Reduced CD8+ T Cell Responses Are Generated by Intravenous Administration of Adenovirus
Secondary lymphoid tissues, e.g. lymph nodes and spleen, are generally considered to be the major sites of activation, proliferation, and differentiation of naïve antigen-specific CD8+ T cells into effector cells. However, the liver has also been implicated as a potential site of extralymphoid induction of CD8+ T cell responses to alloantigens and hepatotropic infectious agents including viruses. Recombinant adenoviruses (rAd) have been demonstrated to direct expression of foreign antigens to the liver when rAd is administered by the intravenous (IV) route. To examine whether intrahepatic expression of foreign antigen delivered by rAd vector would result in the activation of antigen specific CD8+ T cells and the subsequent development of CD8+ effector T cells, we infected C57BL/6 mice with 5×108 PFU of rAd expressing the gene encoding ovalbumin (OVA) protein (Ad-OVA) by IV route. We chose Ad-OVA to probe the CD8+ T cell response to the liver because this protein induces a robust CD8+ T cell response in this mouse strain and OVA antigen-specific T cells from TCR transgenic mouse lines are available for more detailed analysis on antigen-specific CD8+ T cell responses. As a control, we introduced the same inoculum dose of Ad-OVA by subcutaneous (SubQ) route in order to initiate CD8+ T cell response through secondary lymphoid tissue presentation within the lymph nodes draining the site of virus infection.
SubQ inoculation of virus resulted in the accumulation of a significant number of OVA-specific tet+ CD8+ T cells in the liver of inoculated mice at days 7 and 14 p.i. (). Approximately 50% of OVA-specific tet+ CD8+ T cells in the liver produce IFN-γ in response to specific peptide stimulation in vitro (). As demonstrated below, the accumulation of these CD8+ effector T cells in the liver of mice with SubQ infection reflects the activation/proliferation and differentiation of naïve antigen specific CD8+ T cells in the lymph nodes draining the site of virus infection with subsequent migration of the activated cells into the liver at later times. In contrast, the IV administration of virus resulted in a markedly diminished OVA specific CD8+ T cell response in the liver as detected by tetramer staining () with a likewise markedly diminished frequency of CD8+ T cells responding to specific peptide stimulation in vitro with IFN-γ production (). This diminished response to IV adenovirus delivery was not an exclusive property of the Ad-OVA. When we carried out a corresponding comparison of the impact of IV versus SubQ inoculation on the subsequent CD8+ T cell response in the liver using rAd expressing β-galactosidase (Ad-LacZ), we observed a similar diminished CD8+ T cell response to IV administration of Ad-LacZ at days 7 and 14 p.i. as measured by both tetramer staining and IFN-γ production ().
IV adenovirus administration results in the diminished CD8+ T cell responses in the liver.
The finding of the reduced CD8+
T cell response to IV virus administration was not expected and suggested that there may be a diminished induction of antigen specific anti-viral CD8+
T cell responses and/or a defective response of effector T cells generated following IV virus administration when antigen expression is targeted to the liver. It should be noted, however, that IV administration of rAd has been reported to result in immunologic tolerance although the observation is made by administration of rAd virus at 10-100 fold higher doses than that employed in our studies 
. To determine if the diminished CD8+
T cell response observed reflected a high tolerogenic dose of virus delivered by the IV route, we analyzed the effect for various doses of rAd virus on the generation of CD8+
T cell responses in the liver following IV virus inoculation. As demonstrates, the decreased dose of virus inoculum administered by the IV route only served to reduce the overall magnitude of CD8+
T cell response detected in the liver by either tetramer staining or IFN-γ production.
