Antiviral Intrahepatic T-Cell Responses Can Be Restored by Blocking Programmed Death-1 Pathway in Chronic Hepatitis B

Transcription

1 GASTROENTEROLOGY 2010;138: Antiviral Intrahepatic T-Cell Responses Can Be Restored by Blocking Programmed Death-1 Pathway in Chronic Hepatitis B PAOLA FISICARO,* CATERINA VALDATTA,* MARCO MASSARI, ELISABETTA LOGGI, ELISABETTA BIASINI,* LUCA SACCHELLI,* MARIA CRISTINA CAVALLO,* ENRICO M. SILINI, PIETRO ANDREONE, GABRIELE MISSALE,* and CARLO FERRARI* *Laboratory of Viral Immunopathology, Unit of Infectious Diseases and Hepatology, Azienda Ospedaliero-Universitaria di Parma, Parma; Unit of Infectious Diseases, Azienda Ospedaliera S.M.N. di Reggio Emilia; Reggio Emilia; Department of Clinical Medicine, Laboratory of Cellular Immunology, University of Bologna, Bologna; and Anatomia e Istologia Patologica, Dipartimento di Patologia e Medicina di Laboratorio, Azienda Ospedaliero-Universitaria di Parma, Parma, Italy BACKGROUND & AIMS: The antiviral function of peripheral hepatitis B virus (HBV)-specific T cells can be increased in patients with chronic hepatitis B by blocking the interaction of programmed death (PD)-1 with its ligand PD-L1. However, no information is available about the effects of this blockade on intrahepatic lymphocytes. We studied T-cell exhaustion and the effects of PD-1/PD-L1 blockade on intrahepatic and circulating HBV-specific T cells in patients with chronic hepatitis B. METHODS: A total of 42 patients with chronic HBV infection who underwent liver biopsy were studied. The ex vivo phenotype of peripheral and intrahepatic HBVspecific CD8 T cells was assessed by flow cytometry with class I tetramers and antibodies to T-cell differentiation molecules. Functional recovery was evaluated by analyzing expansion and production of interferon (IFN)- and interleukin (IL)-2 after short-term incubation of T cells with HBV peptides in the presence of anti-pd-l1 or control antibodies. RESULTS: Intrahepatic HBV-specific CD8 cells expressed higher levels of PD-1 and lower levels of CD127 than their peripheral counterparts. Blockade of PD-1/PD-L1 interaction increased CD8 cell proliferation and IFN- and IL-2 production by circulating intrahepatic lymphocytes, even though anti-pd-l1 had a stronger effect on intrahepatic compared with peripheral T cells. CONCLUSIONS: T-cell exhaustion by high antigen concentrations promotes HBV-specific T-cell dysfunction by affecting phenotype and function of peripheral and intrahepatic T cells. By restoring antiviral T-cell functions, not only in peripheral but also in intrahepatic lymphocytes, anti-pd-l1 might be a good therapeutic candidate for chronic HBV infection. Hepatitis B virus (HBV) causes self-limited infections in the majority of adult patients; however, in less than 5% of infected adults and in a larger proportion of children, HBV infection evolves into a chronic liver disease, which may lead to severe sequelae. Available data indicate a vigorous and broad T-cell response in patients with acute self-limited hepatitis B, which is in contrast with the much weaker and often barely detectable T-cell response observed during chronic HBV infections, in which HBV-specific T-lymphocytes have been described to be deeply dysfunctional. 1 3 Among the different mechanisms proposed to explain such a functional T-cell impairment, exhaustion by persistent exposure to high antigen concentrations represents a possible cause, 4 6 and T-cell attrition through Bcl2 signalling 7 as well as impaired T-cell receptor signaling through the -chain 3 have been shown to be involved. Recent studies in animal models of virus infection, both in mice and in macaques, indicate that the interaction between the inhibitory receptor programmed death (PD)-1 on lymphocytes and its ligands plays a critical role in T-cell exhaustion by inducing T-cell inactivation, and high PD-1 expression levels on peripheral T-lymphocytes as well as the possibility to improve the T-cell function by blocking PD-1-mediated signaling to confirm the importance of this inhibitory pathway in inducing T-cell exhaustion Moreover, manipulation in vivo of costimulatory pathways to restore the antiviral function of exhausted T cells has successfully been applied in mice persistently infected with lymphocytic choriomeningitis virus to improve therapeutic vaccination. 12 Also, in various human chronic infections, including hepatitis B, high PD-1 levels are expressed by virus-specific T cells, and improvement of the T-cell function has been obtained in vitro by inhibition of the PD-1/PD-ligand 1 (L1) interaction. 6,13 21 Surprisingly, despite an altered function, HBV-specific PD- 1-positive CD8 cells of chronic HBV patients displayed also high levels of the interleukin (IL)-7 receptor CD127, which had previously been described in association with Abbreviations used in this paper: AA, amino acids; APC, allophycocyanin; CMV, cytomegalovirus; CTLA-4, cytotoxic T-lymphocyte-associated antigen-4; FITC, fluorescein isothiocyanate; FLU, influenza; IFN, interferon; IL, interleukin; LIL, liver infiltrating lymphocytes; PBMC, peripheral blood mononuclear cells; PCR, polymerase chain reaction; PD-1, programmed death-1; PD-L, PD-ligand; PE, phycoerythrin; PerCP, peridin chlorophyll protein by the AGA Institute /10/$36.00 doi: /j.gastro

