Hepatitis delta virus infects the cells of hepadnavirus-induced hepatocellular carcinoma in woodchucks

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Hepatitis delta virus (HDV) is a natural subviral agent of human hepatitis B virus (HBV). HDV enhances liver damage during concomitant infection with HBV. The molecular pathogenesis of HDV infection remains poorly understood. To advance our understanding of the relationship between HDV infection and liver cancer, it was determined whether HDV could infect in vivo the cells of hepadnavirus-induced hepatocellular carcinoma (HCC). Woodchucks (Marmota monax) that were chronically infected with HBV-related woodchuck hepatitis virus (WHV) and already developed HCCs were used as an experimental model. The locations of HCCs within the livers were determined using ultrasound imaging followed by open surgery. One week after surgery the WHV carrier woodchucks were superinfected with WHV-enveloped HDV (wHDV). Six weeks later the animals were sacrificed and HDV replication in normal liver tissues and in center masses of HCCs was evidenced by Northern analysis, real-time polymerase chain reaction assay, and immunohistochemistry. Based on accumulation levels of HDV RNAs and numbers of infected cells, the efficiency of wHDV infection appears to be comparable in most HCCs and normal liver tissues. Conclusion: Cells of WHV-induced HCCs are susceptible to HDV infection in vivo, and therefore express functional putative WHV receptors and support the steps of the attachment/entry governed by the hepadnavirus envelope proteins. Because others previously hypothesized that hepadnavirus-induced HCCs are resistant to reinfection with a hepadnavirus in vivo, our data suggest that if such a resistance exists it likely occurs via a block at the post-entry step. The demonstrated ability of HDV to infect already formed HCCs may facilitate development of novel strategies further dissecting the mechanism of liver pathogenesis associated with HDV infection. (HEPATOLOGY 2012;56:76–85)

Hepatitis delta virus (HDV) is a natural subviral agent of hepatitis B virus (HBV). HDV uses the envelope proteins of HBV to form virions and to infect susceptible hepatocytes. HBV and HDV utilize the same so-far-unidentified receptor(s). With the exception of the envelope proteins, the HDV life cycle is independent of HBV. HDV encodes the only protein, delta antigen (δAg), which is essential for HDV replication through the RNA-directed RNA synthesis catalyzed by host RNA polymerase II. The HDV genome is a 1,700 nucleotides-long single-stranded circular RNA. Apart from the envelope proteins and δAg, HDV acquires all factors necessary for its life cycle from the host. In nature, HDV always coexists with HBV in infected liver.1, 2

In a natural setting, only humans can acquire HDV either by coinfection with HBV or by superinfection of HBV carriers with HDV.1, 2 There are about 400 million chronic HBV carriers worldwide, of whom one million die every year from advanced liver disease, including HBV-induced hepatocellular carcinoma (HCC). Chronic HBV infection increases the HCC risk by about 100-fold and causes 50%-80% of all HCCs.3, 4 There are approximately 20 million HDV carriers worldwide. Concomitant HDV infection usually enhances HBV-induced liver pathogenesis. Superinfection of HBV carriers with HDV results in 70%-90% of cases in chronic delta hepatitis that is the more severe form of chronic viral hepatitis, leading to accelerated and more frequent cirrhosis. Persistent HDV replication is a predictive factor for liver-associated mortality. Only a subset of HDV carriers is reported to benefit from interferon-α (including peginterferon) treatment, the only approved anti-HDV therapy. Currently, there are no drugs in use to directly target HDV, and a number of anti-HBV drugs do not block HDV infection.1, 2, 5-8 In Europe, HDV-induced disease is frequent among immigrants from regions of higher HDV endemicity. HDV remains a serious problem in Vietnam, Iran, Pakistan, India, Tajikistan, Mongolia, Tunisia, Brazil, and other South American countries.1, 6, 9 Despite reports suggesting that chronic carriers of HBV/HDV have a 3-fold increased risk of HCC, and develop HCC ≈14 years earlier than carriers of HBV only, there is no consensus opinion on the relationship between HDV infection and liver cancer. The molecular basis of HDV pathogenesis is poorly understood and the role of HDV in HCC induction/development has yet to be elucidated.1, 5, 6, 10-13

