Type I Interferon Signaling Prevents Hepatitis B Virus-Specific T Cell Responses by Reducing Antigen Expression

Hepatitis B virus (HBV) causes acute and chronic infection, and approximately 240 million people are chronically infected with HBV worldwide. It is generally believed that virus-specific CD8⁺ T cell responses are required for the clearance of HBV. However, the relative contributions of genetic variation and innate immune responses to the induction of HBV-specific CD8⁺ T cell responses are not fully understood. In this study, we discovered that different clearance rates between HBV clones after hydrodynamic transduction were associated with the magnitude of HBV-specific CD8⁺ T cell responses and initial HB core antigen expression. Surprisingly, type I interferon signaling negatively regulated HBV-specific CD8⁺ T cell responses by reducing early HBV antigen expression. These results show that the magnitude of the HBV-specific CD8⁺ T cell response is regulated primarily by the initial antigen expression level.

ABSTRACT

Robust virus-specific CD8⁺ T cell responses are required for the clearance of hepatitis B virus (HBV). However, the factors that determine the magnitude of HBV-specific CD8⁺ T cell responses are poorly understood. To examine the impact of genetic variations of HBV on HBV-specific CD8⁺ T cell responses, we introduced three HBV clones (Aa_IND [Aa], C_JPN22 [C22], and D_IND60 [D60]) that express various amounts of HBV antigens into the livers of C57BL/6 (B6) (H-2b) mice and B10.D2 (H-2d) mice. In B6 mice, clone C22 barely induced HBV-specific CD8⁺ T cell responses and persisted the longest, while clone D60 elicited strong HBV-specific CD8⁺ T cell responses and was rapidly cleared. These differences between HBV clones largely diminished in H-2d mice. Interestingly, the magnitude of HBV-specific CD8⁺ T cell responses in B6 mice was associated with the HB core antigen expression level during the early phase of HBV transduction. Surprisingly, robust HBV-specific CD8⁺ T cell responses to clone C22 were induced in interferon-α/β receptor-deficient (IFN-αβR^–/–) (H-2b) mice. The induction of HBV-specific CD8⁺ T cell responses to C22 in IFN-αβR^–/– mice reflects enhanced HBV antigen expression because the suppression of antigen expression by HBV-specific small interfering RNA (siRNA) attenuated HBV-specific T cell responses in IFN-αβR^–/– mice and prolonged HBV expression. Collectively, these results suggest that HBV genetic variation and type I interferon signaling determine the magnitude of HBV-specific CD8⁺ T cell responses by regulating the initial antigen expression levels.

IMPORTANCE Hepatitis B virus (HBV) causes acute and chronic infection, and approximately 240 million people are chronically infected with HBV worldwide. It is generally believed that virus-specific CD8⁺ T cell responses are required for the clearance of HBV. However, the relative contributions of genetic variation and innate immune responses to the induction of HBV-specific CD8⁺ T cell responses are not fully understood. In this study, we discovered that different clearance rates between HBV clones after hydrodynamic transduction were associated with the magnitude of HBV-specific CD8⁺ T cell responses and initial HB core antigen expression. Surprisingly, type I interferon signaling negatively regulated HBV-specific CD8⁺ T cell responses by reducing early HBV antigen expression. These results show that the magnitude of the HBV-specific CD8⁺ T cell response is regulated primarily by the initial antigen expression level.

INTRODUCTION

Hepatitis B virus (HBV) causes acute and chronic infection, and approximately 240 million people are chronically infected with HBV worldwide (1). HBV is categorized into nine genotypes based on a sequence divergence (2). While genetic differences between HBV clones are shown to influence the outcome of HBV infection (3), the immunological basis of the distinct outcomes is poorly defined.

It is generally accepted that cytotoxic CD8⁺ T cell (CTL) responses are required to clear the virus (4, 5). A vigorous, multispecific CD8⁺ T cell response to HBV is induced in patients with acute and self-limited HBV infection. In contrast, HBV-specific CD8⁺ T cells are usually deleted or functionally impaired in patients chronically infected with HBV (6, 7). Despite the well-documented importance of HBV-specific CD8⁺ T cell responses in clearing HBV, it is not known whether genetic variation of HBV has any impact on the magnitude of specific CD8⁺ T cell responses.

Type I interferons (IFN-Is), which include several IFN-α proteins and one IFN-β protein, are key molecules mediating innate immune responses (8). They limit the spread of many RNA and DNA viruses during the early phase of infection (9). Conventional and pegylated IFN-α proteins have been first-line treatments for patients with chronic hepatitis B (CHB) (10, 11). Although the antiviral effect of IFN-Is has been unequivocal (12), the viral clearance rate of IFN-α treatment in CHB patients is rather low (13). The immunological mechanisms that limit the therapeutic efficacy of IFN-α treatment remain to be elucidated, but accumulating evidence suggests that IFN-Is can either promote or inhibit T cell activation, proliferation, differentiation, and survival (14). Surprisingly, the role of IFN-I signaling in the induction of HBV-specific CD8⁺ T cell responses has not been adequately addressed.

