Mechanisms contributing to the development of chronic viral infections, including chronic hepatitis B virus (HBV) infections, are not well understood. We have shown recently that production of IFN-γ, an important antiviral cytokine, by HBV-specific CTLs is rapidly induced when they enter the liver of HBV transgenic mice, and then rapidly suppressed, despite the continued presence of Ag. Suppression of IFN-γ production by the CTLs coincides with the up-regulation of programmed cell death (PD)-1, a cell surface signaling molecule known to inhibit T cell function. To determine whether PD-1 plays a role in the functional suppression of IFN-γ secretion by CTLs, we treated HBV transgenic mice with blocking Abs specific for PD ligand (PD-L)1, the most widely expressed PD-1 ligand, and adoptively transferred HBV-specific CTLs. Treatment with anti-PD-L1 Abs resulted in a delay in the suppression of IFN-γ-producing CTLs and a concomitant increase in the absolute number of IFN-γ-producing CTLs in the liver. These results indicate that PD-1:PD-L1 interactions contribute to the suppression of IFN-γ secretion observed following Ag recognition in the liver. Blockade of inhibitory pathways such as PD-1:PD-L1 may reverse viral persistence and chronic infection in cases in which the CTL response is suppressed by this mechanism.
The clearance of many viral infections, including hepatitis B virus (HBV),3 requires the effector functions of CD8+ CTLs (1, 2). These effector functions include secretion of cytokines, such as IFN-γ and TNF-α, as well as cytolytic activity mediated by perforin and granzyme B (3, 4, 5). Because HBV does not infect mice, an HBV transgenic mouse model was developed to enable the study of HBV-specific CTL effector functions in vivo (6). These mice are transgenic for the entire HBV genome, and their hepatocytes replicate and produce virus at levels comparable to those produced in the infected livers of patients with chronic HBV infection. HBV transgenic mice are immunologically tolerant to HBV proteins (7, 8, 9). Therefore, to examine the adaptive immune response in this system, it is necessary to adoptively transfer HBV-specific CTL lines or splenocytes from nontransgenic mice immunized against HBV proteins (4). Using the HBV transgenic mouse as a model to study the effects of Ag recognition upon virus-specific CTLs in the liver, we have shown that adoptively transferred memory CTLs rapidly secrete IFN-γ upon recognition of Ag in the liver (10). Surprisingly, this initial burst of cytokine secretion is quickly down-regulated, and in vitro stimulation of intrahepatic CTLs with cognate peptide does not restore cytokine secretion until after the clearance of HBV Ags from the liver. Importantly, despite the suppression of IFN-γ secretion, the cytotoxic capabilities of the HBV-specific CTLs increase over time. These results suggest that the mechanism or mechanisms responsible for functional suppression are IFN-γ specific. Due to the importance of IFN-γ in the clearance of many viral infections, including other hepatotropic viruses such as hepatitis C virus, determining the cause(s) of IFN-γ suppression may lead to the development of new therapies for the treatment of these diseases.
Programmed cell death (PD)-1 is a receptor that is expressed on a subset of thymocytes and on activated T and B cells (11, 12). PD-1 has two known ligands: PD ligand (PD-L)1 (B7-H1) and PD-L2 (B7-DC). PD-L1 is expressed in a wide variety of tissues and by a number of different cell types, and its expression is up-regulated by IFN-γ (13, 14, 15, 16, 17, 18, 19). The expression of PD-L2 is much more restricted and appears to be limited to a subset of bone marrow-derived cells, including dendritic cells and macrophages (15, 20). Depending on the model system, PD-L1 and PD-L2 were initially reported to have inhibitory or stimulatory effects on T cell responses (13, 14). Recent reports on PD-L1 using knockout mice or in vivo studies with blocking anti-PD-L1 mAbs have been consistent with an inhibitory role for PD-L1 (13, 16, 21, 22, 23, 24, 25, 26). Recent reports on PD-L2 using knockout mice or blocking mAbs still find opposing functions (27, 28). The basis for these conflicting results is not yet understood, but might be explained by a second, yet to be identified, receptor.
