Temporal Analysis of Early Immune Responses in Patients With Acute Hepatitis B Virus Infection

GASTROENTEROLOGY 2009;137:1289 1300 Temporal Analysis of Early Immune Responses in Patients With Acute Hepatitis B Virus Infection CLAIRE DUNN,* DIMITRA PEPPA,*,, POOJA KHANNA,*, GAIA NEBBIA,* MELERI JONES, NATHAN BRENDISH,* R. MONICA LASCAR,, DAVID BROWN, RICHARD J. GILSON,, RICHARD J. TEDDER,* GEOFFREY M. DUSHEIKO, MICHAEL JACOBS, PAUL KLENERMAN, and MALA K. MAINI*,,, *Division of Infection and Immunity and Centre for Sexual Health and HIV Research, University College London, London, England; Mortimer Market Centre, Camden Primary Care NHS Trust, London, England; Centre for Hepatology, Royal Free Campus, University College London Medical School, London, England; and Nuffield Department of Clinical Medicine, University of Oxford, Oxford, England See related article, Wong VW et al, on page 1113 in CGH. BACKGROUND & AIMS: Hepatitis B virus (HBV) causes more than 1 million deaths annually from immunemediated liver damage. The long incubation period has been difficult to study; by the time most patients present, massive viremia and the majority of viral clearance have already occurred. The aim of this study was to investigate the contribution of innate and adaptive immune mechanisms in early acute HBV through access to an unusual cohort of patients sampled in the preclinical phase and followed up to resolution of their infection. METHODS: Twenty-one patients with acute HBV were studied, 8 of them from before the peak of viremia. Circulating innate cytokines were quantitated by enzyme-linked immunosorbent assay and natural killer (NK) and T-cell effector function by flow cytometry. Results were correlated with temporal changes in viral load, serology, and liver inflammation and compared with healthy controls. RESULTS: Type I interferon (IFN) remained barely detectable throughout, with concentrations no higher than those found in healthy controls. Similarly, interleukin-15 and IFN- 1 were not induced during peak viremia. NK cell activation and capacity for IFN- production were reduced at peak viremia. Early functional HBV-specific CD4 and CD8 T-cell responses were attenuated as viral load increased and recovered again as infection resolved. The transient inhibition of NK and T-cell responses coincided with a surge in the immunosuppressive cytokine interleukin-10 accompanying HBV viremia. CONCLUSIONS: The early stages of acute HBV are characterized by induction of interleukin-10 rather than type I IFN, accompanied by a temporary attenuation of NK and T-cell responses. To view this article s video abstract, go to the AGA s YouTube Channel. About one third of the world s population is estimated to have been infected with the noncytopathic hepatotropic virus hepatitis B virus (HBV). Chronic infection, which can lead to liver cirrhosis and hepatocellular carcinoma, is characterized by dysfunctional innate 1 and adaptive 2 6 immune responses. In this setting, it is not possible to distinguish to what extent these aberrant responses are causal or are instead simply the end result of years of uncontrolled viremia and antigenemia. By contrast, the immune responses generated in the early stages of a viral infection can give clearer insights into the host/pathogen interaction and determinants of viral control. Much evidence implicates a crucial role for innate immune responses in the first line of defense against the prototypic acute viral infection. Such responses can make a major direct contribution to the initial dampening down and containment of viremia. 7 Early innate immune responses such as the production of type I interferons (IFN-I) can also directly assist cellular responses, aiding maturation of natural killer (NK) cells, dendritic cells (DCs), and T cells. 8 10 Support for a potential direct antiviral effect of IFN-I in HBV infection comes from the transgenic mouse model, in which systemic administration of IFN- or induction of IFN- / through injection of poly(i:c) was sufficient to inhibit HBV replication. 11 13 The recently identified IFN- has also been shown to have potential activity against HBV in the transgenic mouse model. 14 IFN- production is elicited by similar mechanisms of viral sensing to IFN-I and is preferentially produced in certain tissues, including the liver. 15 NK and NKT cells have also been shown to have potent antiviral activity against HBV in this model when activated by -galactosylceramide. 16 However, when the early phases of acute infection with HBV were studied in the chimpanzee model, they revealed a surprising lack of detection of immune response Abbreviations used in this paper: DC, dendritic cell; ELISA, enzymelinked immunosorbent assay; FBS, fetal bovine serum; FITC, fluorescein isothiocyanate; HIV, human immunodeficiency virus; IFN-I, type I interferon; IL, interleukin; NK, natural killer; PBMC, peripheral blood mononuclear cell; PE, phycoerythrin; TNF, tumor necrosis factor; TRAIL, tumor necrosis factor related apoptosis-inducing ligand. 2009 by the AGA Institute 0016-5085/09/$36.00 doi:10.1053/j.gastro.2009.06.054

1290 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4 genes in the liver during the entry and expansion phase of viral replication. 17 For example, HBV did not appear to be capable of inducing any of the IFN-I responsive genes 17 that had previously been shown to have anti-hbv activity. 13 These findings have led to the view that HBV is a stealth virus, 18 creeping undetected into the liver, where it can replicate unchecked to an extremely high level. The subsequent immune responses that then reduce viremia are initially noncytolytic, because the majority of viral clearance occurs before the peak in liver damage in both chimpanzees 19 and humans. 20 The reduction in HBV viremia was found to coincide with the production of IFN-, 17,19 a potential source of which could be an early increase in NK cells. 20 However, HBV-specific CD4 and CD8 responses are also present at the time of reduction in viral load and have been shown to play a major role in the noncytolytic clearance of HBV from infected hepatocytes. 