|Home | About | Journals | Submit | Contact Us | Français|
The importance of effective immune responses in recovery from acute hepadnaviral hepatitis has been demonstrated. However, there is no conclusive delineation of virological and immunological events occurring in the liver immediately after hepadnavirus invasion and during the preacute phase of infection. These very early events might be of primary importance in determining the recovery or progression to chronic hepatitis and the intrinsic hepadnaviral propensity to persist. In this study, applying the woodchuck model of acute hepatitis B, the hepatic kinetics of hepadnavirus replication and activation of genes encoding cytokines, cytotoxicity effectors, and immune cell markers were quantified in sequential liver biopsies collected from 1 h postinoculation onward by sensitive real-time cDNA amplification assays. The results revealed that hepadnavirus replication is established in the liver as early as 1 hour after infection. In 3 to 6 h, significantly augmented intrahepatic transcription of gamma interferon and interleukin-12 were evident, suggesting activation of antigen-presenting cells. In 48 to 72 h, NK and NKT cells were activated and virus replication was transiently but significantly reduced, implying that this early innate response is at least partially successful in limiting virus propagation. Nonetheless, T cells were activated 4 to 5 weeks later when hepatitis became histologically evident. Collectively, our data demonstrate that virus replication is initiated and the innate response activated in the liver soon after exposure to a liver-pathogenic dose of hepadnavirus. Nevertheless, this response is unable to prompt a timely adaptive T-cell response, in contrast to infections caused by other viral pathogens.
Hepatitis B virus (HBV) is the prototypic member of the Hepadnaviridae family of small, enveloped, primarily hepatotropic DNA viruses which cause acute and chronic hepatitis and hepatocellular carcinoma (reviewed in reference 9). Hepadnaviruses display highly restricted host specificity. Thus, HBV infection is limited to humans and higher primates, while woodchuck hepatitis virus (WHV), despite sharing significant structural, genomic, and antigenic similarities with HBV, infects the eastern North American woodchuck (Marmota monax) (38). Notwithstanding their limited host ranges, HBV and WHV induce similar courses of hepatitis, which are preceded by a long incubation period ranging from 6 weeks to 6 months for hepatitis B and from 4 to 10 weeks for WHV hepatitis (9, 34). In contrast with the appearance of viremia and adaptive T-cell immunity a few days after invasion with most viral pathogens, a significant increase in HBV replication is not observed until 3 to 5 weeks postinfection (p.i.) in chimpanzees, while activation of HBV-specific adaptive immunity occurs several weeks later (4, 17, 59).
Despite the delay in activation of an HBV-specific T-cell response, a robust, multispecific T-cell reactivity seems to be essential for both induction of acute hepatitis (AH) and the clinical recovery (50, 51), although it is unable to clear the virus entirely (43, 55, 56). Experimental evidence acquired from the chimpanzee and HBV-transgenic mouse systems, as well as the woodchuck model of hepatitis B, has clearly indicated the importance of virus-specific CD4+ and CD8+ T lymphocytes and antiviral cytokines, such as alpha interferon, beta interferon, gamma interferon (IFN-γ), and tumor necrosis factor alpha (TNF-α), in the downregulation of virus replication (21, 35, 36, 48), while bystander recruitment of virus-nonspecific T cells and other immune cell subsets greatly contributes to liver inflammation and hepatocyte destruction (30, 58). In addition, the role of antiviral cytokines, particularly IFN-γ, has been ascribed to the ability to inhibit replication of hepadnavirus in a noncytopathic manner (15, 17). In this regard, elevations in the intrahepatic expression of IFN-γ appeared sufficient in chimpanzees to reduce HBV viremia before the peak onset of hepatitis (17). Also, an increased intrahepatic expression of IFN-γ was consistently detected for many years following recovery from WHV-induced AH, implying its role in control of residual (occult) WHV replication continuing after resolution of acute disease (21).
Natural killer (NK) and NK T (NKT) cells are components of the innate immune system. They are enriched among the resident lymphomononuclear cells in the liver and are capable of producing IFN-γ and other cytokines within minutes following viral invasion (reviewed in reference 8). The results from studies with HBV transgenic mice suggest that activation of NKT cells via CD1d-restricted α-galactosylceramide can downregulate virus replication (22, 23), while nonclassical NKT cells may recognize HBV antigens expressed in the liver (1). It has been suggested that NK- or NKT-derived IFN-γ may be the principal mediator of HBV downregulation in acutely infected chimpanzees (17), while indirect evidence from the woodchuck model suggests that activated, cytotoxic NK or NKT cells may contribute to hepatocyte killing and recovery from WHV infection (20).
Nonetheless, despite findings suggesting that hepadnaviruses may directly activate intrahepatic immune cells capable of producing cytokines promoting antiviral defense and favoring a T-helper-cell type 1 response (17, 59), comprehensive data from a natural model of hepadnaviral infection are lacking. Furthermore, characterization of the liver immune response in the first hours and days following hepadnavirus invasion has not yet been accomplished. The aim of the current study was to recognize, using the woodchuck model of hepatitis B, the nature and the kinetics of intrahepatic immune responses occurring soon after exposure to WHV and during the preacute and acute phases of hepadnaviral infection by quantifying hepatic transcriptional activities of genes encoding critical proinflammatory and antiviral cytokines and markers specifying individual immune cell subsets.