To advance our understanding of the mechanism of HDV infection and its relation to liver carcinogenesis, we determined whether HDV could infect in vivo the cells of hepadnavirus-induced HCCs. To accomplish this goal, we used woodchucks (Marmota monax) chronically infected with woodchuck hepatitis virus (WHV), a hepadnavirus that is closely related to HBV. The WHV carrier woodchucks are a very valuable surrogate animal model to study HBV infection, hepadnavirus-induced HCC, and to test anti-HBV and anti-HCC drugs. Although chronic HBV carrier chimpanzees and wildtype HBV full genome transgenic mice do not develop HCCs, 100% of chronic WHV carrier woodchucks develop HCCs usually within 12-36 months postinfection.4, 14 Furthermore, in the laboratory HDV can be coated with WHV envelope proteins, because the HDV-binding site is conserved within the orthohepadnavirus subfamily. Therefore, numerous in vivo studies on HDV infection were conducted using WHV carrier woodchucks superinfected with HDV.1, 2, 7 For HDV, a putative entry receptor (the receptors for WHV and HBV are currently not identified) and the host range are defined by the origin of the hepadnavirus (i.e., HBV or WHV) envelope proteins forming the virion's coat. For this reason, HDV is often used as a tool to study the mechanism of hepadnavirus attachment and entry.1, 2 Unlike previous studies, we for the first time have superinfected WHV carriers with WHV-enveloped HDV (wHDV) during the late stage of chronic WHV infection, when WHV-induced HCCs were already developed. Three out of three HDV-negative WHV carriers with formed HCCs were successfully superinfected with wHDV. All HCCs harvested upon the completion of the experiment were infected with wHDV. Therefore, one possible mechanism of how HDV may modulate liver carcinogenesis is to infect already formed HCCs and influence their fate by replicating in tumor cells and changing their gene expression profile, although the effect of HDV replication in HCC cells on a fate of a tumor needs to be elucidated in future studies. Our results clearly show that WHV-induced HCCs during chronic hepadnavirus infection are susceptible to wHDV infection and, thus, express functional putative WHV receptors and support the steps of attachment and entry that depend on functioning of the WHV envelope proteins. Therefore, because several studies suggested that HCCs induced by either HBV or WHV are resistant to subsequent reinfection with a hepadnavirus and appear free of hepadnavirus replication markers,15-17 we conclude that if such resistance exists it is mediated by a block at the post-entry step.

Abbreviations

aG, antigenomic; δAg, delta antigen; cccDNA, covalently closed circular DNA; ds, double-stranded; G, genomic; GE, genome equivalent; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HDV, hepatitis delta virus; LL, left liver lobe; LM, left medial liver lobe; MOI, multiplicity of infection; pgRNA, pregenomic RNA; qPCR, quantitative real-time polymerase chain reaction assay; rcDNA, relaxed circular DNA; RL, right lateral liver lobe; RT, reverse transcription; wHDV, HDV coated with the envelope proteins of WHV; WHV, woodchuck hepatitis virus.

Materials and Methods

Superinfection of WHV Carrier Woodchucks With wHDV.

The animals were housed at the animal facilities of Cornell University (Ithaca, NY). All experimental manipulations of animals were performed under protocols approved by the Institutional Animal Care and Use Committee. We used two males and one female WHV carrier (M7724, M7788, and F7807, respectively), which were produced by neonatal infection with the strain WHV7 at 3 days of age.14 Numbers and locations of WHV-induced HCCs within the livers were initially determined by ultrasound imaging. One week before wHDV superinfection an open surgery was performed and biopsies from one HCC (referred further to as HCC1s) and from normal tissue of the left liver lobe (LL) were obtained from each animal. Thus, the findings of the ultrasound imaging were confirmed and the exact locations of established HCCs were determined. Following complete recovery from surgery, each woodchuck was inoculated via the sublingual vein with 0.5 mL of woodchuck-derived serum wHDV, which equaled 8 × 109 HDV genome-equivalents (GE)/animal. The administered HDV dose corresponds to a multiplicity of infection (MOI) of ≈0.27 HDV GE/hepatocyte, because the average liver of an adult woodchuck contains about 3 × 1010 hepatocytes.18 Woodchucks were monitored for 6 weeks after wHDV superinfection and blood was taken weekly. After 6 weeks the animals were sacrificed and blood, normal liver tissues, and HCCs were analyzed for markers of HDV and WHV infections. To avoid the presence of normal tissues, the center masses of HCCs were used for analysis.

RNA Isolation and HDV RNA Standards.

Total RNA from all tissues was isolated using the TRI reagent (Molecular Research Center). For the HDV strand-specific Northern analysis and real-time quantitative polymerase chain reaction assays (qPCRs), the standards of genomic and antigenomic unit-length RNAs were in vitro transcribed and subsequently gel-purified as previously described.19, 20

HDV Strand-Specific Northern Analysis.

Northern analysis was performed essentially as previously described.20 The RNAs were glyoxalated and then resolved using 1.7% agarose gels. The hybridization solution used was Ultrahyb Ultrasensitive Hybridization Buffer, containing 0.1 mg/mL of sheared salmon sperm DNA (Applied Biosystems). The HDV RNAs were detected using 32P-labeled riboprobes, produced by in vitro transcription of either pTW108 or pTW107 plasmid linearized with HindIII. Plasmid pTW108 bears 1.1× unit-length HDV genome (G), whereas plasmid pTW107 contains 1.1× of the HDV antigenome (aG).20-22 Radioactivity was visualized using a biomolecular imager (Typhoon FLA 9000, GE Heathcare). Images were acquired and quantified using ImageQuant TL and prepared for publication using PowerPoint and Adobe Photoshop software.

