis virus (LCMV), Interferon, Immune regulation, Molecular mimicry.

1. Introduction

Multiple sclerosis (MS) is a chronic immune-mediated neurological disease that affects approximately 1.3 million people worldwide. It is generally accepted that both genetic and environmental factors are involved, but disease etiology remains poorly understood after decades of research. Recent advances have identified a number of genetic risk factors, with HLA DRB1*1501 haplotype continuing to provide the strongest association (Gourraud et al, 2012, Wu et al, 2010, Qiu et al, 2011, Link et al, 2012, and Nolan et al, 2012). Yet, with approximately 9% of the world’s population carrying this allele (Solberg et al., 2008) and a global MS prevalence of 0.03% (World Health Organization, 2008), the proportion of DRB1*1501-positive individuals who develop MS is low, on the order of 0.3%. Among environmental factors, a viral agent has been postulated to instigate immune recognition of self-antigens (Fujinami, 2001), hypothetically by means of a mechanism termed “molecular mimicry” (Olson et al., 2001). Under this model, structural similarities between a viral peptide and a self-peptide cause activation of autoreactive T cells. Although a range of viruses has been considered over the past 50 years, including poliovirus, measles, rabies, herpes family viruses, mumps, canine distemper, and retroviruses (Kakalacheva et al., 2011), the identity of a causative virus remains elusive.

A popular candidate for an etiologic role in MS is Epstein–Barr virus (EBV), as antibodies against its nuclear antigen-1 (EBNA1) have been found in MS patients (Kakalacheva et al., 2011). However, the distribution of EBV worldwide does not provide a good fit to the epidemiology of MS. MS is concentrated in the temperate zone, with highest prevalence in Europe, Canada, the United States, and Australia (World Health Organization, 2008). Within the temperate zone, MS shows a gradient of prevalence that increases with latitude (Simpson et al., 2011) and is interspersed with regional pockets of exceptionally high prevalence. Although some of these differences can be explained by genetic factors, MS concordance across monozygotic twin pairs is low, ranging from 13% to 31%, based on studies in Canada, the United States, the British Isles, Finland, and Italy (Willer et al, 2003, Islam et al, 2006, Mumford et al, 1994, Kuusisto et al, 2008, Ristori et al, 2006, and Sadovnick et al, 1993). This suggests that the environmental trigger for MS in genetically susceptible individuals is somewhat uncommon or has low infectivity. In contrast, the EBV seropositivity rate in adults is in excess of 90% (Kakalacheva et al., 2011). Further, exposure to EBV occurs earlier in life among children in developing countries, with universal seroconversion by age 3–4, whereas infection in developed countries often is delayed until adolescence (Hjalgrim et al., 2007). The high seroprevalence of antibodies against EBV, together with earlier exposure in countries with lower MS prevalence, are not fully consistent with the latitudinal gradient and low twin concordance rates seen for MS. In fact, some researchers suggest that EBV is a marker of chronic brain inflammation rather than causative per se (Castellazi et al., 2014).

The ideal candidate for an infectious trigger interacting with DRB1*1501 under the molecular mimicry hypothesis would satisfy both molecular biological and epidemiological criteria. The infectious agent would contain a peptide that binds to HLA in a similar way as the self-antigen and lies in a similar configuration. The bound peptide would activate the same T cell clones as those that recognize the self-antigen. The infectious agent would be somewhat uncommon or have low infectivity in order to explain the low MS concordance across monozygotic twin pairs. The agent would be most prevalent in the temperate geographic zone. Its distribution would be consistent with the latitudinal gradient observed for MS. It would show higher prevalence or infectivity in regions that report elevated incidence or prevalence of MS.

The aim of this investigation was to predict the viral peptide most closely matching these criteria. A set of proteins from viruses capable of causing encephalitis was scanned for regions of sequence similarity to myelin basic protein (MBP) residues 85–99. The peptides with highest sequence homology were then compared to MBP 85–99 on five scales representing characteristics predictive of protein binding and configuration. The highest scoring viral peptides were evaluated for similarity to a binding motif that has been determined experimentally to activate MBP-reactive T cell clones from MS patients with DRB1*1501 haplotype. Finally, the plausibility of the top predicted virus was evaluated through a review of its epidemiology and a comparison to the prevalence patterns observed for MS.

2. Materials and methods

2.1. General

All computations were done with custom programs written in the R language (Hornik, 2014).

2.2. Viral proteins

A list of viruses capable of causing encephalitis was generated from review of medical reference texts. Encephalitogenic viruses endemic to equatorial regions were excluded as unlikely to be causative, since MS is most prevalent in the temperate zone. Protein sequences derived from the viral capsid or envelope or previously observed to be antigenic were selected for testing. Protein sequences were obtained from the UniProt database.

2.3. Reference proteins

Viral homology scores were contrasted with scores from two other comparators. As a negative control for the MBP immunogen, an arbitrary 15-mer from serum albumin (ALB), residues 152–166, was selected. As a negative control for the viral proteins chosen for testing, a set of 17 control proteins was created by generating a chain of randomly selected amino acids whose frequency of occurrence was similar to the amino acid composition of proteins in general. The random proteins were matched for length with the viral proteins.

The segment of MBP that constitutes the immunogenic portion was defined as MBP residues 85–99, based on experimental evidence provided by Wucherpfennig et al. (1994) and Hausmann et al. (1999). Viral protein sequences were scanned for regions of high homology with this peptide. As it is unknown whether high homology is required for the entire peptide length, homology was computed for a range of windows on the MBP 85–99 sequence. All possible windows of length 4–15 that covered the region MBP 89–92, the primary anchor region of the MBP immunogen bound to DRB1*1501, were included. This resulted in 40 windows on the MBP immunogen.

Homology scores were computed using the BLOSUM50 blocks substitution matrix (Henikoff and Henikoff, 1992) and were based on alignments without gaps. For each of the 40 windows, each candidate viral protein was marched through all possible alignments with the MBP sequence that fell within the window. Thus, for a protein of length n and a window width k, a total of (n – k + 1) homology scores were computed. For each viral protein, the peptide location resulting in the maximum score for each window was saved. A limited comparison of results using BLOSUM62 showed that conclusions were not dependent on the BLOSUM matrix used (data not shown).

To allow comparison of homology scores across window locations and widths, scores were normalized to window length by averaging. Average homology scores were scaled to percent of maximum, defined as the homology score obtained when comparing the windowed segment of the MBP immunogen with itself, using the value –1 as minimum of the range. The distribution of homology scores was compared across peptide source (viral or random) using a Wilcoxon rank sum statistic at level 0.05. A Simes test was used to control for multiple comparisons. Since the windows were overlapping and nested, the statistics used are considered exploratory rather than indicating true statistical significance. The program and datasets used to conduct this analysis online in a set of six Supplementary files. Please refer to the README.txt file for usage of the remaining 5 items (homology.txt, MBP.csv, albumin.csv, BLOSUM50.csv, and protein_panel.csv).

2.5. Peptide characteristics profiles

Viral peptides were compared to MBP 85–99 on five scales used to predict protein conformation or binding. These were surface accessibility, antigenicity, flexibility, hydrophobicity, and hydrophilicity (Janin et al, 1978, Welling et al, 1985, Karplus and Schulz, 1985, Eisenberg et al, 1984, and Parker et al, 1986). For each scale, a profile for MBP 85–99 was defined as the sequence of scale scores for the 15 residues. The MBP profile was then compared with that of each candidate homologous viral peptide by summing the positionwise squared deviations in scale scores. Dividing by the peptide length, each homologous viral peptide was given a “mean squared error” (MSE) value that reflected the degree to which the profile of the viral peptide deviated from that of MBP on the given scale. The MSE values were converted to percents by dividing by the maximum possible deviation from MBP that could have been obtained using that scale.

For reference, a set of microbial peptides shown experimentally to be capable of activating a DRB1*1501-restricted MBP-reactive T cell clone (Hausmann et al., 1999) are also included as positive controls.

2.6. Motif for peptide binding and cross-reactivity

A motif was built to codify the criteria a peptide must meet in order to cross-react with MBP in binding to HLA DRA-DRB1*1501 and activating the same T cell clones. These criteria were summarized from X-ray crystallographic data and experimental evidence provided by others, as described here.