Prior to inoculation with WHV, liver biopsies (designated LBx 1) (Fig. (Fig.1)1) were obtained by laparotomy from 29 young-adult, healthy, WHV-naive woodchucks and their fragments were cryopreserved for subsequent DNA and RNA isolations and fixed in formalin for histological examination. All woodchucks were intravenously infected with the same WHV/tm3 inoculum carrying wild-type WHV at a dose of 1.1 × 1010 DNase-protected virus genome equivalents (vge) (GenBank accession number AY334075) (41). The animals were randomly subdivided into groups, each comprised of at least two animals (Table (Table1).1). Beginning at 1 h post-WHV injection, liver biopsies (designated LBx 2) were collected according to the scheme shown in Fig. Fig.1.1. Subsequent liver samples were obtained 6 weeks thereafter (designated LBx 3) and at the end of follow-up (designated LBx 4) (Fig. (Fig.1).1). Serum samples were collected weekly until 10 weeks p.i. and then monthly for up to 6 to 36 months postinoculation. Animals were housed in the Woodchuck Hepatitis Research Facility at Memorial University, St. John's, Newfoundland, Canada. All animal experimental protocols were approved by the Institutional Committee on Animals Bioethics and Care.
Detection of serum WHV surface antigen (WHsAg) and antibodies to WHV core antigen (anti-WHc) was done using specific immunoassays, as previously described (10, 41). Histological examination of paraffin-embedded liver tissue fragments was performed after routine processing and staining, as described elsewhere (21, 42). The morphological assessment of hepatocellular and extrahepatocellular intralobular and periportal lesions and overall degree of hepatitis was graded on a numerical scale from 0 to 3, applying criteria previously described (21, 39, 40).
Total DNA was extracted from 250-μl serum samples or approximately 50 mg of liver tissue using proteinase K digestion and the phenol-chloroform-isoamyl alcohol method as previously described (42). Total RNA was isolated from liver biopsy samples using Trizol reagent (Invitrogen Life Technologies, Burlington, Ontario, Canada), as per the manufacturer's instruction. RNA was treated with DNase to remove potentially contaminating DNA using an RNase-free DNase digestion kit (Sigma Chemical Company, Oakville, Ontario, Canada) prior to reverse transcription to cDNA, as previously described (18).
Quantitative assessment of the WHV DNA load was done by real-time PCR analysis using DNA derived from the equivalent of 25 μl of serum or 50 ng of liver biopsy sample DNA and the Roche LightCycler instrument (Roche Diagnostics, Laval, Quebec, Canada). WHV DNA amplicons were generated using the forward primer 5′-ATGCACCCATTCTCTCGAC and the reverse primer 5′-CTGAGCAGCTTGGTTAGAGT, yielding a 221-bp fragment which was detected by Sybr green I incorporation. Copy numbers of WHV DNA were calculated by extrapolation from a standard curve which was generated using 10-fold serial dilutions of plasmid containing known copy numbers of complete, recombinant WHV DNA. WHV RNA was similarly detected using 50 ng liver total RNA. The specificity of real-time PCR products was always confirmed by nucleic acid hybridization (NAH), i.e., Southern blot hybridization analysis, using complete recombinant WHV DNA as a probe and autoradiography, as previously described (10, 41). The sensitivity of the real-time PCR assay was ~200 vge/μg DNA, ~200 copies/μg RNA, or ~50 vge/ml serum.
WHV covalently closed circular DNA (cccDNA), representing virus genome replicative intermediate, was detected in liver tissue samples by PCR amplification with a sensitivity of 102 vge/μg total DNA, as previously described in detail (26, 41). Briefly, 4 μg of DNA was digested with mung bean nuclease prior to PCR amplification with oligonucleotide primers spanning the nicked region of the partially double-stranded WHV DNA genome. The specificity of PCR amplicons was routinely confirmed by NAH, as previously described (10, 42).
To facilitate analysis of the spectrum and the dynamics of intrahepatic immune response in WHV infection, a number of woodchuck gene sequences encoding markers specifying different immune cell subtypes, immune cell effector molecules, and cytokines were determined, applying a strategy previously reported (18). In general, woodchuck gene sequences were identified by reverse transcription-PCR (RT-PCR) using degenerate oligonucleotide primers whose sequences were deduced through interspecies comparison of the sequences available in GenBank. The resulting amplicons were cloned into the PCRII TOPO TA cloning system (Invitrogen) and the excised fragments sequenced in both directions. Based on the woodchuck sequences determined, the pairs of gene-specific primers were designed. Table Table22 presents the list of woodchuck genes for which hepatic expression was quantified in the current study and the primer pairs used for their quantitative detection.
Real-time RT-PCR assays were developed for each cellular effector molecule and the cytokine gene analyzed, using the Roche LightCycler instrument with Sybr green I detection and PCR primer pairs specified in Table Table2.2. Changes in gene expression levels were determined by comparison to the baseline level of each gene's transcription detected in the liver biopsy sample obtained prior to infection with WHV for each individual animal after normalization to expression of the housekeeping gene β-actin. Following measurement of a given gene's expression in a particular liver sample from an individual animal, the mean expression level was determined for all liver biopsy samples within the experimental groups showed in Table Table11.
To ascertain that the observed changes in the intrahepatic gene expression levels were truly related to infection with WHV but not to injection with serum components present in the WHV inoculum, a control experiment was performed. Thus, the WHV/tm3 inoculum was inactivated by treatment with 50 μg/ml psoralen (Sigma) combined with exposure to 365 nm UV light for 90 min at 4°C. Animals were injected intravenously with 0.5 ml of psoralen-inactivated inoculum containing the equivalent of 1.1 × 1010 vge or with 0.5 ml of similarly treated healthy woodchuck serum prior to euthanasia at 3 days postinjection.
A two-tailed, unpaired Student t test with 95% confidence interval was applied to compare the means of sample groups investigated, and P values of <0.05 were considered statistically significant.
