The prevalence of multiple sclerosis in the United States is 250,000 to 350,000 (Anderson et al 1992), revised to 400,000 by the U.S. National Multiple Sclerosis Society in 2007 to account for population growth. World prevalence is estimated at 1.25 million (Dean 1994). The incidence was 3.2 per 100,000 cases a year in the United States in the 1990s (Jacobson et al 1997), 4.2 in the U.S. in 2007, and 7.5 in Olmsted County, Minnesota, the home of the Mayo Clinic (Mayr et al 2003).
Although some studies show a stable incidence of multiple sclerosis, the number of cases in other locales is increasing. In Olmsted County, the prevalence quintupled and the incidence quadrupled in the past 70 years (Wynn et al 1989), although it appears to have stabilized in high prevalence areas (Mayr et al 2003). In Canada, the increase is largely in females (Orton et al 2006). The prevalence of multiple sclerosis has increased in regions of Scotland, Finland, Norway, Lower Saxony, Sardinia, Italy, Sicily, and the French West Indies (Kurtzke 1991). Allergy, Crohn disease, and type I diabetes show similar geographical distribution and increasing incidence. The increase has been attributed to altered immune regulation as exposure to infectious diseases has diminished (Bach 2002).
Geographical variation in the prevalence of multiple sclerosis is striking. Multiple sclerosis is rare in equatorial countries. The disease becomes more common with distance from the equator in either hemisphere. This variation is partially due to Northern European, especially Scandinavian, ancestry in affected populations, but there is also an environmental influence (Page et al 1993). In the United States, early studies showed that northern areas had a prevalence of over 100 per 100,000, whereas it was only 20 per 100,000 in southern states. This gradient has attenuated over time. Incidence is high (greater than 30 in 100,000) in northern Europe from Iceland to Russia, and in Canada, New Zealand, and southern Australia. Incidence is moderate (5 to 29 in 100,000) in the Mediterranean basin, the southern United States, and southern South America. Incidence is low (less than 5 in 100,000) in East Asia, India, Africa, the Caribbean, Central America, Mexico (especially in Indians and mestizos), and northern South America (Kurtzke 1975; 1993; Pugliatti et al 2002).
Is the cause of multiple sclerosis genetic or environmental? Migration, ethnic, and twin studies suggest that genes and environment both influence the development of multiple sclerosis. Northern European ancestry is a major risk factor for development of multiple sclerosis. Scandinavian ancestry is strongly correlated with multiple sclerosis risk (Pearson product moment correlation = 0.5). English ancestry is negatively correlated in the United States (-0.5) (Page et al 1993). The rate in England is 42 to 80 per 100,000 (Kurtzke 1975). Israeli Jews have a prevalence of up to 62 per 100,000, but Christians (35 per 100,000), Moslem Arabs (15), Druze (11), and Bedouins (17) have lower rates (Alter et al 2006). Genetically similar immigrants have half the rate of native-born Jews, suggesting an environmental factor. Other groups also have a low incidence of multiple sclerosis (Gypsies, Asians, and native black Africans). Five percent of multiple sclerosis patients in the United States are black. Black Americans of African ancestry (often racially mixed) born anywhere in the United States have a relatively high risk compared to native Africans, but half the rate of Caucasians in the United States (Kurtzke et al 1979). They are more likely to have optico-spinal symptoms, larger MRI lesion volumes, and faster disease progression than whites. In contrast, people of Japanese ancestry in the United States have low rates of multiple sclerosis (Detels et al 1977), but much or all of the association disappears when covariates such as socioeconomic status are excluded (Marrie et al 2006).