We believe, therefore, that under the experimental conditions employed here, the diminished CD8+
T cell response observed in the liver after IV virus administration was not easily attributable to high dose tolerance based on the size of the inoculum employed. Rather the results (along with findings presented below) suggest that IV administration of this hepatotropic virus results in direct priming of CD8+
T cells in the liver with consequent alterations in CD8+
T cell activation and/or function. To further test the impact of inoculum dose and route of virus administration on the subsequent CD8+
T cell response, we inoculated mice with 5×108
PFU of rAd by the intranasal (IN) route and evaluated the development of the CD8+
T cell response in the lungs and liver thereafter. As reported by others for virus infection in the respiratory tract 
, virus administration at this site results in induction of CD8+
T cell responses exclusively in the lymph nodes draining the respiratory tract. The IN administration of virus which, like SubQ virus administration, restricts virus to the lymph nodes draining the site of inoculation, results in a vigorous CD8+
T cell response in the liver () and lungs (data not shown).
Induction of CD8+ T Cell Responses Can Occur in the Liver
The IV administration of virus will not only deposit virus in the liver but antigen can be detectable in other sites including the secondary lymphoid tissues by migration of dendritic cells taken up antigen. It was therefore important to demonstrate whether virus deposition in the liver resulted in CD8+ T cell activation in that site and whether altered CD8+ T cell responses in the liver could account for the observed impaired responses observed. One potential explanation for the finding of diminished CD8+ T cell responses in the liver following IV virus inoculation is that there might be a defect in the activation and/or proliferation of naïve CD8+ T cells responding directly in the liver. In order to determine if the liver could serve as a site of naïve CD8+ T cell activation following IV virus administration in addition to secondary lymphoid organs, we employed an adoptive transfer strategy using OT-1 CD8+ TCR transgenic T cells directed to the OVA protein epitope displayed by the Ad-OVA virus. These Thy1.1+ naïve CD8+ TCR transgenic T cells were labeled with the dilution sensitive dye CFSE and then adoptively transferred into congenic Thy1.2+ recipients. Upon 24 hours following transfer, the recipient mice received Ad-OVA virus by SubQ or IV route. Activation and proliferation of the OT-1 T cells in the liver, spleen, and lymph nodes was monitored by CSFE dilution and T cell activation marker expression at 36 and 48 hrs post virus inoculation.
As demonstrates, following SubQ virus administration, proliferation of the adoptively transferred OT-1 T cells was restricted primarily to the inguinal lymph nodes draining at the site of SubQ virus administration. As expected the activated proliferating T cells upregulated expression of CD25. The onset of T cell proliferation occurred between 36 and 48 hrs after virus inoculation. This time frame for proliferation presumably reflects the time required for uptake of virus by tissue antigen presenting cells (APC) and subsequent migration of the APC to the draining inguinal nodes for antigen presentation to naïve T cells. In addition, there was minimal proliferation of T cells present in the liver, spleen or non-draining lymph nodes over this 48 hrs time frame. Thus following SubQ virus administration, viral antigen presentation was restricted to the draining lymph nodes at the site of virus inoculation. We confirmed this conclusion in companion experiments where mononuclear cells were isolated from the liver, spleen, and draining and non-draining lymph nodes at 48 hours following SubQ Ad-OVA administration. These leukocytes were then co-cultured in vitro for 3 days with CSFE-labeled naïve OT-1 T cells. When cultured T cells were evaluated for activation/proliferation, only mononuclear cells isolated from the draining inguinal lymph nodes were able to trigger T cell proliferation in vitro ().
IV adenovirus administration leads to systemic CD8+ T cell proliferation.
When virus was delivered by IV route, by contrast, proliferating T cells were detected in the liver, spleen and lymph node compartments at 36 hours p.i. and proliferation proceeded over the liver and the secondary lymphoid tissues with similar kinetics over the succeeding 12 hours (). This finding suggested that after IV inoculation, virus was distributed to both the liver and secondary lymphoid tissues. In keeping with this suggestion, we found that mononuclear cells isolated from both liver and secondary lymphoid organs could support the in vitro proliferation of naïve OT-1 T cells (). The in vivo analysis of the T cell response to IV virus inoculation also revealed that while the activation marker CD25 was significantly upregulated in T cell responses within the secondary lymphoid organs, CD25 expression was blunted/suppressed in OT-1 T cells responding in the liver (). Minor CD25+CD8+ T cell population appears at 48 hour possibly due to the migration of CD8+ T cells generated in the secondary lymphoid organs into the liver.