The accession numbers for the woodchuck gene sequences established in the course of this study have been submitted to GenBank under the following accession numbers: for the CD4 gene, EF621765; for the CD8 gene, EF621766; for the CD40 ligand (CD40L), EF621170; for CD1d, EF621767; for NKp46, EF621768; and for interleukin-8 (IL-8), EF-126348.
Inoculation of woodchucks with a WHV dose of 1.1 × 1010 vge resulted in transiently serum WHsAg-positive, self-limiting acute infection in all 29 animals. The mean time of WHsAg appearance was 18.5 days, while its average duration in the circulation was 35.8 days. Anti-WHc remained detectable until the end of the observation period (Fig. (Fig.2A).2A). The pattern of serum WHsAg positivity implied, based on the known profiles of serological markers of WHV infection (41, 42), that a self-limited episode of AH had developed in all woodchucks investigated. This was confirmed by histological examination of serial liver biopsy samples collected.
Considering morphological alterations encountered in liver tissue, their characteristics and the timing of their occurrence closely followed those previously reported for experimental acute WHV infection (38, 42). In brief, the appearance of scattered lymphocytes and neutrophils in sinusoids and periportal areas, accompanied by mild proliferations of sinusoidal lining endothelium and bile ducts, was the first manifestation, which occurred at 4 to 5 weeks p.i. In the next few days, scattered lobular infiltrations, consisting mainly of lymphocytes, degenerative or necrotic changes of single hepatocytes, progressing hyperplasia of Kupffer cells, and limited portal infiltrations, occurred. Subsequently, more-severe lobular infiltrations and hepatocyte degenerative and necrotic changes and the appearance of acidophilic bodies were observed. At the peak of AH, usually occurring between weeks 6 and 8 p.i., both lobular and periportal inflammatory changes were evident, the portal infiltrations became most intense and expanded occasionally through limiting plates, and multifocal necroses of parenchyma with varying numbers of acidophilic bodies were found. However, the severity of AH varied, and while the changes in some woodchucks were intense, reaching histological degree of hepatitis 2.5, in others they were mild and hepatitis did not exceed grade 1.5. Following resolution of AH, intermittent minimal lobular and portal inflammatory alterations and periods of normal or nearly normal liver morphology were observed, as previously reported (21, 41, 42).
In previous studies, WHV DNA in the circulation and the virus replication status in the liver were never assessed during the first few hours or days after exposure to the virus. In the current study, WHV DNA was quantified in serum and liver tissue acquired from 1 h p.i. onward. Not surprisingly, WHV DNA was detectable in serum at the first time point of examination, i.e., 1 h p.i. (Fig. (Fig.2A).2A). This reflected the carryover of WHV from the inoculum containing a massive amount of virus, i.e., 1.1 × 1010 DNase-protected vge. Subsequently, the serum level of WHV DNA progressively decreased until 48 to 72 h p.i. and then slowly increased until a sudden expansion at 3 to 4 weeks p.i. (Fig. (Fig.2A),2A), culminating in peak detection at week 7 p.i. (a mean level of 3.2 × 1010 vge/ml). The mean WHV DNA level exhibited minor fluctuations during the peak acute phase of WHV infection, occurring between weeks 6 and 9 p.i., prior to a sudden protraction by more than 7 logs after week 9 p.i. (Fig. (Fig.2A).2A). However, in agreement with our previous findings (10, 21, 42), WHV DNA remained consistently detectable in serum at a mean level of 175 vge/ml during follow-up lasting up to 3 years p.i.
WHV DNA and WHV mRNA transcripts were also detectable in hepatic tissue beginning at 1 h after inoculation with virus. The WHV DNA load transiently but significantly (P = 0.004) and uniformly declined between 48 to 72 h p.i. in all animals, reaching a level approximately 10-fold lower than that detected between 1 and 6 h p.i. (Fig. (Fig.2A).2A). Paralleling WHV DNA detection in serum, the intrahepatic level of WHV DNA exhibited an exponential increase at 3 to 4 weeks p.i. (Fig. (Fig.2A)2A) prior to reaching the mean peak detection of 4 × 109 vge/μg total DNA at 7 weeks p.i. Despite a greater than 4-log reduction in the hepatic WHV DNA load beginning at week 9 p.i., WHV DNA was consistently detectable in the liver for up to 3 years p.i. at a mean level of 2.3 × 104 vge/μg total liver DNA. It remained unclear whether the WHV DNA detected in the liver in the first few hours postexposure reflected the virus originating from the inoculum, which was passing through or was trapped within hepatic tissue, or the virus actively replicating in hepatocytes. However, the finding of a significant lowering in the hepatic load of WHV DNA at 48 to 72 h p.i., which was replenished 24 h later, strongly suggested that the virus was already actively replicating, at least from 96 h p.i. forward.
To directly determine the status of WHV replication in the liver, real-time RT-PCR detecting WHV mRNA was applied. The results showed that low quantities of WHV transcripts (2.6 × 102 copies/μg total RNA) were identifiable as early as 1 h p.i. (Fig. (Fig.2A).2A). In contrast with hepatic WHV DNA, the WHV RNA level progressively increased in the liver and achieved the maximum at a mean level of 7.2 × 108 copies per μg total RNA between weeks 8 and 9 p.i. (Fig. (Fig.2A).2A). A possibility of contamination of cDNA preparations with viral DNA was excluded by DNase treatment of RNA samples prior to the RT reaction and by PCR amplification of both transcribed and nontranscribed mRNA preparations. As shown in Fig. Fig.2B,2B, only RNA samples which were reverse transcribed demonstrated WHV cDNA signals, while those similarly treated in the absence of reverse transcriptase remained negative even when the amplification products were analyzed by NAH. This confirmed strict specificity of the WHV RNA detections. Taken together, the results demonstrated that infection with a massive, liver-pathogenic dose of WHV results in almost immediate establishment, although at a low level, of hepadnavirus replication in the liver, as evidenced by detection of viral transcripts as early as 1 h p.i.