Genetic influences. There are many genetic influences on the development of multiple sclerosis. In studies of twins, the monozygotic concordance rate is 31% (200 times background), the dizygotic rate is 5% after 7.5 years of observation (Sadovnick et al 1993), and the sibling risk is 3.5%, indicating a genetic component to multiple sclerosis. First-degree relatives have a 25-fold increased risk of developing multiple sclerosis compared to the general population (Hogancamp et al 1997), and there may be a greater maternal influence. When both parents are affected, 9% of the children develop multiple sclerosis. In theory, the highest risk monozygotic twin has an affected parent and a twin sister with onset before 21 years of age. Mothers and fathers are equally likely to transmit the disease, with no evidence of a Carter effect-- where the parent who is less likely to be affected is more likely to transmit the disease (Herrera et al 2007). Gender, age at onset, disease course, and severity are more similar than expected among affected patients in a family (Kantarci et al 2002). There are reports of a familial predisposition for a similar course of multiple sclerosis (Hensiek et al 2007), but others believe phenotypes are not concordant (Ebers et al 2000). Children of multiple sclerosis patients (Fulton et al 1999), 10% of first degree relatives, and unaffected twins often have abnormal MRIs (Mumford et al 1994), but their T cell responses to myelin basic protein are normal (Martin et al 1993; Ragheb 1993). The large number of unaffected monozygotic twins (70%) is a strong argument for a significant environmental contribution.
No single Mendelian locus causes multiple sclerosis. However, a limited number of interacting genes might affect susceptibility (Sadovnick 1993). Linkage to DR2 (HLA-DRB1*1501, possibly with DQB1*0602) is strongest in Northern Europeans, but other HLA-DR subtypes are seen in multiple sclerosis patients in the Middle East, Turkey, Sardinia, and Japan. Western forms of multiple sclerosis are linked to DR2 in Japan (Kira et al 1996). In black Americans, African HLA ancestry correlates with disability; DRB1*15 is less likely to be associated with neuromyelitis optica than with typical Western multiple sclerosis. In non-DR2 Japanese patients, multiple sclerosis often resembles Devic disease (eye and spinal cord involvement, seldom with CSF oligoclonal bands). DR4 is linked to a primary progressive course (Kantarci et al 2002). DR2 correlates with the presence of oligoclonal bands in the CSF but not with MRI lesions (Soderstrom et al 1998). This suggests 2 different HLA-linked mechanisms in central nervous system lesions. HLA-Bw4, DRB5 (less progression and severity), DRB1*01, CDR1*14, B*4402, and HLA-C*05 may be resistance factors. HLA-B12 has been linked to multiple sclerosis, and as fate would have it, to vitamin B12 deficiency and myelopathy.
Non-HLA genes that have weak links to multiple sclerosis or to its course include T cell receptors, immunoglobulin allotypes, POU2AF1 (transcriptional coactivator that regulates immunoglobulin expression), complement factors (C6, C7, properdin), the IL-2 receptor beta chain, IL-7 receptor alpha chain, intercellular adhesion molecule-1 (K469E), tumor necrosis factor alleles, the CD45 tyrosine phosphatase, CD24 (a heat stable antigen that may enhance T cell persistence in the brain), synapsin III, Tyk2 in the interferon response pathway, and possibly mitochondrial DNA. Other candidate genes code for myelin basic protein, transketolase, IL-10, chemokines, p53, estrogen and vitamin A receptors, Jagged1 (oligodendrocyte differentiation), and proteolytic enzymes such as calpain. With many of these correlations, a genetic link is possible, but likelihood of a type I error is strong.
Single nucleotide polymorphisms (SNPs) linked to multiple sclerosis appear in the IL-1, IL-2 receptor, IL-7 receptor alpha chain, T cell receptor-related SH2D2A, CD24 costimulatory and antigen presentation molecule, CD58 adhesion molecule, IFN-gamma, IFN-gamma receptor (debated), interferon regulatory factor-5 (IRF-5), MxA, brain-derived neurotrophic factor (BDNF), RAGE, chemokines (CCL3, CCL15, and others), tissue plasminogen activator, GPC5 and KIF1B (axonal transport of mitochondria and synaptic vesicles), mitochondrial complex I, free radical scavengers (paraoxonase I), and anti-glycation (glyoxalase I). SNP links are absent for interferon regulatory factor 1, bax, bcl-2, bcl-x, and p53. The IL-7 receptor alpha link is weak, but altered functions could be important in a subset of patients, as IL-7 enhances immunity and the receptor chain is upregulated by steroids, tumor necrosis factor, and type I interferons.