HDV Strand-Specific qPCR.

The procedure has been described in detail elsewhere.19 The primers used were: (1) forward primer, 312-GGACCCCTTCAGC GAACA-329; and (2) reverse primer, 393-CCTAGC ATCTCCTCCTATCGCTAT-360. The TaqMan probe was 332-AGGCGCTTCGAGCGGTAGGAGTAAGA-357. The HDV numbering was according to Kuo et al.23 To assay G RNA, the above reverse primer was used in the reverse transcription (RT) reaction. To assay aG RNA, the above forward primer was used for the RT reaction. The copy numbers were quantified using a 10-fold dilution series of either G or aG RNA standards (range: 20-200,000 GE of HDV). We deduce that 1 million HDV RNA molecules (G or aG) is equal to 1 pg of the corresponding HDV RNA standard. The 7500 Real Time PCR instrument (Applied Biosystems) was used for all qPCR assays according to the instructions of the manufacturer.

Quantification of WHV Covalently Closed Circular DNA (cccDNA) Copy Numbers Using qPCR.

The cccDNA-enriched fraction that is mostly free of cellular genomic DNA was isolated using previously described protocols24-26 and was further treated with RNase A (Roche), Hpa I (New England Biolabs), and Plasmid-Safe-ATP-dependent DNase (Epicentre Technologies) to eliminate RNA, linear WHV DNA intermediates, double-stranded (ds) linear WHV genomic DNA, residual genomic cellular DNA, and certain forms of relaxed circular (rc) WHV DNA. The resulting DNA was subsequently treated with Mung Bean nuclease (New England Biolabs) to eliminate remaining relaxed circular DNA (rcDNA). The qPCR to quantify cccDNA assayed the dsDNA region unique for cccDNA. Forward primer, 1701-GGTCCGTGTT GCTTGGTCT-1719; reverse primer, 1977-GGACAT GGAACACAGGCAAAAACA-1954; and TaqMan probe, 1846-AATGGGAGGAGGGCAGCATTGATCCT-1871 were used. The numbering corresponds to the WHV7 sequence.27 qPCR was performed with the Applied Biosystems TaqMan Gene Expression Mastermix using each primer at a concentration of 600 nM and the TaqMan probe at a concentration of 250 nM. The reaction conditions were 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C, and 60 seconds at 60°C. To calculate cccDNA copy numbers, a 10-fold dilution series of NheI-linearized plasmid PUC-CMVWHV28 was used (range: 20-200,000 GE of WHV). The cccDNA copy numbers were expressed per μg of total DNA.

Quantification of WHV Pregenome(pg)/Precore RNA Copy Numbers Using qPCR.

For eliminating WHV DNA, total RNA from normal liver tissues or HCCs was treated with Turbo DNase (Ambion) (6 units of DNase/1 μg of RNA) for 2 hours at 37°C. The complementary DNA (cDNA) was synthesized with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) using the reverse primer for qPCR, 2579-TGGCAGATGGAGATTGAGAGC-2559 that is located in a region exclusively present on WHV pg/precore RNAs. For the subsequent qPCR (that we developed) forward primer, 2504-AGAAGACGCACTCCCT CTCCT-2524; reverse primer (also used for the RT step as described above); and a TaqMan probe, 2531-AGAA GATCTCAATCACCGCGTCGCAG-2556 were used. The numbering corresponds to the WHV7 sequence.27 qPCR was carried out with the Applied Biosystems TaqMan Gene Expression Mastermix using each primer at a concentration of 900 nM and the TaqMan probe at a concentration of 250 nM. The reaction conditions were 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C, and 60 seconds at 60°C. To quantify WHV pgRNA copy numbers, a 10-fold dilution series of NheI-linearized plasmid PUC-CMVWHV was used (range: 20-200,000 GE of WHV). The pgRNA copy numbers were expressed per μg of total RNA.

Immunohistochemistry.

Normal liver tissues from LL, left medial liver lobe (LM), and right lateral liver lobe (RL) and HCCs were harvested at the end of the study and were processed together with the samples biopsied 1 week prior to wHDV superinfection. Paraffin sections of formalin-fixed tissues were immunostained with polyclonal rabbit antibodies against recombinant small δAg (1:8,000 dilution) followed by immunoperoxidase detection and hematoxylin-eosin poststaining.29

Results

Superinfection With wHDV of WHV Carrier Woodchucks With HCCs.