In addition, we analyzed expression of WHV cccDNA by a PCR/NAH assay specifically identifying this replicating DNA intermediate. The results revealed that the WHV cccDNA signal was detectable in the liver from 18 h p.i. onward (data not shown).
Since injection with WHV resulted in the initiation of virus replication as early as 1 h p.i., it was of interest to determine whether the virus may at this early stage activate antigen-presenting cells (APC), which should be the first cell type recognizing viral intrusion. In this regard, the liver expression of IL-12, a key cytokine produced by APC involved in innate immune responses (29), IL-8, a strong chemoattractant mediating chemotaxis of phagocytic cells (24), CD1d, a key molecule facilitating antigen presentation by APC to NKT cells (1, 25), and CD40L, involved in activation of APC via CD40-CD40L, was quantified by using real-time RT-PCR assays specifically developed for the purpose of this study.
Figure Figure33 shows that IL-12 achieved a maximum hepatic expression (~20-fold increase over the preinfection level; P = 0.02) between 3 and 6 h p.i. (Fig. (Fig.3A,3A, left panel). Interestingly, in contrast with this early increase, the remaining course of WHV infection, including the period of histologically evident AH with peaking histologically evident liver injury occurring between weeks 6 and 8 p.i. (Fig. (Fig.3A,3A, right panel), was without any noticeable increase in expression of this cytokine.
The transcription level of IL-8 tended to rise (a 12.5-fold increase) within 1 to 3 h p.i. above that detected in the period prior to infection; however, this increase did not reach a statistically significant value (P = 0.1) (Fig. (Fig.3A,3A, left panel). The maximum expression of the cytokine (a 35-fold increase over the preinfection level; P = 0.05) occurred at week 7 p.i., during the phase of acute liver inflammation (Fig. (Fig.3A,3A, right panel).
Since a strong induction of IL-12 expression detected at 3 to 6 h p.i. could be directly associated with activation of APC (29), the intrahepatic expression of CD1d, which facilitates antigen presentation to NKT cells, was also evaluated. The results showed that the level of CD1d mRNA reached a peak (a ~3-fold increase over the preinfection level; P = 0.03) by 48 to 72 h p.i. (Fig. (Fig.3B,3B, left panel). Then, the level subsided until WHV replication was drastically augmented at 3 to 4 weeks p.i. (Fig. (Fig.3B,3B, right panel). Thus, the upregulated transcription of CD1d at 48 to 72 h was associated with a significant one-log decrease in the hepatic WHV DNA load and a decline of WHV DNA in circulation. Furthermore, the concomitant increase (P = 0.02) in CD40L expression (Fig. (Fig.3B,3B, left panel) suggested activation of APC via CD40-CD40L interaction.
WHV infection induced two distinct phases of very early immune activation in the liver, one between 3 and 6 h and a second at 48 to 72 h p.i. Since experimental evidence from other viral infections indicates the ability of NK and NKT cells to respond within minutes to hours to virus by secretion of IFN-γ or by acquisition of cytotoxic function (8), expression of the gene encoding the NK receptor NKp46 was investigated. Furthermore, since IL-12 is recognized as a key cytokine which influences IFN-γ secretion by NK cells and also increases NK-mediated cytotoxicity, which is primarily facilitated by the perforin pathway (44), perforin mRNA levels were also quantified. The results showed that the earliest increase in transcription of IL-12, occurring between 3 to 6 h p.i. (Fig. (Fig.3A,3A, left panel), coincided with significantly augmented transcription of IFN-γ at the same time (3.3-fold; P = 0.018) (Fig. (Fig.4B,4B, left panel). However, apparent increases in the expression of NKp46 (a 2.5-fold induction) and perforin (a 3.4-fold induction) at 3 to 6 h p.i. (Fig. 4A and B, left panels, respectively) did not reach statistically meaningful levels (for both, P = 0.13).
As was shown in Fig. Fig.3B,3B, the early expression of CD1d peaked at 48 to 72 h p.i. This finding raised the possibility that APC may at this stage display an increased capacity to present antigens to CD1d-restricted NKT cells (25). The activation of NKT cells may result in upregulated expressions of IFN-γ and IL-4 (2). As the data indicated, enhanced CD1d expression was indeed associated with an increase in IFN-γ transcription at 48 to 72 h p.i. (fourfold; P = 0.032). Conversely, a twofold increase in IL-4 transcription (Fig. (Fig.4B,4B, left panel) did not achieve the level of statistical significance (P = 0.065).
After week 2 p.i., IFN-γ and perforin mRNA levels reached their maximum, not surprisingly, at week 7 p.i., during the peak of histologically evident liver injury, which is characterized by the most extensive liver lymphomononuclear infiltrations (Fig. (Fig.5A)5A) (34, 38). Intriguingly, intrahepatic transcription of IL-4 and that of the NK cell marker, NKp46, reached a maximum at week 3 p.i., i.e., before the exponential viral expansion in the liver (Fig. (Fig.4B,4B, right panel).
Activation of innate immune effector mechanisms in response to hepadnaviral infection was further supported by an augmented expression of the antiviral 2′,5′-oligoadenylate synthetase (OAS) (Fig. (Fig.4C).4C). Thus, detection of WHV DNA and WHV RNA, as early as 1 to 3 h p.i., was associated with some increase in the OAS mRNA level (1.7-fold), as determined by real-time RT-PCR, but this was not significant (P = 0.14) (Fig. (Fig.4C,4C, left panel). However, at 48 to 72 h p.i., a significant increase (P = 0.05) in OAS (Fig. (Fig.4C,4C, left panel) correlated with a significant (P = 0.03) elevation in expression of IFN-γ (Fig. (Fig.4B,4B, left panel), a known inducer of OAS (45, 60). This occurred at the time when the WHV load significantly decreased in the liver. Taken together, the coincidence of these events supports the notion that the transient upregulation of hepatic IFN-γ occurring in the first 72 h following hepadnaviral exposure was biologically relevant and exerted a strong antiviral effect.