In the current study, we used a mouse model of acute HBV infection (15) to determine whether the HBV clearance rates differ between HBV clones and whether this is associated with HBV-specific CD8⁺ T cell responses. We also examined the role of IFN-I signaling in regulating HBV-specific CD8⁺ T cell responses. Our results reveal a previously unknown impact of HBV genetic variation on the HBV-specific CD8⁺ T cell response, whose magnitude has been directly correlated with the initial HB core antigen (HBcAg) expression level in the liver. Interestingly, the importance of HBcAg expression level in the induction of HBV-specific CD8⁺ T cell responses was dependent on the host major histocompatibility complex (MHC) class I haplotypes. Unexpectedly, our data also suggest that IFN-I signaling suppresses not only antigen expression but also HBV-specific CD8⁺ T cell responses. To our knowledge, this is the first study to demonstrate that the genetic variation of HBV and IFN-I signaling modulate HBV-specific CD8⁺ T cell responses by regulating HBV antigen expression.

RESULTS

The rates of HBV clearance vary between HBV clones and are associated with HBV-specific CD8⁺ T cell responses.

To determine whether HBV clearance rates differ between HBV clones, the HBV Aa_IND clone (Aa; genotype A; GenBank accession no. AB246335), C_JPN22 clone (C22; genotype C; GenBank accession no. AB246344), and D_IND60 clone (D60; genotype D; GenBank accession no. AB246347) (16) were introduced into C57BL/6 (B6) mice by hydrodynamic injection (HDI) of 27 µg of a plasmid containing a replication-competent 1.24-fold HBV DNA copy. The mice were bled on days 1, 4, and 14 to measure the serum levels of HB surface antigen (HBsAg). Because a previous report showed that CD8⁺ T cell responses are required to clear HBV in this model, the mice were sacrificed on day 14 after HDI and the frequency of activated CD8⁺ T cells was analyzed by measuring the fraction of CD69-positive cells to correlate with the HBsAg clearance rate. As shown in Fig. 1A, clone Aa (hatched bars) tended to produce higher serum HBsAg levels than clones C22 (white bars) and D60 (black bars). Clones C22 and D60 produced similar levels of HBsAg on days 1 and 4, although HBsAg levels on day 4 were modestly, but significantly, higher in animals transfected with C22 than in those transfected with D60. On day 14, serum HBsAg was still readily detectable in mice transduced with clones Aa (42.3 ± 30.9 IU/ml) and C22 (213.1 ± 250.2 IU/ml), while it was mostly cleared in D60-transduced mice (0.8 ± 1.0 IU/ml). The frequency of activated CD8⁺ T cells that expressed CD69 in the liver was lowest in the C22-transduced mice and highest in the D60-transduced animals (Fig. 1B). Clone Aa showed an intermediate phenotype between those of clones C22 and D60, regarding the HBsAg clearance rate and CD8⁺ T cell responses after HDI. Collectively, these results suggest that the HBsAg clearance rates differ between HBV clones and are correlated with the activation of intrahepatic CD8⁺ T cells.

FIG 1.

Serum HBsAg clearance and CD8⁺ T cell activation in the liver differ between HBV clones. Three hepatitis B virus (HBV) clones (Aa, C22, and D60) were introduced into C57BL/6 (B6) mice (Aa, n = 7; C22, n = 9; D60, n = 7) by hydrodynamic injection (HDI). (A) The mice were bled on days 1, 4, and 14 after HDI, and the concentration of serum HB surface antigen (HBsAg) was measured. (B) The mice were sacrificed on day 14, and the frequency of CD69⁺ CD8⁺ T cells in the liver (intrahepatic lymphocytes [IHLs]) was analyzed by flow cytometry (n = 5 for each clone). The values are shown after consolidating data from 3 independent experiments. Mean values plus SD are shown. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Because of the apparent differences in HBsAg clearance and CD8⁺ T cell responses between C22 and D60, we compared these two clones in more detail by measuring the amount of the intrahepatic input HBV DNA plasmid and HBV mRNA on days 4 and 14, as well as the intrahepatic HB core antigen (HBcAg) and the serum HB e antigen (HBeAg) levels on day 4. In this system, the input HBV DNA plasmid serves as a template for HBV transcription. Thus, the amount of input HBV DNA on day 4 could be used to normalize the transfection efficacy, while its disappearance from the liver on day 14 indicates HBV clearance. We also compared the intrahepatic interferon-stimulated gene (ISG) expression levels on day 4 to determine whether the innate immune response was activated differently by C22 and D60. In addition, the frequency of HBV-specific CD8⁺ T cells in the liver was analyzed on day 14 using a multimer that binds to the T cell receptor (TCR) specific for Kb-restricted COR93, which is the dominant epitope in B6 mice (MHC haplotype, H-2b) (17). While serum HBsAg on day 4 was slightly higher in the mice transduced with C22 (Fig. 1A) (1,834.1 ± 164.2 IU/ml) than in those transduced with D60 (Fig. 1A) (842.9 ± 129.6 IU/ml), intrahepatic input HBV DNA and HBV mRNA contents on day 4 were similar in the mice transfected with C22 and D60 (Fig. 2A), indicating that the amounts of the DNA template and the transcriptional activities were comparable between these clones. Interestingly, intrahepatic HBcAg expression on day 4 after HDI was weaker in the mice transduced with C22 than in those transduced with D60 (Fig. 2A), while the serum HBeAg level was higher in the C22 group (Fig. 2B) (837.6 ± 82.1 cutoff index [COI]) than in the D60 group (Fig. 2B) (568.8 ± 68.0 COI). These differences were not due to the different abilities of C22 and D60 to induce the innate immune response because the levels of intrahepatic Isg15 expression were similar between these clones on day 4 after HDI (Fig. 2C). On day 14, intrahepatic input HBV DNA and HBV mRNA expression persisted longer in the mice transfected with C22 than in those transfected with D60 (Fig. 2D). Interestingly, COR93-specific CD8⁺ T cells were readily detectable in the D60 group but not in the C22 group (Fig. 2E and F) (P = 0.03). These results suggest that weak HBV-specific CD8⁺ T cell responses were associated with low intrahepatic HBcAg and high serum HBeAg expression during the early phase of HBV transduction.

