Genetic factors influence susceptibility to multiple sclerosis but the responsible genes remain largely undefined, association with MHC class II alleles being the only established genetic feature of the disease. The Nordic countries have a high prevalence of multiple sclerosis, and to further explore the genetic background of the disease, we have carried out a genome-wide screen for linkage in 136 sibling-pairs with multiple sclerosis from Denmark, Finland, Norway and Sweden by typing 399 microsatellite markers. Seventeen regions where the lod score exceeds the nominal 5% significance threshold (0.7) were identified—1q11–24, 2q24–32, 3p26.3, 3q21.1, 4q12, 6p25.3, 6p21–22, 6q21, 9q34.3, 10p15, 10p12–13, 11p15.5, 12q21.3, 16p13.3, 17q25.3, 22q12–13 and Xp22.3. Although none of these regions reaches the level of genome-wide significance, the number observed exceeds the 10 that would be expected by chance alone. Our results significantly add to the growing body of linkage data relating to multiple sclerosis.
Although the exact cause of multiple sclerosis is unknown, epidemiological studies indicate a multifactorial aetiology1 with the involvement of genetic factors suggested by studies in twins,2,3 half-siblings,4 adoptees5 and conjugal pairs.6,7 The role of major histocompatibility complex (MHC) class II alleles is well established,8 but no other candidate genes have consistently been linked or associated with the disease.
Since the early 1990s, developments in the human genome project9,10 have made it possible to employ more systematic approaches to the search for susceptibility genes in complex diseases.11 In multiple sclerosis, six whole genome screens for linkage have now been performed.12,13,14,15,16,17 No one has shown unequivocal linkage, but most have identified more regions of potential linkage than expected by chance and together they provide the best available summary of the likely location of the relevant genes.18
The high prevalence of multiple sclerosis (120/100 000) in Scandinavia19 and the striking correlation between its global geographical distribution and the migration pattern of northern Europeans20 suggest that important susceptibility alleles may have arisen in Scandinavia and been disseminated by their descendants—the so-called ‘Viking hypothesis’.21 The Nordic countries (Denmark, Finland, Norway and Sweden) are thus a promising place in which to search for multiple sclerosis susceptibility genes. In order to facilitate analysis of the disease in this founding population, a collaborative group was formed (the Nordic MS Genetics Study Group) involving seven Nordic research centres. Since its inception in 1994, this group has concentrated on identifying and carefully assessing familial multiple sclerosis in Scandinavia in order to establish a resource for genetic analysis.
Based on results from previous linkage screens, some parts of the genome have already been investigated in these families.22,23 However, the majority of the genome has hitherto remained un-explored. We therefore performed a genome-wide screen for linkage in a large sample of these potentially highly informative Nordic families.
In order to maximise efficiency, we performed our genome screen in two stages. In the first, only the affected siblings (n = 272) from each family were considered, and no parents or unaffected siblings were genotyped. The results of multipoint non-parametric linkage analysis on each chromosome are shown in Figure 1. Although there are no regions where linkage attains genome-wide significance, 17 regions exceed the nominal 5% significance level (lod >0.7) whereas only 10 would have been expected by chance for a marker map of this density. These regions include 1q11–24, 2q24–32, 3p26.3, 3q21.1, 4q12, 6p25.3, 6p21–22, 6q21, 9q34.3, 10p15, 10p12–13, 11p15.5, 12q21.3, 16p13.3, 17q25.3, 22q12–13 and Xp22.3. The marker map provided an average information extraction of 57%, ranging from 15% on the ptelomeric region of chromosome 19 to 86% on chromosome 22 at a particularly informative marker.
In the second stage of our screen, the available parents and unaffected siblings were genotyped for 31 markers from the most promising regions identified in stage 1 in order to increase the information extraction in these regions. In fact, the average increase in information extraction was modest (3%) confirming that little would have been gained by genotyping the parents and unaffected siblings for all 399 markers in stage 1. As expected, the effect on lod scores was also limited with only minor changes for individual regions (Figure 1). In no instance did increasing the information extraction lead to any of the 17 peak lod scores falling below 0.7.