Requirements for the primary anchor positions P1 and P4 are derived from Smith et al. (1998) and Li et al. (2000). Substitution of alanine instead of lysine or arginine at DRβ71 was observed to change the P4 pocket from a small charged pocket to a large hydrophobic one, which then preferred large aromatic residues, specifically phenylalanine or tyrosine. The DRβ substitution G86V is associated with more severe disease (Teutsch et al., 1999) and results in the P1 pocket becoming relatively shallow, preferring small aliphatic amino acids such as leucine or valine.

Requirements for binding to the T cell receptor and activating MBP-specific DRB1*1501-restricted T cell clones are derived from the work of Wucherpfennig et al. (1994) and Hausmann et al. (1999). These researchers demonstrated requirements for the 3 positions that make strongest contact with the T cell receptor. Exact matches to P2H, P3F, and P5K were preferred, with possible substitutions of Y at P3 and R at P5. Peptides that were experimentally determined to cross-react showed that F or Y at P4 was not a strict requirement if P1 was occupied by V.

Requirements for peptide flexibility are based on the findings of Stern et al. (1994) and Li et al. (2000). These studies showed that peptides bound to HLA DR adopt a twisted conformation. Residues with high flexibility, making them conducive to twisting, were identified from the Karplus flexibility scale (1985).

2.7. Viral prevalence reports

To obtain reports on the prevalence of lymphocytic choriomeningitis virus (LCMV) worldwide, PubMed (http://www.ncbi.nlm.nih.gov/pubmed/advanced) was searched for articles matching the criteria [title/abstract = (“LCMV” OR “choriomeningitis”)] AND [keyword = (“prevalence” OR “seroprevalence” OR “seropositive” OR “serological” OR “serologic”)] during the 41-year period from 1974 to 2014. In addition, the collection was supplemented with references from the bibliography of other articles reviewed. Only English-language articles or those that provided prevalence estimates in an English-language abstract were used. Since very few articles met these criteria, none were excluded. No effort was made to evaluate the sampling or assay methods, or to restrict the journals that were included in the survey. One article from South India that was not available in either paper or electronic form in any campus of the University of California library system was not included.

2.8. Prevalence limits corresponding to twin concordance

Assuming that viral exposure is necessary for disease development, the problem was conceptualized using a 2 × 2 table of viral exposure for Twin A versus Twin B; this table is assumed symmetric. Let the (1, 1) cell represent the proportion of monozygotic twin pairs for which both twins have been exposed to virus. The constraint that the sum of all cells is equal to 1 sets an upper bound for the (1, 1) value at the twin concordance rate. The viral prevalence rate for the population was computed as the marginal sum of the first row (or column, being symmetric). An additional constraint was required to restrain the correlation between exposure of one twin and exposure of the other to fall within the interval (0, 1). Correlation was estimated by the φ coefficient (the Pearson correlation coefficient for binary variables).

Viral prevalence requirements were computed to correspond with the lowest and highest twin concordance rates from the literature, 13% (Islam et al., 2006) and 31% (Sadovnick et al., 1993), respectively. For each of these values, the viral prevalence corresponding to between-twin correlations of 0, 0.5, and 0.9 were estimated. Due to intractability of closed-form solutions for the sets of equations, solutions were computed empirically.

2.9. Evaluation of temperate zone concentration

The temperate zone was defined as the range of latitudes from the Tropic of Cancer/Capricorn (23.5°) to the Arctic/Antarctic Circle (66.5°). The contrasting equatorial zone was defined as the region between the Tropic of Cancer (23.5° N) and the Tropic of Capricorn (23.5° S).

Since all LCMV prevalence estimates were obtained within the temperate zone, the analysis objective was to determine whether the number of countries in the temperate zone reporting LCMV prevalence was a significant proportion when the locations of countries not providing prevalence reports were considered.

The following 65 countries lying within the temperate zone were included: Afghanistan; Algeria; Argentina; Australia; Austria; Azerbaijan; Bangladesh; Belarus; Belgium; Bhutan; Bulgaria; Canada; China; Croatia; Czech Republic; Denmark; Egypt; Estonia; Finland; France; Georgia; Germany; Greece; Hungary; Iran; Iraq; Ireland; Israel; Italy; Japan; Jordan; Kazakhstan; Korea; Kyrgyzstan; Latvia; Lebanon; Libya; Lithuania; Mongolia; Morocco; Nepal; Netherlands; New Zealand; Norway; Pakistan; Poland; Portugal; Romania; Russia; Saudi Arabia; Serbia; Slovakia; Slovenia; South Africa; Spain; Sweden; Syria; Tajikistan; Turkey; Turkmenistan; United Kingdom; Ukraine; United States; Uruguay; and Uzbekistan.

To control for the possibility that many countries in the equatorial zone are less developed than those in Europe, Australia, and North America, a country's ability to measure viral prevalence was assessed by determining whether each country had reported on prevalence of EBV. The search for EBV prevalence reports in equatorial regions used the criteria [title/abstract = (“Epstein–Barr”)] AND [title/abstract = (“prevalence” OR “seroprevalence” OR “seropositive” OR “serological” OR “serologic”)] AND [title/abstract = (CountryName)]. Sixty-six country names were tested in this way. These were: Angola; Belize; Benin; Bolivia; Brazil; Burkina; Burundi; Cambodia; Cameroon; Central African Republic; Chad; Colombia; Congo (either one); Costa Rica; Cuba; Dominican Republic; Ecuador; El Salvador; Eritrea; Ethiopia; Gabon; Gambia; Ghana; Guatemala; Guinea; Guinea Bissau; Guyana; Haiti; Honduras; Indonesia; Ivory Coast; Kenya; Laos; Liberia; Macau; Madagascar; Malawi; Malaysia; Mali; Mauritania; Mexico; Mozambique; Myanmar OR Burma; Nicaragua; Niger; Nigeria; Oman; Panama; Peru; Philippines; Rwanda; Senegal; Sierra Leone; Singapore; Somalia; Sudan; Suriname; Tanzania; Thailand; Togo; Uganda; Venezuela; Vietnam; Yemen; Zambia; Zimbabwe. Any country with at least one article meeting these criteria was counted as capable of providing viral prevalence estimates. The articles were not reviewed for content or relevance.

To compare the frequency of prevalence reports on LCMV across the two zones, all temperate zone countries and those equatorial zone countries that had produced at least one article on EBV were included in analysis. The proportion of countries producing prevalence reports on LCMV was compared across zone (temperate or equatorial) using a chi-square test.

2.10. Evaluation of latitudinal gradient

In the analysis of LCMV seroprevalence by latitude, prevalence estimates were assigned to two groups, those ≤5%, and those >5%. The dependence of prevalence values >5% on latitude was then estimated using a linear model that included all data points. The values ≤5% were fit with a mean value only, while the values >5% were fit by intercept and slope as a function of latitude. Incorporation of a factor variable indicating the source organism, human or rodent, did not turn out to be statistically significant, so this factor was not included in the final model. Similarly, year of publication was not significant and therefore also was not included in the final model. Prevalence values with latitudes are listed in Supplementary dataset S2.

2.11. Geographic correspondence of viral prevalence with MS incidence rates

The geographic locations where MS prevalence or incidence is highest were compared to the locations reporting high prevalence of LCMV. Countries with the highest MS prevalence or incidence were identified from an atlas compiled by the World Health Organization (2008). In addition, a literature search was conducted to obtain information on MS incidence for those countries reporting high LCMV prevalence. Comparisons were descriptive and no formal statistical comparison was conducted.

In the comparison of the Gorski Kotar–Kočevje region with Germany, estimates of MS prevalence, MS incidence, DR2 prevalence, and LCMV prevalence were obtained as described below. For Germany, MS prevalence and incidence were taken from the World Health Organization (2008) atlas, and DR2 prevalence was estimated by the prevalence of DRB1*1501 from Solberg et al. (2008). Since there were no LCMV prevalence reports for humans in Germany, the rate was estimated based on the relationship between human and rodent prevalence figures and on observations from other countries in Western Continental Europe. For France, Spain, and Italy, all rodent prevalence estimates were higher than prevalence estimates in humans. Prevalence in the Netherlands was reported for human subjects only. As a conservative approach, human prevalence against LCMV in Germany was estimated by the maximum of rodent prevalence from Germany and human prevalence from the Netherlands, France, Spain, or Italy.

