It should be apparent from the foregoing that the transgenic mouse model system has contributed substantially to our understanding of many aspects of HBV biology, immunobiology and pathogenesis in the past several years. We have learned that HBV can replicate within the mouse hepatocyte, as well as other mouse cell types, suggesting that there are probably no strong tissue or species specific constraints to viral replication once the viral genome enters the cell. However, the failure thus far to detect viral cccDNA in the hepatocyte nucleus in several independently derived transgenic lineages suggests that other, currently undefined, constraints on host range and tissue specificity may also be operative. Thanks to the transgenic mouse model we now understand the pathophysiological basis for HBsAg filament formation and ground glass cell production, and we have learned that at least this viral gene product can be toxic for the hepatocyte, first by compromising its ability to survive the hepatocytopathic effects of LPS and IFN alpha and eventually by causing it to die in the absence of any obvious exogenous stimulus. In recent studies, it has been shown that preformed nucleocapsid particles do not cross the nuclear membrane in either direction at least in the mouse hepatocyte. If this is confirmed, it will have two important implications: first, that nucleocapsid disassembly must occur in the cytoplasm before the nascent viral genome can enter the nucleus; second, that the intranuclear nucleocapsid particles are empty, and therefore serve no currently defined purpose in the viral life cycle. This should stimulate new interest in the analysis of the function of these particles that are a prominent feature of mammalian hepadnavirus infection. The transgenic mouse model has also established definitively that HBV-induced liver disease has an immunological basis, and that the class I-restricted CTL response plays a central role in this process. Additionally, the mouse studies have taught us that when the CTL recognize their target antigen on the hepatocytes they cause them to undergo apoptosis, forming the acidophilic, Councilman bodies that are characteristic of viral hepatitis. Further, we have learned that although the CTL initiate the liver disease, they actually contribute more to disease severity indirectly by recruiting antigen nonspecific effector cells into the liver than by directly killing the hepatocytes themselves. In addition, by releasing IFN gamma when they recognize antigen, the CTL can destroy enough of the liver to cause fulminant hepatitis in mice whose hepatocytes overproduce the large envelope protein and are hypersensitive to the cytopathic effects of this cytokine. We have also learned that the CTL are unable to recognize HBV-positive parenchymal cells outside of the liver, apparently because they cannot traverse the microvascular barriers that exist at most extrahepatic tissue sites. This important new discovery may permit the virus to survive a vigorous CTL response and contribute not only to the maintenance of memory T cells following acute hepatitis but also to serve as a reservoir to reseed the liver in patients with chronic hepatitis. The transgenic mouse model has also revealed that activated CTL and the cytokines they secrete can down-regulate HBV gene expression, and possibly even control viral replication, by noncytotoxic intracellular inactivation mechanisms involving the degradation of viral RNA and, perhaps, the degradation of viral nucleocapsids and replicative DNA intermediates without killing the cell. If HBV replication is indeed interrupted by this previously unsuspected activity, it could contribute substantially to viral clearance during acute infection when the immune response to HBV is vigorous. Alternatively, it could also contribute to viral persistence, by only partially down-regulating the virus during chronic infection when the immune response is weak.