In order to search for evidence of loci interacting with human leukocyte antigen (HLA), each sibling was genotyped for the presence (+) or absence (−) of the DR15 allele. The 136 sib pairs were then considered in three groups on the basis of whether pairs were both positive for the DR15 allele (+/+: n = 82), only one was positive for the allele (+/−: n = 20) or neither was positive (−/−: n = 34). Multipoint lod scores were then calculated for each of the three subgroups independently and summed to generate a HET.LOD score. Under the null hypothesis of no interaction with HLA it would be expected that each subgroup of pairs would maximise to the same sharing probabilities and there would be no difference between the HET.LOD and the lod score obtained for all the families analysed together. The statistical significance of observed differences was judged empirically using simulations (see methods). Table 1 shows the peak HET.LOD observed on each chromosome together with the statistical significance of the difference between this and the lod score observed for all families analysed together. Only chromosome 19 revealed an apparently significant difference; however the calculation of HET.LOD was not constrained to the possible triangle and, the result on chromosome 19 was principally due to the −/− subset maximising outside the possible triangle, indicating that the observed difference is the result of statistical fluctuation and not an interaction. In short, we found no evidence for statistically significant interaction with HLA at any loci.
We have performed a genome-wide screen for linkage in 136 Nordic sib-pairs with multiple sclerosis using 399 dinucleotide microsatellite markers. Although there were no regions of genome-wide significance, 17 regions were identified where the lod score exceeded 0.7 (the 5% nominal significance level). In a genome-wide screen of this size, only 10 such regions would be expected by chance alone. The fact that our results exceed expectation suggests that at least some of these are likely to be genuine.
In genome-wide screening, two sets of significance levels are considered18—point-wise (nominal) and genome wide. Under the null hypothesis of no linked genes the point-wise significance refers to the probability that the maximum lod score (MLS) profile will equal or exceed the given value at a particular (single) point in the genome, while the genome-wide significance refers to the probability that the MLS profile will equal or exceed this value at least once somewhere in the genome. An MLS value of 0.7 has a point-wise (nominal) significance of 5%; that is, the MLS profile at any particular point in the genome will only equal or exceed a value of 0.7 on average once in every 20 genome screens. On the other hand, a genome-wide significance of 5% corresponds to an MLS value of approximately 3.2; this indicates that the MLS profile would be expected to equal or exceed 3.2 at any point in the genome only once in every 20 genome screens. Genome-wide significance levels allow for the multiple testing inherent in screening the whole genome and are thus the best guide to the significance of any single ‘hit’. Although the association of HLA class II alleles with multiple sclerosis is readily demonstrated,8,24 the effect of this locus on susceptibility is modest and therefore evidence for linkage in this region is expected to be difficult to confirm. Previous linkage screens in out-bred populations of northern European origin (American, British and Canadian)12,13,14 showed only modest evidence for linkage to the HLA region. Our study is in keeping with these results in that it also shows only modest evidence for linkage in the HLA-region on chromosome 6p21.
Oturai et al22 have previously investigated a sample of these same sibling-pairs in a region of interest on chromosome 5 and in the HLA region. Differences between our previous results and those presented here principally reflect the fact that the present study includes more families.
The fact that typing parents and unaffected siblings had only a limited effect upon our results is partly due to the small number of families with available parents and unaffected siblings, but also due to the fact that the markers used are highly informative. Typing unaffected siblings or one parent has previously been found to give only a small increase in power except when the PIC-value for a given marker is low.25
While some of our linkage peaks indicate the likely regions containing loci conferring susceptibility to multiple sclerosis specifically, others may indicate regions containing loci conferring susceptibility to autoimmunity in general.26 With this hypothesis in mind there is interesting concordance between our results and those from studies in other autoimmune diseases.
On chromosome 10 we found two regions of potential linkage of which 10p12–13 is located very close to IDDM-10.27 This region has also shown suggestive and potential linkage respectively in recently published studies of sib-pairs with multiple sclerosis from Sardinia16 and Italy.17 Another region of interest in this respect is 6q21 on the long arm of chromosome 6, where our results show a peak lod score of 1.2. This is the same region as another susceptibility locus in insulin dependent diabetes mellitus (IDDM15).28 The peak on the long arm of chromosome 12 overlaps both with a peak in the American linkage screen in multiple sclerosis13 and one in a genome-wide screen for linkage in rheumatoid arthritis.29 The region on the short arm of chromosome 16 coincides with one of the best supported non-HLA regions from the recently published meta-analysis of the original American, British and Canadian linkage screens in multiple sclerosis.30
Our study thus shows both overlap and differences with previous screens in multiple sclerosis and other autoimmune complex traits. The differences with respect to other linkage screens in multiple sclerosis may reflect type I error (false positive findings), or result from differences in the ethnic genetic background of the populations studied.