PD-1 is up-regulated on HBV-specific CTLs following Ag recognition in the liver (10). Based on the known expression patterns and functions of the two PD-1 ligands, PD-L1 seemed the most likely PD-1 ligand in the liver. To determine whether PD-1:PD-L1 interactions contribute to the inhibition of IFN-γ secretion observed following Ag recognition by HBV-specific CTLs, we treated HBV transgenic mice with blocking mAbs specific for PD-L1 and adoptively transferred HBV-specific CTLs. Our results indicate that PD-1:PD-L1 interactions do contribute to the suppression of IFN-γ production by HBV-specific CTLs.
Materials and Methods
HBV transgenic mouse lineage 1.3.32 (inbred C57BL/6, H-2b) has been previously described (6). HBV transgenic mice replicate HBV at high levels in the liver and express all of the HBV Ags. Lineage 1.3.32 C57BL/6 mice or nontransgenic C57BL/6 mice were crossed with B10.D2 (H-2d) mice to produce F1 hybrids (designated B6D2). In all experiments, HBV transgenic mice were matched for age (7–10 wk), sex (male), and serum hepatitis B e Ag (HBeAg) levels. All experiments were approved by The Scripps Research Institute Animal Care and Use Committee.
Cell lines and culture
MMDH3 cells are an immortalized hepatocyte cell line derived from transgenic mice expressing the Met gene (2930). For treatment with CTL-conditioned medium, cells were plated 24 h before treatment in collagen-coated 60-mm dishes. Medium from HBV-specific CTLs cultured for 3 days in the absence (mock) or presence (treated) of peptide (1 μg/ml) was supplemented with insulin growth factor-II (16 ng/ml), insulin (10 μg/ml; Calbiochem), and epidermal growth factor (55 ng/ml; Calbiochem), and diluted 2-fold in hepatocyte culture medium, then added directly to the HBV-Met cells. In some experiments, isotype control, anti-IFN-γ, or anti-TNF-α neutralizing Abs (4) were added to the peptide-stimulated supernatant 30 min before addition to MMDH3 cells. Twenty-four hours later, cells were harvested by trypsinization, washed, and used for flow cytometry.
Isolation of primary hepatocytes, liver sinusoidal endothelial cells (LSECs), and Kupffer cells
Livers were perfused slowly via the portal vein with 20 ml of warm Solution 1 (HBSS without Ca2+ and Mg2+, 0.5 mM EDTA), followed by 20 ml of warm Solution 2 (HBSS with Ca2+ and Mg2+, 10 mM HEPES (pH 7.4)), followed by ∼50 ml of warm Solution 3 (Solution 2 plus 0.75 mg/ml collagenase D). Following complete digestion of the liver (∼15–20 min), the gall bladder was removed and the liver was carefully excised. Cells were collected from the liver by disrupting the liver capsule and swirling the tissue in a petri dish containing Solution 3. Liver nonparenchymal cells (LNPCs) containing LSECs and Kupffer cells were separated from hepatocytes by centrifuging the cell suspension at 100 × g for 2 min at room temperature. The supernatant containing the LNPCs was collected, and the low speed centrifugation step was repeated until no pellet could be seen. The LNPCs were then centrifuged at 500 × g for 5 min at room temperature, and the cell pellet was resuspended in 3 ml of RPMI 1640, layered onto 5 ml of a 12% Percoll:88% Histopaque 1083 solution (Sigma-Aldrich), and then centrifuged at 1500 × g for 10 min at room temperature. The LNPCs were isolated from the interface and washed twice in RPMI 1640 before use. Hepatocytes in the pellet from the initial centrifugation step were washed in DMEM containing 10% FCS and centrifuged at 100 × g for 2 min at room temperature. This washing step was repeated until the supernatant was no longer cloudy. Hepatocyte viability was routinely >80%.