20,21 In this study, we have addressed in humans the paradox of the invisibility of this highly replicating virus to the usual early innate antiviral mechanisms, asking whether this represents a lack of recognition of the virus or an active sabotaging of these responses. We provide data from an unusual cohort of patients, sampled in the early preclinical phase of acute HBV infection, pointing to both a lack of appropriate induction and an active suppression of early immune responses. Patients and Methods Patients Twenty-one patients with acute HBV were included in this study. Table 1 summarizes their clinical characteristics, laboratory parameters, and the number of available sera and peripheral blood mononuclear cells (PBMCs) for each patient. Infection was diagnosed on the basis of a positive HBV DNA test result by polymerase chain reaction and positive serum surface antigen (HBsAg), with the subsequent serologic, biochemical, and clinical evolution of acute HBV infection. HBsAg alone was detectable initially, followed by appearance of hepatitis B e antigen (HBeAg) and then immunoglobulin (Ig) M antibody to hepatitis B core antigen (anti-hbc) on subsequent samples. Recovery from acute infection was defined by the detection of antibody to hepatitis B e antigen and subsequently antibody to hepatitis B surface antigen (anti-hbs) and loss of HBsAg. All patients were human immunodeficiency virus (HIV), hepatitis C virus (HCV), and hepatitis delta virus negative. Within this cohort, 8 patients were recruited before the peak of HBV DNA level, at which stage they were still negative for anti-hbc IgM and in some cases also still negative for HBeAg, the usual markers for acute hepatitis B. Six of the patients from whom only sera were available (patients 4 and 10 14) were part of a previously described cohort of patients who acquired HBV from a common source of accidental inoculation. 20 During the incubation and clinical acute phases of infection, clinical assessment and blood sampling were performed every 2 4 weeks, with informed consent and local ethical board approval. At each time point, alanine aminotransferase (ALT) levels, bilirubin levels, international normalized ratio, HBV serology, and HBV DNA levels were measured and sera stored (Table 1). In addition, for a subset of 12 patients, PBMCs were isolated by gradient centrifugation on Ficoll-Hypaque and frozen or immediately studied as described later. Thirteen patients with acute hepatitis A infection (6 female patients and 7 male patients; median age, 28 years) were diagnosed on the basis of IgM-positive anti hepatitis A antibodies. They had no evidence of acute HBV or HCV infection. They had a median ALT level of 1486 IU/mL, comparable to the median level of ALT in the acute HBV cohort. They all had symptomatic infection with jaundice and a median bilirubin level of 114 mol/l. Antibodies and Reagents The antibodies CD3-Cy5.5/PerCP, CD56/fluorescein isothiocyanate (FITC), IFN- /phycoerythrin (PE), CD107a/PE, tumor necrosis factor (TNF)- /PE, tumor necrosis factor related apoptosis-inducing ligand (TRAIL)/PE, and CD69-APC (BD Biosciences, Cowley, UK) were used for flow cytometric analyses at the manufacturer s recommended concentrations. The anti interleukin (IL)-10 (ebioscience.com, San Diego, CA) and anti IL-10 receptor (BD Biosciences) antibodies for neutralization of IL-10 bioactivity were used at concentrations of 5 g/ml and 10 g/ml, respectively. Recombinant human IL-12 and recombinant human IL-18 (R&D Systems, Abingdon, England) were each used at 500 ng/ml, and recombinant human IL-10 (ebiosciences) was used at 50 ng/ml to inhibit IFN- production by NK cells. Determination of Serum Cytokine Concentrations Standard sandwich enzyme-linked immunosorbent assay (ELISA) kits were used to determine serum concentrations of the cytokines IFN- (High Sensitivity Protocol; PBL Biomedical Laboratories, Piscataway, NJ), IL-15 (R&D Systems), IFN- 1 (ebioscience.com), and IL-10 (Diaclone, Manchester, UK), where 50 L of patient serum or 100 L of supernatants were analyzed according to the manufacturers protocols. Ex Vivo Phenotyping of NK Cells PBMCs isolated from patients with acute HBV and healthy donors were incubated for 30 minutes at 4 C with fluorochrome-conjugated antibodies to CD3, CD56, TRAIL, or CD69. PBMCs were washed twice with phosphate-buffered saline (PBS) plus 1% fetal bovine serum (FBS) and fixed with PBS plus 1% FBS plus 1% paraformaldehyde (PFF buffer) before acquisition on a FACS-

Table 1. Summary of Patient Information Patient no. Maximum ALT level sampled (IU/L) Maximum HBV DNA level sampled (IU/mL) Maximum bilirubin (total) level sampled ( mol/l) Maximum international normalized ratio sampled Preclinical phase samples Symptomatic No. of PBMC samples No. of serum samples No. of samples before peak DNA No. of samples before anti-hbc IgM 1 994 9.3 10 7 16 1.08 Yes Yes 8 10 5 4 20 2 183 1.1 10 7 NA NA Yes Yes 0 15 2 3 38 3 2427 4 10 8 96 1.1 Yes Yes 8 8 3 1 37 4 2557 1.23 10 9 428 NA Yes Yes 0 6 2 2 25 5 3000 8900 407 1.3 No Yes 6 6 0 0 11 6 776 3.27 10 5 379 1.2 No Yes 5 3 0 1 28 7 525 88,000 22 NA No Yes 4 5 0 0 21 8 2376 1.2 10 7 216 1.07 No Yes 4 4 0 0 28 9 2626 1.6 10 6 93 0.99 No Yes 3 3 0 0 5 10 1150 2.5 10 8 78 NA Yes Yes 0 6 2 2 16 11 1502 5.58 10 8 639 NA Yes Yes 0 6 1 1 17 12 1627 2.22 10 8 80 NA No Yes 0 4 0 0 13 13 871 1.92 10 7 18 NA Yes NA 0 5 2 4 20 14 3709 6.54 10 7 257 NA Yes Yes 0 5 1 1 16 15 3531 60,000 172 1.1 No Yes 4 4 0 0 Years 16 78 68,000 7 0.98 No No 1 4 0 1 12 17 59 6500 9 NA No No 1 3 0 1 Years 18 3957 NA 196 NA No Yes 1 1 0 0 NA 19 3335 NA 112 1.06 No Yes 1 1 0 0 NA 20 30 64,000 7 0.9 No No 0 2 0 2 3 21 31 700,000 16 1.2 No No 0 2 0 1 3 Median 1683 1.92 10 7 94.5 1.08 NA, not available. Duration of sampling (wk) October 2009 IMMUNE RESPONSES IN ACUTE HBV 1291

