Control of virus infection of the central nervous system (CNS) 1 is a complicated task for the host because of the potentially damaging consequences of immune responses in the brain and spinal cord. Neurological damage can be associated with cellular cytotoxicity, altered vascular permeability, and the influx of inflammatory cells into an enclosed space containing nonrenewable cells of vital importance to the host. Not surprisingly, therefore, the CNS has a number of mechanisms for controlling and regulating the development of local inflammatory processes. These include the blood-brain barrier, limited capability for antigen presentation, and functional modulation of immune reactions by gangliosides and astrocytes. The blood-brain barrier is composed of nonreactive endothelial cells linked by tight junctions, a basement membrane, and apposing astrocytic foot processes (see Licinio perspective, Fig. 1, in this issue of The Journal). This barrier effectively excludes most circulating soluble factors and circulating leukocytes from the parenchyma of the brain and spinal cord. In addition, cells within the CNS do not normally express class I or class II antigens of the MHC, which are necessary for recognition of antigens by T cells. However, T cells activated in the periphery do cross the blood-brain barrier and enter the CNS in an antigen-nonspecific manner as a part of routine immunologic surveillance. These activated T cells express the adhesion molecules and matrix metalloproteases necessary to traverse resting cerebral capillary endothelial cells and the basement membrane. Only T cells specific for an antigen present in the brain or spinal cord are retained within the CNS. Cells without such specificity leave within a few hours (1). Antigen-presenting cells which promote retention of specific lymphocytes within the CNS are most likely to be perivascular macrophages or microglial cells, both of which can quickly upregulate MHC antigen expression in response to stimulation.

After activated T cells enter the CNS and interact with local antigen-presenting cells, production of cytokines will amplify the inflammatory response. However, the activity of these antigen-specific T cells is strongly regulated by the high concentrations of gangliosides present in the CNS. Gangliosides inhibit T cell proliferation and production of IL-2 after entry into the CNS without decreasing production of the type 2 cytokines IL-4 or IL-10 (2). In addition, astrocytes inhibit production of IL-12, necessary for induction of a type 1 cytokine response (3). These and other data suggest that the CNS, more than other organs, favors the development of local immune responses characterized by the production of type 2 cytokines by T cells (4). These cytokines favor the differentiation of B cells for local antibody production and downregulate macrophage activation by IFN-. Failure of immune regulation within the CNS due to the genetic background of the host or the nature of the stimulus would favor the development of autoimmune neurologic disease which is generally associated with production of type 1 cytokines or a response to infection that is detrimental to the host.