The above kinetic analysis suggested that following IV virus administration, the liver could support the initial activation and early proliferation of naïve antigen specific CD8+ T cells. However, we could not exclude the possibility that following IV virus inoculation, CD8+ T cells were activated in the secondary lymphoid organs, i.e. spleen and lymph nodes, and subsequently migrated to the liver. To further explore these alternatives, we first examined transferred OT-1 T cells resident in the liver for the expression of the early activation marker CD69 following IV or SubQ administration of Ad-OVA virus. As demonstrates, upregulation of CD69 could be detected in the liver resident TCR transgenic CD8+ T cells as early as 4 hours following IV virus administration. As expected, minimal upregulation of CD69 was observed in liver resident OT-1 T cells up to 24 hours following SubQ virus inoculation. Importantly, non-transgenic antigen-nonspecific Thy1.2+ recipient endogenous CD8+ T cells present in the liver demonstrated minimal upregulation of CD69, suggesting that the upregulation of this molecule in the TCR transgenic T cells was antigen driven and was not due to nonspecific inflammatory stimuli in the liver associated with IV virus administration.
The liver supports CD8+ T cell priming with rapid activation of antigen-specific CD8+ T cells.
To further elucidate the liver as the site for priming of naïve CD8+ T cells, we examined the impact of splenectomy on the response of liver-resident OT-1 T cells to IV virus administration in the adoptive transfer model. To further ensure that CD8+ T cells activated within the lymph node compartments were not contributing to the pool of proliferating transgenic T cells in the liver of splenectomized mice, this analysis was carried out in splenectomized and sham splenectomized mice which received monoclonal antibody (Mel-14) to L selectin (CD62L) one day prior to adoptive transfer. This Mel-14 antibody inhibits the migration of circulating naïve T cells into the lymph nodes and prevents T cell activation there. As shown in , the early proliferative response of naïve CD8+ T cells to viral antigen in the liver was unaffected by splenectomy, suggesting that priming of naïve CD8+ T cells can occur in the liver.
CD8+ T Cells Activated in the Liver Lead to Defective Differentiation of Effector T Cells
The above findings suggested that naïve CD8+ T cells activated in the liver are capable of undergoing initial proliferation at this site. However, the expression of at least one marker of T cell activation, i.e. CD25, appeared to be defective in naive T cells activated within the liver environment (, ). To further explore the impact of T cell induction within the liver environment on CD8+ T cell differentiation, we examined the expression of two important CD8+ T cell effector molecules, i.e. IFN-γ and granzyme B, in transferred OT-1 T cells in the liver, spleen and inguinal lymph nodes responding to SubQ or IV Ad-OVA virus administration. As demonstrates, activation/proliferation of naïve T cells following SubQ virus administration was restricted to the inguinal nodes and the proliferating T cells were fully capable of expressing IFN-γ and granzyme B. Following IV virus administration, IFN-γ and granzyme B production were detected in responding T cells stimulated in the spleen and lymph nodes (). By contrast, naïve T cells activated by virus within the liver failed to upregulate expression of either of these T cell effector molecules (). When the influence of splenic and lymph nodes derived CD8+ T cells are removed from these experiments using splenectomized and Mel-14 Ab treated mice, we similarly observe that liver primed CD8+ T cells do not differentiate into potent effectors (). In companion studies, we evaluated the expression of several other cell surface activation markers. We found that, in contrast to CD8+ T cells activated in the spleen or lymph nodes, the majority of naïve CD8+ T cells activated within the liver failed to down regulate CD62L (data not shown).
Liver primed CD8+ T cells do not differentiate into competent effectors.