1292 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4 Calibur flow cytometer (Becton Dickinson) and analyzed using FlowJo analysis software (Treestar Inc, Ashland, OR). Isotype-matched control antibodies were used for defining positive populations stained with the CD69- specific antibody. In Vitro Analysis of NK Cell Effector Function For IFN- production, PBMCs were resuspended in supplemented RPMI (Invitrogen Ltd, Paisley, UK) plus 10% FBS, plated into a round-bottom 96-well tissue culture plate (Fisher Scientific, Pittsburgh, PA) at 3 10 5 cells/well, and incubated with recombinant human IL-12 (500 ng/ml) and recombinant human IL-18 (500 ng/ml) for 21 hours at 37 C; 1 mol/l monensin (Sigma- Aldrich, Gillingham, England) was added for the final 3 hours. PBMCs were then incubated with CD56-FITC and CD3-PerCP/Cy5.5 for 30 minutes at 4 C. Cells were fixed and permeabilized at 4 C with Cytofix/Cytoperm (BD Biosciences) and then stained intracellularly for IFN- /PE in PBS plus 1% FBS plus 0.1% saponin (30 minutes at 4 C; Sigma-Aldrich). PBMCs were then washed with PBS plus 1% FBS and fixed with PFF buffer before acquisition on a FACSCalibur flow cytometer. To assess the effect of IL-10 on IFN- production by NK cells, recombinant human IL-10 (50 ng/ml) or anti IL-10 (5 g/ml) and anti IL-10 receptor (10 g/ml) blocking antibodies were added at the same time as IL-12 and IL-18. For TNF- production, PBMCs were stimulated with phorbol myristate acetate (3 ng/ml) and ionomycin (100 ng/ml; Sigma-Aldrich) for 3 hours; 1 mol/l monensin (Sigma-Aldrich) was added for the final 2 hours. Cells were then stained with the same antibody combination used for phenotyping, before permeabilization and intracellular staining for TNF- as described for IFN-. NK cell degranulation was assessed by detecting the levels of surface-exposed CD107a. PBMCs were incubated with recombinant human IL-12 (500 ng/ml) and recombinant human IL-18 (500 ng/ml) overnight before a further 3-hour coincubation at 37 C with the NK target cell line K562 (5:1 effector:target ratio). CD107a-FITC monoclonal antibody was added at the start of the 3-hour coculture incubation and 1 mol/l monensin for the last 2 hours. Cells were then stained with CD56-FITC and CD3-PerCP/Cy5.5 and fixed before acquisition on a FACSCalibur flow cytometer. NK effector function was determined by subtracting baseline IFN-, TNF-, or CD107 staining in unstimulated samples from that observed after the relevant stimulus. Ex Vivo Determination of HBV-Specific CD4 and CD8 T-Cell Responses PBMCs from all available time points from 2 patients were stimulated in parallel with pools of 15mer peptides overlapping by 10 residues spanning the major proteins of HBV genotype B (Chiron Mimotopes, Victoria, Australia, kind gift of Antonio Bertoletti). The 8 pools comprised precore and core (peptides 1 6 and 1 35), X (1 29), envelope pool 1 (1 38), envelope pool 2 (39 76), polymerase pool 1 (1 42), polymerase pool 2 (43 84), polymerase pool 3 (85 126), and polymerase pool 4 (127 167). Direct ex vivo analysis of IFN- positive virusspecific CD8 and CD4 T cells was performed following stimulation with peptides as previously described (each peptide at 2 g/ml) for 12 hours, with addition of Brefeldin A (10 g/ml) after 1 hour. Cells were stained with anti CD3-PerCPCy5.5 and CD8-APC and then permeabilized for intracellular IFN- /PE staining as previously described. 6 Background IFN- production without peptide stimulation was subtracted before calculating the percent of specific IFN- producing CD8 or CD4 (CD3 CD8 ) T cells. Protein pool-specific responses were calculated by adding responses for different pools within each HBV antigen and total CD4 or CD8 HBV-specific responses by adding responses against all pools at each time point. Statistical Analysis Statistical significance was performed between paired acute samples using the Wilcoxon signed rank test and between patients with acute HBV infection and healthy controls using the Mann Whitney U test. Results The Production of Innate Cytokines Is Impaired in Early Acute HBV Infection A cohort of patients was recruited in the preclinical phase of acute HBV infection and sampled longitudinally through the symptomatic phase to disease resolution (Table 1). Serum concentrations of IFN- were initially measured using a high-sensitivity ELISA. IFN- remained barely detectable in the circulation of these patients during the early incubation phase and throughout the peak of viral replication and subsequent reduction in viremia and onset of symptomatic liver inflammation (indicated by serum ALT level) (Figure 1A, temporal graphs). Eight patients were sampled sufficiently early (while HBV DNA level was still increasing) to be able to assess IFN- levels at the time of peak DNA, and an additional 3 patients still had active viral replication and production of HBeAg ( DNA Hi, Figure 1A, bar chart). These samples were compared with measurements taken at the first time point at which HBV DNA level was reduced to less than 0.1% of peak or undetectable and with levels in healthy donors ( Resolving and Healthy, Figure 1A). Levels of IFN- at peak viremia were not significantly greater (and in fact showed a trend to be lower) than those found after resolution of infection or in healthy donors (Figure 1A). We could not exclude the possibility that increases in IFN- levels were undetectable in HBV infection because