2 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 683 a functionally efficient memory T-cell differentiation However, most of the phenotypic and functional data have been obtained by studying peripheral blood T cells, although characterization of liver infiltrating lymphocytes (LIL) is crucial because the liver is the target of viral replication and the site of cellular injury. Differences between circulating and intrahepatic T cells have been reported in chronic HCV infection, both in phenotype and in response to the manipulation of costimulatory pathways, 18,19,25 but no information is available at present in chronic hepatitis B. This study was thus undertaken to define further the role of T-cell exhaustion in the pathogenesis of chronic HBV infection, by comparing phenotype and function of intrahepatic and circulating HBV-specific T lymphocytes as well as the effect of PD- 1/PD-L1 blockade. Our results indicate that intrahepatic HBV-specific CD8 cells are phenotypically different from their peripheral counterpart because they express higher PD-1 and lower CD127 levels. PD-1/PD-L1 blockade can induce variable levels of functional T-cell restoration both in the liver and in peripheral blood, with a better functional improvement among intrahepatic T cells. This observation opens new potential perspectives for the development of novel immunotherapies for chronic hepatitis B. Materials and Methods Study Subjects A total of 42 patients with chronic hepatitis B were enrolled at the Unit of Infectious Diseases and Hepatology of the Azienda Ospedaliero-Universitaria of Parma, Parma, Italy, and at the Department of Clinical Medicine of the University of Bologna, Bologna, Italy. Diagnosis of chronic hepatitis B was based on the detection of alanine aminotransferase (ALT) elevation lasting for more than 6 months, associated with positive hepatitis B surface antigen (HBsAg), antibody to hepatitis B core antigen (anti-hbc), and HBV-DNA. Patients were never treated with nucleos(t)ide analogues; 7 of 42 patients underwent interferon (IFN)- treatment that was stopped 16 to 2 years before liver biopsy. Twenty-seven patients were male and 15 female, aged from 22 to 68 (median, 41) years; 90.2% were white. At the time of liver biopsy and of the corresponding blood sample, patients ALT levels ranged between 18 to 351 U/L (median, 51 U/L) and HBV-DNA from 112 to IU/mL (median, IU/mL). The great majority of the patients was infected by genotype D HBV, with the exception of 2 patients who were infected by genotype F, 1 by genotype A, and 1 by genotype E. Three patients were hepatitis B e antigen (HBeAg) positive, whereas the remaining patients were HBeAg negative and antibody to hepatitis B e antigen (anti-hbe) positive. Twenty patients were HLA-A2 positive. Screening for HLA-A2 was performed by staining peripheral blood mononuclear cells (PBMC) with a fluorescent anti-hla-a2.01 antibody (AbD Serotec, Kidlington, UK). All patients were negative for anti-hepatitis C virus, virus, human immunodeficiency virus type 1 (HIV-1), and HIV-2 antibodies and for other markers of viral or autoimmune hepatitis. Six healthy uninfected HLA-A2-positive subjects served as controls. The study was approved by the Ethical Committee of the Azienda Ospedaliero- Universitaria of Parma, and all subjects gave written, informed consent. Virologic Assessment The presence or absence of HBsAg, antibody to hepatitis B surface antigen (anti-hbs), total and immunoglobulin M anti-hbc, HBeAg, anti-hbe, anti-hepatitis D virus, anti-hepatitis C virus, anti-hiv types 1 and 2, and anti-cytomegalovirus (CMV) antibodies was determined by using commercial enzime immunoassay kits (Abbott Labs, Abbott Park, IL; Ortho-Clinical Diagnostic, Johnson & Johnson, Raritan, NJ; DiaSorin, Vercelli, Italy). Levels of HBV-DNA in serum samples were quantified by real-time polymerase chain reaction (PCR) (TaqMan; Roche Diagnostics, Indianapolis, IN). HBV genotypes were determined by sequencing of RT and S genes (TRU- GENE HBV genotyping kit; Siemens Healthcare Diagnostics Inc, Deerfield, IL). Synthetic Peptides, Peptide-HLA Class I Tetramers, and Antibodies Two hundred eight 15mer peptides overlapping by 10 residues and covering the entire sequence of the core and polymerase proteins of HBV genotype D were purchased from Mimotopes (Clayton, Victoria, Australia). The 42 peptides corresponding to the core region were pooled in a single mixture, and the 166 peptides covering the polymerase protein were pooled in 2 mixtures. Synthetic peptides and the corresponding phycoerythrin (PE)-labeled tetrameric peptide-hla class I complexes representing HLA-A2-restricted epitopes within core (amino acids [AA], 18 27, FLPSDFFPSV), envelope (AA, , FLLTRILTI; AA, , WLSLLVPFV; and AA, , GLSPTVWLSV), and polymerase (AA, , FLLSLGIHL; and AA, , SLYADSPSV) proteins of genotype D and the HLA-A2 restricted epitopes CMV pp65 (NLVPMVATV) and influenza (FLU) virus matrix (GILGFVFTL) were purchased from Proimmune (Oxford, United Kingdom) and from Beckman Coulter, Inc (Fullerton, CA). Anti-CD8 peridin chlorophyll protein (PerCP), allophycocyanin (APC), or APC-Cy7 conjugated, anti-cd3 PerCP, anti-cd4 PE, anti-pd-1 fluorescein isothiocyanate (FITC), anti-il-2 APC, anti-cd127 Alexa 647, anti-ccr7 PE-Cy7, anti-cd45ra PE-Cy5 were purchased from BD Biosciences-Pharmingen (San Jose, CA). Anti-IFN- -FITC was purchased from Sigma-Aldrich (St.

