The only known route of human infection with JEV is through the bite of an infected mosquito. Virus growing in the mosquito salivary glands (Takahashi and Suzuki, 1979) is inoculated into the skin and directly into capillaries; thousands of infectious particles are probably delivered. The initial site or sites of replication are unknown. The patient is asymptomatic while the virus multiplies and then viremia ensues. Circulating virus penetrates into the central nervous system, probably through defects in the endothelium, although infection through the cribriform plate or olfactory tract has been hypothesized (Albrecht, 1969). Symptoms begin 7 to 14 days after initial infection. JE is a diffuse encephalitis (Miyake, 1964); virus usually can be recovered from most if not all regions of the brain. Neurons containing JE antigens can be demonstrated throughout the brain, and the thalamus typically shows heavy involvement (Johnson et al., 1985). Glial elements are largely spared, and necrotic foci, when present, are of microscopic proportions.

When brain tissues from fatal cases are examined by immunohisto-chemical techniques, the earliest detectable host response is extra-vascular migration of mononuclear phagocytes (Johnson et al., 1985). These cells, accompanied by T-lymphocytes, cluster around infected (antigen-bearing) neurons. Simultaneously, meningeal exudates and perivascular cuffs composed of monocytes, T-cells, and B-cells accumulate. The infected neurons undergo pyknosis and fragmentation, and traces of antigen appear within the mononuclear cells in the nodules. Antibody synthesis by cells within the central nervous system can be detected directly by culture of CSF leukocytes early in the course of infection (Burke et al., 1985a, 1985b). A low or slow antibody response is associated with cultivable virus in the CSF and portends a grave prognosis (Burke et al., 1985c).

When JE occurs in a patient previously immune to another flavivirus (e.g., dengue), the IgG antibody response to JEV is brisk and strong, and an adverse outcome is less likely than in an individual experiencing JE as a first flavivirus infection (Edelman et al., 1975; Hammon, 1969; Sather and Hammon, 1970). The relative contributions of cellular and humoral immunity in these cross-flavivirus anamnestic responses in humans are unknown. Passively administered antibody has a protective effect in JE-challenged mice, even when administered 48 hours after virus challenge (Hammon and Sather, 1973; Ohyama et al., 1959). Inoculation with subimmunizing low doses of inactivated JE vaccine can prime nonhuman primates for an anamnestic antibody response to subsequent challenge, even in the absence of a detectable antibody response to the original immunization (Hoke and Burke, personal communication, 1985).

There is no evidence that the immune response contributes to the pathology in JE. Cyclophosphamide immunosuppression shortens the survival time in JE animal models (Nathanson and Cole, 1970). Little data exist on the role of interferon or that of cell-mediated immunity in JE.