October 2009 IMMUNE RESPONSES IN ACUTE HBV 1293 Figure 1. The production of innate cytokines is impaired during early acute HBV infection. (A) Circulating concentrations of IFN- detected in longitudinal serum samples from 3 representative patients assayed by high-sensitivity ELISA. IFN- concentrations are plotted against liver inflammation (ALT) and viral load (HBV DNA). Changes in 3 major serologic markers of acute HBV infection (anti-hbc IgM, HBeAg, and HBsAg) are depicted as bars above the main graph. Cross-sectional comparison of circulating IFN- levels from healthy controls, patients with acute HBV at the time of peak viremia (DNA Hi) and at the time of resolution of infection (resolving), and patients with acute hepatitis A virus infection (HAV). Circulating concentrations of (B) IL-15 and (C) IFN- 1in sera from patients with acute HBV, patients with acute HAV, and healthy controls. Temporal correlation of HBV viral load and cytokine concentrations detected in longitudinal samples of 3 representative patients with acute HBV. Cross-sectional comparison of healthy donors, patients with acute HBVatthe times of peak viral load and resolution, and patients with acute HAV. Any P values of less than.05 are shown. it was all sequestered at the site of viral replication; liver biopsy was not clinically justifiable in these patients. However, serum concentrations of IFN- were quantitated in patients with another acute hepatotropic virus infection, hepatitis A, and were significantly higher than at any time point in acute HBV infection (Figure 1A). To test for IFN- and for any IFN- subtypes that may not have been detected by the ELISA, a luciferase reporter gene system was used that determines the concentrations of IFN-I based on the IFN-inducible 6-16 promoter in the stably transfected cell line HL116. Significantly less IFN-I was detected in the sera of patients with acute HBV than that of healthy donors (Supplementary Figure 1). An additional functional assay for type I antiviral activity based on its capacity to inhibit the cytopathic effect of encephalomyocarditis virus also showed a lack of induction of IFN-I in acute HBV (data not shown). Along with IFN-I, IL-15 is important for the induction of NK effector function 22 ; production of IL-15 by DCs is induced by IFN-. 23 Using a high-sensitivity ELISA, serum IL-15 was undetectable or at low levels at the point of highest viremia in our HBV cohort (Figure 1B), with concentrations equivalent to those observed in the serum of healthy donor controls (Figure 1B). A significant increase in circulating IL-15 concentrations was observed once the majority of the virus had been cleared (Figure 1B). In view of the lack of IFN-I and IL-15 at the time of viremia, we questioned whether there might instead be induction of IFN- 1 (IL-29, a type III interferon), recently shown to be produced in the liver 15 and to have antiviral potential against HBV. 14 Although IFN- 1 was circulating at higher concentrations than IFN-I, it was not induced above levels in healthy controls (Figure 1C) and, like IL-15, longitudinal analysis suggested some suppression at the time of peak viremia (Figure 1C). As noted for IFN-I, IFN- 1 was significantly elevated in the circulation of patients with hepatitis A compared with those with acute or resolving HBV (Figure 1C).

1294 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4 Figure 2. The NK cell response is delayed during early acute HBV infection. (A) PBMC samples were stained ex vivo for CD3, CD56, and the activation marker CD69, and the proportion of activated CD3 CD56 NK cells was identified by flow cytometry. CD69 NK cells are presented as a percentage of total NK cells. Temporal correlation of NK cell activation, HBV viral load, and ALT level for 3 representative patients. Cross-sectional comparison of ex vivo NK cell activation of healthy controls and patients with acute HBV at time of peak viral load (DNA Hi) and at time of resolution (Resolving). (B) PBMCs were stimulated in vitro with IL-12 and IL-18 and NK cell IFN- production determined by intracellular cytokine staining. Temporal correlation of NK cell IFN- production, HBV DNA levels, and ALT levels for 3 representative patients. Cross-sectional comparison of IFN- production by NK cells from healthy controls and patients with acute HBV at time of peak viral load (DNA Hi) and at the time of resolution (Resolving). Temporal and cross-sectional analysis as in B of (C) NK cell TNF- production and (D) CD107 expression following stimulation protocols described in Patients and Methods. (E) The percent of NK cells expressing TRAIL and (F) the percent of CD56bright NK cells were calculated directly ex vivo and plotted according to viral load kinetics and for cross-sectional comparison.

October 2009 IMMUNE RESPONSES IN ACUTE HBV 1295 Maximal NK Cell Activation and Antiviral Function Are Reduced at Peak Viremia To assess the activation status of NK cells, we analyzed ex vivo expression of the activation marker CD69 on the CD3 CD56 subset in patients in whom longitudinal PBMC samples were available (Table 1). Temporal correlation with disease evolution showed an inhibition of NK cell activation at the time of peak HBV viremia (Figure 2A, temporal graphs), with a significant decrease in activation at peak viral load compared with at the time of resolution of infection (Figure 2A, bar chart). NK activation peaked around or after the time of maximal liver inflammation, as viremia was resolving (Figure 2A). Comparison with levels of activation of circulating Figure 3. Induction of IL-10 in acute HBV. (A) Temporal correlation of circulating concentrations of IL-10 (assayed by sandwich ELISA), liver inflammation (ALT), and viral load (HBV DNA) in longitudinal samples from 3 representative patients. Changes in HBeAg and HBsAg are shown as bars at the top of the graphs. Cross-sectional comparison of circulating IL-10 concentrations obtained for healthy controls and patients with acute HBV at time of peak viremia (DNA Hi) and at time of resolution (Resolving). (B) Temporal correlation of circulating concentrations of IL-10, ALT, and HBV DNA in longitudinal samples from 2 patients with asymptomatic acute HBV infection. Changes in HBeAg, HBsAg, and anti-hbc antibodies are shown as bars at the top of the graphs. (C) Temporal correlation of circulating concentrations of IL-10, IFN- production by NK cells, and viral load (HBV DNA) in longitudinal samples from 3 representative patients. (D) PBMCs from patients with acute HBV were stimulated overnight with IL-12 and IL-18 (500 ng/ml each), with or without IL-10 (50 ng/ml) or anti IL-10 (5 mg/ml) and anti IL-10 receptor (10 mg/ml) blocking antibodies. NK cell derived IFN- was determined by intracellular cytokine staining. NK cells were identified as CD3 CD56 by flow cytometry. Shown are 3 representative dot plots taken from one patient and a summary bar chart.