Rates for the Gorski Kotar–Kočevje region were estimated as follows. LCMV prevalence was estimated by the figure for Vir, Croatia (Dobec et al., 2006). MS prevalence was taken from Peterlin et al. (2006). MS incidence was assumed to be that of Croatia as reported in the World Health Organization (2008) atlas. DR2 prevalence was estimated for the specific population based on Materljan and Sepcic (2002) and Peterlin et al. (2006) according to the following algorithm. A 2 × 2 table of MS status (positive/negative) by DR2 haplotype (present/absent) for the population of this region was created. The marginal total for MS-positive persons was set at the MS prevalence rate of 151.9 per 100,000 according to Peterlin et al. (2006). The proportion of the population who had MS and who also carried the DR2 haplotype was computed by multiplying the rate of 20.6% reported in Materljan and Sepcic (2002) by the MS prevalence rate just cited. The proportion of the population with MS but lacking the DR2 haplotype was obtained by subtraction. The proportion of the population who do not have MS was computed by subtracting the MS prevalence rate from 100%. The proportion of the population who do not have MS but do carry the DR2 haplotype was estimated by multiplying this marginal total by the rate of 8.3% reported for controls by Materljan and Sepcic (2002). The remaining cell value was computed by subtraction.

The resultant 2 × 2 table is shown here for reference. The value of 8.32% from the table below was used for the comparison with Germany. Plausibility of this estimate was confirmed by comparison with the rate of 0.08962 reported for Hvar, Croatia, in Solberg et al. (2008).

DR2 haplotype status
MS status Present Absent All haplotypes
Positive 0.03 0.12 0.15
Negative 8.29 91.56 99.85
All subjects 8.32 91.68 100.00

3. Viral peptide search results

Ten viruses that have the potential to cause encephalitis in regions with a temperate climate were identified from a review of medical reference texts. From these 10 viruses, sequences for 17 proteins were obtained from the UniProt database and searched for regions of homology with MBP 85–99 using overlapping windows of varying lengths. These proteins are listed in Table 1.

Table 1

Viral proteins tested.

 

Virus Name Protein Name UniProt ID
Coxsackievirus A7 Genome polyprotein Q6JKS1
Echovirus 11 (strain Gregory) Genome polyprotein P29813
Enterovirus D68 Genome polyprotein A1E4A3
Epstein–Barr virus (strain B95-8) (HHV-4) (Human herpesvirus 4) Epstein–Barr nuclear antigen 1 P03211
Epstein–Barr virus (strain B95-8) (HHV-4) (Human herpesvirus 4) Major capsid protein P03226
Epstein–Barr virus (strain B95-8) (HHV-4) (Human herpesvirus 4) Capsid protein VP26 P14348
Epstein–Barr virus (strain GD1) (HHV-4) (Human herpesvirus 4) Membrane glycoprotein BILF2 I1YP58
Human cytomegalovirus (strain AD169) (HHV-5) (Human herpesvirus 5) Envelope glycoprotein B P06473
Human cytomegalovirus (strain AD169) (HHV-5) (Human herpesvirus 5) 65 kDa phosphoprotein P06725
JC polyomavirus (JCPyV) (JCV) Major capsid protein VP1 P03089
JC polyomavirus (JCPyV) (JCV) Large T antigen P03072
JC polyomavirus (JCPyV) (JCV) Minor capsid protein VP2 P03095
Lymphocytic choriomeningitis virus (strain Armstrong) (LCMV) Nucleoprotein P09992
Measles virus (strain Edmonston) (Subacute sclerose panencephalitis virus) Nucleoprotein Q89933
Measles virus (strain Edmonston) (Subacute sclerose panencephalitis virus) Hemagglutinin glycoprotein P08362
Poliovirus type 1 (strain Sabin) Genome polyprotein P03301
West Nile virus (WNV) Capsid protein C (position 2–105) P06935

Virtually all encephalitogenic viral proteins in the test set showed some degree of homology with MBP 85–99. The highest relative homology score was seen with the LCMV nucleoprotein (NP) in the region overlapping MBP 89–92, the location of the anchor sequence that binds to the P1–P4 pockets of HLA DRB1*1501. All of the 10 viruses showed their highest homology score within this 4-residue sequence (with one extending to the P5 position). Percent homology scores within the peak homology region ranged from 54 to 86%. The schematic in Fig. 1 displays the peak scoring regions for each virus, listed in descending order of the percent homology score.

[locator:gr1] (link type not set)

Fig. 1

Segments of viral proteins showing highest sequence homology to the immunodominant peptide MBP 85–99. For each virus, the region with highest percent homology was selected. These regions are boxed. The surrounding region, positioned along the full length of the MBP 15-mer, is shown for context. Color coding corresponds to consensus symbols used by Clustal Omega on the UniProt website: orange (*) = exact match; yellow (:) = strongly similar (Gonnet PAM250 score >0.5); green (.) = weakly similar (Gonnet PAM250 score ≤0.5); no color () = dissimilar. Above the MBP immunogen sequence, the location of residues that bind to DRA, DRB1*1501 pockets P1, P4, P6, and P9 are indicated for reference.

 

Median relative homology of the viral proteins increased as the window size narrowed, from 31% for the window spanning all 15 residues to 63% for MBP 89–92. This means that half of the 17 proteins from the encephalitogenic viruses showed at least 63% relative homology to the primary anchor region where MBP 89–92 binds to DRB1*1501.

For comparative purposes, each viral protein was scanned for maximum homology to an arbitrary peptide from serum albumin (ALB) as a negative control. The median percent homology score for viral peptides relative to ALB 156–159 was 64%. This is almost identical to the median against MBP 89–92. Thus, a region of high homology over the 4-residue sequence is not unique to MBP.

Maximum homology scores against MBP and against ALB were also determined for a set of randomly generated sequences of amino acids matched for length. Viral peptides showed significantly (p < 0.05) higher homology to MBP than did random peptides for 4 (10%) of the 40 overlapping windows, while homology scores against ALB did not differ significantly between viral and random peptides for any of the windows. After controlling for multiplicity, a significant difference was seen for the contrast of viral peptides against random sequences in their homology with MBP residues 89–92 (p = 0.006). This is the region that binds to the P1–P4 pockets of DRB1*1501.

3.2. Peptide binding characteristics profiles

The viral peptides in Fig. 1 were compared with MBP 85–99 on five scales: surface accessibility, antigenicity, flexibility, hydrophobicity, and hydrophilicity. The goal was to determine how closely the pattern of structural or binding characteristics of each test peptide matched those of MBP 85–99 on a position-by-position basis. Table 2 presents the comparison of viral peptides to MBP 85–99 on these scales. For reference, a set of microbial peptides shown experimentally to be capable of activating a DRB1*1501-restricted MBP-reactive T cell clone are also included as positive controls.

Table 2

Viral peptide sequence profiles compared to MBP 85–99 (MSE%).

 

% MSE Compared to MBP
Peptide source Surface Accessibility Antigenicity Flexibility Hydrophobicity Hydrophilicity
Test viruses:
LCMV 6 8 4 16 7
JCV 18 5 25 43 20
CMV 24 7 26 47 14
Enterovirus D68 32 16 38 80 26
Poliovirus 19 6 17 32 9
Echovirus 11 32 10 17 68 9
EBV 29 7 22 56 23
Coxsackievirus A7 28 15 42 68 29
Measles virus 25 6 28 57 31
West Nile virus 13 10 22 45 31
Positive controls:
S. aureus 33 15 20 59 25
M. avium 25 7 13 38 5
M. tuberculosis 26 9 12 38 3
B. subtilis 36 19 23 42 20
E. coli / H. influenzae 31 16 14 47 12
Negative control:
Serum albumin 18 21 22 47 18

Abbreviations: LCMV, lymphocytic choriomeningitis virus; JCV, JC polyoma virus; EBV, Epstein–Barr virus; CMV, cytomegalovirus.

Notes: The test virus peptides compared are those presented in Fig. 1. The positive controls are those taken from Hausmann et al. (1999), Table IV. The negative control was an arbitrary peptide from serum albumin 152–166. Values presented are mean squared error when compared to MBP 85–99, scaled as a percent of maximum deviation possible on the specified scale. The most favorable number in the test viral section is bolded.