To determine whether hepadnavirus-induced HCCs are susceptible to HDV infection, three WHV carriers (M7724, M7788, and F7807) were used at the late stage of chronic infection, when HCCs had already developed. WHV carriers were superinfected with wHDV, using a low MOI of 0.27 HDV GE/hepatocyte. Six weeks after wHDV superinfection, woodchucks were euthanized and blood, normal liver tissues, and HCCs were examined for markers of HDV and WHV infections. Serum samples were assayed for HDV genomic RNA and WHV DNA using qPCRs as described previously.19 As shown in Fig. 1, all woodchucks quickly developed HDV viremia, and the serum HDV titers reached the WHV titers within 2 to 4 weeks. The increase of HDV titers coincided with a transient 4 to 10-fold decrease in WHV titers. The serum concentrations of HDV and WHV remained relatively high for the duration of the experiment. Thus, all WHV carriers were successfully superinfected with HDV.

Figure 1.

Titers of WHV and wHDV in sera of WHV carrier woodchucks (M7724, M7788, and F7807) superinfected with wHDV. WHV DNA and HDV RNA were isolated from serum samples collected weekly during 6 weeks after superinfection with wHDV and analyzed by either WHV-specific or HDV-specific qPCR,19 respectively. The WHV titers were also analyzed 1 week prior (week −1) and at the time of wHDV superinfection (week 0). The results represent average numbers based on the analysis of two independent isolations of either RNA or DNA. In each qPCR assay each RNA (or DNA) sample was analyzed in triplicate. The vertical axis represents serum titers (Log10(GE of the virus/mL of serum)) of either WHV (dark circles) or HDV (open rectangles). The horizontal axis represents the time after HDV superinfection (weeks).

Susceptibility of WHV-Induced HCCs to HDV Infection.

Woodchucks were monitored for 6 weeks following HDV superinfection assuming that this period is long enough to develop detectable HDV infection, and short enough so new HCCs likely will not develop. During necropsy at the end of the study one HCC was recovered from the liver of woodchuck M7724, five HCCs from M7788, and two HCCs from F7807. All HCCs with exception of HCC3 of the M7788 and HCC2 of the F7807 were identified in the WHV-infected livers prior to wHDV superinfection via ultrasound imaging and open surgery, and therefore were developed before superinfection.

We examined total RNA from normal liver tissues and HCCs by Northern analysis. For discrimination between G and aG HDV RNAs, HDV strand-specific hybridizations were employed (Fig. 2). The detection of aG RNA is an ultimate proof of infection, because aG RNA is absent from the virions and only appears via RNA replication following HDV infection of hepatocyte.2 At the top panel of Fig. 2, it is shown that all three RNA samples extracted from normal LL tissues of either woodchuck M7724 (lane 1) or F7807 (lane 3), and from normal LM tissue of woodchuck M7788 (lane 2) were positive for aG RNA. Also, all HCCs assayed (i.e., HCC1 from woodchuck M7724 [lane 4]; HCC1, HCC3, HCC4, and HCC5 from woodchuck M7788 [lanes 5, 6, 7, and 8, respectively]; and HCC1 and HCC2 from woodchuck F7807 [lanes 9 and 10, respectively]) were positive for aG RNA, and thus were infected with wHDV. The levels of aG RNA in HCCs and in normal liver tissues were comparable. As anticipated, all RNA samples that tested positive for aG RNA were also positive for G RNA (Fig. 2, bottom panel). No aG and G RNA were detected in total RNA extracted from the HCCs of two control WHV carriers M7746 and M7747, which were not superinfected with wHDV (lanes 11 and 12).

Figure 2.

Detection of HDV G or aG RNAs in normal liver tissues and in WHV-induced HCCs using HDV strand-specific Northern analysis. Total RNAs (50 μg/lane) isolated from either normal liver tissues or HCCs were assayed by Northern analysis for either aG (top panel) or G (bottom panel) HDV RNA. For both panels, the aGst lane was in vitro transcribed gel-purified aG unit-length HDV RNA standard (80 pg), whereas the Gst lane was in vitro transcribed gel-purified genomic unit-length HDV RNA standard (80 pg). In lanes 1-3 are RNAs isolated from normal liver tissues of: lane 1, woodchuck M7724 (RNA from LL, left liver lobe); lane 2, woodchuck M7788 (RNA from LM, left medial liver lobe); lane 3, woodchuck F7807 (RNA from LL). In lanes 4-10 are RNAs isolated from HCCs of WHV carrier woodchucks superinfected with wHDV: lane 4, RNA from HCC1 of woodchuck M7724; lanes 5-8, RNAs from HCC1, HCC3, HCC4, and HCC5, respectively, of woodchuck M7788; lanes 9-10, RNAs from HCC1 and HCC2, respectively, of woodchuck F7807. In lanes 11-12 are RNAs from HCC1 of woodchuck M7746 and HCC1 of woodchuck M7747, respectively. The WHV carriers M7746 and M7747, which were not superinfected with wHDV, served as negative controls. Tumors designated as HCC1s were biopsied prior to the wHDV superinfection.