1296 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4

October 2009 IMMUNE RESPONSES IN ACUTE HBV 1297 NK cells in healthy donors revealed a significant inhibition of in vivo activation of NK cells at all stages of acute HBV infection (Figure 2A). NK cells from our cohort of patients were then stimulated in vitro with IL-12 and IL-18 and assessed for production of the antiviral cytokine IFN-. Longitudinal analysis showed that the maximal potential for NK cells to produce IFN- was reached once viral load began to decrease (Figure 2B). Cross-sectional analysis confirmed a significant increase in the capacity of NK cells to produce IFN- at the time of resolution compared with at peak viremia (Figure 2B). A similar pattern was seen for the capacity of NK cells to produce the antiviral cytokine TNF- (Figure 2C), with an early increase in some patients but a trend for the maximum levels to be reached after reduction in viremia. Cytolytic potential of NK cells, as assessed by the expression of CD107, also increased after reduction of viremia in some patients but overall showed no significant correlation with viral load or liver inflammation (Figure 2D). We then examined NK cell expression of TRAIL, which is induced on CD56bright NK cells during flares of eag-negative chronic HBV infection. 1 Both NK cell TRAIL and the percentage of CD56bright NK cells showed a trend to increase after the peak of viremia, at the time of liver inflammation (Figure 2E and F and data not shown). Overall, peak HBV replication was associated with inhibition of NK cell activation and a predominant reduction in noncytolytic rather than cytolytic potential. Early Production of the Immunosuppressive Cytokine IL-10 The above data suggested that a mechanism of suppression of innate immune responses might be activated by HBV replication. IL-10 is a potent immunosuppressive cytokine that can inhibit both innate and adaptive immunity. 24 Longitudinal assessment of circulating IL-10 levels by a high-sensitivity ELISA showed a close temporal correlation of circulating IL-10 and HBV DNA levels (Figure 3A). IL-10 increased early in the course of infection, in parallel with the rapid increase in HBV viral load and antigenemia, and before the onset of liver inflammation. In some cases, IL-10 remained elevated during the period of liver inflammation (Figure 3A). The reduction in IL-10 coincided with the termination of viremia or with clearance of the secretory form of the HBV nucleocapsid, HBeAg. Cross-sectional analysis of circulating IL-10 concentrations revealed that, unlike the other cytokines measured, the highest concentrations of IL-10 were found at the time of peak viremia, and concentrations were significantly higher than those found in healthy controls (Figure 3A, bar chart). Once infection resolved, IL-10 levels reverted to those seen in healthy donors (Figure 3A). To further investigate the relationship between the severity of infection and IL-10 induction, we analyzed 2 patients who developed a mild asymptomatic course of acute HBV infection. Although the existing literature has focused on patients with classic highly viremic, symptomatic acute HBV infection, most cases of adult infection are mild and asymptomatic. 25 Both patients had only transient low-grade viremia, with HBeAg detected briefly or not at all, and no increase in ALT level (Figure 3B). In contrast to the patients with symptomatic acute HBV characterized by high-level viral replication accompanied by IL-10 induction, these patients had no detectable levels of IL-10 (Figure 3B). In keeping with this, a recent study of 2 patients with a similar asymptomatic course showed no inhibition of NK cell activity at peak viremia. 26 In the patients with symptomatic acute HBV, an inverse temporal correlation was noted between IL-10 levels and NK cell IFN- production (Figure 3C), suggesting a role for IL-10 in the suppression of the NK cell antiviral response. To test this, NK cells were stimulated in vitro with IL-12 and IL-18, with or without the addition of exogenous IL-10 or anti IL-10/IL-10 receptor blocking antibodies. IL-10 induced significant suppression of NK cell derived IFN-, whereas IL-10 blocking restored NK effector function (Figure 3D). HBV-Specific T-Cell Responses Are Detected Early and Attenuated After Induction of IL-10 Among the patients with early acute HBV infection from whom we were able to serially quantitate serum IL-10 were two from whom we had previously reported HBV-specific CD8 responses studied at the same time points. 20 In light of the potential for IL-10 to influence T-cell responses, 24 we investigated the relationship between HBV-specific CD8 responses and serum IL-10 kinetics in these patients. These patients were HLA-A2 4 Figure 4. Correlation of HBV-specific CD4 and CD8 T-cell responses with IL-10 levels in early acute HBV. (A) Correlation of IL-10 with ex vivo CD8 T-cell responses quantitated with HLA-A2 peptide tetramers. Tetramer-staining responses for 2 patients against core 18-27 (left panels) and pol 575-83 (right panels) 20 were plotted against viral load and IL-10. (B) Correlation of IL-10 with ex vivo functional CD8 T-cell responses against overlapping peptides spanning the HBV genome. (Upper panel) Total CD8 responses were calculated by summing IFN- positive responses to each of the 8 pools of overlapping peptides spanning the HBV genome and plotted (as responses per 500 CD8 T cells) against IL-10 and HBV DNA. *ND, not done because of insufficient PBMCs. (Lower panel) IFN- positive responses were calculated according to HBV protein specificity by summing responses for pools covering each protein and plotted (as responses per 1000 CD8 T cells) against IL-10 and HBV DNA. (C) Correlation of IL-10 with ex vivo IFN- positive CD4 T-cell responses against overlapping peptides spanning the HBV genome. (Upper panel) Total responses and (lower panel) breakdown according to HBV protein specificity, both calculated and plotted as for B.