Six viral peptides matched the MBP profile at least as well as the positive controls on the antigenicity scale. These were peptides from LCMV, JC virus (JCV), poliovirus, Epstein–Barr virus (EBV), cytomegalovirus (CMV), and measles. On the other four scales, LCMV matched the MBP profile substantially better than any of the other viral peptides tested. LCMV stood out as exceptionally close to MBP in surface accessibility, flexibility, and hydrophobicity, outperforming even the microbial peptides that cross-reacted with MBP-specific T cell clones. A visual display of the positionwise profile match between LCMV and MBP 85–99 on the flexibility and surface accessibility scales is provided in Fig. 2. Profiles of Mycobacterium avium and serum albumin are shown for comparison, as positive and negative controls, respectively. From the figure it can be observed that the negative control differs from the others in both flexibility and surface accessibility in the residue that would occupy the P1 pocket, and in surface accessibility for the P2 pocket. Yet possibly of more interest is to see that the LCMV peptide matches MBP in the region from P –4 through P –1 more closely on both flexibility and surface accessibility than either the positive or the negative controls. The importance of these positions is highlighted by the experimental observations that the P –1 residue makes contact with the T cell receptor (TCR) and participates in T cell specificity, and truncation of the P –3 and P –4 residues abolish T cell activation (Smith et al., 1998).

[locator:gr2] (link type not set)

Fig. 2

Positionwise peptide flexibility and surface accessibility profiles compared to MBP 85–99. The MBP immunogen profile is depicted in gray. Comparison peptides are overlain in color. The location of MBP residues binding to pockets P1, P4, P6, and P9 of HLA DRA, DRB1*1501 are indicated on the x-axis. The similarity of LCMV NP 400–414 to MBP (top) can be compared with the best-fit positive control from Mycobacterium avium (center) and a negative control taken from an arbitrary segment of serum albumin (bottom).

 

3.3. Comparison to experimentally determined motifs

A motif that specifies the requirements for DRB1*1501-restricted MBP-reactive T cell activation, developed as described in Section 2.6, is displayed in Table 3. Half of the peptides listed in Fig. 1 met the criteria of having HF at positions P2 and P3. These were peptides derived from LCMV, JCV, CMV, poliovirus, and coxsackievirus A7. Of these, only one, LCMV, had a small hydrophobic (isoleucine) at P1. P4 in LCMV is occupied by tyrosine, consistent with cross-reactive peptides that do not have valine at P1. LCMV was also distinguished from the others by the presence of a positively charged residue, arginine, at P5. The final requirement, that of chain flexibility throughout the remainder of the peptide, matches MBP position-by-position more closely for LCMV than for any other peptide evaluated, including those that cross-react experimentally. The MSE% deviation from MBP for the LCMV peptide in the tails outside the primary binding region was 5%, compared to 16–39% for the other viral candidates and 19–32% for peptides shown to cross-react experimentally (Fig. 2; Table 3). This peptide from the LCMV nucleoprotein is the only viral peptide in the test group that met all criteria for cross-reactivity.

Table 3

Peptide profile motif required for DRB1*1501-restricted MBP-specific T cell reactivity.

 

Position Preference MBP LCMV JCV CMV Polio CoxA7
P2 H H H H H H H
P3 F F F F F F F
Primary Either: P1 = V P1 = I P1 = F P1 = Y P1 = S P1 = F
anchor

region
  • P1 occupied by V; or
  • P4 = F or Y, plus concomitant small hydrophobic residue at P1
P4 = F P4 = Y P4 = F P4 = F P4 = Y P4 = K
P5 Positively charged residue such as K or R K R S A D Y
Remaining residues Similar flexibility profile for P –4 to P –1, P6–P11 (MSE %) 0 5% 36% 32% 16% 39%

Abbreviations: MBP, myelin basic protein residues 85–99; LCMV, lymphocytic choriomeningitis virus; JCV, JC polyoma virus; CMV, cytomegalovirus; Polio, poliovirus; CoxA7, coxsackievirus A7.

Residues that meet criteria are bolded.

4. Epidemiological evaluation

LCMV is a zoonotic agent of the family Arenaviridae. Its primary host is the common house mouse, Mus musculus (Lehmann-Grube, 1971), although other rodents, including pets, may become infected (Biggar et al., 1975). The virus persists in asymptomatic carrier mice and is discharged in nasal secretions, saliva, milk, blood, and urine (Lehmann-Grube, 1971, Traub, 1938, and Childs et al, 1992). The distribution of LCMV-seropositive mice is uneven and locally clustered within a mouse population, tending to spread within lineages and tightly linked social groups or territorial subunits, and may infect mice in one household while the neighboring house remains infection-free (Childs et al., 1992). Transmission to humans is primarily by inhalation of aerosolized rodent excreta or by bites or contact with rodent urine, feces, or saliva (Charrel and de Lamballerie, 2010). Although LCMV may result in encephalitis or meningitis, the majority of infected patients develop a mild flu-like illness or are asymptomatic (Bonthius et al., 2007).

The distribution of LCMV was compared to four characteristics observed in epidemiological studies of MS: concordance rates across monozygotic twin pairs, concentration in the temperate zone, latitudinal gradient, and local or regional pockets of increased MS prevalence or incidence.

4.1. Evaluation of consistency with twin concordance rates

Estimates of MS concordance across monozygotic twin pairs range from 13% to 31%. Based on these figures, the corresponding limits on the prevalence of exposure to a hypothetical trigger virus were computed under a range of assumptions. If twins within a pair are exposed to virus independently of one another, an upper limit on viral prevalence that would be consistent with observed twin concordance would be 23–47%. If twins are more likely to be exposed together, the upper limit for prevalence would be decreased correspondingly. For between-twin correlations in the range of 50–90%, viral prevalence of 34–40% would be consistent with an MS concordance rate of 31%, and viral prevalence of 13–17% would be consistent with 13% MS concordance. If a median MS concordance rate of 22% and a between-twin correlation of 50% are assumed, the upper limit on viral prevalence would be 29%.

A literature search over the past 40 years yielded 36 reports on LCMV prevalence, containing 21 estimates in humans and 36 in rodents (Supplementary Table S1; Supplementary dataset S2). Prevalence ranged from 0 to 37.5% in humans and 0–67% in rodents. The frequency distribution of viral prevalence estimates was highly skewed. Median prevalence in humans was 3% and in rodents, 6.3%. Prevalence in humans was bimodal, with 85% of the values falling below 6% and the remaining 3 values above 30%. Comparison of observed LCMV prevalence rates with prevalence limits based on MS twin concordance shows that the observed rates are generally in agreement with pre-specified limits.

4.2. Evaluation of concentration in temperate zone

All 36 LCMV prevalence reports fell within the temperate zone. To evaluate whether this differed significantly from the contrasting equatorial zone, the proportion of countries providing LCMV prevalence reports was compared across zone (temperate or equatorial). To address the concern that less developed countries in the equatorial zone were less capable of assessing viral prevalence, each equatorial country was assessed to determine the existence of reports on a more commonly studied virus, EBV. Out of 66 countries in the equatorial zone, 33 had at least one report on EBV. Among the 65 countries within the temperate zone, 16 (25%) had provided LCMV prevalence reports. The difference in LCMV reporting rates for the 65 temperate zone countries versus the 33 equatorial countries that had produced reports on EBV prevalence was statistically significant (p = 0.0047).

4.3. Evaluation of latitudinal gradient

LCMV prevalence estimates were obtained from latitudes of 24–54° from the equator (north or south) (Supplementary dataset S2). The highest prevalence values in each of the 10-degree latitudinal tiers were: 11.5% from 24 to 34° latitude; 51% from 35 to 44° latitude; and 67% from 45 to 54° latitude. These values were all obtained from rodents. Within human subjects, the prevalence values by latitudinal tier were 3.5%, 36%, and 37.5%, respectively. In contrast, prevalence values of 0–5% were found at all latitudes. Linear regression applied to the 24 prevalence estimates that were above 5% showed an increasing trend of 0.84% prevalence per degree of latitude (p = 0.007; Fig. 3).

[locator:gr3] (link type not set)

Fig. 3

Relationship of LCMV exposure to latitude. Reported seroprevalence values in humans (black triangles) and rodents (gray circles) is plotted against distance from the equator. For prevalence values above 5%, the dashed line shows the dependence of prevalence on latitude as estimated by linear regression.