Peptides, plasmids, and recombinant vaccinia viruses
Immunization of mice and adoptive transfer
B6D2 mice were immunized against the HBV envelope protein using a plasmid DNA prime/vaccinia virus boost regimen exactly as described (7). Greater than 95% of HBV-specific CTLs generated by this procedure are specific for the ENV28 epitope (10). Two weeks after the vaccinia virus boost, mice were sacrificed and their splenocytes were harvested and RBCs were lysed, and 5–8 × 107 splenocytes containing ∼3–4 × 106 ENV28-specific CD8+ T cells were i.v. injected into the tail vein of nontransgenic and HBV transgenic B6D2 recipients. Groups of three to four mice were sacrificed at various time points following transfer, and their livers and spleens were harvested for further analysis.
In vivo blocking of PD-L1
The PD-L1 blocking Ab 10F.9G2 (rat IgG2b) has been previously described (19, 35). HBV transgenic recipients were injected i.p. with 200 μg of 10F.9G2 in 200 μl of PBS every 3 days beginning the day before adoptive transfer. Control mice received similar injections of PBS alone (untreated) or 200 μg of isotype control rat IgG2b Abs (LTF-2; BioExpress). Thus, mice analyzed on days 3 and 5 posttransfer were treated on days −1 and 2.
Lymphomononuclear cell preparation
Splenocytes were prepared by pressing the spleens through a 70-μm cell strainer (BD Biosciences) with the plunger of a 1-ml syringe, followed by two washes in HBSS. Intrahepatic lymphocytes (IHLs) were isolated as previously described (10). Briefly, livers were perfused with 10 ml of PBS via the portal vein to remove circulating lymphocytes, and the liver was minced and pressed through a 70-μm cell strainer. The resulting cells were digested in 10 ml of RPMI 1640 (Mediatech) containing 0.02% (w/v) collagenase (Sigma-Aldrich) and 0.002% (w/v) DNase I (Sigma-Aldrich), for 30 min at 37°C. Cells were washed with RPMI 1640 and then overlaid on a Percoll/Histopaque solution consisting of 12% Percoll (Pharmacia) and 88% Histopaque-1083 (Sigma-Aldrich). Following centrifugation for 10 min at 1500 × g, the IHLs were isolated from the interface and washed twice in HBSS.
Immunofluorescent staining and FACS analysis
Recombinant soluble dimeric H-2Ld:IgG1 fusion protein (BD Pharmingen) was complexed with HBV ENV28 peptide (Ld+CD11b−) and Kupffer cell (CD54+CD11b+) populations. Ex vivo intracellular cytokine staining was performed following a 5-h incubation at 37°C in the presence of brefeldin A and IL-2, without peptide. Cells were acquired on a digital LSRII flow cytometer (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).
A20 cells (unloaded or loaded with 10 μg/ml ENV28 peptide for 1 h at 37°C) were used as targets in a standard 51chromium release assay. Briefly, 5 × 103 51chromium-labeled target cells (100 μl) per well of a 96-well plate were incubated with 5 × 106 donor splenocytes or 1 × 106 recipient IHLs (100 μl) for 5 h at 37°C, followed by measurement of 51chromium release in the supernatant. E:T ratios were determined by the frequency of Ld:ENV28+CD8+ T cells, as measured by flow cytometry. Target cells were incubated with rat IgG2b isotype control mAb (irrelevant mAb (Irr mAb)) or PD-L1 mAb (20 μg/ml) for 1 h on ice, and then added directly to the CTLs without washing. In some experiments, CTLs were incubated in 100 μl of medium containing 100 nM concanamycin A (Sigma-Aldrich), an inhibitor of perforin-mediated cytotoxicity (36), for 2 h at 37°C before addition of target cells (100 μl).