The prominent reduction in hepadnavirus replication and resolution of AH appear reliant upon a strong and multispecific antiviral T-cell response, which is characterized by secretion of IFN-γ and TNF-α (15, 16). To assess whether the observed initial elevations in intrahepatic expression of IFN-γ (Fig. (Fig.4B,4B, left panel) could be due to the presence of activated T cells, hepatic transcription of genes encoding CD4, CD8, and CD3 T-cell markers was quantified. It was found that CD4 and CD8 mRNA levels tended to increase (by 20 to 25%) in the first 3 h post-WHV exposure, but these increases did not reach a statistically significant difference (P = 0.18 and 0.12, respectively) over the preinfection levels. The CD4 and CD8 mRNA levels remained consistently low (10 to 15% of the preinfection levels; P = 0.3 and 0.2, respectively) until week 4 p.i. (Fig. (Fig.5A,5A, left panel). However, not surprisingly, very prominent increases in the mRNA levels of CD4 (8.3-fold; P = 0.017) and CD8 (6.7-fold; P = 0.03) were detected during the peak of histologically evident liver inflammation, which also correlated with detection of the maximal level of IFN-γ mRNA (a 31.3-fold increase; P = 0.003). On the other hand, the accompanied elevation (a 17-fold increase) in TNF-α mRNA did not achieve a statistically different level (P = 0.07) (Fig. (Fig.5A,5A, right panel).
In contrast to CD4 and CD8, the level of CD3 mRNA showed a distinctive peak (a 2.9-fold increase; P = 0.03) at 48 to 72 h p.i. (Fig. (Fig.5A,5A, left panel). Subsequently, the CD3 expression level became again augmented beginning at week 3 p.i., preceding the rise in CD4 and CD8 transcription by approximately 1 week. Then, the expression profiles of CD3, CD4, and CD8 paralleled each other both during and after the acute phase of hepatitis (Fig. (Fig.5A,5A, right panel). These results suggested that upregulated intrahepatic expression of IFN-γ (Fig. (Fig.4B,4B, left panel) detected at 48 to 72 h p.i. reflected activation of cells of the innate immune system rather than conventional T-cell subsets; however, a contribution of the latter cannot be completely excluded. In the same time period, a tendency to elevate transcription of TNF-α (a 2.2-fold increase over the preinfection level) was also observed, but this increase did not achieve a statistically meaningful difference (P = 0.08) (Fig. (Fig.5A,5A, left panel). The augmented coincident expression of CD3, CD1d (Fig. (Fig.3B,3B, left panel), and IFN-γ (Fig. (Fig.4B,4B, left panel) at 48 to 72 h p.i. implied that NKT cells also became activated soon after exposure to pathogenic hepadnavirus.
In addition, the enhanced expression of CD3, preceding the rise in CD4 and CD8 expression during the acute phase of WHV hepatitis (Fig. (Fig.5A,5A, right panel) and coinciding with the augmented transcription of NKp46, may suggest that the activation of both NKp46-positive NK cells and CD3-positive NKT cells was taking place before major infiltration of the liver with CD4+ and CD8+ T lymphocytes (Fig. (Fig.5A,5A, right panel). In addition, it appears that CD3-positive NKT cells at this relatively late stage of infection, i.e., week 3 p.i., were responding by producing IL-4 (Fig. (Fig.4B,4B, right panel).
The intrahepatic expression of perforin was significantly augmented (1.9-fold; P = 0.05) at 48 to 72 h p.i. (Fig. (Fig.5B,5B, left panel). To ascertain whether an increase in the local cytotoxic activity in the liver might be responsible for the transient decrease in hepatic WHV DNA, expression of CD95L, an effector molecule capable of inducing death of hepatocytes, which are naturally endowed with CD95 (49), was quantified by real-time RT-PCR. As shown in Fig. Fig.5B5B (left panel), CD95L was significantly (P = 0.03) upregulated (2.6-fold) at 48 to 72 h p.i., suggesting that indeed augmented hepatic cytotoxicity, possibly mediated by both perforin and CD95L, might contribute to the transient depletion of WHV DNA seen at 48 to 72 h p.i. In the time period after week 2 p.i., CD95L expression in the liver became noticeably enhanced, reaching a maximum (a 4.8-fold increase; P = 0.01) at week 4 p.i. (Fig. (Fig.5B,5B, right panel). Surprisingly, this CD95L mRNA peak preceded both the expansion of virus in hepatic tissue and the rise in CD4 and CD8 mRNA levels coinciding with the appearance of histologically evident AH (Fig. (Fig.5A,5A, right panel). On the other hand, this peak of CD95L expression occurred together with the upregulated transcription of CD3, suggesting a possible contribution of NKT cells to the increased intrahepatic detection of CD95L. However, we have previously reported that hepatocytes also constitutively express CD95L (18). Thus, it cannot be completely excluded that an increase in the intrahepatic CD95L mRNA level was due at least in part to the elevated transcription of CD95L in hepatocytes. It is potentially possible, since IFN-γ upregulates CD95L transcription in hepatocytes and intrahepatic expression of this cytokine was augmented (P = 0.032) between 3 and 6 h p.i. (Fig. (Fig.4B,4B, left panel).