3 684 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Louis, MO). Anti-PD-L1 and the isotype control antibodies were purchased from ebioscience (Boston, MA). Isolation of PBMC and LIL PBMC were isolated from fresh heparinized blood by Ficoll-Hypaque density gradient centrifugation and resuspended in RPMI 1640 supplemented with 25 mmol/l HEPES, 2 mmol/l L-glutamine, 50 g/ml gentamicin, and 8% human serum. LIL were isolated from excess liver tissue not needed for diagnostic purposes. The liver biopsy fragment was extensively washed in RPMI to remove contaminating blood and digested with collagenase (Sigma Chemical Co, St. Louis, MO; 1 g/ ml) and DNAase (Sigma Chemical Co; 25 g/ml) for 1 hour, at 37 C. The mononuclear cell suspension was then washed and resuspended in complete medium. After isolation, lymphocytes were either stained immediately for phenotypic analysis or cultured in a round-bottomed 96-well plate (at least 20,000 LIL/well), in the presence of autologous irradiated (4000 rads) PBMC ( /well) and HBV peptides; after 3 days, activated T cells were expanded by addition of IL-2 (50 U/mL). Cell Surface and Intracellular Staining For immunophenotyping, PBMC were incubated for 15 minutes at room temperature with individual PE-labelled tetramers or with a pool of HBV tetramers; LIL were incubated with the HBV-tetramer pool or with the CMV and FLU tetramers. Fluorescent antibodies were then added for 15 more minutes. After a wash with phosphate-buffered saline (PBS)/0.1% fetal calf serum, cells were immediately analyzed on an FACSCanto (Becton Dickinson, San Jose, CA) by the BD FACSDiva software (Becton Dickinson). For intracellular cytokine staining after in vitro expansion, cells were incubated in medium alone (control) or with viral peptides (1 mol/l) for 1 hour, and brefeldin A (10 g/ml) was added for an additional 4 hours. After washing, cells were stained with anti-cd8, anti-cd4, and anti-cd3 fluorescent antibodies; fixed; permeabilized; and marked with anti-ifn- and anti-il-2 antibodies, as already described. 6 In Vitro Expansion and Blockade PD-1/PD-L1 blocking experiments on peripheral T cells were performed as previously described. 6 LIL were cocultured with autologous irradiated PBMC as feeder cells that were previously incubated for 45 minutes at 37 C with the anti-pd-l1 or isotype control antibody (10 g/ml) and with the 15mer or HLA-A2 restricted peptides at a 1 mol/l final concentration. Cells were finally analyzed by intracellular cytokine staining after 2 weeks of culture. Statistical Analysis Wilcoxon paired test or Mann Whitney U test was used to compare frequencies of different T-cell subsets in peripheral blood and liver, as well as to evaluate differences in cytokine production between anti-pd-l1 or control antibody-treated lymphocyte cultures. Correlations between HBV-DNA values and frequencies of intrahepatic HBV-specific CD8 cells and between PD-1 expression on intrahepatic tetramer-positive cells or their functional recovery and serum ALT or Ishak score were evaluated by the Spearman rank correlation test. Results Frequency of HBV-Specific CD8 T Cells To compare the ex vivo frequency and phenotype of HBV-specific CD8 T cells from liver and peripheral blood, 17 HLA-A2- positive patients with chronic hepatitis B were studied with a panel of 6 major histocompatibility complex (MHC) class I tetramers containing frequently detected HBV epitopes (core AA, 18 27; envelope AA, , AA , and AA ; and polymerase, AA and AA ). Because of limitations in the number of cells available from liver biopsies, staining of LIL was carried out with a mixture of all tetramers, whereas blood cells were stained with the tetramer mixture and also with individual HBV tetramers. Only core AA 18 to 27-specific T cells were detected in the blood, and no relevant phenotypic difference with respect to PD-1 and CD127 expression was observed between HBV-specific cells detected with the tetramer pool and with the core tetramer (data not shown). Frequencies of HBV-specific CD8 cells were significantly higher among intrahepatic (range, 0.09% 10.5%) than among circulating (range, 0.02% 0.04%) T cells (P.0002; Figure 1A). As a result of this, HBV-specific CD8 T cells were always detectable in the liver (with the exception of a single case in which too few liver lymphocytes were purified to allow analysis), whereas peripheral HBV-specific cells were detected only in 10 patients. Frequencies of intrahepatic HBV-specific CD8 cells tended to decline as a function of serum HBV-DNA increase, with the lowest frequencies of HBV-specific CD8 cells corresponding to viremia levels above 10 5 IU/mL (Figure 1B). A significant inverse correlation was detected between percentages of intrahepatic HBV tetramer-positive cells and serum HBV-DNA titers (Spearman rank correlation test, P.0023; Figure 1C). Phenotypic Profile of HBV-Specific and Total CD8 T Cells HBV-specific and total CD8 T cells from liver and blood were evaluated for PD-1 and CD127 expression; CMV-specific CD8 cells were assessed for comparison. Although PD-1 expression was high on circulating HBVspecific CD8 cells (range, 63.6% 89%; mean, 76.3%), it was significantly higher on HBV-specific CD8 cells from the liver (range, 76.2% 100%; mean, 93%; Supplementary Figure 1) (P.0039; Figure 2A). Moreover, PD-1 ap-