As demonstrated above (), viral antigen administration by IV route results in a diminished number of antigen-specific tet+ CD8+ T cells at days 7 and 14 p.i. compared to SubQ virus inoculation where there is a substantial accumulation of antigen-specific tet+ CD8+ T cells in the liver. While this discrepancy would most easily argue for a defect in the activation/proliferation of responding CD8+ T cells following IV virus administration, our observations on the early response of CD8+ T cells in the liver (e.g. and ) do not suggest such a defect. Rather the data suggest that IV virus administration and liver priming likely result in a defect in CD8+ T cell differentiation into effector cells. To further examine the basis for this potential discrepancy, we analyzed the response of Thy1.1+ transgenic CD8+ T cells transferred into congenic Thy1.2+ recipients at day 7 after administration of virus by SubQ or IV route. Consistent with our earlier findings in , a large fraction of transgenic Thy1.1+CD8+ T cells capable of responding to antigenic stimulation with IFN-γ production accumulated in the liver on day 7 following SubQ virus delivery. More importantly, a corresponding large number of Thy1.1+ transgenic T cells accumulated in the liver following IV virus delivery but these T cells, like activated CD8+ T cells analyzed at early time points after antigenic stimulation in the liver (, ), failed to produce IFN-γ (). In addition, the observed CD8+ T cell dysfunction was independent of viral doses administered into mice (). It suggests that the impaired CD8+ T cell function as seen in the IV virus delivery is not due to the high dose of viral antigen administered into the liver.
IV adenovirus infection causes a robust defect in CD8+ T cell effector function with elevated PD-1 expression in the liver.
We next examined whether CD8+
T cell dysfunction is associated with the upregulation of PD-1 expression. Functionally exhausted CD8+
T cell populations and CD8+
T cell death in the liver impair viral clearance during persistent infections including HCV and HBV 
. The PD-1/B7-H1 inhibitory pathway has been demonstrated to play an important role in the regulation of anti-viral immune responses in the liver 
. To this end, Thy1.1+
T cells were adoptively transferred into naïve Thy1.2+
C57BL/6 mice that were then infected with Ad-OVA via either SubQ or IV immunization route. At day 7 post IV infection, the majority of OVA-specific CD8+
T cells in the liver express PD-1 while PD-1 is not expressed by OVA-specific CD8+
T cells that migrate to the liver following SubQ immunization (). In addition, antigen-specific and bulk intrahepatic CD8+
T cells generated during IV adenovirus infection express elevated levels of active caspase-3, which serves as an indicator of apoptosis while the liver can support viable antigen-specific CD8+
T cells when virus is delivered to other peripheral sites (data not shown). However, treatment of anti-B7-H1 Ab prior to adenovirus infection did not restore the IFN-γ production by CD8+
T cells in the liver although it slightly increased the number of IFN-γ+
T cells in the spleen compared to mice treated with control Ab (). It suggests that the PD-1 negative costimulatory pathway might not be directly involved in the induction of CD8+
T cells with impaired function in the liver.
Antigen Dependence of CD8+ T Cell Dysfunction in the Liver Environment
Our observations to this point strongly suggest that the liver is a major site of naïve CD8+ T cell activation when hepatotropic antigen such as rAd is administered by IV route. Naïve CD8+ T cell priming in the liver results in the alteration of CD8+ T cell differentiation and expression of effector activities. However, activated CD8+ T cells, which accumulate in the liver following SubQ virus administration, do not exhibit a dysfunctional phenotype. Furthermore, IV virus delivery does result in the activation of CD8+ T cells within secondary lymphoid organs and the activated CD8+ T cells, which remain in that site, appear to be functionally normal (). Yet if these CD8+ T cells within secondary lymphoid tissues, which have been activated in response to IV virus administration and then migrate to the liver, they would appear to also acquire a dysfunctional phenotype.
We therefore wanted to determine if, in the liver environment, previously activated functional effector CD8+ T cells can be rendered dysfunctional and whether this process was dependent on specific antigen recognition within the liver. To this end, we transferred naïve Thy1.1+ OT-1 CD8+ T cells into congenic Thy1.2+ recipient animals and infected the recipient animals with Ad-OVA virus by SubQ route. Four days after infection, we isolated Thy1.1+ T cells from the secondary lymphoid tissues of the recipient mice and then transferred these activated CD8+ T cells into Thy1.2+ recipient mice, which had been infected with either Ad-OVA or Ad-LacZ by IV route (). By employing this strategy, we determined whether the effector response of fully activated effector CD8+ T cells could be suppressed within the liver environment and whether specific antigen recognition (mediated by IV Ad-OVA administration) or simply acute liver inflammation (orchestrated by IV Ad-LacZ administration) was required to suppress the activated effector CD8+ T cells. Activated Thy1.1+ OT-1 CD8+ T cells adoptively transferred into uninfected recipient mice served as a background control. Upon 24 hours after cell transfer, the Thy1.1+ OT-1 CD8+ T cells were isolated from the livers of the recipient mice and were analyzed for specific tetramer staining directly ex vivo and for the expression of IFN-γ and TNF-α in response to specific antigen stimulation by the in vitro intracellular cytokine assay.