To further ascertain that WHV induced activation of an intrahepatic immune response as early as 72 h p.i., two additional healthy woodchucks were intravenously injected with 1.1 × 1010 vge of the WHV/tm3 inoculum and analyzed at 3 days p.i. In agreement with the first set of data, significant elevations in CD3 (P < 0.0001), NKp46 (P = 0.03), and CD1d (P < 0.0001) expression levels were detected 72 h after injection with virus, suggesting again an accumulation and/or local activation of NK and NKT cells (Fig. (Fig.6A).6A). However, transcription of CD4 and CD8 was also meaningfully upregulated (by 47% and 21%; P = 0.001 and 0.003, respectively) compared with the hepatic expression of the genes prior to WHV infection (Fig. (Fig.6A).6A). Despite these increases, we did not notice any morphological evidence of lymphomononuclear infiltrations, suggesting that this augmented expression could be due to the activation of the cells already residing within the liver.
Quantification of the cytokine expression at 72 h p.i. also confirmed the data obtained from the first experiment demonstrating significant elevation of IFN-γ and a trend toward augmented expression of IL-4, possibly reflecting activation of NKT cells (Fig. (Fig.6B).6B). In this supplementary experiment, both IFN-γ (P = 0.004) and IL-4 (P < 0.0001) transcription was significantly augmented over the preinfection levels. In addition, a meaningful (P < 0.0001) increase in TNF-α may reflect activation of intrahepatic macrophages, as was previously suggested in regard to a transient elevation in IL-12 mRNA (Fig. (Fig.3A,3A, left panel) at 3 to 6 h p.i., which had subsided by 72 h p.i. (Fig. (Fig.3A,3A, left panel, and 6B). Similarly, as shown in Fig. Fig.4C4C (left panel), detection of OAS mRNA was significantly (P = 0.0017) elevated at 72 h (Fig. (Fig.6B),6B), confirming that the invading WHV is recognized by effector immune cells. Finally, expression of both the cytotoxic effector molecules CD95L (P < 0.0001) and perforin (P = 0.047) was significantly elevated at 72 h p.i. (Fig. (Fig.6C),6C), raising again the possibility that antiviral cytotoxic mechanisms also are activated in the liver at this very early stage of hepadnavirus infection.
To exclude the possibility that exposure to components of serum carrying the virus but not to WHV itself might be responsible for activating the intrahepatic immune response, additional control experiments were performed using an inactivated WHV inoculum or healthy woodchuck serum. For this purpose, a sample of woodchuck serum serving as WHV/tm3 inoculum and containing 1.1 × 1010 WHV vge and serum from a healthy woodchuck were treated with psoralen under UV light. As shown in Fig. Fig.6D,6D, inactivated WHV inoculum or similarly treated control normal woodchuck serum was unable to upregulate expression of the hepatic genes, which have been previously found to be significantly augmented at 72 h p.i. Specifically, there were no increases in IFN-γ or CD3 mRNA levels and no induction of OAS (Fig. (Fig.6D).6D). These results confirmed that productive WHV infection but not exposure to serum components or circulating viral antigens was responsible for activation of intrahepatic immune responses as found in our study.
A unique feature of hepadnaviral hepatitis is a prolonged incubation period where no apparent clinical symptoms or biochemical manifestations of liver injury are evident. Previous studies, using both the woodchuck model of hepatitis B (11, 34, 35) and HBV-infected chimpanzees (47, 62), have suggested that viral replication in the liver remains largely undetectable until 3 to 4 weeks p.i., at which time exponential virus expansion leads to infection of all or almost all hepatocytes. These studies have also demonstrated that antiviral immunity, mediated predominantly by virus-specific CD8+ cytotoxic T lymphocytes via both noncytolytic and cytotoxic mechanisms, is finally responsible for downregulation of hepadnaviral replication and clinical recovery from hepatitis. Furthermore, manipulations of the antiviral immune response in HBV-transgenic mice have suggested an involvement of innate immune cell subsets in inhibition of viral replication (22, 23). Previous observations with HBV-infected chimpanzees (62) and WHV-infected woodchucks (20) have also suggested an involvement of the innate immune system in controlling hepadnavirus infection. However, these studies commenced evaluation of virus replication and intrahepatic immune responses not earlier than 1 week postinfection (62).
Since the half-life of HBV in serum is estimated to be as short as 4 h (46), although it might even be shorter (13), it could be expected that infection of hepatocytes may occur promptly after exposure to virus, at least in situations when the host is exposed to liver-pathogenic doses of virions (41). Furthermore, it is acknowledged that cells of the innate immune system are activated within minutes to hours following invasion with viral pathogens (8). Taken together, we hypothesized that inoculation with WHV doses known to induce serologically and histologically evident AH, i.e., exceeding 103 to 104 virions (41), should result in activation of the hepatic innate immunity, although one pertinent study using three chimpanzees infected with HBV, in which expression of the innate immune response-affiliated genes was assessed by cDNA microarray analysis (61), suggested otherwise. To investigate these issues, a large cohort of WHV-naive woodchucks was infected with a well-characterized WHV inoculum containing 1.1 × 1010 DNase-protected vge per dose. This large cohort of animals permitted reliable determination of viral kinetics and the status of intrahepatic immune activation starting as early as 1 h p.i.
