Interest in pathogenesis of progressive multifocal leukoencephalopathy (PML) followed the observation of the high risk for the disease in HIV infection and the recent observation of an association with a variety of newer therapeutic modalities, eg, natalizumab, an α4β1 integrin inhibitor, and efalizumab, an anti-CD11a monoclonal antibody. Any hypothesis of PML pathogenesis must account for a number of facts. Firstly, the causative agent JC virus is ubiquitously present, yet only a vanishingly small number of infected persons develop the disease. Secondly, disorders of cell-mediated immunity increase the risk of the disease, particularly HIV infection. Impaired innate immunity is not a risk for PML, and antibodies against JC virus are not protective. Thirdly, a latent period of several months appears necessary following the administration of natalizumab and efalizumab before PML develops. Fourthly, restoration of the immune system can arrest the PML. It is possible that infection with JC virus occurs with a form of the virus shed in the urine of as many as 40% of all adults and present in sewage worldwide. Once acquired, perhaps through an oropharyngeal route, it may replicate and disseminate. A neurotropic form of JC virus that replicates in glial tissues causes PML when immunosurveillance is impaired. There are many unanswered questions with respect to PML pathogenesis. How is virus acquired? What tissues are infected? What is the origin of the neurotropic form? When does virus enter brain? What is the role of central nervous system immunosurveillance? The lack of an animal model has made answering these questions challenging.