4 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 685 Figure 1. Frequency of HBV-specific CD8 T cells in liver and peripheral blood of chronic HBV patients. Liver infiltrating and peripheral blood HBV-specific CD8 cells from 17 HLA-A2 chronic HBV patients were stained with a pool of 6 HLA-A2 tetramers. Panel A illustrates percentages of tetramer-positive cells on total CD8 T cells. Differences between intrahepatic and peripheral tetramer-positive cell frequencies were assessed by the Wilcoxon paired test. Panel B represents mean frequency values of intrahepatic tetramer-positive CD8 cells subdivided in relation to different ranges of HBV-DNA concentration. P values according to the Mann Whitney test are reported. Panel C shows the inverse correlation between percentages of intrahepatic tetramer-positive CD8 cells and serum HBV-DNA titers, assessed by the Spearman test. peared to be more expressed on intrahepatic than peripheral HBV-specific CD8 cells also at individual cell level, as assessed by mean fluorescence intensity (P.013) (Supplementary Figure 2). A lower percentage of intrahepatic HBV-specific T cells were CD127 positive (range, 10.1% 53.7%; mean, 28.2%; Supplementary Figure 1) compared with peripheral blood cells (range, 28.8% 100%; mean, 71%) (P.0039; Figure 2A). Thus, intrahepatic HBVspecific CD8 cells are phenotypically different from their peripheral counterparts because they express higher levels of PD-1 and lower levels of CD127, as a likely expression of a deeper condition of exhaustion. We then tried to correlate frequencies of intrahepatic PD-1 /tetramer cells with biochemical and histologic parameters of liver disease. A weak positive correlation by the Spearman correlation test was found with ALT values (P.0094) and with Ishak fibrosis score (P.01; Figure 2B), suggesting a possible role for intrahepatic PD-1 expression as a marker of disease progression. PD-1 and CD127 expression was then analyzed on total CD8 cells (Figure 2C). PD-1 was more expressed on total intrahepatic than peripheral CD8 cells (P.001; Figure 2C), and, in both liver and blood compartments, more PD-1 was detected on HBV-specific than on total CD8 cells (P.002 and P.015, respectively; Figure 2C). Also, among the overall CD8 population, CD127 expression was significantly lower in the intrahepatic than in the peripheral compartment (P.001; Figure 2C), with no significant difference between HBV-specific and total CD8 cells. To understand further whether compartmentalization could induce a higher PD-1 expression on CD8 T cells within the liver irrespective of their specificity, intrahepatic and peripheral cells from 4 patients were stained with a CMV-specific tetramer and from 1 patient with a FLU tetramer. Detectable frequencies of CMV-specific CD8 cells were detected both in the liver and in the peripheral blood (frequencies of CMV- and FLU-specific CD8 T cells ranging from 0.13% to 1.9% of total CD8 cells in the liver and from 0.02% to 1.08% in the blood). Levels of PD-1 and CD127 expression were comparable on liver infiltrating and circulating CMV- or FLU-specific cells (Figure 2D). We then asked whether PD-1 was selectively more expressed at distinct T-cell differentiation stages. To answer this question, CD45RA and CCR7 expression on HBV-specific and total CD8 cells were analyzed. Most CD8 cells in both liver and blood compartments were CD45RA /CCR7 effector/memory cells (Figure 3). All intrahepatic CD8 subsets among HBV-specific and total CD8 cells showed higher PD-1 expression compared with their peripheral blood counterparts (Figure 3A and B). Moreover, intrahepatic PD-1 expression was significantly higher on CD45RA CCR7 and on CD45RA CCR7 CD8 cell subsets among both HBV-specific and total CD8 cells (Supplementary Figure 3). Whereas the CD45RA CCR7 profile defines a central/memory subset, CD45RA CCR7 CD8 cells cannot be defined as naïve because they express high levels of the B2 integrin CD11a, which has been shown to be expressed on antigen experienced cells 26 (data not shown). Effect of PD-1/PD-L1 Blockade on the HBV- Specific T-Cell Function Whereas the effect of PD-1/PD-L1 blockade was previously characterized on the function of HBV-spe-

5 686 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Figure 2. PD-1 and CD127 expression by tetramer-positive and total CD8 T cells. Panel A: Percentages of PD-1 and CD127 cells among HBV-tetramer CD8 lymphocytes isolated from liver or blood of chronic HBV patients. Dot plots on the right show a representative example of intrahepatic and peripheral blood HBV-specific CD8 cells. Panel B: Correlation between PD-1 intrahepatic tetramer CD8 cells and Ishak hepatic fibrosis score (P.01 by the Spearman test). Panel C: PD-1 (left) and CD127 (right) expression on HBV-specific and total CD8 cells derived from liver and blood. Panel D: Expression of PD-1 and CD127 on intrahepatic and peripheral CMV-specific CD8 cells; all CMV-responsive patients were positive for serum IgG but not for IgM anti-cmv antibodies. A representative example is illustrated in the right dot plots. Statistical analysis in panels A and C was carried out on a paired set of data (intrahepatic and peripheral frequencies of HBV-specific and total CD8 cells) by the Wilcoxon paired test. cific peripheral blood cells, 6 no information is available at present on intrahepatic HBV-specific T cells at the site of hepatitis lesions where PD-1 is more expressed than in the periphery and where also PD-L1 has been reported to be up-regulated following viral infection. 27 To test the effect of PD-1/PD-L1 blockade on both intrahepatic and peripheral blood CD4 and CD8 cells, lymphocytes from liver and blood of 22 patients were stimulated in vitro with 15mer peptide mixtures corresponding to the HBV-core (16 patients) or HBVpolymerase (6 patients) regions of genotype D for 14 days in the presence of either an anti-pd-l1 or an isotypic control antibody, and cytokine production by CD4 and CD8 cells was analyzed by intracellular cytokine staining. Cells from all patients infected with HBV genotypes other than D (4 patients) were stimulated with core peptides because core is a very conserved HBV protein 28 with a minimal intergenotype variability, and all showed cytokine production in response to stimulation. IFN- production was significantly enhanced in cells treated with anti-pd-l1 antibody in comparison with control cultures (Figure 4A). An improved IFN- production, from 1.5- to folds, was detected in 12 of 24 tests (50%) for intrahepatic CD8 cells (P.0035) and from 1.8- to 8.2-folds in 8 of 25 tests (32%) for peripheral CD8 T cells (P.028). In 8 of 24 tests (33%), anti-pd-l1 increased the frequency of liver-derived IFN- -producing CD4 T cells (fold increase ranging from 1.7 to 46; P.02), whereas an increased frequency of peripheral IFN- producing CD4 cells was detected in 13 of 25 tests (52%) (fold increase ranging from 2.2 to 4.25; P.01).