Our quantitative analysis, applying a real-time PCR assay, demonstrated WHV DNA in serum (~4 × 106 vge/ml) and liver tissue (~6 × 105 vge/μg total DNA) at 1 h after injection with virus. However, these high levels of the WHV genomes detected almost immediately after inoculation almost certainly originated to a large degree, if not entirely, from carryover of the inoculum injected. On the other hand, quantification of WHV RNA provided a direct insight into the status of virus replication in hepatic tissue. At 1 h p.i., WHV transcripts were found at levels approximating 200 to 400 copies per μg total liver RNA. This indicated that the virus was able to enter hepatocytes, repair its partially double-stranded DNA, and transcribe DNA to mRNA within 1 h after injection. In this regard, our study is the first where the evidence of such early replication of hepadnavirus in vivo has been documented. In our work, viral mRNA was detected by sensitive real-time RT-PCR (sensitivity of <200 vge/μg RNA). In one pertinent but not fully compatible study, in which HBV RNA was evaluated by an RNase protection assay with HBV-infected chimpanzees, HBV pregenomic RNA transcripts were detected in the liver beginning at 3 to 4 weeks p.i., with subsequent exponential expansion between weeks 4 and 6 p.i. (61). The difference between ours and the study mentioned is almost certainly due to the greater sensitivity of the PCR-based WHV mRNA detection. Our finding of WHV replication in the liver as early as 1 h p.i. is not unique and appears to be compatible with replication kinetics delineated for some other viruses. For example, evidence of de novo synthesis of measles virus RNA in HeLa cells was found as early as 2 h p.i. when analyzed by real-time RT-PCR; however, the earlier time points postinoculation were not examined in this study (52). Also, transcription of the nucleopolyhedrovirus in its host was detected at 1 h p.i. by RT-PCR (14).
Our attempt to detect WHV cccDNA in the liver in the first few hours after inoculation was not successful. This replicative intermediate of the WHV genome, which constitutes an obligate prerequisite for the generation of hepadnaviral mRNA transcripts (32), was for the first time identified at 18 h p.i. The discrepancy between the time of the earliest detection of WHV mRNA and WHV cccDNA was not surprising and was most certainly related to a greater sensitivity of the WHV mRNA RT-PCR assay than the PCR assay available for WHV cccDNA detection (at least 5- to 10-fold) and to naturally occurring lower copy numbers of WHV cccDNA than WHV mRNA in infected cells. In one related study, in which early kinetics of duck HBV replication were investigated using in vitro-infected primary duck hepatocytes examined by classical Southern blot hybridization methods, virus cccDNA and single-stranded DNA, both indicative of active replication, were detected at 48 h p.i. (53).
While the hepatic load of WHV mRNA transcripts progressively increased starting from 1 h p.i., there was only a slight parallel increase in the WHV DNA level until 3 weeks p.i. compared with that detected at 48 to 72 h p.i. From week 3 to 6 weeks p.i., a strong coordinated expansion of levels of WHV DNA and mRNA was apparent, suggesting exponential viral replication. This result is in agreement with previous findings from HBV infection showing that exponential viral replication includes proportional increases in expression of both hepadnaviral genomes and replicative intermediates (47, 62). Subsequently, a parallel increase in viral RNA transcripts and DNA, although of a lower magnitude, continued until 8 to 9 weeks p.i. However, there was a noticeable transient, but not statistically meaningful, decrease in the detection of both nucleic acid forms in the liver and WHV DNA in serum around week 6 p.i. From week 10 p.i. forward, a progressive decline in hepatic loads of WHV RNA and DNA occurred, lasting until 30 weeks p.i. Nonetheless, traces of WHV DNA and RNA remained detectable in hepatic tissue until the end of the observation period, which was as long as 3 years p.i. in some animals. This finding was consistent with the results of previous studies where the life-long persistence of low-level replication of infectious WHV after seemingly complete serological and biochemical resolution of AH was documented (10, 21, 42; reviewed in reference 37).
The microenvironment of the liver displays unique immunological properties which have been ascribed to hepatic APC, including liver sinusoidal endothelial cells and Kupffer cells, and to the disproportionate occurrence of NK and NKT cells (reviewed in reference 54). Furthermore, recruitment of NK and NKT cells into the liver from the splenic compartment has been observed following viral infection (57). Our results suggested that WHV infection resulted in apparent sequential activation of APC and NK or NKT cells within the liver, leading to a temporary decrease in the viral load. Specifically, a significant (P < 0.004) one-log decrease in hepatic WHV DNA content was detected in all animals whose livers were sampled between 48 to 72 h after infection. The hepatic WHV DNA level returned to that seen prior to this decrease approximately 24 h later, implying that the rebound was due to active WHV replication. This temporary reduction of WHV DNA was associated with significantly augmented hepatic transcription of antiviral mediators, such as IFN-γ and OAS, cellular markers CD1d and CD3, and cytolytic effector molecules CD95L and perforin. This strongly argues that an activation of intrahepatic innate immunity caused this transient but significant decline in the viral load, although without apparent modification in the level of virus transcription. These results are the first of this kind, and they do not conform with those reported for experimental HBV infection in chimpanzees, which suggested that innate immunity is not activated by hepadnaviral infection (61). However, the aforementioned study began examination of the gene expression in serial liver biopsies from 1 week p.i. and applied less-sensitive microarray analysis, while evaluation of the host response in our study was commenced within 1 h after WHV inoculation and applied 5- to 10-fold more-sensitive real-time RT-PCR assays.
Experimental evidence accumulated from infections with other viral pathogens clearly indicates that NK and NKT cells have the ability to respond to virus by production of IFN-γ or by acquisition of a cytotoxic function within minutes or hours after infection (8). In this regard, recognition of viral antigens by the activating NK receptor NKp46 has been implicated as a key stimulus during influenza virus infection (31). Our data demonstrated that as early as 3 to 6 h after exposure to WHV, there was a significant increase in intrahepatic expression of IFN-γ. This was accompanied by trends, although without reaching statistically significant difference, toward elevated expression of NKp46 (a 2.5-fold induction) and perforin (a 3.4-fold induction), a key effector molecule mediating NK cell cytotoxicity. This may suggest that NK cells are activated almost immediately following WHV infection, although the current data are not conclusive in this regard. These events occurred in parallel with a significant upregulated expression of IL-12, implying that the initial production of IFN-γ may have augmented the activation of APC.
