The lesions of multiple sclerosis are most often held to be caused by an immune attack on CNS myelin.1 The nature of the antigens involved remains obscure, but peptides from at least two major myelin proteins—myelin basic protein and proteolipid protein—are known to cause cell mediated demyelination in animals (experimental allergic encephalomyelitis) which in some ways resembles human multiple sclerosis.1 Minor myelin components may presumably also act as autoantigens. Myelin oligodendrocyte glycoprotein (MOG) is a quantitatively minor myelin protein localised to oligodendrocyte cell bodies and processes and to the outer layer of CNS myelin sheaths. Originally detected by a mouse monoclonal antibody to rat cerebellar glycoproteins, the amino acid sequence of MOG has been deduced from cDNAs of rat, mouse, human, and bovine species and shows a high degree of evolutionary conservation and sequence motifs identifying it as a member of the immunoglobulin superfamily. In common with several other members of this family, MOG has a membrane spanning domain and an extracellular glycosylated N-terminus and its presence on the outermost surface of myelin and the oligodendrocyte plasma membrane may make MOG accessible to the immune system. The evidence that MOG may act as an autoantigen in multiple sclerosis includes reports that anti-MOG antibodies cause extensive CNS demyelination both in vivo and in vitro, that peptides from the primary sequence of MOG can produce experimental allergic encephalomyelitis, that the predominant T cell response in a population of patients with multiple sclerosis is to MOG and that antibodies to MOG can be demonstrated in the CSF and serum of patients with multiple sclerosis. Additionally the human MOG gene has been localised to a region of the major histocompatibility complex on chromosome 6, the significance of this in immune theories of the causation of multi- ple sclerosis being unclear.2 Against this background, analysis of the human MOG gene has defined its intron/exon structure and provided evidence that a 1.9 kb Taq 1 restriction fragment length polymorphism (RFLP) of the human MOG gene could be linked to multiple sclerosis in a significant fashion in a population of Australian patients with multiple sclerosis.3 We have examined this Taq 1 polymorphism in a population of patients with multiple sclerosis and controls from the Southampton area.

Venous blood samples were obtained from 40 patients with multiple sclerosis in the Southampton area and from 80 age and sex matched controls. The patients were 29 women and 11 men diagnosed as having clinically definite multiple sclerosis by the Poser criteria.4 The age range was 23–43 years with a mean of 33.9 years. DNA was extracted by standard procedures,5digested with Taq 1, electrophoresed, and subjected to Southern blotting using the same human MOG cDNA as previously.3 Six polymorphic bands of 4.9, 4.4, 3.25, 2.4, 1.9, and 1.65 kb were found, plus three invariant bands of 1.75, 1.25, and 1.05 kb. This contrasts with the Australian study, in which Taq 1 digestion of human genomic DNA and hybridisation with the same MOG cDNA probe showed five polymorphic bands of 5.25, 4.65, 2.4, 1.9, and 1.05 kb and only one invariant band of 1.7 kb. In that study the 1.9 kb band was present in 12.2% of patients with multiple sclerosis and 3.7% of controls, a difference which was significant (p<0.05). The table shows the distribution of the Taq 1 digest polymorphic bands between patients with multiple sclerosis and controls in the present study; none of these distribution differences were statistically significant. We assume that the 1.9 kb band found here is the same as in the previous study, in which case the incidence in the two sets of patients with multiple sclerosis is very similar—12.5% and 12.2%. However the incidence of this RFLP in the control group in the present study was much higher than in the previous study—8.8% v3.7%—thus denying significance. We assume our finding of extra invariant bands in Taq 1 digests of the human MOG gene and largely different sized RFLPs are due to genetic differences in the two populations. This illustrates the difficulties of extrapolating between two geographically widely separated cohorts. It is also important to have comparable groups in terms of clinical types of multiple sclerosis, age and sex structures, and nature of the control populations used.