6 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 687 Figure 3. PD-1 expression is higher on intrahepatic than on peripheral lymphocytes at all T- cell differentiation stages. Frequencies of intrahepatic and peripheral HBV-specific (panel A) and total (panel B) CD8 cell subsets with different CD45RA and CCR7 expression. Histograms illustrate the percentage of PD-1 T cells in each T-cell subset. Differences between intrahepatic and peripheral PD-1 expression were assessed by the Wilcoxon paired test. Reported values derive from 12 patients. When the overall effect of anti-pd-l1 on CD4 and CD8 cells was compared in peripheral and intrahepatic T cells, a significantly better increase in IFN- production was detected among liver cells, as shown by fold increase in IFN- T cells, which is the ratio between the percentage of IFN- producing CD4 and CD8 cells upon anti-pd-l1 and control antibody treatment (P.0088; Figure 4B). As far as IL-2 production is concerned, a significant improvement was induced by anti-pd-l1 treatment on intrahepatic CD8 (P.027) and CD4 (P.0068) T cells, whereas PD-1/PD-L1 blockade was less effective in periphery, where increased IL-2 production upon PD-1/ PD-L1 blockade was observed in only 3 of 17 and 6 of 17 tests for CD8 and CD4 T cells, respectively (P.0085 and P ns, respectively; Figure 5A). Intrahepatic CD8 and CD4 T cells showed enhanced anti-pd-l1 induced IL-2 production in 7 of 16 and 8 of 16 tests, respectively. A better effect of anti-pd-l1 treatment on IL-2 production was observed for intrahepatic compared with peripheral T cells, as a fold increase in IL-2-producing CD4 and CD8 cells (P.01; Figure 5B). Moreover, when the overall effect of anti-pd-l1 treatment was assessed irrespective of compartment and T-cell subset, analyzing together CD4 and CD8 cells from liver and blood to have a representation of the global effect of anti-pd-l1 on antiviral responses, an increased production of IFN- or IL-2 was detected in 86.4% (19/22) and in 62.5% (10/16) of the patients tested, respectively. The effect of PD-1/PD-L1 blocking was also tested on intrahepatic and peripheral lymphocytes from 4 HLA- A2 patients, after stimulation with a pool of the HLA- A2 restricted 9mer peptides contained in the HBV tetramers (core peptide, AA 18 27; envelope peptides, AA , AA , and AA ; and polymerase peptides, AA and AA ) and with CMV or FLU peptides as a control. Intrahepatic lymphocytes from all chronic patients expanded well, and HBV-specific CD8 cells derived from 2 of 4 patients expanded more vigorously in the presence of the anti-pd-l1 antibody (from 8% to 22% of the total CD8 cells in one case and from 0.46 to 45% in the other patient) and produced 2.5 and 1.5 times more IFN-, respectively; IL-2 production increased in 3 patients (from 5.6% to 13.5%, from 0.1% to 5.8%, and from 1% to 6.5%, respectively) (Figure 6). Peripheral CD8 cells did not expand in 1 of the 4 HLA-A2-positive patients upon stimulation with HBV peptides; in 2 other cases, tetramer-positive cells expanded 2 times better upon anti-pd-l1 treatment and produced 2- to 3-fold more IFN- but not IL-2 (data not shown). In the fourth patient, CD8 cells expanded upon peptide stimulation, but no improvement was induced by anti-pd-l1. CMV- and FLU-specific CD8 T cells from both liver and blood expanded efficiently with a variable but frequently positive effect of anti-pd-l1 on T-cell proliferation and function (Figure 7 and Supplementary Figure 4). CMV- and FLU-specific T-cell responses of 6 healthy uninfected HLA-A2-positive subjects were tested as additional control and were minimally or not affected at all by anti-pd-l1 treatment.

7 688 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Figure 4. Anti-PD-L1 can enhance the capacity of HBV-specific T cells to produce IFN-. Lymphocytes from liver and peripheral blood were cultured in vitro in the presence of anti- PD-L1 or an isotypic control antibody and IFN- production was analyzed by intracellular cytokine staining upon stimulation with a mixture of overlapping 15mer peptides covering the overall HBV-core or HBV-polymerase regions. Panel A illustrates the effect on cytokine production of liver and blood CD8 and CD4 cell induced by anti- PD-L1. Panel B depicts changes in frequency of IFN- producing CD8 and CD4 cells analyzed together induced by anti-pd-l1, assessed as the ratio between the percentage of IFN- -positive T cells detected in cultures treated with anti-pd-l1 and the percentage detected in control cultures (fold increase). Responsive patients with fold increase 1.5 are illustrated. Statistics were done by the Wilcoxon paired test. IFN- production by liver-infiltrating CD8 and CD4 T cells was then analyzed in relation to patients HBV-DNA titers to determine the influence of viremia on functional T-cell restoration by anti-pd-l1 treatment. By dividing the patient population based on serum HBV-DNA titers below or above 10 6 copies/ml, a better functional restoration upon anti-pd-l1 treatment was observed in patients with lower viremia (Figure 8). A weak functional recovery was detectable also in 2 of the 3 HBeAg-positive patients with extremely elevated HBV-DNA levels 10 9 copies/ml but selectively among liver cells; recovery in peripheral blood cells started to be seen for viremia levels between 10 6 and 10 9 copies/ml (data not shown). Finally, we looked for possible correlations between cytokine production improvement upon anti-pd-l1 treatment and the level of liver inflammation as expressed by the Ishak score, but no statistical correlations were found according to the Spearman rank correlation test. Discussion Studies in animal models of virus infections show that persistent exposure to high antigen concentrations can affect the antiviral T-cell function by causing different degrees of functional impairment, up to physical T-cell deletion, as a function of the quantity of antigen to which T cells are exposed. 29,30 T-cell exhaustion by this mechanism may play an important role in the pathogenesis of the HBV-specific T-cell hyporesponsiveness of chronic HBV infection because high antigen concentrations are constantly present at all stages of infection. The importance of T-cell exhaustion in chronic HBV infection is confirmed by the high expression of PD-1 on circulating HBV-specific CD8 cells and by the possibility to partially improve the peripheral blood T-cell function by blocking PD-1/PD-L1 interaction with anti-pd-l1 antibodies. 6 In line with this, functional improvement of HBV envelope-specific T cells has been reported to be more difficult to induce by PD-1/PD-L1 blockade, as a possible effect of deeper exhaustion resulting from the large excess of HBV envelope antigens secreted by infected cells during HBV infection. 6 Moreover, additional support to exhaustion is provided by the inverse correlation between function of HBV-specific T cells and viremia levels because T cells are more profoundly inhibited in the presence of high viremia. 6,31,32 Because T cells are an essential component of antiviral protection, restoration