1298 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4 positive and had responses detectable directly ex vivo to 2 well-defined HLA-A2 restricted HBV epitopes (core 18-27 and pol 575-83) using HLA-A2/peptide tetramers, peaking after clearance of the majority of HBV. 20 Plotting these data against serum IL-10 showed that CD8 specific for these 2 epitopes peaked in one patient while IL-10 levels were still elevated, while in the other, tetramerbinding CD8 only increased once IL-10 levels decreased (Figure 4A). We had sufficient PBMCs available from 2 HLA-A2 negative patients sampled from early in the course of infection to perform a more comprehensive analysis of functional HBV-specific T-cell responses using overlapping peptides spanning the whole HBV genome. CD8 and CD4 T-cell responses were determined by production of IFN- after 12 hours of stimulation with overlapping peptides divided in 8 pools and analyzed at 7 time points for each patient according to changes in viral load and IL-10 levels (Figure 4B and C and Supplementary Figure 2). Total functional HBV-specific CD8 responses (calculated by summing individual IFN- positive responses to each peptide pool, upper panels, Figure 4B) peaked at the first sample points and again after reduction in viral load and IL-10 levels, showing a transient attenuation at the time of peak viremia and IL-10. Differences in temporal evolution of CD8 responses directed to peptides spanning the different HBV proteins are shown in the lower panels of Figure 4B. In both patients, the predominant responses seen at the first time point tested (when HBV viremia was just starting to increase) were directed against the nonstructural proteins polymerase and X, whereas responses to the structural proteins core and envelope correlated with the resolution of infection (lower panels, Figure 4B). Total HBV-specific functional CD4 responses also showed an inverse correlation with HBV load and IL-10 levels (upper panels, Figure 4C). As with the CD8 responses, specificities within the nonstructural protein X dominated the CD4 response seen before peak DNA, whereas responses to core or envelope became more obvious after most HBV DNA had been cleared (lower panels, Figure 4C). Discussion The immune responses generated in the primary stages of a viral infection are believed to be critical determinants of subsequent disease outcome. Acute infection with HBV has a prolonged, clinically silent incubation phase; by the time of typical presentation with jaundice, the majority of viral clearance has already occurred. In this study, we took advantage of a cohort of patients who had been opportunistically sampled in the presymptomatic phase of initial viremia to examine the potential primary antiviral role of innate immune responses. We found that IFN-I, IL-15, and IFN- 1 were not appropriately induced in response to HBV infection and that the peak of viremia coincided with an inhibition of NK cell activation and effector potential. Functionally active virus-specific CD4 and CD8 T-cell responses were detectable from the first sampling point, were transiently attenuated, and then recovered as viral load decreased. The period of reduced NK and T-cell responses coincided with maximal HBV replication and a surge of IL-10 production. Thus, we propose that, in addition to failing to induce some immune mechanisms, HBV uses an immunosuppressive strategy to actively inhibit others. The lack of a noteworthy IFN-I response in acute HBV infection is in keeping with the findings in chimpanzees experimentally infected with HBV 17 and contrasts with its marked early induction in HIV infection. 27 This may be partially attributable to specific features of the replication strategy of HBV, allowing it to escape recognition by the cellular sensing machinery. 28 Our finding that IFN/IL-15 levels tended to be lower at peak viremia than at disease resolution raises the possibility of an additional active inhibition of production. A number of viruses have evolved strategies to evade the IFN-I response, inhibiting its production or its downstream effects. 29 Inhibition of Toll-like receptor triggering of IFN-I from plasmacytoid DC (pdc) can be mediated by the antiinflammatory cytokine IL-10. 30 Attenuated pdc production of IFN-I in chronic HCV has been attributed in part to the action of IL-10, produced by monocytes in response to the core protein of HCV. 31 Similarly, liver macrophages (Kupffer cells) can exhibit TLR-triggered, IL-10 dependent cross talk with innate immune responses, resulting in the down-regulation of NK cell IFN- production. 32 An IL-10 mediated inhibition of NK cell activation and noncytolytic potential in the patients we studied is supported both by the tight temporal correlation between these variables directly ex vivo and by the impact of blocking IL-10 on the restitution of NK cell effector function in vitro. IL-10 can also inhibit specific T-cell responses both directly 33 and via effects on DCs, 24 consistent with its temporal correlation with attenuated functional HBV-specific T-cell responses. Other factors, such as the likely increased propensity of HBV-specific T cells to undergo PD-1 mediated anergy 34 or activationinduced cell death at the peak of antigenic stimulation, 35 may also play a role. Potential cellular sources of IL-10 include Kupffer cells 32 and virus-specific T cells; 36 38 recent work suggests IL-10 may be derived from multiple different cell types in the context of a human chronic viral infection. 39 Although IL-10 may play an anti-inflammatory role, it is unlikely to be solely triggered in response to tissue damage because its induction can precede the onset of liver inflammation in some patients. Instead, it closely parallels the onset of viremia and production of viral nucleocapsid antigens. The fact that the patients with milder and more transient asymptomatic acute HBV viremia/antigenemia had no detectable IL-10