The overlap with other autoimmune complex diseases supports the growing body of evidence that linkage screens will undoubtedly lead to the identification of chromosomal regions containing susceptibility genes: some will be disease-specific; others will prove to be more important in promoting the process of autoimmunity rather than a specific disease phenotype reflecting selective organ specificity. More specifically, our study contributes to the steady accumulation of linkage data relating to multiple sclerosis. Our observation that significant linkage cannot be demonstrated even in a relatively large number of affected sibling pairs from a putative founder population has implications for the size of the effect attributable to any one susceptibility gene. The result indicates the need to add new data sets from related and ethnically diverse populations incrementally so that meta-analysis can in time reliably locate the chromosomal regions encoding susceptibility genes and steer the application of complementary strategies aimed at identification of functional polymorphisms.
Materials and methods
A total of 136 families with affected sibling pairs were included in the study (Table 2). All patients met the criteria for clinically definite multiple sclerosis according to the Poser criteria31 and gave informed or written consent to take part in genetic analysis. Where available, DNA was also obtained from consenting parents, and in those families where parents were unavailable, from consenting unaffected siblings.
Families with more than two affected siblings were not included since such families are more likely to include parents who are homozygous for susceptibility alleles and therefore uninformative for linkage. In addition we chose our families exclusively from the general out-bred Nordic population and specifically avoided families from sub-isolate populations such as the Sami (where the disease is uncommon).32
A total of 399 dinucleotide microsatellite markers from the Applied Biosystems Medium Density Linkage Mapping Set (LMS-MD10) were used to screen the genome. The microsatellite markers were highly informative showing a mean heterozygosity of 78% and mean polymorphism information content (PIC) of 76%. The average separation between these markers was 9.7 cM.
DNA was extracted from venous blood using standard techniques. Each marker was amplified by polymerase chain reaction (PCR) using one of the following methods:
‘Touch Down’ PCR on either a GRI Tetrad or a Hybaid Omnigene thermal-cycler, in 10 μl reaction volumes as described in Sawcer et al.12
Using an 877 Integrated Thermal Cycler (Applied Biosystems) in 5 μl reaction volumes with up to two markers in each reaction.
‘TrueAllele’ (Applied Biosystems) PCR in 15 μl reaction volumes on a GRI Tetrad machine using ABI recommended conditions.
PCR products from each panel of markers were pooled prior to electrophoresis on either a 373A Sequencing Machine (Applied Biosystems) using 6% acrylamide gels or a 3700 Genetic Analyser (Applied Biosystems) which performs capillary electrophoresis. All genotyping was performed in the laboratory in Cambridge. In order to ensure accurate allele calling across the various gels, alleles were normalised against the CEPH-individual 1347–02. Genotyping was performed using the semi-automated GENESCAN and GENOTYPER software (Perkin Elmer).
Each of the affected individuals was tested for the presence or absence of the DR 15 alleles by amplifying genomic DNA using PCR and sequence-specific primers (5′-CCG CGC CTG CTC CAG GAT-3′ and 5′-TCC TGT GGC AGC CTA AGA G-3′). A positive control for the PCR was provided by primers amplifying the human growth hormone locus. The PCR assay was performed in a final volume of 13 μl, containing 30 ng genomic DNA, 20 mM ammonium sulphate, 75 mM Tris HCL (pH 9.0), 0.01% Tween, 2 mM Magnesium Chloride, 200 μM each of dATP, dCTP, dGTP, dTTP and 0.125 units Taq. Cycle conditions were 94°C for 2 min, followed by 10 cycles at 94°C for 20 s, and at 65°C for 60 s, and then 20 cycles at 94°C for 20 s, 61°C for 50 s and 72°C for 30 s. Amplified products were visualised under ultraviolet light after running in a 2% agarose gel containing TBE buffer and 0.5 mg/μl ethidium bromide for 30 min at 100 V. Gel interpretation assigned the presence or absence of DR 15 alleles according to the 1998 nomenclature report.33 Each individual was typed twice and typing was repeated if discordant results were obtained.