8 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 689 Figure 5. Anti-PD-L1 can improve IL-2 production by HBV-specific T cells. Liver-derived and peripheral blood T cells were stimulated with HBV-core or HBV-polymerase peptides in the presence of an anti- PD-L1 or an isotypic control antibody and were tested by intracellular cytokine staining for IL-2 production. Panel A illustrates the changes in IL-2 production detected in anti-pd-l1 compared with control antibody treated cultures. Panel B shows the overall effect of anti-pd-l1 on CD8 and CD4 cells represented together, as assessed by fold increase in IL-2 producing cells. Responsive patients with fold increase 1.5 are illustrated. of an efficient T-cell function may represent a strategy to cure infection, and manipulation of costimulatory pathways involved in T-cell activation may be a rational approach to achieve this goal. Results obtained in natural HBV infection with peripheral blood T cells are encouraging, 6 and recovery of intrahepatic HBV-specific T-cell responses by PD-1/PD-L1 blockade has been reported in transgenic mice. 9 However, in natural infection, peripheral T cells may only partially reflect the intrahepatic T-cell compartment where pathogenetically relevant T cells are recruited. Thus, the aim of the study was to extend functional T-cell analysis so far limited to the peripheral blood compartment to liver-infiltrating T cells to define better the real potential of PD-1/PD-L1 blockade as a strategy to restore antiviral protection. Our results provide clear evidence that the function of liver-infiltrating HBV-specific T cells can be improved by PD-1/PD-L1 blockade and that this effect is greater on intrahepatic than on peripheral T cells. This indicates that functional impairment by exhaustion is a crucial inhibitory mechanism especially for HBV-specific T cells recruited at the site of infection where contact with large amounts of antigen presented by infected and noninfected liver cells takes place. A comparable frequency of IFN- - and IL-2-producing T cells was detected upon in vitro expansion by peptide stimulation of intrahepatic and peripheral T cells, although the ex vivo frequency of HBV-specific T cells detected by tetramer staining was much higher within the liver. This suggests that intrahepatic HBV-specific T cells are able to expand and to produce cytokines less efficiently than their peripheral counterpart, in line with the higher PD-1 expression within the liver, as shown by the frequency of intrahepatic PD-1-positive CD8 cells and by the level of PD-1 expression at individual T-cell level. High intrahepatic PD-1 expression was found on all differentiation subsets of HBV-specific CD8 cells detected at the site of virus replication and cellular injury, where all virus-specific cells are expected to be antigen experienced. Liver-infiltrating CD8 cells expressed more PD-1 and less CD127 than peripheral CD8 cells, indicating a more exhausted phenotype of T cells at the site of viral replication where persistent exposure to high antigen concentrations probably contributes to down-regulate CD127 expression and to maintain elevated PD-1 levels. PD-1 increase within the liver was greater among total than HBV-specific CD8 cells. This may be due to the higher frequency of HBVspecific CD8 cells within the liver. Moreover, lack of recruitment of naïve CD8 cells into the liver and the more activated phenotype typical of liver-infiltrating T cells may further contribute to the higher PD-1 expression on total intrahepatic CD8 cells compared with their peripheral blood counterpart. Our results are in keeping with the findings of Radziewicz et al 18 in chronic HCV infection, which showed high PD-1 and low CD127 expression on liver-infiltrating HCV-specific CD8 cells, in contrast with the higher CD127 levels on circulating HCV-specific CD8 cells. The better functional recovery detected in our study among liver-infiltrating than among peripheral blood HBV-specific T cells upon PD-1/PD-L1 blockade is instead different from what was previously reported by Nakamoto et al 21 in chronic HCV infection, where no effect of anti- PD-L1 on liver-derived HCV-specific CD8 cells was detected. This difference may reflect the specific features of the patient cohort in the Nakamoto et al study, 21 which was composed of patients with more advanced disease at which stage antiviral T-cell reactivity may be more profoundly affected by the longer duration of infection and by the concurring effect of multiple inhibitory mechanisms. Moreover, accumulation of CD4 CD25 Foxp3 T regulatory cells with high PD-1 expression has recently been reported into the liver of chronically HCV infected individuals. 33 Because PD-1/PD-L1 blockade can increase both expansion and suppressor function of these regulatory cells, the final effect of blockade on liver-infiltrating T cells may be lack of improvement of effector functions caused by enhanced suppression by T regulatory cells. In addition, coexpression of different inhibitory receptors, including PD-1, LAG-3, CD160, and cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4), has recently been re-

9 690 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Figure 6. Effect of anti-pd-l1 on IFN- and IL-2 production by intrahepatic HBV-specific T cells analyzed by tetramer staining. Intrahepatic lymphocytes from 4 HLA-A2 chronic HBV patients were cultured in the presence of anti-pd-l1 or a control antibody and stimulated with a pool of HLA-A2 restricted HBV peptides (core, 18 27; envelope , , and ; polymerase, and ). The Figure illustrates in vitro expansion and cytokine production upon anti-pd-l1 or control antibody treatment by tetramer-positive CD8 cells. ported in relation to the severity of chronic lymphocytic choriomeningitis virus infection. 34 These at least partially nonoverlapping and nonredundant pathways may synergistically but differentially contribute to T-cell exhaustion. For example, an effect of CTLA-4 has been observed on exhausted HIV- and HCV-specific CD4 and CD8 cells, whereas no effect of this inhibitory receptor has been reported in chronic lymphocytic choriomeningitis virus infection. 8,35,36 Thus, the degree of efficacy of PD-1/ PD-L1 blockade may be affected by the simultaneous expression of additional inhibitory receptors, which may provide a variable contribution to T-cell exhaustion in