These findings obtained by Northern analyses were confirmed and extended using HDV strand-specific qPCR (Table 1). The RNAs from all normal and HCC tissues obtained from the superinfected animals during necropsy tested positive for both strands of HDV RNA. As anticipated, the G RNA levels were ≈7 to 27-fold higher than those of aG.2 In animal M7724, the HDV RNA levels in HCC1 were higher than those in normal LL, LM, and RL tissues. In woodchuck M7788 in the HCC3, HCC4, and HCC5 the levels of HDV RNA accumulation were higher than in surrounding normal liver tissues. The HCC1 (M7788) had HDV RNA levels comparable to those of normal tissues. Only HCC2 from the same animal had the lowest HDV RNA levels among all tissues analyzed. This may reflect a difference in the susceptibility to infection related to the differentiation state of HCC. The levels of G RNA accumulation in HCCs of woodchuck M7807 were ≈4-fold lower than the average G RNA level in normal tissues. As expected, total RNA from HCC1s of WHV carriers M7746 and M7747 that were not superinfected with wHDV, and also the RNAs from the normal liver tissues and HCC1s of woodchucks M7724, M7788, and F7807, which were obtained during laparotomy prior to wHDV superinfection, were negative for HDV RNAs (data not shown). All HCCs including those identified prior to wHDV superinfection became HDV-positive following superinfection, and most of them appeared to be infected at least as efficiently as normal liver tissues.

Table 1. Quantitative Analysis of HDV Genomic (G) and Antigenomic (aG) RNAs in Normal Liver Tissues and WHV-Induced HCCs Using HDV Strand-Specific Real-Time PCR Assays
WoodchuckNormal Liver TissueHDV RNA(x104 GE/μg Total RNA)HCCHDV RNA (x104 GE/μg Total RNA)
GaGG/aG ratioGaGG/aG ratio
  1. LL, left liver lobe; LM, left medial liver lobe; RL, right lateral liver lobe. The number of HDV genome equivalents (GE) was determined using HDV strand-specific qPCR assays. To calculate copy numbers of G and aG HDV RNA, we used the in vitro transcribed gel-purified unit-length HDV RNA standard of genomic or antigenomic polarity, respectively. The letters a, b, c, or d indicate the average number based on the analysis of two, three, four, or five independent RNA isolations, respectively. In each qPCR assay each RNA sample was analyzed in triplicate.

M7724LL113d7.73a14.6HCC1252d15.8b15.9
 LM81.1b6.69a12.1    
 RL102b9.30a11.0    
M7788LL48.7d1.99a24.5HCC178.1b5.07a15.4
 LM107c11.7a9.15HCC20.87b0.10b8.4
     HCC3354d13.0a27.2
     HCC4312c16.8c18.6
     HCC5743b37.0a20.1
F7807LL1,090c73.9a14.7HCC1215c23.1c9.3
 LM827a59.7a13.9HCC2278c39.1a7.1
 RL1,020b67.4a15.1    

Next, formalin-fixed sections of normal liver tissues and HCCs were examined by immunohistochemistry, using antibodies against δAg. As anticipated, all normal liver tissues from woodchucks superinfected with wHDV displayed δAg-positive hepatocytes. Figure 3 demonstrates the staining for newly synthesized δAgs in several normal liver tissues of woodchucks M7724, M7788, and F7807. The intracellular δAg distribution is typical for HDV infection. The staining is mainly nuclear with apparent nucleolar exclusion (Fig. 3B-E). Occasionally, HDV-infected hepatocytes display additional cytoplasmic staining (Fig. 3F) that was observed previously and likely represents the appearance of δAgs with mutation(s) in the nuclear localization signal and/or the formation of complexes between δAgs and WHV envelope proteins.19, 30, 31 Similarly, all the HCCs except HCC2 of woodchuck M7788 displayed HDV-positive cells. This observation correlates with the finding that HCC2 had the lowest level of HDV replication (Table 1). Figure 4 demonstrates HDV-positive cells from five different HCCs. The intracellular δAg distribution patterns (Fig. 4B-F) are virtually the same as those observed in normal hepatocytes (Fig. 3B-F). The unperturbed morphology of normal hepatocytes and the tumorous phenotype of HCC cells (including HDV-positive cells in both kinds of tissues) were confirmed by a pathologist (I.T.), thus verifying that both normal hepatocytes (Fig. 3) and HCC cells (Fig. 4) were infected with wHDV. The samples of normal liver tissues and HCCs that were obtained via surgery prior to wHDV superinfection were negative for δAgs (Figs. 3A, 4A, respectively), as expected. The number of HDV-positive cells (with the exception of the HCC2 of woodchuck M7788) ranged between 0.09 and 6.7 cells/per 1,000 analyzed cells (Table 2) that reflects the consequences of low MOI infection (0.27 HDV GE/hepatocyte) as early as 6 weeks postinoculation. Within each set of tissues obtained from an individual animal a higher number of HDV-positive cells did not always correspond to the higher level of HDV RNA accumulation (Tables 1, 2), likely because different subpopulations of hepatocytes may support the same efficiency of wHDV entry, but different levels of HDV RNA replication. Based on the numbers of HDV-infected cells (Table 2) and the qPCR data (Table 1), we confirmed that normal hepatocytes and HCC cells have comparable susceptibilities to HDV infection in vivo.

