This study was designed to analyze timing and mechanisms of CD8 T cell recruitment to the liver in acute HCV infection. The prospective in vivo data indicate a role for CXCL10, CXCL11, CCL4 and CCL5, but not CXCL16 in CD8 T cell recruitment, and a role for type I IFNs in the induction of these chemokines. The latter was confirmed by in vitro experiments in which intracellular HCV RNA induced the expression of CXCL10, CXCL11, CCL4 and CCL5 in human hepatoma cells in a type I IFN-dependent manner. Thus, HCV RNA-induced type I IFN recruits CD8 T cells to the liver via induction of chemokines.
Based on our findings, we propose the following sequence of events in acute HCV infection (). The earliest detectable event is an intrahepatic type I IFN induction in response to intracellular HCV RNA. The source of type I IFN is currently undefined in vivo
, but in vitro
experiments suggest that it may be produced by either hepatocytes (this study and 14
or plamacytoid dendritic cells 17
). This is followed by production of chemokines 2-8 weeks after HCV infection. Chemokine expression is not limited to hepatocytes but also found on sinusoidal cells, as demonstrated for CXCL11 () with the caveat that we did not have established immunohistochemistry techniques for the other chemokines that were quantitated by realtime PCR in the liver and EIA in blood. Adhesion molecules ICAM-1 and VCAM-1 which facilitate T cell entry into the liver are detectable at the same time (). However, HCV-specific tetramer+
T cells become detectable in the circulation not earlier than 8-12 weeks after infection. They express CXCR3 and/or CCR5 and are immediately recruited into the liver as evidenced by the selective disappearance of CXCR3+
cells from circulation. Amplification of the intrahepatic lymphocytic infiltrate results in liver injury, as evidenced by serum ALT elevation, which is followed by a decrease of HCV RNA titer and/or HCV clearance.
Model of the sequence of events in liver and blood culminating in the acute phase of hepatitis C
This specific sequence of events that we observed in HCV infection differs from that in other hepatotropic viral infections. In acute hepatitis B virus (HBV) infection, for example, intrahepatic type I IFN responses appear to be absent as determined by a lack of detectable 2,5-OAS-1 mRNA induction 18
. Accordingly, there is no increase in the expression of CXCL10 and CXCL11 during the incubation phase of HBV infection 18
. CXCL10 and CXCL11 are only induced in the later, acute phase of HBV infection, concomitant to an increase in intrahepatic IFN-γ mRNA 18
, which is a well known inducer of CXCL10 and CXCL11. These results suggest that the mechanisms of chemokine induction and T cell recruitment differ in HCV and HBV infection.
Because the generation of Patr-tetramers required the mapping of T cell epitopes for the each chimpanzee, these reagents were not yet available at the time the liver biopsies were obtained. Therefore, our study had to rely on molecular methods to quantitate T cell subsets and chemokine receptor expression and it was not possible to differentiate between HCV-specific T cells and non-specific bystander cells in the liver. Nevertheless, several observations support the notion that the early intrahepatic immune response was antigen-specific. First, the population of HCV-specific, tetramer+
CD8 T cells that expressed CXCR3 and/or CCR5 selectively decreased in the blood at the same time as CD8β mRNA levels increased in the liver, suggesting recruitment of HCV-specific, tetramer+
CD8 T cells from the circulation to the liver. Second, adoptive transfer studies demonstrated in a mouse model of acute hepatitis B that virus-specific T cells are the initiators of a cascade that leads to amplification of the intrahepatic infiltrate by non-specific bystander cells and ultimately, to increase of serum ALT levels and acute hepatitis 19
Chemokine-dependent mechanisms of T cell recruitment also operate in chronic HCV infection as shown in several cross-sectional studies that examined the expression of chemokine receptors on liver-infiltrating T cells 20, 21
, and intrahepatic expression of chemokines 20, 22
. Although these cross-sectional studies did not establish a temporal relation between the expression of chemokine receptors on circulating T cells, the induction of chemokines in the liver and the recruitment of T cells to the liver, they implicate that CXCR3 20-22
, CCR5 21
, CCR7 23
, CXCR6 21
, and CXCR4-binding chemokines 24
contribute to chronic HCV infection. Our study demonstrates that CCR5- and CXCR3-binding chemokines but not the CXCR6-binding chemokine CCL16 play a role in acute HCV infection. The continuing induction of CCR5-binding rather than the CXCR3-binding chemokines 6 months after infection in those animals that progress to chronic HCV infection suggests a shift in intrahepatic chemokine production as acute hepatitis turns into chronic hepatitis.
In this context it is of interest that an antagonist to CXCL10 has recently been proposed to contribute to nonresponse to interferon-based therapy in chronic hepatitis C 25
. We do not think that this mechanism explains our findings for the following reasons: First, the current study relates to the spontaneous outcome of acute HCV infection not to treatment-induced outcome of chronic HCV infection. Second, CXCL10 levels do not predict outcome of acute HCV infection whereas they are negative predictors of treatment outcome 25
. Third, the late appearance of HCV-specific T cells in the circulation in acute hepatitis C in the current study coincides precisely with the onset of the intrahepatic CD8 T cell infiltrate suggesting impairment of T cell priming rather than recruitment.
Importantly, different outcomes of HCV infection were observed even though the induction of HCV-specific T cells, the expression of chemokine receptors on HCV-specific T cells, the expression of chemokines in the liver and the recruitment of CD8 T cells to the liver occurred at comparable time points in all chimpanzees. These results differ from those of a previous study, which suggested an association between impaired T cell recruitment to the liver and a chronic course of HCV infection 9
. This study, however, tracked the appearance of HCV-specific T cells in the blood with in vitro assays that required several weeks of T cell expansion rather than using an ex vivo analysis with tetramers. This methodical difference is important because HCV-specific CD8 T cells have been described as being “stunned” with impaired proliferative capacity 9, 26
and thus might not have been detected by functional assays. Other differences between the two studies are the nature of the HCV inoculum and the route of HCV infection. Chimpanzees were inoculated with different HCV strains via either an intrahepatic or intravenous route of infection in the previous study 9
, and it is possible that the heterogeneity in the inoculum and the route of infection altered T cell priming and/or their recruitment to the liver. It is therefore an important finding of our study that even HCV infections where viral factors are kept uniform and where the timing of induction and recruitment of HCV-specific T cells do not vary may result in divergent outcomes, i.e. in HCV clearance or persistence. Future studies should therefore address whether host factors that influence antigen presentation and recognition in the HCV-infected liver contribute to the divergent outcomes of HCV infection.
Intriguingly, defects in antigen presentation may also contribute to the late induction of HCV-specific immune responses that we propose in this study (Figure 8). It is known that the effectiveness of antigen-transport to and T cell priming in perihepatic lymph nodes depends on the number and viability of infected cells and the inflammatory environment. Cell death is known to enhance the immunogenicity of cell-associated viral antigens via the crosspriming pathway 27
HCV may neither kill enough hepatocytes to release antigen nor cause sufficient inflammation to induce crosspriming. In addition, a strong natural killer cell response, resulting from the HCV RNA-induced type I IFN reported in this study, may reduce the number of antigen expressing dendritic cells in a perforin-mediated manner 28
. If the dendritic cell-mediated transport of HCV antigens from the liver to the lymph nodes was reduced the priming of effective T cell responses and the appearance of HCV-specific T cells in the circulation and the liver would be delayed as observed in this study. Thus, we propose that the delayed appearance of HCV-specific CD8 cells in the liver during acute HCV infection is due to a delayed induction rather than delayed recruitment of HCV-specific T cells.