FIG 2.

HBV clearance rate is correlated with HBV-specific CD8⁺ T cell responses in the liver. A plasmid encoding a 1.24-fold-length C22 or D60 clone was introduced into B6 mice by HDI. (A) The amounts of input HBV DNA (upper panel), HBV mRNA (middle panel), and HB core antigen (HBcAg) expression (lower panel) in the liver were monitored on day 4 after HDI by Southern, Northern, and Western blotting, respectively (n = 3 for each clone). (B) The concentration of serum HB e antigen (HBeAg) was measured on day 4 (n = 4 for each clone). (C) Intrahepatic interferon-stimulated gene 15 (Isg15) expression on day 4 was analyzed by quantitative real-time PCR (n = 4 for each clone). Values relative to those for untransduced mice are represented. (D) Input HBV DNA (upper panel) and HBV mRNA (lower panel) contents in the liver were monitored on day 14 as described above (n = 4 for each clone). (E and F) The frequency of COR93-specific CD8⁺ T cells in the liver was analyzed on day 14 (n = 5 for each clone). Mean values plus SD are shown. n.s., not significant; *, P < 0.05; **, P < 0.01.

The intrahepatic HBcAg expression level is correlated with HBV-specific CD8⁺ T cell responses.

To test whether intrahepatic HBcAg expression and/or the serum HBeAg level determine the magnitude of HBV-specific CD8⁺ T cell responses, we constructed a 1.3-fold-length HBV DNA plasmid which encodes an additional enhancer and X protein compared to the 1.24-fold length of the construct. As shown in Fig. 3A and B, 1.3-fold-length C22 expressed more 3.5-kb intrahepatic mRNA, HBcAg, and serum HBeAg than the 1.24-fold-length C22 on day 4 after HDI. In contrast, the expression levels of intrahepatic HBsAg and the 2.4-kb and 2.1-kb intrahepatic mRNA contents were comparable between these clones (Fig. 3A). We then analyzed the intrahepatic input HBV DNA and HBV mRNA contents, as well as the fraction of activated and COR93-specific CD8⁺ T cell responses on day 14 after HDI of 1.24-fold- and 1.3-fold-length HBV C22. As shown in Fig. 3C, intrahepatic input HBV DNA and HBV mRNA were readily cleared in the mice that received the 1.3-fold-length C22 clone but not in those that received the 1.24-fold-length counterpart. Importantly, the clearance of HBV was associated with stronger activation of global CD8⁺ T cells and a higher frequency of COR93-specific CD8⁺ T cells in the liver (Fig. 3D) (P = 0.005 and P = 0.03, respectively). These results suggest that the magnitude of COR93-specific CD8⁺ T cell responses and the rate of HBV clearance were determined by the intrahepatic HBcAg expression level rather than the serum HBeAg level.

FIG 3.

The intrahepatic HBcAg expression level is associated with CD8⁺ T cell responses. C22 clones at 1.24-fold and 1.3-fold lengths were introduced into B6 mice by HDI. (A) The amounts of input HBV DNA (upper panel) and HBV mRNA (middle panel), as well as HBcAg and HBsAg expression (lower panel), in the liver were monitored on day 4 after HDI (n = 3 for each plasmid). (B) The concentration of serum HBeAg was measured on day 4 (n = 4 for each plasmid). (C) Intrahepatic input HBV DNA (upper panel) and HBV mRNA (lower panel) content were monitored on day 14 (n = 4 for each plasmid). (D) The frequencies of CD69⁺ CD8⁺ T cells (left panel) and COR93-specific CD8⁺ T cells (right panel) in the liver were analyzed on day 14 (n = 5 for each plasmid). Mean values plus SD are shown. *, P < 0.05; **, P < 0.01.

To confirm that a high level of HBcAg expression is important for the induction of HBV-specific CD8⁺ T cell responses and HBV clearance, we complemented HBcAg expression in the mice that received the 1.24-fold-length C22 plasmid. Specifically, we mixed 27 μg of the HBV DNA plasmid containing 1.24-fold-length C22 with 2 μg of the plasmid encoding HBcAg under the control of the cytomegalovirus (CMV) promoter or the control CMV vector and hydrodynamically injected the plasmid mixture into B6 mice. As shown in Fig. 4A, intrahepatic HBcAg expression on day 4 was clearly complemented in the mice injected with both HBV DNA and the HBcAg plasmid. As expected, the HBsAg expression level was not affected by HBcAg complementation. Importantly, intrahepatic input HBV DNA and HBV mRNA were almost completely cleared from the mice that received the HBcAg plasmid in addition to the HBV DNA plasmid (Fig. 4B). As expected, this clearance was associated with strong induction and activation of HBV-specific CD8⁺ T cell responses (Fig. 4C). These results suggest that a high level of intrahepatic HBcAg expression facilitates the induction of HBV-specific CD8⁺ T cells and HBV clearance.