Figure 3.

Immunohistochemical detection of the δAgs in normal liver tissues from WHV carriers superinfected with wHDV. The paraffin sections of formalin-fixed normal liver tissues from WHV carriers superinfected with wHDV were immunostained with rabbit polyclonal antibodies against δAg (1:8,000 dilution) using the immunoperoxidase technique and then poststained with hematoxylin-eosin. The unperturbed morphology of normal hepatocytes (including HDV-positive cells) was confirmed by a pathologist (I.T.). The magnification used was 200×. (A,B) The liver lobe (LL) tissues of the woodchuck M7724 collected prior to (negative control) and after wHDV superinfection (at the end of the study), respectively. Images (C-F) are of normal tissues collected during necropsy at the end of the study. (C,D) The LL and left medial liver lobe (LM) tissues, respectively, of woodchuck M7788. (E,F) The LL and LM tissues, respectively, of woodchuck F7807. The cells infected with wHDV are stained in brown for δAgs. The insertions (2-fold amplifications) provide detailed patterns of δAg subcellular distribution.

Figure 4.

Immunohistochemical detection of the δAgs in WHV-induced HCCs from WHV carriers superinfected with wHDV. The sections of the WHV-induced HCCs from WHV carriers superinfected with wHDV were fixed with formalin, embedded in paraffin, and immunostained to detect δAgs as described in Fig. 3. The tumorous morphology of HCC cells (including HDV-positive cells) was confirmed by a pathologist (I.T.). The magnification used was 200×. (A,B) Samples of HCC1 from woodchuck M7724 collected prior to (negative control) and after wHDV superinfection (at the end of the study), respectively. Images (C-F) are of HCCs collected during necropsy at the end of the study. The HCCs from woodchuck M7788 are in (C) (HCC4) and (D) (HCC5). The HCCs from woodchuck F7807 are presented in (E) (HCC1) and (F) (HCC2). The cells infected with wHDV are stained in brown for δAgs. The insertions (2-fold amplifications) provide detailed patterns of δAg localization.

Table 2. Numbers of the Cells Infected with HDV in Normal Liver Tissues and in WHV-Induced HCCs
WoodchuckNormal Liver TissueNumber of Cells Positive for δAg /1,000 CellsHCCNumber of Cells Positive for δAg /1,000 Cells
  1. LL, left liver lobe; LM, left medial liver lobe; RL, right lateral liver lobe. Paraffin sections of formalin-fixed tissue were stained by immunoperoxidase technique using antibodies against delta antigen (δAg). The numbers of δAg-positive cells per total number of analyzed cells are shown in parentheses.

M7724LL0.29 (7/24,096)HCC10.45 (9/20,232)
 LM0.25 (3/12,177)  
 RL0.42 (6/14,315)  
M7788LL0.7 (18/25,531)HCC12.0 (65/31,760)
 LM0.09 (17/189,300)HCC20.0 (0/28,000)
   HCC31.7 (16/9,600)
   HCC46.7 (116/17,423)
   HCC53.1 (60/19,360)
F7807LL3.1 (33/10,750)HCC10.4 (9/21,677)
 LM2.2 (23/10,450)HCC20.8 (34/44,160)
 RL4.4 (67/15,120)  

Markers of WHV Infection in Normal Liver Tissues and HCCs.

Detection of cccDNA, which is not present in virions, is the ultimate evidence of hepadnavirus infection. The quantification of WHV cccDNA by qPCR is summarized in Table 3. Overall, in HCCs the levels of cccDNA accumulation were ≈5 to 360-fold lower than in normal liver tissues. Several recent studies analyzed HCCs induced by HBV or WHV using immunostaining for the core antigen and in situ hybridization for viral DNAs, and concluded that HCCs were free of viral replication markers.15-17 Our data suggest that during chronic infection WHV replication is not completely blocked, but is significantly suppressed in the majority of HCCs. The only exception, HCC3 of woodchuck M7788, had 4.8 times less cccDNA as compared with normal liver tissues. Examination of tissue sections by a pathologist (I.T.) did not detect normal (nontumorous) hepatocytes within HCC tissues, indicating that we indeed quantified the accumulation of cccDNA in HCCs and not in normal hepatocytes that could have been engulfed in HCCs. Overall, the susceptibility of the normal liver tissues and HCCs to HDV infection apparently does not correlate with the extent of WHV replication, because no correlation was observed between the levels of WHV and HDV replications (Tables 1, 3).

