Based on our CTL studies of over 44 persons with chronic HCV infection, we are able to arrive at a number of conclusions. Clearly this cellular immune response is heterogeneous among infected persons. We have not identified any specific HCV protein which appears to be immunodominant for CTL responses, but rather we have detected diverse responses to both structural and non-structural proteins. Using an identical stimulation strategy for all persons studied, we have been able to detect responses in only approximately one third of persons with chronic infection. Among these persons, the responses among liver-infiltrating lymphocytes are greater than those detected in fresh peripheral blood, suggesting that the CTL are homing to the site of maximal viral burden in these persons. Some viral proteins contain overlapping epitopes presented by more than one HLA class I molecule, and we have also found cases where peptides in the same HLA superfamily, such as the HLA A3 superfamily which contains A11, for which the same peptide can be presented by both alleles (manuscript in preparation). Although sequence variation between the infecting strain and the vaccinia constructs used to test for responses may lead to non-recognition of some variants, even the highly conserved core protein appears to be an inconsistent and actually infrequent target for detectable CTL responses. The magnitude of the CTL response appears to vary greatly, from being undetectable to being so vigorous that it an be detected in stimulated peripheral blood. The breadth of the response also varies widely, ranging from the detection of a response to a single epitope in some persons, to the simultaneous recognition of up to five different epitopes in others. Even in persons of the same HLA type, we have not seen consistent targeting of the same epitopes except in rare cases. Despite the detection of over 20 epitopes and their restricting class I alleles using CTR derived from liver-infiltrating lymphocytes, we have identified only one epitope that has been shown to be targeted by more than one person of the same HLA type. These findings lead us to speculate that the CTL response may be submaximal in the majority of infected persons. The reasons for this are presently obscure, but could relate to a number of factors. The epitopes targeted are found within variable regions of the virus, such that immune escape from established CTL responses has to be considered a real possibility. Sequence variation may also lead to antagonism of CTL responses, as has been demonstrated for both HIV and HBV infections. Furthermore, sequence variation either within or adjacent to regions containing CTL epitopes can lead to altered antigen processing, either due to alteration of proteolytic processing of the viral peptides in the cytoplasm or to altered transport and altered association with class I molecules. A number of issues regarding the CTL response in HCV infection still require substantial attention. The apparent inability of CTL to clear this virus needs to be addressed, as does the potential role for viral immunomodulatory molecules in HCV persistence. Although we and others have shown CTL responses to be present in persons with chronic infection, the role of CTL in acute HCV infection needs to be determined. The best studied chronic human viral infection is HIV infection, in which expanding data indicate that the early events following primary infection predict the subsequent course of illness. Viral load in the first 1-2 years after infection is highly predictive of the subsequent disease course in HIV infection, and recent experimental data in humans suggest that early immune responses may be predictive of subsequent disease course. Such studies in HCV infection have been difficult to achieve, since primary HCV infection is often asymptomatic, and transfusion-related cases are now rare. (ABSTRACT TRUNCATED)