Tissue RNA analyses
Total liver RNAs were analyzed for HBV transcripts by Northern blot exactly as previously described (4, 6). Quantitative analyses were performed using Optiquant phosphor imaging analysis software (Packard Instrument).
The extent of hepatocellular injury was monitored by measuring serum alanine aminotransferase (sALT) activity in the peripheral blood at multiple time points following adoptive transfer, as previously described (6).
PD-L1 is expressed on many cell types in the inflamed liver
Within 1 day following adoptive transfer into HBV transgenic mice, HBV ENV28-specific CTLs begin to express PD-1, a cell surface receptor known to have inhibitory effects upon IFN-γ secretion (Fig. 1⇓) (10). To determine which cell populations in the liver might be able to provide a negative signal to these cells via PD-L1, we isolated primary hepatocytes and LNPCs from the livers of nontransgenic recipients, HBV transgenic recipients, and nonrecipient HBV transgenic mice before adoptive transfer and on days 1, 3, and 5 posttransfer, and examined PD-L1 surface expression. Primary murine hepatocytes do not express PD-L1 in unmanipulated or nontransgenic recipients, but PD-L1 expression was up-regulated on hepatocytes isolated from HBV transgenic recipients on all days examined (Fig. 2⇓A). LNPCs consist of LSECs, which line the hepatic sinusoids, and Kupffer cells, which are liver resident macrophages. Both LSECs and Kupffer cells expressed easily detectable levels of PD-L1 in the absence of CTLs (Fig. 2⇓, B and C), as previously reported (37). Within 24 h following adoptive transfer of ENV28-specific CTLs into HBV transgenic recipients, PD-L1 expression increased significantly on both LSECs and Kupffer cells, but was unchanged in nontransgenic mice receiving CTLs (Fig. 2⇓, B and C). Additional examination of other cell populations in the liver of HBV transgenic recipients indicated that PD-L1 was also expressed on B220+, CD8+, CD11b+, and CD11c+ cells in unmanipulated mice, and was increased on CD8+, CD11b+, and CD11c+ cells in HBV transgenic recipients (Fig. 2⇓, D–G).
IFN-γ induces PD-L1 expression on hepatocytes
IFN-γ mediates up-regulation of PD-L1 on a variety of cell types (19, 20). To determine whether IFN-γ produced by CTLs is responsible for the up-regulation of PD-L1 on hepatocytes, we treated MMDH3 cells, an immortalized murine hepatocyte cell line, with culture medium conditioned by peptide-stimulated HBV-specific CTLs. Following 24 h of culture, MMDH3 cells significantly up-regulated surface expression of PD-L1 compared with MMDH3 cells cultured with medium from unstimulated CTLs (Fig. 3⇓A). Addition of neutralizing Abs against IFN-γ and TNF-α to the peptide-stimulated supernatant indicates that IFN-γ alone is responsible for PD-L1 up-regulation (Fig. 3⇓B). These results are consistent with those recently published regarding expression of PD-L1 on human hepatocytes (38).