Table 3. Quantitative Analysis of Covalently Closed Circular WHV DNA (cccDNA) in Normal Liver Tissues and in WHV-Induced HCCs Using WHV cccDNA-Specific Real-Time PCR Assay
WoodchuckNormal Liver TissueWHV cccDNA(GE/μg Total DNA)HCCWHV cccDNA (GE/μg Total DNA)
  1. LL, left liver lobe; LM, left medial liver lobe; RL, right lateral liver lobe. The numbers of WHV cccDNA genome equivalents (GE) were determined using WHV cccDNA-specific real-time PCR assay and represent the average for two independent cccDNA preparations. For each animal, the fold reduction of cccDNA content in HCCs has been calculated by dividing the average number of cccDNA GE in normal tissues by the number of cccDNA GE in a particular HCC. In each qPCR assay each DNA sample was analyzed in triplicate.

M7724LL1.45 x105HCC17.85 x103 (15.7 fold lower)
 LM1.31 x105  
 RL9.29 x104  
M7788LL3.37 x104HCC11.57 x103 (17.9 fold lower)
 LM2.25 x104HCC25.15 x102 (54.6 fold lower)
   HCC35.86 x103 (4.8 fold lower)
   HCC44.00 x102 (70.2 fold lower)
   HCC57.84 x101 (358.6 fold lower)
F7807LL1.10 x105HCC11.38 x103 (71.5 fold lower)
 LM8.34 x104HCC22.24 x103 (44.0 fold lower)
 RL1.02 x105  

Next, RNAs extracted from normal tissues and HCCs were treated with DNase to eliminate WHV DNAs, and subsequently were assayed for WHV pg/precore RNAs (pgRNA) using qPCR (Table 4). We found that pgRNA accumulation in HCCs was diminished within a 2.5 to 120-fold range as compared with corresponding normal liver tissues. The HCCs from woodchucks M7724 and F7807 accumulated ≈16-36 times less pgRNA than normal tissues, whereas HCC2, HCC3, and HCC5 from woodchuck M7788 accumulated only ≈2.5-7.0 times less pgRNA. The HCC1 and HCC4 from woodchuck M7788 accumulated 120.5 and 23.1 times less pgRNA, respectively (Table 4). No correlation was found between the levels of pgRNA and cccDNA in either normal liver tissues or in HCCs. Based on the cccDNA levels, WHV replication was significantly suppressed in most HCCs. At the same time, pgRNA accumulation in HCCs seemed to be less profoundly decreased (Tables 3, 4). Similar to cccDNA levels, there was no apparent correlation between the accumulation levels of HDV RNAs and WHV pgRNA (Tables 1, 4).

Table 4. Quantitative Analysis of WHV Pregenome/Precore RNA (pgRNA) in Normal Liver Tissues and in WHV-Induced HCCs Using WHV pgRNA-Specific Real-Time PCR Assay
WoodchuckNormal Liver TissueWHV pgRNA (GE/μg Total RNA)WHV pgRNA/cccDNAHCCWHV pgRNA(GE/μg Total RNA)WHV pgRNA/cccDNA
  1. LL, left liver lobe; LM, left medial liver lobe; RL, right lateral liver lobe. The numbers of WHV pgRNA genome equivalents (GE) were determined using WHV pgRNA-specific real-time PCR assay and represent the average for two independent RNA preparations. In each qPCR assay each DNA sample was analyzed in triplicate. The ratio of WHV PgRNA/cccDNA was calculated by dividing the number of the pgRNA copies/per μg of total RNA by the number of cccDNA copies/ per μg of total DNA, and it indicates rate of the WHV pgRNA transcription.

M7724LL4.30 x1082.95 x103HCC11.12 x107 (36.4 fold lower)1.43 x103
 LM2.61 x1082.00 x103   
 RL5.38 x1085.79 x103   
M7788LL6.69 x1071.98 x103HCC14.94 x105 (120.5 fold lower)3.15 x102
 LM5.22 x1072.32 x103HCC22.43 x107 (2.5 fold lower)4.72 x104
    HCC38.03 x106 (7.4 fold lower)1.37 x103
    HCC42.58 x106 (23.1 fold lower)6.46 x103
    HCC51.47 x107 (4.1 fold lower)1.87 x105
F7807LL4.25 x1083.84 x103HCC12.43 x107 (16.4 fold lower)1.76 x104
 LM2.85 x1083.42 x103HCC21.85 x107 (21.5 fold lower)8.27 x103
 RL4.84 x1084.77 x103   

In addition, we performed immunostaining for WHV core antigen. Similarly to previous reports,15-17 we found that normal liver tissues contained a considerable number of strongly core-positive hepatocytes along with hepatocytes that displayed core staining of reduced intensity (Supporting Fig. S1). The HCCs demonstrated an overall negative core staining, with only light positivity in rare neoplastic subpopulations (Fig. S2). No core-positive cells were observed in HCC1 of woodchuck M7724 and in HCC2 of woodchuck F7807.