Based on detection of WHV transcription shortly after exposure and the fact that hepadnaviral envelope proteins may activate NKT cells (1), it is reasonable to suggest that early activation of NK cells could lead to enhanced presentation of WHV antigens to CD1d-restricted NKT cells, culminating in elevations of intrahepatic IFN-γ and IL-4 at 48 to 72 h p.i. In support of this possibility, CD3 expression was also found to exhibit a distinct peak (a 2.9-fold increase comparing to the preinfection level) at 48 to 72 h p.i. (see Fig. Fig.5A,5A, left panel). Since CD1d-restricted NKT cells express a T-cell receptor (TCR) comprised of TCRα/β chains in combination with the CD3 receptor complex (25), significant elevations in CD3 but not CD4 or CD8 T-cell markers could be interpreted in support of the idea that WHV infection also activated intrahepatic NKT cells, resulting in the increased expression of IFN-γ. It has been shown that activation of CD1d-restricted NKT cells inhibits viral replication in HBV-transgenic mice via noncytopathic mechanisms mediated by IFN-γ (22). Overall, the results obtained suggest that WHV infection shortly after invasion may first activate NK cells and subsequently NKT cells, with the latter possibly contributing to a transient decrease in viral DNA. The activation of the intrahepatic innate immunity was transient, waning by 72 h p.i. Therefore, it might not be surprising that this innate immune response was undetected in the liver in previous studies which commenced evaluation of expression of the genes affiliated with this response at 1 week after inoculation of chimpanzees with HBV (61).
Infection with lymphocytic choriomeningitis virus has been shown to induce NKT-cell activation, resulting in IFN-γ and IL-4 expression, which is immediately followed by a decrease in their levels and a subsequent increase several weeks later (19, 28). Interestingly, peak expression of the NK marker, NKp46, occurred in our study at 3 weeks p.i., coinciding with significant elevations in IL-4 expression, in the absence of upregulated transcription of IFN-γ. Although we cannot determine the cellular site of augmented expression of IL-4, the histological analysis showed a lack of lymphomononuclear inflammatory infiltrations in the liver at that time. Thus, NKT cells may represent the predominant cell type responsible for the increased hepatic expression of IL-4 in our study. Since specific antibodies for detection of woodchuck NKT cells or IL-4 are currently lacking, the explanation of this possibility will require further investigation when such reagents become available. Furthermore, WHV infection skewing toward the T-helper-cell type 2 response, characterized by IL-4 and the absence of IFN-γ expression, is another enticing possibility.
Despite activation of innate immune cell subsets in the liver during days following WHV infection, coinciding with significant reductions in viral load, this initial antiviral response waned and failed to promptly induce a CD4+ and CD8+ T-cell response until 5 to 6 weeks later. These findings are in contrast with those encountered during infections with other viral pathogens, which tend to induce timely sequential activation of innate and adaptive immune cell subsets, leading normally to the self-resolution of acute infection. The prolonged period of antigen-specific immunological ignorance to hepadnaviral infection, as depicted in our study by a lack of hepatic expression of CD4 and CD8 T-cell marker genes and an absence of liver inflammation following activation of the innate immune response or in the studies by others as a lack of hepadnavirus-specific T-cell responses (4, 35), may be partially explained by the tolerance-inducing effect of the liver. The capacity of the liver to induce tolerance to oral or allogeneic antigens is now well recognized (3, 5, 27). It is thought that this occurs via several mechanisms, including suboptimal T-cell priming and induction of T-cell anergy (reviewed in references 6 and 12). In our study, the expression of CD4 and CD8 was transiently elevated immediately following infection. In addition, inflammatory mediators, including IFN-γ, have been shown to influence the expression of adhesion molecules on endothelial cells, which have been implicated in mediating T-cell trapping in the liver (33). Furthermore, transient activation of antigen-specific T cells has been observed in TCR-transgenic mouse models, wherein cognate antigen presentation was restricted to hepatocytes, leading to dysfunctional priming of naive T cells (5, 7). Thus, initial trapping of CD4 and CD8 T cells, followed by suboptimal priming or deletion, may potentially facilitate hepadnaviral subversion of the adaptive immune response.
In summary, our findings provide new insights into the features of the immune responses associated with hepadnaviral infection. The data obtained indicate that infection induced by a liver-pathogenic dose of hepadnavirus rapidly initiates viral replication in hepatic tissue and activates the local immune system. The infection appears to almost immediately stimulate intrahepatic innate immune cells, including APC, NK, and NKT cells, which coincides with a transient but profound suppression of the hepatic virus load. However, in contrast with other viral infections, this very early immune activity does not precipitate a swift adaptive T-cell immune response. The reason behind this is unknown, but it could be due to as yet unidentified viral factors or a consequence of the liver's ability to induce immune tolerance.
We thank Colleen L. Trelegan for her expert technical assistance, Luke Grenning for assistance during woodchuck laparotomies, and Judy Foote for histological processing of liver biopsies. We also thank Julia Pohling for her contribution to identification of woodchuck IL-8 gene sequence.
This research was supported by grant MOP-14818 from the Canadian Institutes of Health Research (to T.I.M.). C.S.G. was supported in part by a Canadian Liver Foundation doctoral fellowship. T.I.M. is the Canada Research Chair (Tier 1) in Viral Hepatitis/Immunology, supported by the Canadian Institutes of Health Research and the Canada Foundation for Innovation.
Published ahead of print on 2 July 2008.