October 2009 IMMUNE RESPONSES IN ACUTE HBV 1299 suggests that there may be a threshold effect for its induction. A limitation of our study is the lack of access to samples from the intrahepatic compartment, the site of viral replication; liver biopsies are not generally clinically justifiable in acute HBV infection in humans. We therefore cannot exclude some induction of IFN-I or localized early activation of NK cells in the liver of these patients. However, changes in the circulation have been found to closely mirror those in the liver in acute HBV in chimpanzees 19 ; in line with this, we have previously been able to detect IFN- and NK cell activation in the circulation of patients with hepatic flares of chronic HBV infection. 1 The possibility of complete sequestration of IFN-I in the liver in acute HBV is also not supported by our ability to detect IFN- at high levels in the circulation of patients with hepatitis A, another acute hepatotropic virus infection. Another caveat to our study is that we were not able to sample immune events in the first month after infection, before the logarithmic rise in HBV DNA levels. During this initial phase after viral entry, there appears be a temporary block to HBV replication and spread. 40 It remains possible that this is partially mediated by innate immune mechanisms, as suggested by some of our patients with elevated IL-15 and NK cell activation/function just before the expansion phase. Recent data are in line with this hypothesis, showing activation of NK and NKT cells within 72 hours of experimental infection with woodchuck hepadnavirus, resulting in transient suppression of viral replication. 41 What are the clinical implications of the immunologic changes documented in the preclinical stages of acute HBV infection? The lack of IFN-I induction and temporary disarming of other immune responses may help to explain how HBV can replicate unchecked to such extreme levels of viremia; in a typical case of acute symptomatic HBV, the majority of hepatocytes are likely to be infected 21 and peak viral load is at least a log higher than in acute HCV or HIV. 27 Despite the lack of dampening down of viremia by an IFN response, more than 95% of adults infected with HBV manage to control the infection. Although IFN-I are conventionally believed to promote the development of adaptive immune responses by limiting viral spread 7 and enhancing cross-priming, 42 they conversely also have the potential to impair them and promote viral evasion by acting through STAT2 to inhibit DC development and expansion. 43 Thus, the lack of an initial IFN-I response may paradoxically allow more effective DC priming of virus-specific T-cell responses, 44,45 consistent with the early detection of HBV-specific T-cell responses. Our data support the concept that clearance of acute HBV is heavily dependent on successful adaptive responses, in line with chimpanzee 21 and human 20,46 data. The transitory decline in NK cell and T-cell responses may provide a window of opportunity for viral escape but does not prevent ultimate viral control in the majority of acutely infected patients. Instead, the induction of IL-10 may represent a mechanism that delays and limits the extent of liver damage, analogous to the IL- 10 mediated reduction in lung pathology in acute influenza virus infection, 38 such that only a small minority of patients develop lethal fulminant hepatitis. However, if IL-10 continues to be induced in patients with established chronic HBV infection, this could combine with other described tolerizing mechanisms 4 6 to contribute to the maintenance of viral persistence, as shown in murine models. 47,48 Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at doi: 10.1053/j.gastro.2009.06.054. References 1. Dunn C, Brunetto M, Reynolds G, et al. Cytokines induced during chronic hepatitis B virus infection promote a pathway for NK cell-mediated liver damage. J Exp Med 2007;204:667 680. 2. 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;191:1269 1280. 3. Reignat S, Webster GJ, Brown D, et al. Escaping high viral load exhaustion: CD8 cells with altered tetramer binding in chronic hepatitis B virus infection. J Exp Med 2002;195:1089 1101. 4. 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;81:4215 4225. 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;205:2111 2124. 6. 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;118:1835 1845. 7. Muller U, Steinhoff U, Reis LF, et al. Functional role of type I and type II interferons in antiviral defense. Science 1994;264:1918 1921. 8. Andoniou CE, van Dommelen SL, Voigt V, et al. Interaction between conventional dendritic cells and natural killer cells is integral to the activation of effective antiviral immunity. Nat Immunol 2005;6:1011 1019. 9. Biron CA. Interferons alpha and beta as immune regulators a new look. Immunity 2001;14:661 664. 10. Le Bon A, Durand V, Kamphuis E, et al. Direct stimulation of T cells by type I IFN enhances the CD8 T cell response during cross-priming. J Immunol 2006;176:4682 4689. 11. McClary H, Koch R, Chisari FV, et al. Relative sensitivity of hepatitis B virus and other hepatotropic viruses to the antiviral effects of cytokines. J Virol 2000;74:2255 2264. 12. Wieland SF, Guidotti LG, Chisari FV. Intrahepatic induction of alpha/beta interferon eliminates viral RNA-containing capsids in hepatitis B virus transgenic mice. J Virol 2000;74:4165 4173. 13. Wieland SF, Vega RG, Muller R, et al. Searching for interferoninduced genes that inhibit hepatitis B virus replication in transgenic mouse hepatocytes. J Virol 2003;77:1227 1236. 14. Robek MD, Boyd BS, Chisari FV. Lambda interferon inhibits hepatitis B and C virus replication. J Virol 2005;79:3851 3854. 15. Sommereyns C, Paul S, Staeheli P, et al. IFN-lambda (IFNlambda) is expressed in a tissue-dependent fashion and primarily acts on epithelial cells in vivo. PLoS Pathog 2008;4:e1000017.