FIG 4.

HBV clearance is facilitated in mice that transcomplemented cognate HBcAg expression. Twenty-seven micrograms of plasmid encoding 1.24-fold-length C22 was mixed with 2 µg of plasmid encoding HBcAg under the control of the cytomegalovirus (CMV) promoter or the same amount of an empty CMV vector, and the plasmid mixture was hydrodynamically injected into B6 mice. (A) Mice were sacrificed on day 4 after HDI (empty vector, n = 2; HBcAg complementation, n = 3), and the amount of input HBV DNA content (upper panel), as well as HBcAg and HBsAg expression (lower panel), in the liver was monitored. (B) Mice were sacrificed on day 14 after HDI (n = 5 for each group), and the amounts of input HBV DNA (upper panel) and HBV mRNA (lower panel) in the liver were monitored. (C) The frequencies of activated CD8⁺ T cells (left panel) and COR93-specific CD8⁺ T cells (right panel) in the liver were analyzed on day 14 (n = 5 for each group). Mean values plus SD are shown. *, P < 0.05; **, P  < 0.01.

The impact of genetic variation on the magnitude of HBV-specific CD8⁺ T cell responses and HBV clearance rate is dependent on MHC haplotypes.

To examine whether the induction of HBV-specific CD8⁺ T cell responses to clone C22 is also weak in the mice with a different MHC haplotype, we hydrodynamically introduced 1.24-fold-length C22 or D60 into B10.D2 mice whose MHC haplotype is H-2d. In this lineage, the dominant CD8⁺ T cell epitope is Ld-restricted ENV28, which is located between amino acid residues 28 and 39 of HBsAg (18). We compared serum HBsAg levels on days 1, 7, and 14 as well as the frequency of activated CD8⁺ T cells on day 14 after HDI. As shown in Fig. 5A, similar levels of HBsAg were produced in C22- and D60-transduced mice on day 1, and HBsAg levels on day 7 were modestly higher in the animals transfected with C22 than in those transfected with D60. Notably, serum HBsAg similarly disappeared on day 14 in the mice transduced with C22 and D60 (Fig. 5A). Importantly, clone C22 and D60 strongly activated CD8⁺ T cells in B10.D2 mice (Fig. 5B) compared with B6 mice (Fig. 1B), suggesting that the differences between clones C22 and D60 were dependent on the MHC haplotypes of host mice.

FIG 5.

The magnitude of HBV-specific CD8⁺ T cell responses is dependent on MHC haplotypes. C22 or D60 at a 1.24-fold length was introduced into B10.D2 mice by HDI. (A) The concentration of serum HBsAg was measured on days 1, 7, and 14 (n = 4 for each clone). (B) The frequencies of CD69⁺ CD8⁺ T cells in the liver were analyzed on day 14 (n = 4 for each clone). Mean values plus SD are shown. n.s., not significant; **, P < 0.01.

HBV-specific CD8⁺ T cell responses are negatively regulated by IFN-I signaling.

To further investigate the impact of HBV antigen expression on the induction of HBV-specific CD8⁺ T cells and HBV clearance, we blocked type I interferon (IFN-I) signaling, which is known to suppress HBV antigen expression. Specifically, we introduced 1.24-fold-length C22 into IFN-α/β receptor-deficient (IFN-αβR^–/–) mice and MHC-matched controls (B6 mice) by HDI. Various virological parameters were measured on days 4 and 14 after HDI and associated with CD8⁺ T cell responses in the liver on day 14. As shown in Fig. 6A, intrahepatic HBV mRNA expression was higher in IFN-αβR^–/– mice than in B6 mice on day 4. Accordingly, intrahepatic HBcAg and HBsAg expression (Fig. 6A) as well as serum HBeAg and HBsAg levels (data not shown) on day 4 were also higher in IFN-αβR^–/– mice than in B6 mice. Strikingly, robust COR93-specific CD8⁺ T cell responses were induced in IFN-αβR^–/– mice (Fig. 6B) (P = 0.007). In association with the T cell responses, HBV DNA input and HBV mRNA were mostly cleared in IFN-αβR^–/– mice by day 14 after HDI, while they persisted in B6 mice (Fig. 6C). These results suggest that IFN-I signaling negatively regulates COR93-specific CD8⁺ T cell responses and HBV clearance.

FIG 6.

Genetic ablation of IFN-I signaling enhances HBV-specific CD8⁺ T cell responses. C22 at a 1.24-fold length was introduced into interferon-α/β receptor-deficient (IFN-αβR^–/–) and B6 mice by HDI. (A) The amounts of input HBV DNA (upper panel) and HBV mRNA (middle panel) and HBcAg and HBsAg expression (lower panel) in the liver on day 4 were monitored (n = 4 for each strain). (B) The frequency of COR93-specific CD8⁺ T cells in the liver was analyzed on day 14 (n = 4 for each strain). (C) Intrahepatic input HBV DNA (upper panel) and HBV mRNA (lower panel) contents on day 14 were monitored (n = 4 for each strain). Mean values plus SD are shown. **, P < 0.01.