Discussion

In the present study, using Northern analysis, qPCR, and immunohistochemistry, we established for the first time that HDV infects WHV-induced HCCs in vivo, which advances our understanding of the infection mechanisms of HDV and hepadnavirus, and of the relationship between the ongoing infection and development/progression of HCC. The levels of HDV replication (Table 1) and numbers of HDV-infected cells (Table 2) demonstrate that HCC cells in vivo express functional putative hepadnavirus receptors and support the steps of the attachment/entry governed by the hepadnavirus envelope proteins and also trafficking and replication of HDV with an efficiency comparable to that of normal hepatocytes. Taken together with detection of WHV cccDNA in tumors, which suggests WHV replication and therefore a production of the envelope proteins in HCC cells, our data lead to the speculation that HCC cells may support the entire HDV life cycle in vivo, including the assembly and egress of new virions. The fact that only HCC2 (M7788) was inefficiently infected with HDV (Table 1) may reflect differences in the differentiation status of HCCs. By inference, it is very likely that HBV-induced HCCs are susceptible in vivo to infection with HBV-enveloped HDV. Overall, it appears that the loss of hepadnavirus receptors is not an essential feature of HCC development.

A single previous study has reported detection of HDV RNA in HCCs of superinfected WHV carrier woodchucks. However, that article did not address the mechanism of how and when HDV appeared in HCCs.32 We can envision two such mechanisms. One is that HDV persists in the hepatocyte from the moment of the initial infection (i.e., before the infected hepatocyte becomes malignant) and therefore may influence the induction and development of HCC. The second is that HDV infects already established hepadnavirus-induced HCC. Previously, it has not been investigated whether HDV could infect hepadnavirus-induced HCCs in vivo. In the present study we clearly demonstrated that in vivo HDV is able to infect WHV-induced HCCs in WHV carrier woodchucks. Because the HDV genome can accumulate up to ≈300,000 copies/infected cell,33 our data suggest that the efficient HDV replication may change the gene expression profile in HCC cells following infection of the tumor, and therefore may influence further HCC development. However, the effect of HDV replication in HCC cells on further development/progression of a tumor has to be elucidated in future studies.

Previously, several reports described HCCs induced either by HBV or by WHV as apparently hepadnavirus-free (based on negative staining for the core antigen and negative in situ hybridization for viral DNAs) regardless of ongoing viremia.15-17 Accordingly, a hypothesis was proposed that HCCs, which originated from hepadnavirus-infected hepatocytes (as evidenced by presence of integrated hepadnavirus DNA in HCCs), are resistant to new reinfections with a hepadnavirus.15, 16 Our results are not consistent with the absence of WHV replication, but rather suggest its significant suppression in most HCCs. Our data suggest that WHV reverse transcription and/or conversion of the rcDNA into cccDNA is suppressed, or cccDNA stability is compromised in HCCs. However, this hypothesis needs to be tested further using a larger number of HCCs and matching liver tissues from WHV carriers. Interestingly, based on the copy number ratio of pgRNA/cccDNA, which may indicate the efficiency of cccDNA-directed transcription of pgRNA, and considering that it is unlikely that pgRNA could have been transcribed from integrated WHV DNA, it seems that rates of pgRNA transcription in HCCs were either comparable to or higher than those in normal liver tissues. The only exception was HCC1 (M7788), for which the pgRNA/cccDNA ratio was 6.9-fold lower than the average ratio calculated for normal liver tissues (Table 4). Furthermore, our data demonstrating the susceptibility of HCCs to HDV infection in vivo suggest that if the above proposed superinfection exclusion for hepadnavirus occurs in HCCs, it is mediated by a block at the post-entry step. Currently, superinfection of hepadnavirus-induced HCCs with either HBV or WHV has yet to be demonstrated.

The observed absence of a correlation between HDV and WHV replication levels in both normal liver tissues and HCCs suggests that, because HDV requires only the envelope proteins from the helper hepadnavirus,2 then as long as a sufficient supply of the envelope proteins is available, the extent of HDV infection is not directly dependent on the rate of hepadnavirus replication. This may explain at least in part why some anti-HBV drugs do not inhibit HDV infection.

Acknowledgements

Woodchucks used in the study were bred, infected with WHV, and maintained as chronic WHV carriers under the NIH contract NIAID N01-AI-05399 until the development of HCC. We thank Eva Permaul and Deborah Berry from the histology laboratory of the Lombardi Cancer Center at Georgetown University for excellent assistance with the immunohistochemistry of infected tissues. We also thank William Mason for encouragement, Bud Tennant for support, and Igor Prudovsky and Steven Weinman for constructive comments.

Ancillary