1300 DUNN ET AL GASTROENTEROLOGY Vol. 137, No. 4 16. Kakimi K, Guidotti LG, Koezuka Y, et al. Natural killer T cell activation inhibits hepatitis B virus replication in vivo. J Exp Med 2000;192:921 930. 17. Wieland S, Thimme R, Purcell RH, et al. Genomic analysis of the host response to hepatitis B virus infection. Proc Natl Acad Sci U S A 2004;101:6669 6674. 18. Wieland SF, Chisari FV. Stealth and cunning: hepatitis B and hepatitis C viruses. J Virol 2005;79:9369 9380. 19. Guidotti LG, Rochford R, Chung J, et al. Viral clearance without destruction of infected cells during acute HBV infection. Science 1999;284:825 829. 20. Webster GJ, Reignat S, Maini MK, et al. Incubation phase of acute hepatitis B in man: dynamic of cellular immune mechanisms. Hepatology 2000;32:1117 1124. 21. Thimme R, Wieland S, Steiger C, et al. CD8( ) T cells mediate viral clearance and disease pathogenesis during acute hepatitis B virus infection. J Virol 2003;77:68 76. 22. Nguyen KB, Salazar-Mather TP, Dalod MY, et al. Coordinated and distinct roles for IFN-alpha beta, IL-12, and IL-15 regulation of NK cell responses to viral infection. J Immunol 2002;169:4279 4287. 23. Santini SM, Lapenta C, Logozzi M, et al. Type I interferon as a powerful adjuvant for monocyte-derived dendritic cell development and activity in vitro and in Hu-PBL-SCID mice. J Exp Med 2000;191:1777 1788. 24. Moore KW, de Waal Malefyt R, Coffman RL, et al. Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol 2001;19: 683 765. 25. McMahon BJ, Alward WL, Hall DB, et al. Acute hepatitis B virus infection: relation of age to the clinical expression of disease and subsequent development of the carrier state. J Infect Dis 1985; 151:599 603. 26. Fisicaro P, Valdatta C, Boni C, et al. Early kinetics of innate and adaptive immune responses during hepatitis B virus infection. Gut 2009;58:974 982. 27. Stacey AR, Norris PJ, Qin L, et al. Induction of a striking systemic cytokine cascade prior to peak viremia in acute human immunodeficiency virus type 1 infection, in contrast to more modest and delayed responses in acute hepatitis B and C virus infections. J Virol 2009;83:3719 3733. 28. Guidotti LG, Chisari FV. Immunobiology and pathogenesis of viral hepatitis. Annu Rev Pathol 2006;1:23 61. 29. Katze MG, He Y, Gale M Jr. Viruses and interferon: a fight for supremacy. Nat Rev Immunol 2002;2:675 687. 30. Duramad O, Fearon KL, Chan JH, et al. IL-10 regulates plasmacytoid dendritic cell response to CpG-containing immunostimulatory sequences. Blood 2003;102:4487 4492. 31. Dolganiuc A, Chang S, Kodys K, et al. Hepatitis C virus (HCV) core protein-induced, monocyte-mediated mechanisms of reduced IFN-alpha and plasmacytoid dendritic cell loss in chronic HCV infection. J Immunol 2006;177:6758 6768. 32. Tu Z, Bozorgzadeh A, Pierce RH, et al. TLR-dependent cross talk between human Kupffer cells and NK cells. J Exp Med 2008;205: 233 244. 33. Biswas PS, Pedicord V, Ploss A, et al. Pathogen-specific CD8 T cell responses are directly inhibited by IL-10. J Immunol 2007; 179:4520 4528. 34. Zhang Z, Zhang JY, 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: 1938 1949, 1949 e1 3. 35. Maini MK, Boni C, Ogg GS, et al. Direct ex vivo analysis of hepatitis B virus-specific CD8( ) T cells associated with the control of infection. Gastroenterology 1999;117:1386 1396. 36. Abel M, Sene D, Pol S, et al. Intrahepatic virus-specific IL-10- producing CD8 T cells prevent liver damage during chronic hepatitis C virus infection. Hepatology 2006;44:1607 1616. 37. Kaplan DE, Ikeda F, Li Y, et al. Peripheral virus-specific T-cell interleukin-10 responses develop early in acute hepatitis C infection and become dominant in chronic hepatitis. J Hepatol 2008; 48:903 913. 38. Sun J, Madan R, Karp CL, et al. Effector T cells control lung inflammation during acute influenza virus infection by producing IL-10. Nat Med 2009;15:277 284. 39. Brockman MA, Kwon DS, Tighe DP, et al. IL-10 is upregulated in multiple cell types during viremic HIV infection and reversibly inhibits virus-specific T cells. Blood 2009;114:346 356. 40. Bertoletti A, Ferrari C. Kinetics of the immune response during HBV and HCV infection. Hepatology 2003;38:4 13. 41. Guy CS, Mulrooney-Cousins PM, Churchill ND, et al. Intrahepatic expression of genes affiliated with innate and adaptive immune responses immediately after invasion and during acute infection with woodchuck hepadnavirus. J Virol 2008;82:8579 8591. 42. Le Bon A, Etchart N, Rossmann C, et al. Cross-priming of CD8 T cells stimulated by virus-induced type I interferon. Nat Immunol 2003;4:1009 1015. 43. Hahm B, Trifilo MJ, Zuniga EI, et al. Viruses evade the immune system through type I interferon-mediated STAT2-dependent, but STAT1-independent, signaling. Immunity 2005;22:247 257. 44. Bahl K, Kim SK, Calcagno C, et al. IFN-induced attrition of CD8 T cells in the presence or absence of cognate antigen during the early stages of viral infections. J Immunol 2006;176:4284 4295. 45. Robbins SH, Bessou G, Cornillon A, et al. Natural killer cells promote early CD8 T cell responses against cytomegalovirus. PLoS Pathog 2007;3:e123. 46. Urbani S, Boni C, Amadei B, et al. Acute phase HBV-specific T cell responses associated with HBV persistence after HBV/HCV coinfection. Hepatology 2005;41:826 831. 47. Brooks DG, Trifilo MJ, Edelmann KH, et al. Interleukin-10 determines viral clearance or persistence in vivo. Nat Med 2006;12: 1301 1309. 48. Ejrnaes M, Filippi CM, Martinic MM, et al. Resolution of a chronic viral infection after interleukin-10 receptor blockade. J Exp Med 2006;203:2461 2472. Received December 19, 2008. Accepted June 18, 2009. Reprint requests Address requests for reprints to: Mala K. Maini, PhD, FRCP, Division of Infection and Immunity, Department of Immunology and Molecular Pathology, Windeyer Building, 46 Cleveland Street, University College London, London, England W1T 4JF. e-mail: m.maini@ucl.ac.uk; fax: (44) 0 20 7679 9652. Acknowledgments The authors thank the staff and patients at the Mortimer Market Centre, Camden Primary Care NHS Trust, and the Royal Free Hospital Centre for Hepatology for clinical samples, as well as George Webster and Antonio Bertoletti for agreeing to use of their published T-cell data in Figure 4A. Conflicts of interest The authors disclose no conflicts. Funding Supported by the Medical Research Council (New Investigator Award G0501132 to C.D. and Clinician Scientist Award G108515 to M.K.M.). D.P. was supported by fellowship E005 from the UCLH Clinical Research and Development Committee. P.K. was supported by the Wellcome Trust, the NIHR Biomedical Research Centre Programme, and the James Martin 21st Century School.