Potential conflict of interest: Nothing to report.
The strength of antiviral T cell responses correlates with clearance of hepatitis B virus (HBV) infection, but the immunological mechanisms mitigating or suppressing HBV-specific T cells are still poorly understood. In this study, we examined the role of CD4+ Foxp3+ regulatory T cells (Tregs) in a mouse model of acute HBV infection. We initiated HBV infection via an adenoviral vector transferring a 1.3-fold overlength HBV genome (AdHBV) into transgenic DEREG mice, where Tregs can be transiently but selectively depleted by injection of diphtheria toxin. The effect of Treg depletion on the outcome of HBV infection was characterized by detailed virological, immunological, and histopathological analysis. Numbers of Tregs increase in the liver rapidly after initiation of HBV replication. Initial depletion of Tregs revealed their complex regulatory function during acute infection. Tregs mitigated immunomediated liver damage by down-regulating the antiviral activity of effector T cells by limiting cytokine production and cytotoxicity, but did not influence development of HBV-specific CD8 T cells or development of memory T cells. Furthermore, Tregs controlled the recruitment of innate immune cells such as macrophages and dendritic cells to the infected liver. As a consequence, Tregs significantly delayed clearance of HBV from blood and infected hepatocytes. Conclusion: Tregs limit immunomediated liver damage early after an acute infection of the liver, thereby contributing to conservation of tissue integrity and organ function at the cost of prolonging virus clearance. (HEPATOLOGY 2012;56:873–883)
Hepatitis B is a major global health problem caused by the human hepatitis B virus (HBV). Up to 10% of adults and >90% of neonates infected with this virus develop chronic infection, correlating with T cell dysfunction and hyporesponsiveness.1 Currently, about 350 million people worldwide are chronic HBV carriers, and >0.5 million die every year due to HBV-associated liver disease and hepatocellular carcinoma. Our knowledge about the mechanisms resulting in HBV persistence and disease pathogenesis is limited due to the lack of suitable model systems and appropriate patient material for immunological studies. Chronically infected patients are not able to launch strong and polyclonal CD8+ and CD4+ T cell responses, which are essential for clearance of viral infection from the liver.1 Several possible explanations for this lack of antigen-specific cellular immunity against chronic viral infection in the liver have been put forward. Negative selection of HBV-specific T cells, immunological ignorance or anergy—as a result of continuous exposure to high levels of viral antigens—or impaired activation of innate immunity may result in T cell hyporesponsiveness. Furthermore, the liver provides an inherently tolerizing immunological environment,2 where antigen-presenting cells skew immune responses generated in the liver toward tolerance and limit effector function of T cells by expression of coinhibitory molecules such as B7H1. It remains an open question, however, whether regulatory T cells (Tregs) that have been defined as key cell population in limiting antigen-specific immunity,3 contribute to severity of liver damage and influence the outcome of acute infection. CD4+Foxp3+ Tregs were originally shown to be a specialized T cell subpopulation suppressing autoreactive T cell responses and maintaining immunological tolerance.4 This T cell subset was first characterized by constitutive surface expression of the interleukin-2 receptor alpha (CD25),5, 6 but CD25 is also expressed on activated conventional effector T cells limiting specific depletion.7 Identification of the transcription factor Foxp3 controlling function in natural as well as in induced Tregs, provided a specific marker for Tregs in mice and humans.8, 9
Studies in viral and bacterial infections, tumor rejection and autoimmunity demonstrated that Tregs suppress proliferation, cytokine production and cytotoxic activity of naïve and antigen-specific CD4+ and CD8+ effector T cells and are able to interfere with the activity of antigen-presenting cells as well as B cells.3
Studies addressing the role of Tregs in HBV infection mostly rely on correlation of Treg frequencies in peripheral blood of patients with different disease stages and have been somewhat contradictory.10-12 Therefore, we aimed at defining the overall effect that Tregs impose on the adaptive and innate immune response against HBV and at determining how they may influence the outcome of infection. For our study, we used DEREG mice. DEREG mice are transgenic C57BL/6 mice that express an enhanced green fluorescent protein-human diphtheria toxin receptor fusion protein under control of the foxp3-promotor.13 Foxp3+ Tregs can be depleted in DEREG mice by injecting diphtheria toxin (DTX) systemically and specifically, albeit only transiently.13 Because HBV cannot infect murine hepatocytes, we used an adenoviral vector transferring a 1.3-fold overlength HBV-genome (AdHBV) across the species barrier.14 Following Ad-HBV infection, HBV replicates in hepatocytes and infectious HBV virions are secreted into the bloodstream. Depending on the dose of the inoculum, induction of T cell responses leads to an acute, self-limiting or a persistent HBV infection.14, 15 This study investigates the regulatory effects of Tregs on the intrahepatic HBV-specific T cell and innate immune response, and on the B cell response in the early phase of HBV infection.