PD-L1 blockade delays suppression of IFN-γ production and increases the total number of IFN-γ-producing CTLs
To determine whether signaling events through PD-1 might be responsible for the observed suppression of IFN-γ secretion by ENV28-specific CTLs, we adoptively transferred ENV28-specific memory cells into HBV transgenic mice that were treated with PD-L1 blocking Abs or with PBS or isotype control IgG (Irr mAb) every 3 days beginning the day before transfer. Mice were sacrificed on days 3 and 5 posttransfer, and IHLs were isolated and analyzed for IFN-γ production directly ex vivo, without additional peptide stimulation. Only IHLs were analyzed because HBV-specific CTLs are not detectable in the blood, spleen, and lymph nodes at these time points (10). This experiment was performed on three separate occasions with three to four mice per group, and the results from all experiments are shown in Fig. 4⇓. Treatment of recipients with PD-L1 blocking Abs resulted in a statistically significant (p ≤ 0.05) 1.6- to 2.5-fold increase in the frequency of ENV28-specific CTLs producing IFN-γ ex vivo on day 3 after transfer (Fig. 4⇓A), whereas the total number of ENV28-specific CTLs in the liver was identical (Fig. 4⇓B). However, by day 5 following transfer, the frequencies of IFN-γ-producing CTLs in both control and PD-L1 mAb-treated recipients decreased substantially (Fig. 4⇓A). Interestingly, the total number of ENV28-specific CTLs in the PD-L1 mAb-treated recipients on day 5 was approximately twice that of the control recipients (Fig. 4⇓B), resulting in the presence of 1.7- to 4.3-fold more IFN-γ+ ENV28-specific CTLs in the livers of PD-L1 mAb-treated mice, depending on the experiment (Fig. 4⇓C). Importantly, PD-L1 mAb-treated mice contained equivalent or fewer HBV transcripts compared with untreated mice (Fig. 4⇓D), suggesting that differences in Ag levels are not responsible for the observed effects of PD-L1 mAb on the frequency of IFN-γ-producing CTLs. These results suggest that interfering with PD1:PD-L1 interactions delays the suppression of IFN-γ production after Ag recognition, but is unable to completely prevent it. However, the effect of PD-L1 blockade upon CTL numbers resulted in a net increase in the number of IFN-γ+ CTLs on day 5, effectively maintaining an increased production of IFN-γ in the liver.
The increase in IFN-γ production was accompanied by a 40–60% increase in the total number of inflammatory cells present in the livers of PD-L1 mAb-treated recipients on day 5 (Fig. 4⇑E). This is consistent with previous results from our laboratory showing that secretion of IFN-γ in the liver by CTLs results in the production of the chemokines IFN-γ-inducible protein-10 (IP-10) and monokine induced by IFN-γ (MIG) which in turn promote the recruitment of inflammatory cells into the liver (39). sALT activity in PD-L1 mAb-treated mice was also ∼2-fold higher than untreated controls (Fig. 4⇑F), indicating increased liver damage.
PD-L1 blockade has no effect on cytolytic effector function
To date, the literature regarding the effects of PD-L1:PD-1 interactions on the cytolytic capability of CTLs has been contradictory, with at least one report suggesting that PD-L1 inhibits cytotoxicity (19), and at least one report suggesting PD-L1 has no effect (40). To determine whether the PD-L1 mAb-induced increase in sALT activity levels could be due to enhanced ENV28-specific CTL cytotoxicity, we performed cytotoxicity assays using IHLs isolated from Irr mAb-treated or PD-L1 mAb-treated mice as effector cells. Donor splenocytes were also tested for comparison. The E:T ratios were calculated based on the number of Ld:ENV28 dimer-positive cells (measured by flow cytometry) added to each well. Because A20 target cells express high levels of PD-L1, we preincubated them with Irr mAb or PD-L1 mAb and then added them directly to effector cells. As expected, an increase in cytotoxicity (Fig. 5⇓, A, B, and D) and granzyme B staining (Fig. 5⇓C) was observed in ENV28-specific CTLs following transfer into transgenic recipients. No differences in specific cytotoxicity or granzyme B expression were observed between CTLs isolated from Irr mAb and PD-L1 mAb-treated recipients. Additionally, preincubation of target cells with PD-L1 mAb had no effect. To determine whether PD-L1 affects Fas ligand-mediated cytolysis, effector cells were preincubated with concanamycin A, a powerful inhibitor of perforin-mediated cytolysis. Whereas cytotoxicity was greatly decreased, there were no differences between IHLs isolated from Irr mAb or PD-L1 mAb-treated mice (Fig. 5⇓D). Pretreatment of the target cells with PD-L1 mAb also had no effect, indicating that PD-L1 does not affect the perforin or Fas ligand-mediated cytolytic capabilities of effector CD8+ T cells in this system. Instead, the increase in sALT activity is most likely due to the large number of ENV28-specific CTLs and total inflammatory cells in the liver on day 5 in PD-L1 mAb-treated recipients.