IFN-I signaling suppresses HBV-specific CD8⁺ T cell responses by reducing the HBV antigen expression level in the liver.

The negative impact of IFN-I signaling on HBV-specific CD8⁺ T cell responses might reflect direct suppression through the IFN-α/β receptor on CD8⁺ T cells themselves or through an indirect mechanism, such as reduction of HBV antigen after HDI transduction. To distinguish these alternatives, we harvested spleen cells from TCR-transgenic mice expressing a TCR specific for COR93 (lineage BC10.3, IFN-αβR^+/+) (17) and adoptively transferred them into B6 mice and IFN-αβR^–/– mice. Four days later, both groups were hydrodynamically transduced with 1.24-fold-length HBV C22 and the frequency and the number of COR93-specific CD8⁺ T cells in the liver and spleen were analyzed on day 7 after HDI (Fig. 7A). As shown in Fig. 7B and C, the adoptively transferred COR93-specific CD8⁺ T cells expanded vigorously in the livers and spleens of HBV-transduced B6 mice (white bars), compared to levels in HBV negative-control animals (hatched bars). Importantly, the transferred CD8⁺ T cells expanded more vigorously in IFN-αβR^–/– mice (Fig. 7B and C, black bars) than in B6 mice. Because IFN-I signaling is intact in the transferred COR93-specific CD8⁺ T cells, these results suggest that HBV-specific CD8⁺ T cell responses might be enhanced in IFN-αβR^–/– mice independently of IFN-I signaling in HBV-specific CD8⁺ T cells.

FIG 7.

IFN-αβR-expressing HBV-specific T cells proliferate after HBV transduction in IFN-αβR^–/– mice. (A) COR93-specific CD8⁺ T cells obtained from T cell receptor (TCR) transgenic mice (IFN-αβR^+/+) were adoptively transferred to IFN-αβR^–/– mice (n = 4) or B6 mice (n = 6). Four days later, clone C22 was introduced into these mice by HDI. Two B6 mice that received specific T cells were not transfected with HBV C22 to serve as controls. (B and C) The frequency (left panel) and the number (right panel) of the transferred CD8⁺ T cells in the liver (B) and the spleen (C) were analyzed by flow cytometry on day 7 after HDI. Mean values plus SD are shown. *, P < 0.05; **, P < 0.01.

To determine whether the induction of robust HBV-specific CD8⁺ T cell responses to C22 in IFN-αβR^–/– mice reflect the enhanced antigen expression levels in the liver, we introduced 1.24-fold-length C22 into IFN-αβR^–/– mice, with HBV-specific small interfering RNAs (siRNAs) (siHBVs) or negative-control siRNA (siCTRL). As shown in Fig. 8A, the serum HBsAg levels on days 1 and 4 were suppressed in the mice treated with siHBVs. Importantly, the frequency of COR93-specific CD8⁺ T cells was significantly decreased in the mice with siHBVs compared to their frequency in the control mice (Fig. 8B) (P = 0.003). Interestingly, the serum HBsAg level, as well as the amount of input HBV DNA plasmid and HBV mRNA content in the liver, were higher on day 14 in the mice treated with siHBVs than in the siCTRL-treated animals (Fig. 8A and C). These results indicate that IFN-I signaling exerts a detrimental effect on HBV clearance and CD8⁺ T cell responses by reducing HBV antigen expression during the early phase of HBV transduction.

FIG 8.

Suppression of the initial antigen expression inhibits HBV-specific CD8⁺ T cell responses in IFN-αβR^–/– mice. IFN-αβR^–/– mice were transduced with clone C22 DNA and HBV-specific siRNAs (siHBVs) or with clone C22 DNA and negative-control siRNA (siCTRL) by HDI. (A) The concentration of serum HBsAg was measured on days 1, 4, 7, and 14 (n = 5 for each group). (B) The frequency of COR93-specific CD8⁺ T cells in the liver was analyzed on day 14 (n = 5 for each group). (C) Intrahepatic input HBV DNA (upper panel) and HBV mRNA (lower panel) contents on day 14 were monitored (n = 4 for each group). Mean values plus SD are shown. **, P < 0.01; ***, P < 0.001.

DISCUSSION

The current study was initiated to determine the impact of virological factors on HBV-specific CD8⁺ T cell responses and HBV clearance in an immunologically naive host. To this end, we used previously described clinical isolates of HBV (16) whose clinical and virological characteristics differ (3) in combination with a well-defined mouse model of HBV infection (15, 19) that has been used to study the role of immune responses in eliminating HBV (20, 21). As with experimentally infected chimpanzees (5), the clearance of HBV after hydrodynamic transfection is dependent on CD8⁺ T cell responses (22). However, virological factors that influence the magnitude of HBV-specific CD8⁺ T cell responses have not been fully elucidated. Our results indicate that the intrahepatic HB core antigen (HBcAg) expression level at an early stage of HBV transduction is associated with the magnitude of HBV-specific CD8⁺ T cell responses in B6 (H-2b) mice. The role of HBcAg expression level in the induction of HBV-specific CD8⁺ T cells largely diminished in B10.D2 (H-2d) mice, indicating that the influence of HBV genetic variation on HBV-specific CD8⁺ T cells is dependent on MHC haplotypes. More importantly, our data revealed that IFN-I signaling negatively regulated HBV-specific CD8⁺ T cell responses by reducing antigen expression. To our knowledge, this is the first study that demonstrates an unexpected link between HBV genetic variation, IFN-I signaling, and HBV-specific CD8⁺ T cell responses (Fig. 9).