Non-parametric affected pair methods are commonly used to study complex diseases such as multiple sclerosis where no clear model of inheritance can be identified. These methods are based upon the principle that affected individuals will be more similar in those parts of the genome linked to a disease susceptibility gene than expected by chance.34 We performed multipoint non-parametric (model-free) linkage analysis using the MAPMAKER/SIBS program (version 2.0).35 This calculates a MLS value at each point in the genome on the basis of all the available genotypes, and thereby generates MLS profiles along each chromosome. Allele frequencies used in the analysis were estimated from data using the SPLINK program (version 1.07).25 Any positive MLS value indicates potential linkage but the peaks in these profiles, which may be referred to as ‘hits’, identify the most likely locations for disease susceptibility genes. The MAPMAKER/SIBS program was also used to calculate the information extraction along each chromosome. This information indicates how much of the available identity by descent (IBD) information has been extracted by the markers typed. On the X-chromosome MAPMAKER/SIBS stratifies sib-pairs on basis of gender and performs linkage analysis independently in brother-brother, sister-brother and sister-sister pairs. It then generates a composite lod score based on these three results. As no parents were typed in the first stage of the screen, the SIBERROR program (version 1.0)36 was used to test for the presence of pedigree errors. One sib-pair was identified as unrelated and one pair as identical twins. These two pairs were therefore excluded, leaving a total of 136 affected sib-pairs used in the analysis.
Interactions between HLA and postulated disease-susceptibility loci were tested as follows. Affected individuals were denoted as “+” if they had at least one DR15 allele. Affected sibling pairs therefore could be classified into one of three groups: +/+ (n = 82), +/− (n = 20) or −/− (n = 34). Following the suggestion of Rice,37 the probability P that an affected sib pair inherited a given parental allele IBD was modelled as a logistic regression, with the HLA group included as a categorical variable. The significance of the resulting lod score was assessed by random permutations of HLA groups among the affected pairs and repeating the analysis.
Hogancamp WE, Rodriguez M, Weinshenker BG . The epidemiology of multiple sclerosis Mayo Clin Proc 1997 72: 871–878
Sadovnick AD, Armstrong H, Rice GP et al. A population based study of multiple sclerosis in twins: update Ann Neurol 1993 33: 281–285
Mumford CJ, Wood NW, Kellar-Wood H, Thorpe JW, Miller DH, Compston DA . The British Isles survey of multiple sclerosis in twins Neurology 1994 44: 11–15
Sadovnick AD, Ebers GC, Dyment DA, Risch NJ . Evidence for genetic basis of multiple sclerosis. The Canadian Collaborative Study Group Lancet 1996 347: 1728–1731
Ebers GC, Sadovnick AD, Risch NJ . A genetic basis for familial aggregation in multiple sclerosis. Canadian Collaborative Study Group Nature 1995 377: 150–151
Robertson NP, O’Riordan JI, Chataway J et al. Offspring recurrence rates and clinical characteristics of conjugal multiple sclerosis Lancet 1997 349: 1587–1590
Ebers GC, Yee IML, Sadovnick AD, Duquette P the Canadian Collaborative Study Group. Conjugal multiple sclerosis: population-based prevalence and recurrence risks in offspring Ann Neurol 2000 48: 927–931
Hillert J . Human leukocyte antigen studies in multiple sclerosis Ann Neurol 1994 36: S15–17
Weber JL, May PE . Abundant class of human DNA polymorphisms which can be typed using the polymerase chain reaction Am J Hum Genet 1989 44: 388–396
Gyapay G, Morissette J, Vignal A et al. The 1993–94 Genethon human genetic linkage map Nat Genet 1994 7: 246–339
Davies JL, Kawaguchi Y, Bennett ST et al. A genome-wide search for human type 1 diabetes susceptibility genes Nature 1994 371: 130–136
Sawcer SJ, Jones HB, Feakes R et al. A genome screen in multiple sclerosis reveals susceptibility loci on chromosomes 6p21 and 17q22 Nat Genet 1996 13: 464–468
Haines JL, Ter-Minassian M, Bazyk A et al. A complete genomic screen for multiple sclerosis underscores a role for the major histocompatibility complex Nat Genet 1996 13: 469–471