ALT, alanine aminotransferase; BFA, brefeldin A; DC, dendritic cell; DTX, diphtheria toxin; HBc, HBV core protein; HBeAg, hepatitis B e antigen; HBs, HBV small surface protein; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; IFNγ, interferon-γ; i.u., infectious units; k/o, knockout; LAL, liver-associated lymphocyte; MHC, major histocompatibility complex; NK, natural killer; PCR, polymerase chain reaction; RPMI 1640, Roswell Park Memorial Institute 1640; TNF, tumor necrosis factor; Tregs, regulatory T cells.
Materials and Methods
AdHBV and AdHBV knockout (k/o) were constructed, produced, purified, and titrated as described.14, 15 All animal experiments were approved by the local authorities and animals received human care in accordance to the National Institutes of Health guidelines. Eight-week-old female DEREG mice received a single injection of 109 i.u. AdHBV intravenously. For depletion of Tregs, 1 μg diphtheria toxin (Merck, Darmstadt, Germany) was injected intraperitoneally on 3 consecutive days. Mice were sacrificed at indicated time points. To isolate splenocytes, spleens were minced through 100 μm cell strainers and washed with Roswell Park Memorial Institute 1640 (RPMI 1640)/10% fetal bovine serum medium after erythrocyte lysis. To prepare liver-associated lymphocytes (LALs), livers were perfused with phosphate-buffered saline before mincing through 100-μm cell strainers. After washing, cells were sedimented at 300g for 5 minutes at 4°C. Cell pellets were suspended in 50-mL collagenase-medium (William's medium E + 70 μL 2.5 M CaCl2 + 220 U/mL collagenase type IV; Worthington, Lakewood, CO) and digested for 20 minutes at 37°C. Resulting cell suspension was layered onto 8 mL Biocoll separating solution (Biochrom, Berlin, Germany) and centrifuged at 300g for 17 minutes at 4°C without breaks. Lymphocytes contained in the supernatant were collected and washed and cultivated in RPMI 1640/10% fetal bovine serum. All cell culture media and supplements were obtained from Invitrogen (Carlsbad, CA).
Alanine aminotransferase (ALT) activity was measured in 32 μL murine serum using a Reflovet Plus reader (Roche Diagnostics, Mannheim, Germany). Hepatitis B surface antigen (HBsAg), hepatitis B e antigen (HBeAg) and HBV small surface protein (HBs) antibodies (anti-HBs) were quantified in 1:20 dilutions of murine serum using AXSYM assays (Abbott Laboratories, Abbott Park, IL). For quantification of serum HBV titers, DNA was extracted from 50 μL murine serum using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) and subjected to real-time polymerase chain reaction (PCR) quantification as described.16
Liver tissue samples were fixed in 4% buffered formalin for 48 hours and embedded in paraffin. Tissue sections (4 μm) were stained with CD3-specific or HBV core protein (HBc)–specific antibodies (Diagnostic Biosystems, Pleasanton, CA). Semiquantitative analysis of stained sections was performed by counting localization, intensity, distribution, and percentage of positive cell staining throughout the whole tissue specimen.