 

4.4. Correspondence of regional differences

To examine regional differences in MS incidence and compare them to prevalence patterns for LCMV, MS incidence and prevalence reports were sought from the literature. Three geographic regions with high LCMV prevalence values also had reports on MS incidence. These were Croatia, northern Italy, and southern Australia. Croatia was the country with the highest MS incidence worldwide in 2008, followed by Iceland, Hungary, and Slovakia. The highest values of LCMV prevalence in humans, over 35%, were seen in Slovakia and Croatia (Supplementary Table S1; Supplementary dataset S2). In northern Italy, five cities along the Po River basin report increasing MS prevalence. In the Trentino forested area just north of the mouth of the Po River, LCMV in rodents varied but levels up to 28.6% were found. In Australia, a latitudinal gradient for MS was reported, with highest prevalence in Hobart on the southern island of Tasmania. A moderately high LCMV prevalence rate of 9.6% was seen in rodents in the adjacent province of the mainland, New South Wales.

In order to control for the effect of HLA DR2 haplotype on MS in Croatia, the high-risk Gorski Kotar–Kočevje region was contrasted with Germany, the ancestral homeland of many of its residents. Estimates of MS incidence and prevalence, DR2 prevalence, and LCMV prevalence were compared across the two locations (Table 4). With only half the DR2 prevalence, this region of Croatia is estimated to have 10 times the MS incidence and 10 times the LCMV prevalence as Germany.

Table 4

Comparison of Gorski Kotar–Kočevje region with Germany in rates of DR2, LCMV, and MS.

 

Parameter Germany Gorski Kotar–Kočevje Ratio (Gorski Kotar–Kočevje to Germany)
Estimated DR2 prevalence 0.1724 0.0832 0.483
MS prevalence per 100,000 149 151.9 1.019
MS incidence per 100,000 2.85 29 10.175
LCMV prevalence (%) 3.6 36 10.000

Abbreviations: LCMV, lymphocytic choriomeningitis virus; MS, multiple sclerosis.

DR2 haplotype refers to DRA in combination with either DRB1*1501 or DRB5*0101. Estimates obtained as described in Section 2.11.

5. Discussion

While rapid advances have been made in identifying genetic factors associated with MS, the identity of an infective agent triggering the disease in genetically susceptible individuals remains uncertain. Many agents have been proposed, based primarily on their correlation with disease exacerbation or their isolation from CSF or brain tissue (Kakalacheva et al., 2011). Yet none of these have been linked to a specific causative mechanism, nor do they explain the prevalence patterns long observed for MS. Identification of an infectious agent that explains how the disease is initiated would be of great benefit. To that end, the present investigation is the first to employ a wide-ranging homology search and detailed motif analysis in combination with an evaluation of viral epidemiology to propose a viral trigger for MS.

Results of the homology search that best match the motif for DRB1*1501-restricted MBP-reactive T-cell activation predict a segment of the LCMV nucleoprotein as most likely to mimic MBP in binding to HLA-DR. In support of this prediction, LCMV prevalence reported in humans is consistent with MS twin concordance rates. Like MS, the virus is concentrated in the temperate zone geographically and shows evidence of a latitudinal gradient for prevalence values in excess of a background level of 5%.

The correlation of regional pockets of high MS incidence with exposure to LCMV is more difficult to assess. Studies linking the presence of LCMV directly to prevalence or incidence of MS have not been conducted. In addition, estimates of LCMV prevalence have not been acquired in a systematic manner, and studies varied in target population, sampling methods, and type of assay. Further, since infected rodents may live in close proximity to uninfected rodents, and virus is transmitted to humans by wind, it is not possible to correlate the precise location of rodent-borne infections with the range of locations within which humans may be exposed. Thus, a comparison of the literature reports describing the geographic distribution of LCMV with patterns observed for MS can be considered suggestive only.

Nevertheless, a few geographic regions where high LCMV prevalence occurs in proximity to high or increasing MS prevalence suggest locations where a systematic study may yield results. Most striking among these is a region of southeastern central Europe, including the countries of Croatia, Slovenia, Hungary, and Slovakia. This region reported the highest LCMV prevalence in humans (Dobec et al, 2006 and Reiserová et al, 1999) and also leads the world in MS incidence (World Health Organization, 2008). Rodents in waste dumps in this area also carried LCMV at an extremely high rate of 47% (Duh et al., 2014). After controlling for the frequency of the DRB1*1501 allele, an area on the Croatia–Slovenia border noted for its continuing high prevalence of MS (Peterlin et al, 2006 and Percović et al, 2010) was estimated to have 10 times the MS incidence and 10 times the LCMV prevalence as Germany (Table 4). Other regions with elevated prevalence of LCMV in rodents (Tagliapietra et al, 2009 and Smith et al, 1993) lie near geographic regions where MS prevalence is high or increasing, such as northern Italy (Granieri et al., 2007) and southeastern Australia (Hammond et al, 1988 and McLeod et al, 2011). The lack of person-to-person transmission of LCMV is consistent with low MS concordance across monozygotic twin pairs, while transmission by wind has the potential to explain the exceptionally high MS prevalence in some windy locations such as the Orkney-Shetland Islands, Sardinia, and New Zealand (Visser et al, 2012, Cocco et al, 2011, and Miller et al, 1990).

These findings challenge the conclusions drawn elsewhere that implicate EBV as a causative agent in MS. In contrast to EBV, LCMV provides a better match both to the protein binding and conformational requirements of the peptide-HLA complex and to the geographic distribution of MS. LCMV NP 400–414 fits the binding motif identified experimentally to activate MBP-specific T cell clones from MS patients far more closely than any region of EBNA1. The distribution of LCMV as a low-level presence with a preference for the temperate zone, a possible latitudinal gradient, and regional pockets of high infectivity that may be transmitted more efficiently by wind provides a more credible explanation for the observed distribution of MS than does EBV. EBV is ubiquitous and its distribution follows a reverse latitudinal gradient.

Experimental evidence in support of a link between EBV and MS includes an increased T cell response to EBNA1 and the observation that EBNA1-specific T cells recognize myelin antigens more frequently than other autoantigens (Lünemann et al., 2008). It was believed that such evidence provided support for a role of EBV under the molecular mimicry hypothesis. However, experimental results using a panel of microbial peptides revealed that a single MBP-specific T cell clone from an MS patient was able to recognize several peptides when bound to DRA-DRB1*1501, not all of which shared the motif predicted to cross-react with MBP (Hausmann et al, 1999 and Wucherpfennig and Sethi, 2011). Results of the present analysis are in general accord with this observation. Each of the 10 viruses in the test set was able to provide a peptide with high homology to MBP. Half of these viral peptides matched MBP exactly in residues of the P2 and P3 positions, which are the primary contacts for the T cell receptor, and the entire set of 10 peptides matched MBP on surface accessibility and antigenicity as well as the set of positive controls.

This seems to imply that a range of microbial agents may trigger the self-reactivity that is characteristic of autoimmunity and raises the possibility that clonal expansion of self-reactive T cells could involve several microbial peptides. Yet such a hypothesis fails to explain how a large group of potential pathogens, many of them ubiquitous, instigate a disease that affects only 0.3% of persons carrying the DRB1*1501 haplotype. It is apparent that cross-reactivity is not sufficient to explain causality and that another factor must be involved.

Theories regarding the mechanism by which autoimmunity is initiated fall into two broad classes. The first, based on disease initiation by one or more infectious triggers, includes molecular mimicry, epitope spreading, interaction between 2 or more viruses, and dual TCR theory (Cusick et al., 2013). A second class of mechanistic hypotheses proposes an aberration in immune regulation, such as disturbance of the balance between IFN and TNF (Cantaert et al., 2010). If LCMV is confirmed as a causative agent, it may suggest that both mechanisms are involved.

One hypothesis suggested by the present analysis is that at least a subset of autoimmune cases is initiated by virally-induced dysregulation of the immune system that allows both viral presence and the inflammatory process to persist. In individuals with a high-risk HLA haplotype, similarity between a viral peptide that binds to HLA and a self-peptide may result in clonal expansion of T cells reactive to both. If viral exposure occurs at a younger age, when thymic involution is less advanced, self-reactive T cell clones might escape negative selection due to cross-reactivity between self-peptides and exogenous viral antigens. Support for this hypothesis comes from experimental observations on MS and rheumatoid arthritis (RA).