Ebers GC, Kukay K, Bulman DE et al. A full genome search in multiple sclerosis Nat Genet 1996 13: 472–476
Kuokkanen S, Gschwend M, Rioux JD et al. Genomewide scan of multiple sclerosis in Finnish multiplex families Am J Hum Genet 1997 61: 1379–1387
Coraddu F, Sawcer S, D’Alfonso S et al. A genome screen for multiple sclerosis in Sardinian multiplex families Eur J Hum Genet 2001 9: 621–626
Broadley S, Sawcer S, D’Alfonso S et al. A genome screen for multiple sclerosis in Italian families Genes Immun 2001 2: 205–210
Sawcer SJ, Goodfellow PN, Compston DAS . The genetic analysis of multiple sclerosis TIGs 1997 13: 234–239
Celius EG, Vandvik B . Multiple sclerosis in Oslo, Norway: prevalence on 1 January 1995 and incidence over a 25-year period Eur J Neurol 2001 8: 463–469
Compston A, Ebers G, Lassmann H, McDonald I, Matthews B, Wekerle H . McAlpine’s Multiple Sclerosis Churchill Livingstone: London 1998 pp 63–100
Poser CM . The dissemination of multiple sclerosis: a Viking saga? A historical essay Ann Neurol 1994 36: S231–S243
Oturai A, Larsen F, Ryder L et al. Linkage and association analysis of susceptibility regions on chromosome 5 and 6 in 106 Scandinavian sibling pair families with multiple sclerosis Ann Neurol 1999 46: 612–616
Larsen F, Oturai A, Ryder L et al. Linkage analysis of a candidate region in Scandinavian sib pairs with multiple sclerosis reveals linkage to chromosome 17 Genes Immun 2000 1: 456–459
Masterman T, Ligers A, Olsson T, Andersson M, Olerup O, Hillert J . HLA-DR15 is associated with lower age at onset inmultiple sclerosis Ann Neurol 2000 48: 211–219
Holmans P, Clayton D . Efficiency of typing unaffected relatives in an affected sib-pair linkage study Am J Hum Genet 1995 57: 1221–1232
Becker KG, Simon RM, Bailey-Wilson JE et al. Clustering of non-major histocompatibility complex susceptibility candidate loci in human autoimmune diseases Proc Natl Acad Sci USA 1998 95: 9979–9984
Reed P, Cucca F, Jenkins S et al. Evidence for a type 1 diabetes susceptibility locus (IDDM10) on human chromosome 10p11–q11 Hum Mol Genet 1997 6: 1011–1016
Delepine M, Pocoit F, Habita C et al. Evidence of a non-MHC susceptibility locus in type 1 diabetes linked to HLA on chromosome 6 Am J Hum Genet 1997 60: 174–187
Jawaheer D, Seldin MF, Amos CI et al. A genomewide screen in multiplex rheumatoid arthritis families suggests genetic overlap with other autoimmune diseases Am J Hum Genet 2001 68: 927–936
The Transatlantic Multiple Sclerosis Genetics Cooperative. A meta-analysis of genomic screens in multiple sclerosis Multiple Sclerosis 2001 7: 3–11
Poser CM, Paty DW, Scheinberg L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols Ann Neurol 1983 13: 227–231
Gronlie SA, Myrvoll E, Hansen G, Gronning M, Mellgren SI . Multiple sclerosis in North Norway, and first appearance in an indigenous population J Neurol 2000 247: 129–133
Bodmer JG, Marsh SG, Albert ED et al. Nomenclature for factors of the HLA system, 1998 Hum Immunol 1999 60: 361–395
Holmans P . Affected sib-pair methods for detecting linkage to dichotomous traits: review of the methodology Hum Biol 1998 70: 1025–1040
Kruglyak L, Lander ES . Complete multipoint sib-pair analysis of qualitative and quantitative traits Am J Hum Genet 1995 57: 439–454
Ehm M, Wagner M . A test statistic to detect errors in sib-pair relationships Am J Hum Genet 1998 62: 181–188
Rice JP . The role of meta-analysis in linkage studies of complex traits Am J Med Genet 1997 74: 112–114
We are grateful to neurologists from the Nordic countries who identified the families involved in this study. We thank J Gray, B Smillie and M. Maranian for expert technical assistance.
This project has received financial support from the European Commission (project number CT97–2422), the Multiple Sclerosis Society of Great Britain and Northern Ireland (grant 591/00), the Wellcome Trust (grant 057097), the Danish Multiple Sclerosis Society, the Danish Medical Research Council, the Swedish Medical Research Council (project numbers: 11023 and 11220), and the Swedish Association of Neurologically Disabled, Karolinska Institute.
About this article
The Pharmacogenomics Journal (2019)
Biology Bulletin Reviews (2016)
Metabolic Brain Disease (2016)
HLA allele frequencies in Iranian opticospinal multiple sclerosis patients<br>——HLA in Opticospinal MS
Journal of Biomedical Science and Engineering (2011)
Genes & Immunity (2009)