To detect gene transcripts, 50 mg of liver tissue were homogenized in 1 mL Trizol reagent (Invitrogen), and RNA was extracted, and 750 ng of total liver RNA was reverse-transcribed into complementary DNA using the Super Script III First-Strand Synthesis Super Mix (Invitrogen). Real-time PCR was performed on a Light Cycler 480 II using the SYBR Green I Master Mix (Roche Diagnostics). Immune marker gene transcripts were analyzed relative to glyceraldehyde 3-phosphate dehydrogenase transcripts using exon-exon spanning primers and normalized to expression levels in liver tissue of naïve mice as described.15
LALs were stimulated in the presence of brefeldin A (Invitrogen) for 5 hours using HBc- and HBs-peptide libraries (15-mers overlapping by 11 amino acids; Thinkpeptides, Oxford, UK) spanning the whole protein sequence (genotype D). Each library was divided into three pools (HBsP1, S peptides 1-18; HBsP2, 19-36; HBsP3, 37-54; HBcP1, core peptides 1-15; HBcP2, 16-30; HBcP3, 31-43). Because HBcP3 proved to be immunodominant, it was used for further experiments. To assess CD137 expression, LALs were stimulated for 24 hours with recombinant HBc (kindly provided by P. Pumpens and A. Dishlers, Riga, Latvia), HBs protein (kindly provided by Rhein Biotech AG, Düsseldorf, Germany) or ovalbumin. For flow cytometry analysis, cells were stained with ethidium monoacide (Invitrogen) and anti–mCD3-V500, anti–mCD4-eFluor450, anti–mCD8-eFluor780, anti–mNK1.1-PerCP-Cy5.5, anti–mTNFα-PE-Cy7, anti–mIL2-APC, anti–mIFNγ-PE, anti–mFoxp3-AlexaFluor647, anti–mCD62L-PE-Cy7, anti–mCD127-APC, anti–m33D1-PE, anti–mF4/80-APC, anti–mMHCII-eFluor450, anti–mNK1.1-APC, anti–mCD137-PE, respectively (eBioscience, San Diego, CA). HBV multimers HBc93-100 and HBs190-197 were produced as described.17 For intracellular staining, cells were permeabilized and fixed after surface staining using the BD Cytofix/Cytoperm Kit (BD Biosciences, Heidelberg, Germany). Flow cytometrical analysis was performed on a FACSCanto II (BD Biosciences).
Data are expressed as the mean and SD. Results are analyzed using the Student t test. A P value of ≤0.05 was considered significant.
Tregs Limit T Cell Infiltration into the Liver and Liver Damage.
Infection with AdHBV but not with control AdHBV k/o induced a rapid increase of Treg frequencies (day 3) in the liver and subsequently (day 7) an increase in Treg numbers (Fig. 1A). Increased Treg frequencies were first observed in the liver and only at day 21 postinfection in the spleen (Fig. 1B) where we detected no antigen expression (data not shown). This indicated a local expansion of Tregs in the liver as the site of antigen expression before recruitment of additional Tregs. To study Treg function during experimental AdHBV-infection, we injected DEREG mice intraperitoneally with DTX shortly before and on 2 days following intravenous infection with AdHBV (Fig. 1C) efficiently depleting Tregs from liver and spleen in AdHBV-infected DEREG mice (Supporting Fig. 1A). Shortly after depletion (day 7), Tregs started to re-expand (Supporting Fig. 1B) and frequently lost green fluorescent protein expression, indicating selection of transgen-negative Tregs (Supporting Fig. 1C). We systematically analyzed other time points of Treg elimination, but neither depletion 1 week before nor 1 to 5 weeks after infection significantly altered any of the parameters studied here (data not shown). This led us to choose depletion of Tregs during infection with AdHBV as shown in Fig. 1C for all experiments shown. Whereas systemic Treg frequencies normalized after week 3 (data not shown), frequencies in the liver remained elevated for more than 2 months if HBV antigens were expressed (data not shown).
HBV-specific T cell responses against virus-infected hepatocytes result in inflammatory liver disease and hepatocyte death, which can be detected by increased ALT activity in the serum of infected individuals. Around day 7 postinfection, serum ALT levels peaked in AdHBV-infected mice and remained elevated until day 21 (Fig. 1D). Treg-depleted mice showed two-fold higher ALT activity on day 7 but reduced liver damage on day 21, indicating stronger but more short-lived liver inflammation. In control mice infected with AdHBV k/o serum ALT activity was not increased (Fig. 1D). Inflammatory activity in liver histology as well as CD3 T cell infiltration were only observed in AdHBV but not in AdHBV k/o infected mice (Supporting Fig. 2). Correlation with the induction of HBc-specific T cells (Fig. 1E) was consistent with the notion that immunomediated liver damage detected here is HBV specific.15 Treg activation, however, was not antigen-specific (Fig. 1F). Taken together, experimental infection with HBV by use of adenoviral gene transfer leads to rapid increase in Treg frequencies locally in the liver that restrict early immunomediated liver damage directed against HBV-infected hepatocytes.