Some genes modify the course of multiple sclerosis (This does not equate to linkage with the presence of the disease, or susceptibility). These genes affect immune regulation and glial or neuronal vulnerability. Three percent of Europeans have a homozygous deletion of ciliary neurotrophic factor (CNTF), a growth factor for neurons (Linker et al 2001). In this group, the disease is more severe and onset is earlier (Giess et al 2002). Mice lacking ciliary neurotrophic factor or leukemia inhibitory factor (LIF) have worse experimental allergic encephalomyelitis. ApoE4 may be more common in progressive forms of multiple sclerosis and auger cognitive impairment, a faster rate of disability progression, and more MRI destruction, although some studies and a large meta-analysis find no link. Monocytes expressing the chemokine receptor, CCR5, accumulate in multiple sclerosis lesions, and CCR5+ T cells correlate with MRI lesions. A mutation of the receptor, CCR5-delta 32, (homozygous in 1% and heterozygous in 13% of Caucasians) protects against HIV infection as well as severe rheumatoid arthritis (Marmor et al 2001). This mutation is associated with multiple sclerosis (Favorova et al 2002) and slows progression (Kantor et al 2003). Other putative or unconfirmed genetic links to the course of multiple sclerosis include the IL-1beta receptor and IL-1 receptor antagonist, transforming growth factor-beta, immunoglobulin Fc receptors, CD24, CTLA-4, alpha B-crystallin (linked to progression), and phenylethanolamine N-methyl transferase (converts norepinephrine to epinephrine).
Genetic or environmental control of response to target antigens. Antibody response to certain viruses, particularly measles, is increased. Antiviral responses are probably not specific for a single inciting agent, as they vary among plaques and among patients (Mattson et al 1980).
Excessive antibody responses may be part of the immune dysregulation that characterizes multiple sclerosis. Nonspecific activation of B cells and exposure to CNS antigens are potential driving factors. The apparent increase in anti-virus antibody responses could simply be due to a nonspecific rise in all titers, making it easier to detect the antibodies. There is a significant increase of autoantibodies to 2,’3’ cyclic nucleotide 3’ phosphodiesterase (IgM), alpha B- and alpha A-crystallin (anti-inflammatory), aquaporin-4 (Devic variant), cardiolipin, contactin-2/TAG-1 or contactin/TIP30 of the juxtaparanodal domain (rats), DNA, galactocerebroside, gangliosides, glial fibrillary acidic protein (GFAP; in secondary progressive multiple sclerosis, strong correlation with clinical deficits), glyceraldehyde-3-phosphate dehydrogenase (GAPDH, linked to fatigue), glycopeptides, heat shock proteins (60 and 90), myelin proteins (MAG, MBP, MOG, OSP, phosphatidylcholine, and PLP), neurofilament light chains (axons), neutrophil cytoplasmic antigen, NG-2 (AN-2), Nogo (debated), nuclear antigens, proteasomes, transaldolase, thyroid microsomal antigens, smooth muscle, and thyroglobulin (Reindl et al 2006). Elevations are most common in progressive forms of multiple sclerosis (Spadaro et al 1999). There are reports of excessive responses to chlamydia. IFN-beta does not induce autoantibodies, but interferon therapy on a background of autoantibodies is more likely to lead to neutralizing antibodies to interferon.
Excessive T-cell reactions to a variety of brain antigens approach the threshold of statistical significance. This might be expected in a chronic inflammatory disease of the central nervous system and does not prove causation. Myelin basic protein-reactive T cells are more common than in controls, especially when high avidity cells are detected with low, physiologically relevant levels of myelin basic protein (Bielekova et al 2004). However, myelin basic protein-reactive cells are equivalent between patients and their normal family members (Fredrikson et al 1994), maybe reflecting familial HLA-regulated responses to antigens.
Cytokine production in both innate and adaptive immunity is hereditary and predicts the type of multiple sclerosis. Th1 responses are strongly linked in families, with 0.8 to 0.9 hereditability. In healthy family members of patients, lipopolysaccharide-stimulated IL-10 is reduced by 12%, and tumor necrosis factor-alpha is increased by 10% compared to multiple sclerosis-free families. Low IL-10 plus high tumor necrosis factor-alpha in a family predicts a 4-fold increased risk of developing relapsing or remitting multiple sclerosis, and an 8-fold increase of relapsing-remitting over primary progressive multiple sclerosis (de Jong et al 2000).
