Multiple sclerosis (MS) is a major cause of disability and handicap in young adults. The etiology of MS is still unknown, but an immune-mediated attack on myelinated axons caused by environmental factors in a genetically susceptible individual is considered to be of pathogenetic importance. In the majority of patients the disease course is initially characterized by the occurrence of attacks of disease such as acute optic neuritis (ON) and other isolated syndromes. In this stage, termed relapsing-remitting MS (RRMS) there is often good recovery but in most patients incomplete recovery or the conversion to a secondary progressive disease course eventually leads to the development of irreversible impairment. The presumed role of immune mechanisms has led to many attempts to treat MS with immunomodulatory drugs, including glucocorticoids. However, when the studies included in the present thesis were initiated there was little evidence supporting the common use of glucocorticoids in MS. We conducted two double-blind, randomized, placebo-controlled studies of oral high-dose methylprednisolone treatment in ON and attacks of MS. These studies showed that whereas methylprednisolone had only marginal efficacy in ON, treatment resulted in better recovery after an attack of MS for a period of at least eight weeks. The lower efficacy in ON may reflect that these patients have less intrathecal inflammation, as assessed by magnetic resonance imaging using Gadolinium (Gd)-enhancement as a measure of inflammation, than do patients with RRMS. Indeed, patients with enhancing lesions on MRI had prolonged attack duration and a better response to methylprednisolone treatment compared to patients without Gd-enhancement on MRI. Oral high-dose methylprednisolone treatment also resulted in resolution of Gd-enhancement and lowering of the cerebrospinal fluid (CSF) concentration of myelin basic protein (MBP), a biochemical marker of demyelination. Autoreactive CD4 T cells are considered to be pivotal in the pathogenesis of RRMS. We found that patients with possible onset symptoms of MS (POSMS) and RRMS had evidence of systemic CD4 T cell activation as evidenced by increased expression of the CD26 molecule, a putative marker of proinflammatory T cells that secrete T helper type 1 (Th1) cytokines. Th1 cytokines are considered to be instrumental in driving intrathecal inflammation in MS. Patients with POSMS or RRMS had a lower percentage of CD4 T cells in CSF expressing the HLA-DR molecule, and this was especially pronounced in patients carrying the MS-associated DRB1*1501 HLA allele. We also found a lower percentage of CD4 T cells in CSF that co-expressed CD25 and HLA-DR in POSMS and RRMS. HLA-DR and CD25 expression has been associated with regulatory T cell activity, and we hypothesize that changes in CD4 T cell expression of these molecules could reflect alterations in the activity of regulatory T cells in MS. The expression of CD26 decreased and coexpression of CD25 and HLA-DR increased after treatment with oral high-dose methylprednisolone. These changes may be related to the induction of the immunomodulatory cytokine transforming growth factor (TGF)-β after methylprednisolone treatment. Intrathecal IgG synthesis is a hallmark of MS. We found that intrathecal IgG synthesis levels correlated with demyelination as assessed by the CSF concentration of MBP in patients with RRMS. Furthermore, in RRMS the number of cells in CSF secreting anti-MBP autoantibodies of high relative affinity correlated with disease activity. In contrast, in POSMS, the number of cells in CSF secreting anti-proteolipid protein antibodies correlated with disease activity. We hypothesize that these findings reflect a change in pathogenetically relevant myelin autoantigens during the course of MS. The precise role of myelin autoantibodies, as effector molecules in immune-mediated demyelination or merely as markers of pathogenetically relevant T cell reactivity, remains to be established. However, systemic and intrathecal IgG synthesis was suppressed after treatment with oral high-dose methylprednisolone, consistent with a pathogenetic role of IgG at least in some patients with MS. The CSF leukocyte count and the CSF activity of matrix metalloproteinase (MMP)-9 correlated with Gd-enhancement in MRI studies prior to treatment with oral high-dose methylprednisolone. CSF pleocytosis and MMP-9 activity was not influenced by methylprednisolone treatment in spite of marked suppression of MRI disease activity, but patients with MMP-9 activity in CSF had shorter relapse-free periods. We hypothesize that this reflects the multi step nature of intrathecal inflammation in MS. Subarachnoid and perivascular inflammation may be necessary but not sufficient to initiate parenchymal inflammation with infiltration by phagocytic macrophages and other effector cells. Indeed, we found that patients carrying an allele encoding a deficient chemokine receptor, CCR5 Δ32, which is expressed by the majority of blood-derived macrophages in MS lesions, had prolonged attack-free periods.