Initial Treg Depletion Increases Early CD8+ T Effector Function in the HBV-Infected Liver.
To determine which cells may contribute to liver damage by killing infected hepatocytes, we analyzed the immune cell population in the liver on day 7 at the peak of liver inflammation using flow cytometry. Importantly, we isolated significantly more LALs from livers of AdHBV-infected mice than from AdHBV k/o–infected mice or noninfected control mice. Numbers of CD4+ and CD8+ T cells as well as NK1.1+ natural killer (NK)/NK T cells among LALs increased in the liver of AdHBV infected mice (Fig. 2A). In contrast, control infection with AdHBV k/o resulted only in a minor increase of intrahepatic CD8+ T cell numbers (Fig. 2B). Treg depletion resulted in a further significant increase in numbers and the frequencies of liver-associated CD8+ and CD4+ T cells while not affecting NK1.1+ (NK and NK-T) cells (Fig. 2D, Supporting Table 1). Importantly, Treg depletion led to an increase in Lamp1+ effector T cells, indicating an increase in their cytotoxic function (Fig. 2E).
To characterize in more detail the role of Tregs in the regulation of the antiviral CD8+ T cell response, during the course of infection we followed HBc-specific T cell responses following ex vivo peptide restimulation. We isolated LALs from AdHBV-infected, Treg-depleted, and nondepleted DEREG mice and monitored interferon-γ (IFNγ), interleukin 2, and tumor necrosis factor (TNF) production by CD8+ T cells by intracellular cytokine staining. On day 7 and day 21 postinfection, Treg-depleted mice exhibited a significantly increased virus-specific CD8+ T cell response (Fig. 3A,B). The overall frequency of HBc-specific IFNγ-producing CD8+ T cells was still low at the peak of liver inflammation at day 7, but increased to 6%-8% of total CD8+ T cells until day 70 (Fig. 3A). Depletion of Tregs lead to a significant increase in total numbers of HBc-specific IFNγ- and TNF-producing CD8+ T cells already at day 7 postinfection (Fig. 3A,B). Interestingly, TNF was produced by a large number of CD8+ T cells after stimulation with HBc but also with control peptide (Supporting Fig. 3). We found CD8+ T cells that produced TNF ex vivo only in AdHBV but not in AdHBV k/o infected or noninfected mice and only at day 7, indicating that HBV infection caused a unique but transient activation state in some CD8+ T cells leading to TNF production under the conditions of the experiment. Interestingly, the numbers of these TNF-expressing CD8+ T cells were further enhanced by Treg depletion (Fig. 3B). Accordingly, numbers of polyfunctional CD8+ T cells increased shortly after Treg depletion, but this effect was not sustained (Fig. 3C).Taken together, these results demonstrate early induction of TNF and IFNγ-producing CD8+ T cells during HBV infection that is controlled by Tregs.
Short-Term Depletion of Tregs Does Not Impair the Development of HBV-Specific CD8 T Cells and Establishment of Memory T Cells.
To investigate whether initial depletion of Tregs influences the establishment of HBV-specific CD8 T cells and ultimately the development of memory T cells, we quantified the numbers of intrahepatic HBV-specific CD8+ T cells during the course of infection using Kb multimer staining (Fig. 4A) and characterized their differential expression of the survival factor CD127 and the homing receptor CD62L (Fig. 4B). Multimer staining revealed that HBc- and S-protein–specific CD8+ T cell populations expanded continuously in the liver, increasing from below 0.5% of all CD8+ LALs on day 7 to 2%-6% on day 70 (Fig. 4A). To our surprise, no differences were found when Treg-depleted and nondepleted mice were compared (Fig. 4A) clearly demonstrating that Tregs did not impair development of HBV-specific CD8+ T cells following AdHBV infection.