Environmental influences. Migration studies show that environment determines some of the risk for multiple sclerosis (Hogancamp et al 1997). Migrants to a low incidence area have a smaller risk of developing multiple sclerosis than if they had remained in situ (Ebers and Sadovnick 1993). Asians and Latinos maintain their low risk after migration (Detels et al 1977; Ebers and Sadovnick 1993). People who migrate from a low incidence area to a high incidence area before the age of 15 years have a high risk of developing multiple sclerosis. After the age of 15, migration does not affect the risk of developing multiple sclerosis, although not all studies agree.
The ratio of “Asian” (prominent optic nerve and spinal cord demyelination) to “Western” clinical phenotypes has changed in Japan from 2:1 in patients who were born in the 1920s to 1:4 in patients born in the 1970s. This suggests environmental alterations have modified the form of multiple sclerosis (Kira et al 1999). Canine distemper virus, related to measles, or other viruses carried by small house pets were once implicated, but the association was a result of recall bias. In the Faroe Islands, 4 epidemics of multiple sclerosis appeared after British troops occupied the islands in 1940 through 1944. Multiple sclerosis onset was attributed to a virus carried by the British. The virus required prolonged exposure (2 years) in people 11 to 45 years old, and the agent is theorized to cause multiple sclerosis 5 to 8 years after exposure (Kurtzke et al 1993). It is also possible that contact with the multiple sclerosis agent at an early age (0 to 3 years old) is protective; Faroese born between 1941 and 1945 do not have multiple sclerosis (Cooke 1990).
Virus infections sometimes trigger exacerbations of multiple sclerosis. One third of patients with upper respiratory infections will have an exacerbation (Sibley et al 1985; Panitch et al 1991; Correale et al 2006) and one third develops new MRI lesions during the “at risk” period, especially in early multiple sclerosis (Sibley 2001). Picornaviruses and perhaps all rhinoviruses may be the most potent triggers. Nonetheless, virus infections decrease by 20% to 50% in multiple sclerosis (Sibley et al 1985), especially when the disease becomes progressive. Faster progression correlates with fewer virus infections (Sibley 2001). Dysregulated interferon responses are suspect (Feng et al 2002a). High antibody titers to canine distemper virus correlate with a 5-fold increased risk of multiple sclerosis, but many patients have not been exposed. Epstein-Barr virus titers are also elevated, but Epstein-Barr virus in brain lesions has not been confirmed.
Bacterial infections increase exacerbations by 3-fold (Rapp et al 1995; Correale et al 2006) although some believe virus infections are more likely to trigger true exacerbations. Smokers may induce multiple sclerosis in themselves and their children. Smokers have a 60% increase in exacerbations, possibly from chronic bronchitis and immune activation, but there is no increase in progression. Interferon therapy does not reduce virus infection rates, but it prevents virus infections from triggering exacerbations (Panitch 1994). There seems to be no effect of interferon-neutralizing antibodies on infections.
Multiple sclerosis is not transmitted vertically (breast milk), through transfusions, or conjugally. Doctors, nurses, and spouses of patients do not have an increased incidence of multiple sclerosis. This lack of transmission argues against known viral or retroviral infections.
Environmental antigens and age shape immune responses. A dirty environment (ie, viral and bacterial exposure) allows transgenic mice (with V-beta-8.2, V-alpha-2.3, myelin basic protein-specific T cell antigen receptor genes) to develop experimental allergic encephalomyelitis. However, no lesions develop in transgenic mice raised in a clean, specific pathogen-free facility (Goverman et al 1993). Moreover, it is very difficult to induce experimental allergic encephalomyelitis in wild mice from a non-laboratory environment. In humans, all environments contain pathogens, but the type and timing of exposure could affect immunity as well as tolerance. Parasites and microbiome richness are associated with balanced selection of interleukin polymorphisms that are effective against viruses and bacteria. Exposure to infant siblings during the first 6 years of life decreases the incidence of multiple sclerosis by up to 8-fold (Ponsonby et al 2005). Multiple sclerosis patients generally have more education and higher socioeconomic status than average. As a corollary, sanitation may be better, and childhood infections occur later in patients than in the general population (Alter et al 1986; Delasnerie-Laupretre and Alperovitch 1990). Regions with a high incidence of multiple sclerosis have a low incidence of hepatitis B and schistosomes. This suggests that a strong immune response to less-frequent viruses triggers an autoimmune or bystander reaction in multiple sclerosis. "Clean" environments (in man this may mean exposure to pathogens late in childhood) seem to predispose one to multiple sclerosis, rheumatoid arthritis, and inflammatory bowel disease—the “hygiene hypothesis.” Antibiotics can have a prolonged effect on some taxa of gut microbiota. They reduce complexity of the microbiome in mice and shift immunity from a mix of Th17 and T-regulatory cells to one of Th17 predominance.