FIG 9.

Effects of genetic variation and IFN-I signaling on antigen expression and T cell responses. (A) Clones that express small amounts of HBcAg induce weak HBV-specific CD8⁺ T cell responses, resulting in delayed HBV clearance, while clones that express large amounts of HBcAg induce robust CD8⁺ T cell responses, resulting in HBV clearance. (B) IFN-I signaling suppresses HBV-specific CD8⁺ T cell responses by reducing HBV antigen expression.

HBV is categorized into nine genotypes (A to H and J) based on a sequence divergence of over 8% in the entire genome (2, 23). While clinical manifestations appear significantly different between genotypes (3, 24, 25), the cellular and molecular basis of these differences have been poorly defined. The use of a genetically well-defined small-animal model allowed us, for the first time, to examine the impact of genetic differences between clones on immune-mediated HBV clearance. The results of this study provide the first proof of concept that the genetic variation of HBV strongly influences the magnitude of HBV-specific CD8⁺ T cell responses, which in turn determine the outcome of HBV infection.

The importance of antigen expression in HBV clearance was implied in experimentally infected chimpanzees, in whom low antigen expression during the early phase of HBV infection was associated with weak HBV-specific CD8⁺ T cell responses and delayed clearance (26). Tian et al. also showed that the HBV clearance rate was influenced by the amount of HBV input plasmid in the hydrodynamic transfection system (27). In our system, intrahepatic HBcAg expression, but not HBsAg or HBeAg, appears to determine the magnitude of HBV-specific CD8⁺ T cell responses and the HBV clearance rate in B6 mice, as illustrated by the difference between 1.24-fold-length C22 and D60 (Fig. 2) and between 1.24-fold and 1.3-fold-length C22 (Fig. 3). These differences do not reflect the ability of each HBV clone to induce an innate immune response because levels of intrahepatic Isg15 induction were not different between C22- and D60-transduced animals (Fig. 2C). Moreover, although both HBsAg and HBeAg were implicated in suppressing HBV-specific CD8⁺ T cell responses (28), neither antigen seemed to influence HBV clearance or CD8⁺ T cell responses in this model (Fig. 1, 3, 6, and 8). Rather, the importance of HBcAg expression presumably reflects the presence of the dominant CD8⁺ T cell epitope. The dominant HBV-specific CD8⁺ T cell response in B6 mice is elicited to HBcAg (17), and the activation of global CD8⁺ T cells is closely correlated with the frequency of COR93-specific CD8⁺ T cells (Fig. 1B and Fig. 2E and F). Furthermore, the impact of HBcAg expression on the induction of HBV-specific CD8⁺ T cell responses largely diminished in B10.D2 mice (Fig. 5), whose dominant CD8⁺ T cell epitope is ENV28, derived from HBsAg (18). On the other hand, previous studies with the HDI system indicated that core antigen expression and capsid formation were necessary for the clearance of HBV from B6 mice (29, 30), implying that capsid formation provided an unknown stimulus required for HBV-specific CD8⁺ T cell responses. Identifying virological factors that regulate HBV-specific CD8⁺ T cell responses is clinically important to understand the basis of distinct outcomes among HBV genotypes.

It is generally believed that the immune system is so sensitive that a minute amount of antigen is sufficient to elicit CD8⁺ T cell responses. It has been shown that as few as 20 peptide-MHC class I complexes on an antigen-presenting cell are required to fully activate a CD8⁺ T cell (31). In contrast to this notion, our data suggest that a relatively large amount of antigen should be expressed in the liver to trigger HBV-specific CD8⁺ T cell expansion. The requirement of high antigen expression for the induction of HBV-specific CD8⁺ T cell responses in this study may reflect the nature of the antigen-expressing cells. In this system, as with natural HBV infection, HBV is expressed almost exclusively in the hepatocytes because the endogenous HBV promoters drive the antigen expression. Therefore, HBV antigen presentation by professional antigen-presenting cells depends on the cross-presentation machinery, which is less efficient than endogenous antigen presentation (32).

Surprisingly, IFN-I signaling suppresses HBV-specific CD8⁺ T cell responses and delays HBV clearance after HDI (Fig. 6). The negative impact of IFN-I signaling on HBV clearance was also demonstrated in the hydrodynamic system (27), but the immunological basis for those findings was not addressed. The role of IFN-Is in virus-specific CD8⁺ T cell responses is contentious. Generally, it is believed that IFN-Is promote virus-specific T cell expansion and differentiation (33, 34). In contrast, recent reports showed that blockade of IFN-I signaling enhanced virus-specific CD4⁺ T cell responses by decreasing the expression of negative immune regulatory molecules, such as interleukin 10 (IL-10) and PD-L1, and facilitated viral clearance in mice infected with lymphocytic choriomeningitis virus (LCMV) (35, 36). A more recent study using human immunodeficiency virus type 1 (HIV-1)-infected humanized mice showed that blockade of IFN-I signaling during combined antiretroviral therapy (cART) could reverse HIV-induced T cell dysfunction and reduce HIV reservoirs (37). The current study provides the first evidence that IFN-I signaling prevents virus-specific CD8⁺ T cell responses to a virus expressed in the liver.