Although on day 7 postinfection most HBV-specific CD8+ T cells were of the effector or effector memory phenotype, a growing population of HBc93-100 (Fig. 4C, upper panel) and HBs190-197-specific (Fig. 4C, lower panel) CD8+ T cells with a central memory T cell phenotype (i.e., CD62LhighCD127+) emerged over time. In the late phase of infection (day 70), 70%-90% of HBV-specific CD8+ T cells in the liver were CD62LhighCD127+ central memory cells (Fig. 4C), indicating that virus-specific central memory T cells reside not only in lymphoid tissues, but also in the liver. Although leading to a slightly reduced frequency of HBc93-100-specific CD8+ T cells on day 44, depletion of Tregs during the early phase of infection did not influence the establishment of long-term HBV-specific memory CD8+ T cells.
Initial Elimination of Tregs Improves Recruitment of Macrophages and Dendritic Cells Into the Infected Liver.
The first line of defense against viral infections is the innate immune system, in which activation of macrophages and dendritic cells (DCs) plays a prominent role. Cytokines released by these cells contribute to inflammation and may suppress viral replication. To find out whether Tregs also exert regulatory effects on macrophages and DCs, we quantified F4/80+ macrophages and 33D1+MHCII+ DCs during the course of infection by flow cytometry and analyzed their IFNγ and TNFα secretion. We found a pronounced recruitment of macrophages into the liver until day 7 postinfection, which was enhanced at day 3 postinfection after depletion of Tregs (Fig. 5A). Upon Treg depletion, there was no change in the dynamics or relative numbers of macrophages producing IFNγ or TNFα spontaneously ex vivo. Compared with macrophages, much smaller numbers of intrahepatic DCs were recruited into the infected liver (Fig. 5B), but DC numbers also increased at day 7 postinfection when Tregs were depleted. In contrast to hepatic macrophages, more than 80% of intrahepatic DCs produced IFNγ, and a substantial population also produced TNFα. Again, absence of Tregs early during infection did neither modify the dynamics nor the relative numbers of DCs producing IFNγ (Fig. 5B) nor cause a change in the major histocompatibility complex (MHC) II expression pattern of macrophages (Fig. 5C).
Depletion of Tregs Improves Early Immune Control of Acute HBV Infection.
To define the impact of Tregs on the establishment and course of HBV infection, we followed infection parameters in Treg-depleted and nondepleted, AdHBV-infected DEREG mice. Although mice were infected with equal efficiency (indicated by equal levels of HBeAg up to day 7 postinfection), HBeAg and HBsAg were cleared significantly faster from the serum of Treg-depleted animals (Fig. 6A,B). Strikingly, HBV serum titers were reduced by ≈90% in the absence of Tregs and HBV viremia was rapidly cleared (Fig. 6C). Production of anti-HBs antibodies (anti-HBs), finally resulting in seroconversion from HBsAg to anti-HBs, is a hallmark of HBV immune control. From day 44 onward, we detected an anti-HBs response, which was significantly increased after initial Treg depletion (Fig. 6D). Immunohistochemical analysis of AdHBV-infected liver tissue at the peak of liver inflammation confirmed that significantly more CD3+ T cells infiltrated the liver after initial Treg depletion (Fig. 6E). This led to a marked reduction of the number of HBc-positive hepatocytes at the later stages of infection (Fig. 6E, lower panel). Together, these results suggest that Tregs mitigate, but do not abrogate, the early antiviral immune response against HBV infection.
Using a mouse model of acute hepatitis B, we found that Tregs accumulate in the liver at the same time when activated effector T cells infiltrate the infected liver tissue. We found that initial elimination of Tregs improves antiviral effector function, cytokine production, and cytotoxicity of HBV-specific T cells, but did not substantially affect the development of long-term T cell memory. Increased T cell activity early after Treg depletion boosted immunomediated tissue damage, supporting the notion that Tregs reduced immunomediated tissue damage at the cost of delaying HBV clearance. Along this line, Tregs delayed but did not prevent HBsAg/anti-HBs seroconversion. Finally, Tregs delayed the influx of macrophages and DCs into the liver during the early phase of infection without affecting their cytokine secretion.