TCR engagement is required for impairment of CD8+ T cell effector function in the liver.
The OT-1 T cells isolated from the control (uninfected) livers and from the inflamed livers of mice infected IV with Ad-LacZ virus uniformly bound specific tetramer and expressed IFN-γ and TNF-α in response to peptide stimulation. In contrast, T cells isolated from the livers of mice with the expression of specific antigen by Ad-OVA infection exhibited reduced tetramer staining and defective expression of proinflammatory cytokines compared to the inflammation only and control groups (). These results indicate that the liver environment per se or even the inflamed liver environment resulting from virus infection is not sufficient to induce an abortive CD8+ T cell response. Rather, the development of this defective response requires specific antigen recognition within the liver.
Antigen Presentation by Liver Parenchyma Cells Induces Suboptimal Differentiation of CD8+ T Cells
Our observations to this point suggest that the liver can serve as a site for the induction of CD8+
T cell responses outside the secondary lymphoid tissues. However, naïve T cell activation by antigen within the liver or the encounter of previously activated CD8+
T cells with antigen displayed in the liver results in the dysregulation of CD8+
T cell effector function. Although the liver contains a variety of CD45+
cell types which could serve as APC for naïve or activated antigen specific CD8+
T cells, hepatotropic agents like rAd virus may preferentially infect and express antigen in hepatocytes 
. This consideration prompted us to inquire whether T cell activation in the liver leading to dysregulation of effector function of the activated CD8+
T cells may be attributed to the interaction of T cells with antigen displayed on liver parenchymal cells. To explore this possibility, we constructed bone marrow chimeras in which lethally irradiated C57BL/6 mice were reconstituted with allogeneic BALB/c bone marrow. Following reconstitution, these animals contain hematopoietic cells expressing H-2d
haplotype MHC class I molecules and liver parenchyma expressing the H-2b
haplotype MHC class I molecules. Reconstitution of bone marrow chimera was verified by flow cytometry analysis: the intrahepatic CD11b+
cells isolated from Balb/c→B6 chimeras express high levels (>95%) of H2-Kd
, and only a small fraction (~4%) express H2-Kb
at levels that are slightly above background. In these mice, only liver parenchyma cells should be able to present OVA257-264
epitope to H2-Kb
T cells during Ad-OVA infection. In contrast, the bone marrow derived APCs in the Balb/c→B6 mice chimera mice should express only H2-Kd
class I molecules and be unable to present cognate OVA antigen to OT-1+
T cells during Ad-OVA infection.
To examine the ability of liver parenchyma cells to present antigen and induce CD8+ T cell proliferation, CFSE labeled Thy1.1+OT-1+CD8+ T cells were adoptively transferred into Balb/c→B6 and control B6→B6 mice that were then infected with Ad-OVA by IV immunization route. At 2 days p.i., Thy1.1+OT-1+CD8+ T cells had proliferated to similar levels in both Balb/c→B6 and control B6→B6 mice (). Thus liver parenchyma cells presenting foreign antigen are able to induce the proliferation of naïve CD8+ T cells. Furthermore, CD8+ T cells primed by the antigen-presenting liver parenchyma cells fail to differentiate into competent effectors. The dividing T cells do not acquire the ability to produce IFN-γ or GrB following antigen recognition on liver parenchyma cells (). This suggests that antigen presentation by liver parenchyma cells results in incomplete differentiation of CD8+ T cells.
Antigen presentation by liver parenchyma cells induces suboptimal differentiation of CD8+ T cells.