We have shown previously that HBV ENV28-specific CTLs responding to Ag in the liver of HBV transgenic mice immediately secrete IFN-γ, and then over the course of the next 3–5 days become suppressed in their ability to produce IFN-γ, despite the continued presence of Ag (10). In this study, we provide evidence that one factor contributing to this suppression is the interaction of the inhibitory PD-1 receptor on CTLs with its ligand PD-L1. Treatment of HBV transgenic recipients with PD-L1 blocking Abs delayed the suppression of IFN-γ production, as evidenced by a ∼2-fold increase in the frequency of ENV28-specific CTLs able to produce IFN-γ directly ex vivo on day 3 after transfer. Importantly, this delay resulted in a ∼2- to 4-fold increase in the total number of IFN-γ-producing HBV-specific CTLs in the liver on days 3 and 5 posttransfer. This is the first report we are aware of indicating that prevention of PD-L1:PD-1 interactions increases the fraction of Ag-specific CTLs able to produce IFN-γ in response to physiological concentrations of peptide:MHC complexes. These results are particularly interesting in light of a recent report suggesting that IFN-γ acts directly on CD8+ CTLs to increase their abundance during viral infection (41). Thus, it is possible that the effect of PD-L1 blockade upon IFN-γ production by CTLs indirectly enhances CTL accumulation at the site of Ag recognition, thereby determining the magnitude and duration of the inflammatory response with a corresponding increase in tissue destruction and, potentially, the control of infection. This interpretation is consistent with a report by Barber et al. (24) showing increased accumulation of lymphocytic choriomeningitis virus (LCMV)-specific T cells and immunopathology in LCMV-infected PD-L1 knockout mice.
Similar to a natural infection, HBV replicates and produces progeny virus in the hepatocytes of HBV transgenic mice, but not in the nonparenchymal cells in the liver (6). ENV28-specific CTLs home to the liver immediately following adoptive transfer (10), suggesting that hepatocytes are responsible for the presentation of Ag. Interestingly, hepatocytes do not normally express detectable levels of cell surface PD-L1. However, PD-L1 expression is significantly up-regulated within 24 h on hepatocytes and LNPCs following adoptive transfer of CTLs specific for Ags produced by hepatocytes, and this up-regulation is dependent on IFN-γ. Thus, it is possible that the hepatocytes themselves are providing the negative signal. However, many other cell populations in the liver also express PD-L1 following CTL adoptive transfer; therefore, we cannot definitively say whether any one specific cell type is responsible. Interestingly, PD-L1 expression on hepatocytes appeared to reach a peak on day 3 and was subsiding on day 5 (Fig. 2⇑A). Similarly, PD-L1 expression on B220+, CD8+, CD11b+, and CD11c+ populations appeared to peak on day 1 (Fig. 2⇑, D–G). In contrast, LSECs and Kupffer cells maintained higher levels of PD-L1 expression at all time points examined (Fig. 2⇑, B and C). More experiments will be necessary to determine the mechanisms behind this, but based on our results and those of others (19, 20), it seems likely that sensitivity to IFN-γ regulates PD-L1 surface expression.
One of the characteristics of adoptively transferred ENV28-specific CTLs is that their cytolytic capabilities and granzyme B expression increase as their ability to produce IFN-γ decreases (10). We noted a 2-fold increase in the sALT activity of HBV transgenic recipients treated with PD-L1 mAb, indicating an increase in the destruction of hepatocytes in the liver. Further examination of the intrahepatic ENV28-specific CTLs isolated from untreated and PD-L1 mAb-treated mice revealed that they had identical cytolytic capabilities and granzyme B expression. In addition, treatment of target cells with PD-L1 mAb had no effect on cytolytic function. Together these results show that PD-1:PD-L1 signaling has no effect on the development or execution of the cytolytic function of CTLs in this system.