In our study using AdHBV for establishing HBV infection in murine hepatocytes in vivo, Tregs limited the number of IFNγ-producing HBV-specific CD8 T cells, confirming the function of Tregs to control virus-specific T cell immunity also for HBV infection. In AdHBV-infected but not in AdHBV k/o–infected or noninfected animals, we observed the rapid development of a large number of TNF-producing CD8 T cells. Because TNF expression by T cells correlated with the onset of hepatitis in our model system, it is likely that this T cell population contributes to immunomediated damage during viral hepatitis. Although the molecular mechanisms determining this unique responsiveness of liver-associated CD8 T cells to produce TNF during HBV infection remains to be identified in future studies, it is important to note that Tregs control the number of these TNF-producing T cells and thus contribute to protecting the liver from overzealous immunity.
Tregs, however, did not influence the priming of HBV-specific CD8 T cells following AdHBV infection. This finding indicates that in our model, Tregs acted locally in the liver to prevent liver damage inflicted by CD8 T cells rather than in lymphatic tissue to prevent priming and expansion of virus-reactive T cells. This is consistent with earlier studies in autoimmunity that a main feature of Tregs is restraining inflammation and maintaining organ integrity.18 Beneficial immunoregulatory functions of Tregs have been suggested from other viral infection models.19 The molecular mechanisms involved in the observed protection of the liver from immunomediated damage remains to be identified but very likely entail the regulatory molecules interleukin-10 and/or transforming growth factor-β.18 Because under noninflammatory conditions the liver harbored few Tregs and numbers rapidly increased after infection, our results indicate that recruitment of natural Tregs into the virus-infected liver was operational in the control of CD8 T cell effector function in the liver. It is of interest to note that CXCR3 mediates Treg recruitment to inflamed human liver tissue via hepatic sinusoidal endothelium, which is also used by activated effector CD8 T cells.20 This indicates a fine balance in the recruitment of effector and regulatory T cells that may operate to limit immunomediated liver damage.
The role of Tregs during acute viral infection is multifaceted: they can mitigate virus-specific immune responses and delay virus clearance,21, 22 but in a murine model of mucosal herpes simplex virus infection prevented fatal infection by allowing a timely entry of immune cells into infected tissue.23 Our study in acute viral hepatitis clearly demonstrates that Treg depletion improved early antiviral immunity against infected hepatocytes, albeit at the cost of increased liver immunopathology, and thus implies that Treg function may differ between organs. Notwithstanding, differences in the infecting viruses such as replication strategies and particularities of virus-specific immune responses may be responsible for distinct outcomes after Treg depletion. The model of experimental HBV infection used here, which eventually results in clearance of HBV from the infected mouse liver,15 does not allow any notion on the consequences of Treg depletion for prevention of viral persistence in the liver.
Mouse studies have suggested that Treg depletion may improve DNA vaccination against certain viruses, including HBV.24 Our results support the notion that Treg depletion accelerates viral clearance. One may conclude that depletion of Tregs or modulation of Treg function could serve as a valuable tool for immunotherapy, but several obstacles remain. First, a specific target structure on human Tregs for selective depletion is still missing.7, 8 Second, Treg depletion may trigger autoimmune reactions.25 Third, our data indicate that depletion of Tregs might cause side effects in patients and especially increase immunomediated liver damage by TNF-secreting T cells or innate immune cells recruited into the liver. Finally, it is questionable whether Tregs indeed enhance the protective effect of vaccination, since we found no influence of Tregs on the development of HBV-specific central memory T cells.
Taken together, our study demonstrates that intrahepatic Tregs have a crucial influence on immunopathology during acute HBV infection. Our results indicate that Tregs not only suppress HBV-specific adaptive immune responses, but also influence innate immunity in the early phase of acute HBV infection by regulating influx of macrophages and DCs. Thus, Tregs apparently have liver-protective functions during acute viral infection, whereas their role in promoting viral immune escape and persistent infection needs to be addressed in future studies.
We thank Ingo Drexler, Tanja Bauer, Sarah Kutscher, and Martin Sprinzl for valuable discussions and input.