Our data do not contradict the universally accepted antiviral effect of IFN-Is (12, 38, 39). Indeed, genetic ablation of IFN-I signaling resulted in increased expression of intrahepatic HBV mRNA, as well as intrahepatic and serum HBV antigen (Fig. 6A). Paradoxically, however, the antiviral ability of IFN-Is appears to suppress HBV-specific CD8⁺ T cell responses by reducing HBV antigen expression (Fig. 6 and 8). This notion is consistent with the fact that different T cell responses among the various clones were closely associated with HBcAg expression levels (Fig. 2). However, our results do not exclude the possibility that IFN-I signaling prevents HBV-specific CD8⁺ T cell responses by other mechanisms, such as the induction of IL-10 or PD-1 (35, 36) or the suppression of cross-priming.

Because the HDI system does not completely replicate natural HBV infections, and because HBV is known for its weak ability to induce IFN-I signaling (40–42), the physiological importance of IFN-I-mediated T cell suppression during natural HBV infection should be evaluated with an abundance of caution. Indeed, the kinetics of HBV protein expression in naturally infected patients is very different from that of hydrodynamically transfected mice. Nevertheless, it is still possible that IFN-I expression affects HBV-specific CD8⁺ T cell responses and the outcome of natural HBV infection. Although HBV per se does not activate IFN-I signaling during natural HBV infection (40–42), IFN-I signaling might be provided by other viral or bacterial infections. Importantly, a recent study demonstrated that the clearance of HBV was delayed in chimpanzees that had been chronically infected with hepatitis C virus (HCV) (43). Because ISGs were strongly induced in chimpanzees chronically infected with HCV, it is possible that IFN-I signaling induced by HCV infection suppressed HBV antigen expression and HBV-specific CD8⁺ T cell responses.

In summary, the results described herein suggest that genetic variation and IFN-I signaling determine HBV-specific CD8⁺ T cell responses by regulating the initial antigen expression level in the liver. Similar mechanisms may influence the outcome of HBV infection in humans.

MATERIALS AND METHODS

Ethical statement.

All experiments involving mice were performed at the Center for Experimental Animal Science at Nagoya City University according to a protocol approved by the Institutional Animal Care and Use Committee of the Nagoya City University Graduate School of Medical Sciences (approval number H25M_46). The experiments were carried out in accordance with the International Guiding Principles for Biomedical Research Involving Animals (44).

Mice, hydrodynamic injection, and adoptive transfer.

C57BL/6 (B6) mice (MHC haplotype, H-2b) and B10.D2 mice (H-2d) were obtained from the breeding colonies of Japan SLC, Inc. (Shizuoka, Japan). IFN-α/β receptor-deficient (IFN-αβR^–/–) (H-2b) mice (45) were obtained from Shizuo Akira (Osaka University, Japan) through Oriental BioService Inc. (Kyoto, Japan) and bred in our animal center. Ten- to 12-week-old male mice were used in all the experiments. To transduce HBV DNA into the mouse liver, hydrodynamic injection (HDI) was performed as described previously (15, 19). Briefly, 27 μg of HBV DNA plasmid in a large volume (equivalent to 8% of the body weight) of phosphate-buffered saline (PBS) was intravenously injected into mice within 5 to 8 s. In selected experiments, spleen cells were isolated from HBV-specific T cell receptor (TCR) transgenic mice (lineages BC10.3 and CD45.1; H-2b) that express a TCR specific for the Kb-restricted COR93 epitope (MGLKFRQL, located between residues 93 and 100 of the hepatitis B core antigen [HBcAg] [17]), kindly provided by Francis V. Chisari (The Scripps Research Institute, USA), and 5 × 10⁵ spleen cells per mouse were adoptively transferred into B6 and IFN-αβR^–/– mice (CD45.2; H-2b) before HDI of an HBV DNA plasmid.

SiRNA treatment.

To suppress intrahepatic HBV antigen expression, a mixture of three HBV-specific siRNAs (50 μg of each siHBV, namely, 251, 1804, and 2312 [46]) was transduced into the mouse livers by HDI, simultaneously with an HBV DNA plasmid. Control mice were transduced with 150 μg of negative-control siRNA by HDI.

Peptide and plasmids.

The peptide of the Kb-restricted cytotoxic T lymphocyte (CTL) epitope COR93 was purchased from Scrum Inc. (Tokyo, Japan). Plasmids containing a replication-competent 1.24-fold- or 1.3-fold-length HBV DNA genome of C_JPN22 (C22; GenBank accession no. AB246344), D_IND60 (D60; GenBank accession no. AB246347), and Aa_IND (Aa; GenBank accession no. AB246335) (16) were used for hydrodynamic transfection. Because the COR93 epitope in the original C22 clone (MGLKIRQL) differs from that in D60 (MGLKFRQL), a single nucleotide mutation was introduced into clone C22 so that both clones express the same CTL epitope. The same mutation was inserted into a plasmid encoding the entire HB core protein of clone C22 under the control of the cytomegalovirus (CMV) promoter, and the plasmid was used to transcomplement HB core antigen expression in vivo.