The intestinal microflora is a diverse ecosystem, and most of its bacterial RNA sequences are from novel uncultivated microorganisms. Toll-like receptors on immune cells in the gut are activated by intestinal microbes, leading to tolerance of food antigens, less inflammatory bowel disease, and possibly to altered immune responses in multiple sclerosis. Some microbiota induce IL-10-secreting regulatory T cells, others induce Th17 or Th1 proinflammatory cells (Ivanov et al 2008). Stress, antibiotics, opiates, and ischemia all increase virulence of gut flora (Alverdy et al 2005) and induce IL-17 and tumor necrosis factor.
Dental caries correlate with higher incidence of multiple sclerosis. Periodontal disease bacteria increase the severity of experimental allergic encephalomyelitis. Some bacteria, “probiotics,” as well as parasitic infestation with helminths induce Th2 responses and reduce the severity of experimental colitis as well as human ulcerative colitis and Crohn disease. Infestation with the parasites may prevent multiple sclerosis from developing (Fleming and Cook 2006). A trial is in progress using ova from Trichuris trichiuria, the pork whipworm. Parasites induce eosinophilia, but also regulatory T and B cells and regulatory macrophages, plus secretion of IL-4 and IL-10 and transforming growth factor-beta. Parasites also reduce secretion of IFN-gamma and IL-12 and reduce new MRI lesions, clinical progression, and attack frequency 20-fold (Correale and Farez 2007). Oral tolerance with myelin basic protein antigens does not affect the course of multiple sclerosis, but the richer antigenic repertoire of parasites and probiotics could have benefit.
Seasonal variation in activity of multiple sclerosis differs in various locales (Goodkin and Hertsgaard 1989), perhaps related to vitamin D levels or virus exposure. In autumn and winter, IFN-gamma production by activated mononuclear cells is increased in multiple sclerosis patients compared to controls (Balashov et al 1998), but this does not correlate with MRI lesions (Killestein et al 2002b). MRI activity peaks in late springtime (Auer et al 2000).
Vitamin D affects the onset and the course of multiple sclerosis. The provitamin 7-dehydrocholesterol is synthesized in the skin. Ultraviolet sunlight converts it to vitamin D3 (cholecalciferol). This compound leaves the skin to be further activated in liver and then in renal mitochondria to calcitriol (1,25(OH)2; cholecalciferol; vitamin D2). Foods that contain vitamin D are fatty seafood, egg yolks, and chanterelle mushrooms. Serum vitamin D levels fluctuate with the seasons, possibly linked to the May/November birth-month ratio of 1.43 for development of multiple sclerosis (Sadovnick et al 2007). More people with multiple sclerosis are born in the springtime than in the fall, suggesting a vitamin D effect on the fetus. Vitamin D activates cAMP, which inhibits inflammation. It also induces regulatory T cells (Ghoreishi et al 2009); regulatory T cells migrate from the mother into fetal lymph nodes (Mold et al 2008), so an effect of sunlight during pregnancy is possible.
Tasmanian children exposed to large amounts of sunlight, especially in winter, as well as white United States soldiers are one third as likely to develop multiple sclerosis later in life (van der Mei 2003). Subjects with darker buttock skin (presumably not exposed to sunlight, thus, a measure of genetic background) were less likely to develop multiple sclerosis. Whites have higher vitamin D levels than blacks, and levels in whites correlate with resistance to development of multiple sclerosis. Fish consumption, outdoor work, and rural life lowers multiple sclerosis mortality. Nurses taking vitamin D supplements (greater than or equal to 400 IU daily) are 40% less likely to develop multiple sclerosis (Munger et al 2004; Ponsonby et al 2005), though their lifestyle could differ from those without vitamin D supplements. Similar correlations are seen in Norway (Kampman and Steffensen 2010). Women shrouded for religious reasons develop osteoporosis, especially in Northern European countries. Lack of sunlight could increase the incidence of multiple sclerosis in these women.