Arenaviruses such as LCMV block induction of type 1 interferon (IFN) by preventing translocation of interferon regulatory factor 3 (IRF-3) to the nucleus where it can activate transcription (Pythoud et al, 2012, Martínez-Sobrido et al, 2009, and Ortiz-Riaño et al, 2011). Consequently, high levels of viremia may persist with little IFN response. If virally induced inhibition of IFN synthesis contributes to inflammation in MS, one would expect that supplementation by exogenous IFN would provide improvement. This is indeed what is observed. IFN-β is an approved MS treatment for relapsing-remitting MS (RRMS) that decreases the relapse rate and slows progression in a subset of patients (Jacobs et al., 2000). Among responder RRMS patients, IFN-β treatment induces a strong and significant expression of genes predominantly or selectively induced by type I IFNs (Comabella et al, 2009 and van Baarsen et al, 2008). Gene expression profiles indicate that expression of immune defense genes in a subset of RRMS patients is consistent with response to a viral, rather than bacterial, pathogen (van Baarsen et al., 2006). In contrast, IFN-β non-responders demonstrate a baseline gene expression profile reflecting a saturated IFN activation pathway (van Baarsen et al., 2008), suggesting that IFN production was not inhibited in this group. The subset of RRMS patients characterized as responders exhibit characteristics consistent with a deficiency in endogenous IFN-β, a state that hypothetically could be induced by infection with LCMV.

In an analysis parallel to this one, the virus predicted as trigger for RA blocks STAT1 phosphorylation, enabling persistent infection in macrophages and polarization toward a Th1-mediated inflammatory response (data not shown). If virally induced JAK-STAT inhibition contributes to inflammation in RA, one would expect that IFN levels would be normal or elevated but that IFN effector functions would be reduced. Experimental observations support this prediction. In comparison to patients with osteoarthritis or reactive arthritis, synovial biopsies from RA patients showed significantly higher levels of IFN-β (van Holten et al., 2005a). Unlike treatment of MS, the use of IFN-β therapy in RA has not been effective. Although IFN-β improved arthritic symptoms in animal models, a randomized placebo-controlled dual-dose trial of IFN-β for RA showed no evidence of efficacy for either dose (van Holten et al., 2005b). In fact, the high-dose IFN-β group had a higher rate of discontinuation than the other groups, and a significantly higher proportion of patients receiving IFN-β reported aggravation of RA symptoms compared to controls. Instead, effective treatments for RA include antagonists of tumor necrosis factor (TNF) or other cytokines in the pro-inflammatory cascade (Siebert et al., 2015). Since one of the effector functions of IFN-β is to reduce TNF levels (Jungo et al., 2001) and, in so doing, to maintain a homeostatic equilibrium between pro-inflammatory and anti-inflammatory processes (Cantaert et al., 2010), these experimental results suggest a dysregulation between pro-inflammatory and anti-inflammatory cytokines in RA.

IFN-β profoundly inhibits HLA class II expression, a process that requires IFN-stimulated gene factor 3 (ISGF3) (Lu et al., 1995). Since IFN-β inhibition by LCMV occurs upstream of ISGF3, it would be expected that upregulation of class II HLA molecules would result from LCMV infection, providing more binding sites for viral or self-antigens. Similarly, the virus predicted to trigger RA establishes persistent infection in macrophages, aided by inhibition of STAT1 phosphorylation, and induces Th1 polarization. Taken together, these two studies suggest that viral perturbation of immune regulation, including such actions as IFN inhibition or other features enabling a virus to establish persistent infection, may combine with molecular mimicry and increased numbers of HLA binding sites to generate a chronic inflammatory process such as is seen in autoimmune disorders. Whether viral perturbation of immune regulation in fact plays a role in the initiation of autoimmunity is a matter for further experimental investigation. Elucidation of the mechanistic details involved would increase scientific understanding of autoimmune processes and could potentially provide new targets for vaccines or therapeutic interventions.

Acknowledgments

The author thanks Elizabeth A. Holly of the University of California – San Francisco, Brad Efron and Susan Holmes of Stanford University, Betz Halloran of the Fred Hutchinson Cancer Research Center, and Norm Breslow of the University of Washington for helpful comments and suggestions. Acknowledgments also go to the many biologists, physicians, epidemiologists, virologists, and other scientists whose meticulous work provided the foundation for this analysis, and to the patients, whose cause inspired this investigation.

Appendix A. Supplementary data

The following are Supplementary data to this article:

[locator:mmc1] (link type not set)