Lymphomononuclear cell preparation.

Intrahepatic lymphocytes (IHLs) and spleen cells were prepared as described previously (17). Livers were perfused with 10 ml of PBS via the portal vein to remove circulating lymphocytes, and the liver cell suspension was pressed through a 70-μm cell strainer (Corning Inc., Corning, NY) with the plunger of a 1-ml syringe and digested with 10 ml of RPMI 1640 medium (Sigma-Aldrich Co. LLC, St. Louis, MO), containing 0.02% (wt/vol) collagenase IV (Sigma-Aldrich) and 0.002% (wt/vol) DNase I (Sigma-Aldrich), for 40 min at 37°C. The cells were washed with RPMI 1640 and then laid over Percoll-Histopaque solution consisting of 12% Percoll (GE Healthcare UK Ltd., Little Chalfont, UK) and 88% Histopaque-1083 (Sigma-Aldrich). After centrifugation for 20 min at 750 × g, the IHLs were isolated at the interface. The lymph mononuclear cells were washed twice with RPMI 1640 medium and used for further analysis. Spleen cells were isolated by pressing them through a 70-μm cell strainer, washed three times with RPMI 1640 medium containing 5% fetal bovine serum (FBS), and used for further analysis.

Flow cytometric T cell analysis.

Lymphomononuclear cells isolated from the liver and spleen were incubated with a mixture containing the COR93-Kb dimer, peridinin chlorophyll protein (PerCP)-Cy5.5-conjugated anti-mouse CD8, and fluorescein isothiocyanate (FITC)-conjugated anti-mouse CD69 for 1 h on ice. After being washed, the cells were incubated for 30 min with allophycocyanin (APC)-conjugated anti-mouse IgG1 on ice to detect dimer-positive cells. The Kb dimer without peptide was used as a control. In adoptive T cell transfer experiments, V450-conjugated anti-mouse CD45.1 was added to the antibody mixture so that CD45.1⁺ donor TCR transgenic T cells could be followed in the CD45.2⁺ recipient mice. All antibodies were purchased from BD Biosciences (San Jose, CA). The cells were acquired using FACS Canto II (BD Biosciences), and the data were analyzed using FlowJo (FlowJo, LLC).

Tissue DNA, RNA, and protein analyses.

Total DNA was isolated from liver tissue exactly as described previously (18, 47). Total RNA was isolated from liver tissue using Isogen (Nippon Gene, Tokyo, Japan) by following the manufacturer’s instructions. Total liver DNA and RNA were analyzed for input HBV DNA content by Southern blotting and for HBV mRNA content by Northern blotting, as described previously, with minor modifications (17, 47). Briefly, total DNA and RNA were separated in 1% and 1.2% agarose gels and transferred to a positively charged nylon membrane (Roche Diagnostics GmbH, Mannheim, Germany). The membrane was then hybridized with a digoxigenin (DIG)-dUTP-labeled full-length HBV DNA fragment, which was generated using the DIG High Prime DNA labeling and detection starter kit II (Roche Diagnostics GmbH), and then detected by the alkaline phosphatase-labeled anti-DIG antibody according to the manufacturer’s instructions. The signals were analyzed using an ImageQuant LAS 4000mini (GE Healthcare UK Ltd.). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was used to normalize the amounts of mRNA among the samples. Total liver protein was extracted with a mixture of surfactants, followed by tissue homogenization to analyze the amounts of HBcAg and surface antigen (HBsAg) by Western blotting, as described previously (48, 49). Briefly, proteins from the liver lysate were separated by SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Millipore, Temecula, CA). The membrane was stained by monoclonal antibodies against HBcAg and HBsAg (19C1-8 and T5124A, respectively; Institute of Immunology Co., Tokyo, Japan), washed, and stained again by a horseradish peroxidase (HRP)-linked anti-mouse IgG (Cell Signaling Technology Inc., Beverly, MA) to HBcAg or an IgM (Abcam plc, Cambridge, UK) to HBsAg. The membrane was incubated with the Immobilon Western chemiluminescent HRP substrate (Millipore). The signals were analyzed using an ImageQuant LAS 4000mini. The blots were stripped and probed again for GAPDH to serve as a loading control.

Quantitative real-time PCR.

Quantitative mRNA levels were determined using quantitative real-time PCR (qRT-PCR) with StepOnePlus real-time PCR systems (Applied Biosystems, Foster City, CA). Complementary DNA (cDNA) was synthesized using the High Capacity RNA-to-cDNA kit (Applied Biosystems). We used the predesigned TaqMan gene expression assay (Applied Biosystems) for the analyses of mRNA levels of mouse Isg15 and mouse Gapdh, which was used for normalization.

Biochemical analyses.

Serum HBsAg and hepatitis B e antigen (HBeAg) were measured by a chemiluminescent enzyme immunoassay (CLEIA) using a commercial assay kit, LumipulseG1200 (Fujirebio, Tokyo, Japan), as described previously (16), and results are expressed as international units per milliliter and cutoff indexes (COIs), respectively.

Statistics.

Student’s t test (unpaired) was performed using Microsoft Excel. Data are depicted as means ± standard deviations (SD), and P values of <0.05 were considered significant.