Multiple sclerosis patients have low vitamin D levels and demineralized bone due to a combination of fear of the sun’s heat, treatment with antiepileptics for pain, immobility, and possibly alterations in interferon responses and sympathetic innervation of bone (interferon therapy enhances bone formation; depression and sympathetic hyperactivity is linked to osteoporosis). Vitamin D metabolic pathways genes do not correlate with multiple sclerosis risk. Vitamin D increases intestinal calcium uptake and promotes bone mineralization. Receptors for vitamin D, vitamin A, thyroid hormone, numerous “orphan receptors,” and the peroxisome proliferator-activated receptor are similar. Vitamin D receptors increase on T cells that are activated with mitogens plus vitamin D3. Vitamin D inhibits CD4 T cell proliferation. CD8 cells express 2- to 3-fold more receptors than CD4 cells, but effects of vitamin D on CD8 cells are unknown.
Vitamin D is immunosuppressive. It induces IL-4, transforming growth factor-beta, and regulatory CD4 T cells, and it inhibits production of IL-2, IL-12, IFN-gamma, TNF-alpha, and function of Th1 and type I dendritic cells. It inhibits onset (more in females) and relapses of experimental allergic encephalomyelitis via an IL-10 pathway. Ultraviolet radiation and sunburn are strong inducers of IL-10 and Th2 responses, and sunburn can block immune responses to vaccination and various antigens. It is unknown whether rapid fluctuations in IL-10 affect the course of multiple sclerosis. In contrast, melatonin induces IL-12 and IL-18, leading to Th1 responses and worsening experimental allergic encephalomyelitis.
Other putative environmental factors that modify multiple sclerosis activity include in vitro fertilization (LHRH agonists), exposure to wool or sheep, and consumption of smoked sausage or fresh cow milk (the milk protein, butyrophilin, shares antigens with myelin oligodendrocyte glycoprotein). A high socioeconomic status confounds some of these factors (Ben-Shlomo 1992). Regular smoking doubles the risk of having multiple sclerosis; men are more susceptible than women, and adolescence may be a critical period of susceptibility to smoking effects on multiple sclerosis. A diet low in saturated fats (the Swank diet), or treatment with evening primrose oil (rich in linoleic and gamma-linolenic acids) may modestly lower the rate of exacerbations (Dworkin et al 1984). It appears to add to the benefit of interferon and glatiramer therapy. Leptin, made by adipocytes, is proinflammatory and is increased in multiple sclerosis (Matarese et al 2010).
Association with autoimmune diseases. Most autoimmune diseases are not associated with multiple sclerosis (Reder and Arnason 1985; Wynn et al 1989). There are scattered reports of multiple sclerosis coexisting with Crohn disease or ulcerative colitis and possibly with myasthenia gravis, type I diabetes, narcolepsy (also HLA-DR2-linked), and thyroid disease, but other associations are lacking. This suggests that the etiology of multiple sclerosis differs from most autoimmune diseases. Therapy with alemtuzumab (anti-CD52, Campath-1H) induces antithyroid antibodies, presumably by altering immune regulation. Importantly, the demyelinating variant of multiple sclerosis, Devic disease/neuromyelitis optica, is highly associated with autoimmune diseases (10-fold increase). Epidemiological mixing of multiple sclerosis with this variant leads to spurious associations with other autoimmune diseases.
Some diseases are infrequent in multiple sclerosis. Asthma and allergies are half as common as in the general population. Cancer is reduced to two thirds or three fourths in multiple sclerosis compared to controls (Sadovnick et al 1992; Koch-Henriksen et al 1998; Bahmanyar et al 2009). During trials of subcutaneous IFN-beta-1a, the reported to expected ratio of cancer was 1:11, with 50% more cases in the placebo groups than in the interferon groups (Sandberg-Wollheim et al 2011). Multiple sclerosis patients have low uric acid levels and rarely develop gout. Uric acid ameliorates experimental allergic encephalomyelitis. Therapeutic attempts to raise uric acid in multiple sclerosis are underway. There is a strong negative association with Down syndrome, possibly because chromosome 21 codes for type I interferon receptors and S100b (Weilbach and Toyka 2002). Lupus is rare in multiple sclerosis. Lupus has excessive interferon production and responses to interferon compared to low serum levels and interferon resistance in multiple sclerosis (Feng et al 2009b; Javed and Reder 2006).