References

  • Biggar et al., 1975 R.J. Biggar, J.P. Woodall, P.D. Walter, G.E. Haughie. Lymphocytic choriomeningitis outbreak associated with pet hamsters. Fifty-seven cases from New York State. JAMA. 1975;232(5):494-500
  • Bonthius et al., 2007 D.J. Bonthius, R. Wright, B. Tseng, L. Barton, E. Marco, B. Karacay, et al. Congenital lymphocytic choriomeningitis virus infection: spectrum of disease. Ann. Neurol.. 2007;62:347-355
  • Cantaert et al., 2010 T. Cantaert, D. Baeten, P.P. Tak, L.G.M. van Baarsen. Type I IFN and TNFα cross-regulation in immune-mediated inflammatory disease: basic concepts and clinical relevance. Arthritis Res. Ther. 12. 2010;219
  • Castellazi et al., 2014 M. Castellazi, C. Contini, C. Tamborino, F. Fasolo, G. Roversi, S. Seraceni, et al. Epstein–Barr virus-specific intrathecal oligoclonal IgG production in relapsing-remitting multiple sclerosis is limited to a subset of patients and is composed of low-affinity antibodies. J. Neuroinflammation. 2014;11:188 188
  • Charrel and de Lamballerie, 2010 R.N. Charrel, X. de Lamballerie. Zoonotic aspects of arenavirus infections. Vet. Microbiol.. 2010;140(3–4):213-220
  • Childs et al., 1992 J.E. Childs, G.E. Glass, G.W. Korch, T.G. Ksiazek, J.W. Leduc. Lymphocytic choriomeningitis virus infection and house mouse (Mus musculus) distribution in urban Baltimore. Am. J. Trop. Med. Hyg.. 1992;47(1):27-34
  • Cocco et al., 2011 E. Cocco, C. Sardu, R. Massa, E. Mamusa, L. Musu, P. Ferrigno, et al. Epidemiology of multiple sclerosis in south-western Sardinia. Mult. Scler.. 2011;17(11):1282-1289
  • Comabella et al., 2009 M. Comabella, J.D. Lünemann, J. Río, A. Sánchez, C. López, E. Julià, et al. A type I interferon signature in monocytes is associated with poor response to interferon-β in multiple sclerosis. Brain. 2009;132:3353-3365
  • Cusick et al., 2013 M.F. Cusick, J.E. Libbey, R.S. Fujinami. Multiple sclerosis: autoimmunity and viruses. Curr. Opin. Rheumatol.. 2013;25(4):496-501
  • Dobec et al., 2006 M. Dobec, B. Dzelalija, V. Punda-Polic, L. Zoric. High prevalence of antibodies to lymphocytic choriomeningitis virus in a murine typhus endemic region in Croatia. J. Med. Virol.. 2006;78(12):1643-1647
  • Duh et al., 2014 D. Duh, E. Varljen-Buzan, S. Hasic, R. Charrel. Increased seroprevalence of lymphocytic choriomeningitis virus infection in mice sampled in illegal waste sites. Parasites Vectors. 2014;7(Suppl 1) O31
  • Eisenberg et al., 1984 D. Eisenberg, R.M. Weiss, T.C. Terwilliger. The hydrophobic moment detects periodicity in protein hydrophobicity. Proc. Natl. Acad. Sci.. 1984;81:140-144
  • Fujinami, 2001 R.S. Fujinami. Can virus infections trigger autoimmune disease?. J. Autoimmun.. 2001;16(3):229-234
  • Gourraud et al., 2012 P.A. Gourraud, H.F. Harbo, S.L. Hauser, S.E. Baranzini. The genetics of multiple sclerosis: an up-to-date review. Immunol. Rev.. 2012;248:87-103
  • Granieri et al., 2007 E. Granieri, N.T. Economou, R. De Gennaro, M.R. Tola, L. Caniatti, V. Govoni, et al. Multiple sclerosis in the province of Ferrara: evidence for an increasing trend. J. Neurol.. 2007;254(12):1642-1648
  • Hammond et al., 1988 S.R. Hammond, J.G. McLeod, K.S. Millingen, E.G. Stewart-Wynne, D. English, J.T. Holland, et al. The epidemiology of multiple sclerosis in three Australian cities: Perth, Newcastle and Hobart. Brain. 1988;111(1):1-25
  • Hausmann et al., 1999 S. Hausmann, M. Martin, L. Gauthier, K.W. Wucherpfennig. Structural features of autoreactive TCR that determine the degree of degeneracy in peptide recognition. J. Immunol.. 1999;162:334-338
  • Henikoff and Henikoff, 1992 S. Henikoff, J.G. Henikoff. Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sci.. 1992;89:10915-10919
  • Hjalgrim et al., 2007 H. Hjalgrim, J. Friborg, M. Melbye, et al. The epidemiology of EBV and its association with malignant disease. A. Arvin, G. Campadelli-Fiume, E. Mocarski (Eds.) Human Herpesviruses: Biology, Therapy, and Immunoprophylaxis (Cambridge University Press, Cambridge, 2007) Available from: http://www.ncbi.nlm.nih.gov/books/NBK47424/ (accessed 12.14). chapter 53
  • Hornik, 2014 Hornik K., 2014. The R FAQ. http://cran.r-project.org/doc/FAQ/R-FAQ.html#Citing-this-document (accessed 10.11.14.).
  • Islam et al., 2006 T. Islam, W.J. Gauderman, W. Cozen, A.S. Hamilton, M.E. Burnett, T.M. Mack. Differential twin concordance for multiple sclerosis by latitude of birthplace. Ann. Neurol.. 2006;60:56-64
  • Jacobs et al., 2000 L.D. Jacobs, R.W. Beck, J.H. Simon, R.P. Kinkel, C.M. Brownscheidle, T.J. Murray, et al. Intramuscular interferon beta-1a therapy initiated during a first demyelinating event in multiple sclerosis. CHAMPS Study Group. N. Engl. J. Med.. 2000;343(13):898-904
  • Janin et al., 1978 J. Janin, S. Wodak, M. Levitt, B. Maigret. Conformation of amino acid side-chains in proteins. J. Mol. Biol.. 1978;125:357-386
  • Jungo et al., 2001 F. Jungo, J.M. Dayer, C. Modoux, N. Hyka, D. Burger. IFN-β inhibits the ability of T lymphocytes to induce TNF-α and IL-1β production in monocytes upon direct cell-cell contact. Cytokine. 2001;14(5):272-282
  • Kakalacheva et al., 2011 K. Kakalacheva, C. Münz, J.D. Lünemann. Viral triggers of multiple sclerosis. Biochim. Biophys. Acta. 2011;1812:132-140
  • Karplus and Schulz, 1985 P.A. Karplus, G.E. Schulz. Prediction of chain flexibility in proteins—a tool for the selection of peptide antigens. Naturwissenschafren. 1985;72:212-213
  • Kuusisto et al., 2008 H. Kuusisto, J. Kaprio, E. Kinnunen, T. Luukkaala, M. Koskenvuo, I. Elovaara. Concordance and heritability of multiple sclerosis in Finland: study on a nationwide series of twins. Eur. J. Neurol.. 2008;15:1106-1110
  • Lehmann-Grube, 1971 F. Lehmann-Grube. Lymphocytic choriomeningitis virus. (Springer-Verlag, New York, 1971)
  • Li et al., 2000 Y. Li, H. Li, R. Martin, R.A. Mariuzza. Structural basis for the binding of an immunodominant peptide from myelin basic protein in different registers by two HLA-DR2 proteins. J. Mol. Biol.. 2000;304:177-188
  • Link et al., 2012 J. Link, I. Kockum, A.R. Lorentzen, B.A. Lie, E.G. Celius, H. Westerlind, et al. Importance of human leukocyte antigen (HLA) class I and II alleles on the risk of multiple sclerosis. PLoS One. 2012;7(5):e36779
  • Lu et al., 1995 H.T. Lu, J.L. Riley, G.T. Babcock, M. Huston, G.R. Stark, J.M. Boss, et al. Interferon (IFN) β acts downstream of IFN-γ-induced class II transactivator messenger RNA accumulation to block major histocompatibility complex class II gene expression and requires the 48-kD DNA-binding protein, ISGF3-γ. J. Exp. Med.. 1995;182:1517-1525
  • Lünemann et al., 2008 J.D. Lünemann, I. Jelčić, S. Roberts, A. Lutterotti, B. Tackenberg, R. Martin, et al. EBNA1-specific T cells from patients with multiple sclerosis cross react with myelin antigens and co-produce IFN-γ and IL-2. J. Exp. Med.. 2008;205(8):1763-1773
  • Martínez-Sobrido et al., 2009 L. Martínez-Sobrido, S. Emonet, P. Giannakas, B. Cubitt, A. García-Sastre, J.C. de la Torre. Identification of amino acid residues critical for the anti-interferon activity of the nucleoprotein of the prototypic arenavirus lymphocytic choriomeningitis virus. J. Virol.. 2009;83(21):11330-11340
  • Materljan and Sepcic, 2002 E. Materljan, J. Sepcic. Epidemiology of multiple sclerosis in Croatia. Clin. Neurol. Neurosurg.. 2002;104:192-198
  • McLeod et al., 2011 J.G. McLeod, S.R. Hammond, J.F. Kurtzke. Migration and multiple sclerosis in immigrants to Australia from United Kingdom and Ireland: a reassessment. I. Risk of MS by age at immigration. J. Neurol.. 2011;258:1140-1149
  • Miller et al., 1990 D.H. Miller, S.R. Hammond, J.G. McLeod, G. Purdie, D.C. Skegg. Multiple sclerosis in Australia and New Zealand: are the determinants genetic or environmental?. J. Neurol. Neurosurg. Psychiatry. 1990;53(10):903-905
  • Mumford et al., 1994 C.J. Mumford, N.W. Wood, H. Kellar-Wood, J.W. Thorpe, D.H. Miller, D.A. Compston. The British Isles survey of multiple sclerosis in twins. Neurology. 1994;44(1):11-15
  • Nolan et al., 2012 D. Nolan, A. Castley, M. Tschochner, I. James, W. Qiu, D. Sayer, et al. Contributions of vitamin D response elements and HLA promoters to multiple sclerosis risk. Neurology. 2012;79(6):538-546
  • Olson et al., 2001 J.K. Olson, J.L. Croxford, M.A. Calenoff, C. Dal, M.C. anto, S.D. Miller. A virus-induced molecular mimicry model of multiple sclerosis. J. Clin. Invest.. 2001;108(2):311-318
  • Ortiz-Riaño et al., 2011 E. Ortiz-Riaño, B.Y.H. Cheng, J.C. de la Torre, L. Martínez-Sobrido. The C-terminal region of lymphocytic choriomeningitis virus nucleoprotein contains distinct and segregable functional domains involved in NP-Z interaction and counteraction of the type I interferon response. J. Virol.. 2011;85(24):13038-13048
  • Parker et al., 1986 J.M.R. Parker, D. Guo, R.S. Hodges. New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites. Biochem. 1986;25:5425-5432
  • Percović et al., 2010 O. Percović, A. Jurjević, I. Antoncić, S. Dunatov, M. Bralić, S. Ristić. The town of Cabar, Croatia, familiar pseudocluster for multiple sclerosis—descriptive epidemiological study. Coll. Antropol.. 2010;34(Suppl 2):141-144
  • Peterlin et al., 2006 B. Peterlin, S. Ristić, J. Sepčić, B.K. Vračko, A. Rako, L. Lovrečić, et al. Region with persistent high frequency of multiple sclerosis in Croatia and Slovenia. J. Neurol. Sci.. 2006;247:169-172
  • Pythoud et al., 2012 C. Pythoud, W.W.S.I. Rodrigo, G. Pasqual, S. Rothenberger, L. Martínez-Sobrido, J.C. de la Torre, S. Kunz. Arenavirus nucleoprotein targets interferon regulatory factor-activating kinase IKKϵ. J. Virol.. 2012;86(15):7728-7738
  • Qiu et al., 2011 W. Qiu, I. James, W.M. Carroll, F.L. Mastaglia, A.G. Kermode. HLA-DR allele polymorphism and multiple sclerosis in Chinese populations: a meta-analysis. Mult. Scler.. 2011;17(4):382-388
  • Reiserová et al., 1999 L. Reiserová, M. Kaluzová, Š. Kaluz, A.C. Willis, J. Závada, E. Závodská, et al. Identification of MaTu-MX agent as a new strain of lymphocytic choriomeningitis virus (LCMV) and serological indication of horizontal spread of LCMV in human population. Virol. 1999;257:73-83
  • Ristori et al., 2006 G. Ristori, S. Cannoni, M.A. Stazi, N. Vanacore, R. Cotichini, M. Alfò, et al. Multiple sclerosis in twins from continental Italy and Sardinia: a nationwide study. Ann. Neurol.. 2006;59(1):27-34
  • Sadovnick et al., 1993 A.D. Sadovnick, H. Armstrong, G.P. Rice, D. Bulman, L. Hashimoto, D.W. Paty, et al. A population-based study of multiple sclerosis in twins: update. Ann. Neurol.. 1993;33:281-285
  • Siebert et al., 2015 S. Siebert, A. Tsoukas, J. Robertson, I. McInnes. Cytokines as therapeutic targets in rheumatoid arthritis and other inflammatory diseases. Pharmacol. Rev.. 2015;67:280-309
  • Simpson et al., 2011 S. Simpson, L. Blizzard, P. Otahal, I. Van der Mei, B. Taylor. Latitude is significantly associated with the prevalence of multiple sclerosis: a meta-analysis. J. Neurol. Neurosurg. Psychiatry. 2011;82:1132-1141
  • Solberg et al., 2008 O.D. Solberg, S.J. Mack, A.K. Lancaster, R.M. Single, Y. Tsai, A. Sanchez-Mazas, et al. Balancing selection and heterogeneity across the classical human leukocyte antigen loci: a meta-analytic review of 497 population studies. Hum. Immunol.. 2008;69(7):443-464
  • Smith et al., 1993 A.L. Smith, G.R. Singleton, G.M. Hansen, G. Shellam. A serological survey for viruses and Mycoplasma pulmonis among wild house mice (Mus domesticus) in southeastern Australia. J. Wildl. Dis.. 1993;29(2):219-229
  • Smith et al., 1998 K.J. Smith, J. Pyrdol, L. Gauthier, D.C. Wiley, K.W. Wucherpfennig. Crystal structure of HLA-DR2 (DRA1*1010. DRB1*1501) complexed with a peptide from human myelin basic protein. J. Exp. Med.. 1998;188(8):1511-1520
  • Stern et al., 1994 L.J. Stern, J.H. Brown, T.S. Jardetzky, J.C. Gorga, R.G. Urban, J.L. Strominger, D.C. Wiley. Crystal structure of the human class II MHC protein HLA-DR1 complexed with an influenza virus peptide. Nature. 1994;368:215-221
  • Tagliapietra et al., 2009 V. Tagliapietra, R. Rosà, H. Hauffe, J. Laakkonen, L. Voutilainen, O. Vapalahti, et al. Spatial and temporal dynamics of lymphocytic choriomeningitis virus in wild rodents, northern Italy. Emerg. Infect. Dis.. 2009;15(7):1019-1025
  • Teutsch et al., 1999 S.M. Teutsch, B.H. Bennetts, M.M. Buhler, R.N. Heard, G.J. Stewart. The DRB1 Val86/Val86 genotype associates with multiple sclerosis in Australian patients. Hum. Immunol.. 1999;60(8):715-722
  • Traub, 1938 E. Traub. Factors influencing the persistence of choriomeningitis virus in the blood of mice after clinical recovery. J. Exp. Med.. 1938;68(2):229-250
  • van Baarsen et al., 2006 L.G.M. van Baarsen, C.T.C.M. van der Pouw Kraan, J.J. Kragt, J.M.C. Baggen, F. Rustenburg, T. Hooper, et al. A subtype of multiple sclerosis defined by an activated immune defense program. Genes Immun.. 2006;7:522-531
  • van Baarsen et al., 2008 L.G.M. van Baarsen, S. Vosslamber, M. Tijssen, J.M.C. Baggen, L.F. van der Voort, J. Killestein, et al. Pharmacogenomics of interferon-β therapy in multiple sclerosis: baseline IFN signature determines pharmacological differences between patients. PLoS One. 2008;3(4):e1927
  • van Holten et al., 2005a J. van Holten, T.J.M. Smeets, P. Blankert, P.P. Tak. Expression of interferon β in synovial tissue from patients with rheumatoid arthritis: comparison with patients with osteoarthritis and reactive arthritis. Ann. Rheum. Dis.. 2005;64(12):1780-1782
  • van Holten et al., 2005b J. van Holten, K. Pavelka, J. Vencovsky, H. Stahl, B. Rozman, M. Genovese, et al. A multicentre, randomised, double blind, placebo controlled phase II study of subcutaneous interferon beta-1a in the treatment of patients with active rheumatoid arthritis. Ann. Rheum. Dis.. 2005;64(1):64-69
  • Visser et al., 2012 E.M. Visser, K. Wilde, J.F. Wilson, K.K. Yong, C.E. Counsell. A new prevalence study of multiple sclerosis in Orkney, Shetland and Aberdeen city. J. Neurol. Neurosurg. Psychiatry. 2012;83(7):719-724
  • Welling et al., 1985 G.W. Welling, W.J. Weijer, R. van der Zee, S. Welling-Wester. Prediction of sequential antigenic regions in proteins. FEBS Lett.. 1985;188(2):215-218
  • Willer et al., 2003 C.J. Willer, D.A. Dyment, N.J. Risch, A.D. Sadovnick, G.C. Ebers. Twin concordance and sibling recurrence rates in multiple sclerosis. PNAS. 2003;100(22):12877-12882
  • World Health Organization, 2008 World Health Organization. Atlas: Multiple Sclerosis Resources in the World 2008. (WHO Press, Geneva, 2008)
  • Wu et al., 2010 J.S. Wu, I. James, W. Qiu, A. Castley, F.T. Christiansen, W.M. Carroll, et al. Influence of HLA-DRB1 allele heterogeneity on disease risk and clinical course in a West Australian MS cohort: a high-resolution genotyping study. Mult. Scler.. 2010;16(5):526-532
  • Wucherpfennig and Sethi, 2011 K.W. Wucherpfennig, D. Sethi. T cell receptor recognition of self and foreign antigens in the induction of autoimmunity. Semin. Immunol.. 2011;23(2):84-91
  • Wucherpfennig et al., 1994 K.W. Wucherpfennig, A. Sette, S. Southwood, C. Oseroff, M. Matsui, J.L. Strominger, D.A. Hafler. Structural requirements for binding of an immunodominant myelin basic protein to DR2 isotypes and for its recognition by human T cell clones. J. Exp. Med.. 1994;179:279-290