10 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 691 Figure 7. Effect of anti-pd-l1 on IFN- and IL-2 production by intrahepatic CMV- and FLU-specific T cells analyzed by tetramer staining. Liver infiltrating lymphocytes from 4 HLA-A2 patients were cultured in the presence of anti-pd-l1 or a control antibody and stimulated with HLA-A2 restricted CMV or FLU peptides. T-cell expansion and cytokine production are illustrated. relation to stage and type of infection and perhaps to the type of compartment analyzed in the same infected host. In conclusion, the possibility to improve the function not only of peripheral blood 6 but also of intrahepatic HBV-specific T cells is very promising in the perspective of a therapeutic stimulation of protective T-cell responses to induce control of infection and to improve the effect of available therapies. A note of caution is raised by the finding that PD-1/PD-L1 blockade can enhance also responses of T cells specific for other HBV-unrelated pathogens, such as CMV and FLU, by pathogen-specific T cells recruited into the liver as a result of inflammation. Moreover, a large fraction of the total intrahepatic CD8 cells is PD-1 positive, confirming that PD-1 blockade in vivo could probably also affect responses to HBV-unrelated antigens, enhancing the likelihood of increasing pathogenic instead of protective T-cell responses. This concern is partially mitigated by the lack of detectable toxicity upon in vivo PD-1 blockade in patients with advanced hematologic malignancies 37 and in macaques, which were extensively analyzed after blockade for several serum parameters including autoantibody levels. 10 In any

11 692 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Figure 8. IFN- production by liver infiltrating T cells is more efficiently improved by anti-pd-l1 at low HBV-DNA titers. IFN- production by liver infiltrating CD8 and CD4 cells after anti-pd-l1 treatment was analyzed in relation to patients HBV-DNA titers. Statistical analysis was done by the Wilcoxon paired test to compare the effect of the anti-pd-l1 and the control antibodies and by the Mann Whitney test to assess the differences between cytokine recovery in patients with an HBV- DNA titer below or above copies/ml. case, manipulation of costimulatory pathways should be considered with extreme caution in vivo if associated with immunomodulatory agents able to enhance the overall T-cell reactivity, such as IFN-, because of the potential risk to enhance also undesired T-cell responses. As suggested by vaccination studies in animal models of chronic viral infection, 12 PD-1/PD-L1 blockade should be preferentially proposed as a strategy to further enhance virus-specific T-cell responses specifically amplified by direct and selective stimulation induced by in vivo antigen administration. Supplementary Materials Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at and at doi: /j.gastro References 1. Bertoletti A, Ferrari C. Kinetics of the immune response during HBV and HCV infection. Hepatology 2003;38: Rehermann B, Nascimbeni M. Immunology of hepatitis B virus and hepatitis C virus infection. Nat Rev Immunol 2005;5: Das A, Hoare M, Davies N, et al. Functional skewing of the global CD8 T-cell population in chronic hepatitis B virus infection. J Exp Med 2008;205: Chang JJ, Thompson AJV, Visvanathan K, et al. The phenotype of hepatitis B virus-specific T cells differs in the liver and blood in chronic hepatitis B virus infection. Hepatology 2007;46: Zhang Z, Zhang J-Y, Wherry EJ, et al. Dynamic programmed death 1 expression by virus-specific CD8 T cells correlates with the outcome of acute hepatitis B. Gastroenterology 2008;134: Boni C, Fisicaro P, Valdatta C, et al. Characterization of hepatitis B virus (HBV)-specific T-cell dysfunction in chronic HBV infection. J Virol 2007;81: Lopes AR, Kellam P, Das A, et al. Bim-mediated deletion of antigen-specific CD8 T cells in patients unable to control HBV infection. J Clin Invest 2008;118: Barber DL, Wherry EJ, Masopust B, et al. Restoring function in exhausted CD8 T cells during chronic viral infection. Nature 2006;439: Maier H, Isogawa M, Freeman GJ, et al. PD-1:PD-L1 interactions contribute to the functional suppression of virus-specific CD8 T lymphocytes in the liver. J Immunol 2007;178: Velu V, Titanji K, Zhu B, et al. Enhancing SIV-specific immunity in vivo by PD-1 blockade. Nature 2009;458: Finnefrock AC, Tang A, Fengsheng L, et al. PD-1 blockade in rhesus macaques: impact on chronic infection and prophilactic vaccination. J Immunol 2009;182: Ha SJ, Mueller SN, Wherry EJ, et al. Enhancing therapeutic vaccination by blocking PD-1-mediated inhibitory signals during chronic infection. J Exp Med 2008;205: Freeman GJ, Wherry EJ, Ahmed R, et al. Reinvigorating exhausted HIV-specific T cells via PD-1-PD-1 ligand blockade. J Exp Med 2006;203: Day CL, Kaufmann DE, Kiepiela P, et al. PD-1 expression on HIV-specific T cells is associated with T-cell exhaustion and disease progression. Nature 2006;443: Trautmann L, Janbazian L, Chomont N, et al. Up-regulation of PD-1 expression on HIV-specific CD8 T cells leads to reversible immune dysfunction. Nat Med 2006;12: Petrovas C, Casazza JP, Brenchley JM, et al. PD-1 is a regulator of virus-specific CD8 T cell survival in HIV infection. J Exp Med 2006;203:

12 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION Urbani S, Amadei B, Tola D, et al. PD-1 expression in acute hepatitis C virus (HCV) infection is associated with HCV-specific CD8 exhaustion. J Virol 2006;80: Radziewicz H, Ibegbu CC, Fernandez ML, et al. Liver-infiltrating lymphocytes in chronic human hepatitis C virus infection display an exhausted phenotype with high levels of PD-1 and low levels of CD127 expression. J Virol 2006;81: Golden-Mason L, Palmer B, Klarquist J, et al. Up-regulation of PD-1 expression on circulating and intrahepatic HCV-specific CD8 T cells associated with reversible immune dysfunction. J Virol 2007;81: Penna A, Pilli M, Zerbini A, et al. Dysfunction and functional restoration of HCV-specific CD8 responses in chronic hepatitis C virus infection. Hepatology 2007;45: Nakamoto N, Kaplan DE, Coleclough J, et al. Functional restoration of HCV-specific CD8 T cells by PD1 blockade is defined by PD1 expression and compartmentalization. Gastroenterology 2008;134: Van Leeuwen EMM, de Bree GJ, Remmerswaal EBM, et al. IL-7 receptor chain expression distinguishes functional subsets of virus-specific human CD8 T cells. Blood 2005;106: Kaech SM, Tan JT, Wherry EJ, et al. Selective expression of the interleukin 7 receptor identifies effector CD8 T cells that give rise to long-lived memory cells. Nat Immunol 2003;4: Wherry EJ, Barber DL, Kaech SM, et al. Antigen-independent memory CD8 T cell do not develop during chronic viral infection. Proc Natl Acad Sci U S A 2004;101: Bengsch B, Spangenberg N, Kersting C, et al. Analysis of CD127 and KLRG1 expression on hepatitis C virus-specific CD8 T cells reveals the existence of different memory T-cell subsets in the peripheral blood and liver. J Virol 2007;81: Heydtmann M, Hardie D, Shields PL, et al. Detailed analysis of intrahepatic CD8 T cells in the normal and hepatitis C-infected liver reveals differences in specific populations of memory cells with distinct homing phenotypes. J Immunol 2006;177: Muhlbauer M, Fleck M, Schutz C, et al. PD-L1 is induced in hepatocytes by viral infection and by interferon- and - and mediates T-cell apoptosis. J Hepatol 2006;45: Chain BM, Myers R. Variability and conservation in hepatitis B virus core protein. BMC Microbiol 2005;5: Wherry EJ, Blattman JN, Murali-Krishna K, et al. Viral persistence alters CD8 T-cell immunodominance and tissue distribution and results in distinct stages of functional impairment. J Virol 2003; 77: Wherry EJ, Ha SJ, Kaech SM, et al. Molecular signature of CD8 T-cell exhaustion during chronic viral infection. Immunity 2007; 27: Webster GJM, Reignat S, Brown D, et al. Longitudinal analysis of CD8 T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy. J Virol 2004;78: Maini MK, Boni C, Lee CK, et al. The role of virus-specific CD8 cells in liver damage and viral control during persistent hepatitis B virus infection. J Exp Med 2000;191: Franceschini D, Paroli M, Francavilla V, et al. PD-L1 negatively regulates CD4 CD25 Foxp3 Tregs by limiting STAT-5 phosphorylation in patients chronically infected with HCV. J Clin Invest 2009;119: Blackburn SD, Shin H, Haining WN, et al. Coregulation of CD8 T cell exhaustion by multiple inhibitory receptors during chronic viral infection. Nat Immunol 2009;10: Kaufmann DE, Kavanagh DG, Pereyra F, et al. Up-regulation of CTLA-4 by HIV-specific CD4 T cells correlates with diseases progression and defines a reversible immune dysfunction. Nat Immunol 2007;8: Nakamoto N, Cho H, Shaked A, et al. Synergistic reversal of intrahepatic HCV-specific CD8 T-cell exhaustion by combined PD- 1/CTLA-4 blockade. PLoS Pathog 2009 [Epub ahead of print]. 37. Berger R, Rotem-Yehudar R, Slama G, et al. Phase I safety and pharmacokinetic study of CT-011, a humanized antibody interacting with PD-1, in patients with advanced hematologic malignancies. Clin Cancer Res 2008;14: Received March 6, Accepted September 17, Reprint requests Address requests for reprints to: Carlo Ferrari, MD, Professor, Laboratory of Viral Immunopathology, Unit of Infectious Diseases and Hepatology, Azienda Ospedaliero-Universitaria di Parma, Via Gramsci, 14, Parma, Italy. cafer@tin.it; fax: (39) Conflicts of interest The authors disclose no conflicts. Funding Supported by the VIRGIL EC grant QLK2-CT ; by Fondazione Cassa Risparmio di Parma, Parma, Italy; by grant 85/2007, Ricerca Finalizzata, Progetto Ordinario, Ministry of Health, Italy; and by Fondo per gli Investimenti della Ricerca di Base (grant No. RBNE013PMJ); and by the Ministry of Instruction, University and Research, Italy.

13 693.e1 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Supplementary Figure 1. PD-1 and CD127 expression by intrahepatic HBV-specific CD8 T cells. Representative dot plots showing tetramerpositive cell frequency as well as PD-1 and CD127 expression on HBV-specific CD8 cells isolated from the liver of 8 chronic HBV patients.

14 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 693.e2 Supplementary Figure 2. PD-1 mean fluorescence intensity (MFI) is higher on intrahepatic than peripheral CD8 cells. Representative examples of PD-1 MFI of HBV-specific (upper panel) and total (lower panel) CD8 cells from liver and blood are illustrated on the left of each panel. Histograms on the right of each panel illustrate mean values of PD-1 MFI from all patients. Statistical analysis was performed by the Wilcoxon paired test.

15 693.e3 FISICARO ET AL GASTROENTEROLOGY Vol. 138, No. 2 Supplementary Figure 3. PD-1 expression on intrahepatic T-cell subsets. Frequencies of PD-1-positive CD8 cells among T-cell subsets identified by different expression of CD45RA and CCR7. Panel A illustrates HBV-specific CD8 cells, whereas total CD8 cells are depicted in panel B. The highest PD-1 expression is displayed by CD45RA CCR7 and CD45RA CCR7 cells.

16 February 2010 T-CELL RESTORATION IN CHRONIC HBV INFECTION 693.e4 Supplementary Figure 4. Effect of anti-pd-l1 on IFN- and IL-2 production by peripheral blood CMV- and FLU-specific CD8 cells analyzed by tetramer staining. Circulating lymphocytes from 3 HLA-A2 chronic HBV patients were cultured in the presence of anti-pd-l1 or a control antibody and stimulated with HLA-A2 restricted CMV or FLU peptides. Expansion and cytokine production are illustrated.