There have been several reports suggesting that PD-1 and PD-L1 inhibit the immune response during viral infections. Iwai et al. (37) infected wild-type and PD-1 knockout mice with adenovirus and showed that clearance of adenovirus from the liver was achieved much more quickly in PD-1 knockout animals. This was attributed to increased proliferation of CD4+ and CD8+ T cells in the absence of PD-1 signaling, thereby increasing the effector:infected target ratio. More recently, Barber et al. (24) showed that viral titers in mice chronically infected with the clone 13 strain of LCMV were undetectable in the spleen and liver and significantly reduced in the lungs and kidneys following 2 wk of treatment with the same PD-L1 mAb that we used in our experiments. Examination of the LCMV-specific CTL response in the spleens of PD-L1 mAb-treated mice compared with untreated mice revealed increases in the total numbers of LCMV-specific CTLs, as well as in the frequency of CTLs able to produce cytokines upon in vitro stimulation with peptide. Unfortunately, cytokine production in PD-L1 mAb-treated animals was not monitored before clearance of virus from the spleen, so it is difficult to determine whether blocking of PD-L1 directly resulted in increased frequencies of cytokine-producing cells, or whether the observed increases reflected the clearance of virus from the spleens of PD-L1 mAb-treated mice, such that the CTLs were rested and thus more able to produce cytokines upon restimulation with peptide. In support of the latter explanation, it has been reported previously that CD8 T cells cannot reinitiate cytokine production immediately following Ag removal (42). In contrast, the CTLs in our study, which were examined directly ex vivo without the addition of exogenous, nonphysiological concentrations of peptide, demonstrated delayed suppression of IFN-γ production and increased frequencies of IFN-γ-producing cells in mice treated with PD-L1 mAb compared with untreated controls. These results suggest that, in addition to proliferative and/or survival effects, PD-L1 blockade also increases the ability of CTLs to produce IFN-γ in vivo, even in the presence of lower levels of Ag. Perhaps PD-1:PD-L1 signaling does not suppress the maximum IFN-γ-producing capacity of CTLs, but rather it increases the signaling threshold for IFN-γ production. This hypothesis is supported by evidence that PD-1 down-modulates TCR signaling through the inhibition of both PI3K activation and Zap70 phosphorylation (43, 44).
It is noteworthy that PD-L1 blockade appears to only delay the suppression of IFN-γ production by HBV-specific CTLs, because the frequencies of IFN-γ-producing cells were decreased in both untreated and PD-L1 mAb-treated mice on day 5 following transfer compared with day 3. This suggests that one or more additional mechanisms are responsible for the suppression of IFN-γ production by HBV-specific CTLs in the liver.
We thank Drs. Arlene Sharpe, Luca Guidotti, and Jared Purton for helpful comments and suggestions, and Alana Althage and Christina Whitten for excellent technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by National Institutes of Health Grants R01 CA40489-21 (to F.V.C.) and P01 AI56299 and a Gates Grand Challenges in Global Health award (to G.J.F.). H.M. was supported by National Institutes of Health Grant T32 AI07244-22 and a Cancer Research Institute postdoctoral fellowship.
↵2 Address correspondence and reprint requests to Dr. Francis V. Chisari, Department of Molecular and Experimental Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, SBR10, La Jolla, CA 92037. E-mail address:
↵3 Abbreviations used in this paper: HBV, hepatitis B virus; IHL, intrahepatic lymphocyte; Irr mAb, irrelevant mAb; LCMV, lymphocytic choriomeningitis virus; LNPC, liver nonparenchymal cell; LSEC, liver sinusoidal endothelial cell; PD, programmed cell death; PD-L, PD ligand; sALT, serum alanine aminotransferase.
- Received June 8, 2006.
- Accepted December 12, 2006.
- Copyright © 2007 by The American Association of Immunologists