Footnotes

Independent Researcher, Albany, CA 94706, USA


Search this site

Stay up-to-date with our monthly e-alert

If you want to regularly receive information on what is happening in MS research sign up to our e-alert.

Subscribe »

About the Editors

  • Prof Timothy Vartanian

    dsc_0787_400x400.jpg Timothy Vartanian, Professor at the Brain and Mind Research Institute and the Department of Neurology, Weill Cornell Medical College,...
  • Dr Claire S. Riley

    headshotcsr1_185x250.jpg Claire S. Riley, MD is an assistant attending neurologist and assistant professor of neurology in the Neurological Institute, Columbia...
  • Dr Rebecca Farber

    picforelsevier.jpg Rebecca Farber, MD is an attending neurologist and assistant professor of neurology at the Neurological Institute, Columbia University, in...

This online Resource Centre has been made possible by a donation from EMD Serono, Inc., a business of Merck KGaA, Darmstadt, Germany.

Note that EMD Serono, Inc., has no editorial control or influence over the content of this Resource Centre. The Resource Centre and all content therein are subject to an independent editorial review.

The Grant for Multiple Sclerosis Innovation
supports promising translational research projects by academic researchers to improve understanding of multiple sclerosis (MS) for the ultimate benefit of patients.  For full information and application details, please click here

Journal Editor's choice

Recommended by Prof. Brenda Banwell

Causes of death among persons with multiple sclerosis

Gary R. Cutter, Jeffrey Zimmerman, Amber R. Salter, et al.

Multiple Sclerosis and Related Disorders, September 2015, Vol 4 Issue 5