Hepatitis B Virus Deregulates the Cell Cycle To Promote Viral Replication and a Premalignant Phenotype

Hepatitis B virus (HBV) infection is a major health problem with high risk of developing hepatocellular carcinoma (HCC). By using a biologically relevant system of HBV infection of primary human hepatocytes (PHHs), we studied how HBV perturbs gene expression and whether these effects are relevant to HBV-associated HCC. HBV induced a distinct profile of growth factor production, marked particularly by significantly lower levels of the transforming growth factor β (TGF-β) family of proteins. Transcriptome profiling revealed multiple changes in cell proliferation and cell cycle control pathways. Cell cycle analysis demonstrated that HBV-infected PHHs are enriched in the G₂/M phase. HBV proviral host factors were upregulated upon infection and particularly enriched in cells in the G₂/M phase. Together, these results support the notion that HBV deregulates cell cycle control to render a cellular environment that is favorable for productive infection. This may coincidently induce a premalignant phenotype that predisposes infected hepatocytes to subsequent malignant transformation.

ABSTRACT

Hepatitis B virus (HBV) infection is a major health problem worldwide, and chronically infected individuals are at high risk of developing cirrhosis and hepatocellular carcinoma (HCC). The molecular mechanisms whereby HBV causes HCC are largely unknown. Using a biologically relevant system of HBV infection of primary human hepatocytes (PHHs), we studied how HBV perturbs gene expression and signaling pathways of infected hepatocytes and whether these effects are relevant to productive HBV infection and HBV-associated HCC. Using a human growth factor antibody array, we first showed that HBV infection induced a distinct profile of growth factor production by PHHs, marked particularly by significantly lower levels of the transforming growth factor β (TGF-β) family of proteins in the supernatant. Transcriptome profiling next revealed multiple changes in cell proliferation and cell cycle control pathways in response to HBV infection. A human cell cycle PCR array validated deregulation of more than 20 genes associated with the cell cycle in HBV-infected PHHs. Cell cycle analysis demonstrated that HBV-infected PHHs are enriched in the G₂/M phase compared to the predominantly G₀/G₁ phase of cultured PHHs. HBV proviral host factors, such as PPARA, RXRA, and CEBPB, were upregulated upon HBV infection and particularly enriched in cells in the G₂/M phase. Together, these results support the notion that HBV deregulates cell cycle control to render a cellular environment that is favorable for productive HBV infection. By perturbing cell cycle regulation of infected cells, HBV may coincidently induce a premalignant phenotype that predisposes infected hepatocytes to subsequent malignant transformation.

IMPORTANCE Hepatitis B virus (HBV) infection is a major health problem with high risk of developing hepatocellular carcinoma (HCC). By using a biologically relevant system of HBV infection of primary human hepatocytes (PHHs), we studied how HBV perturbs gene expression and whether these effects are relevant to HBV-associated HCC. HBV induced a distinct profile of growth factor production, marked particularly by significantly lower levels of the transforming growth factor β (TGF-β) family of proteins. Transcriptome profiling revealed multiple changes in cell proliferation and cell cycle control pathways. Cell cycle analysis demonstrated that HBV-infected PHHs are enriched in the G₂/M phase. HBV proviral host factors were upregulated upon infection and particularly enriched in cells in the G₂/M phase. Together, these results support the notion that HBV deregulates cell cycle control to render a cellular environment that is favorable for productive infection. This may coincidently induce a premalignant phenotype that predisposes infected hepatocytes to subsequent malignant transformation.

INTRODUCTION

Hepatitis B virus (HBV) infection is a major health problem worldwide, with more than 240 million chronically infected individuals despite the availability of an effective vaccine. Without curative treatment, these subjects are at high risk of developing end stage liver diseases, like cirrhosis and hepatocellular carcinoma (HCC) (1). The primary treatment goals for patients with HBV infection are to prevent progression of the disease, particularly to cirrhosis and HCC. Nucleoside and nucleotide analogues are effective antivirals that suppress HBV replication. However, the drawbacks are lifelong treatment and risk of developing drug resistance. Interferon alpha is the only licensed immunomodulatory drug for hepatitis B. Although current treatments for hepatitis B have limitations, suppression of HBV replication by antivirals is associated with a reduction in the incidence of HCC (2).

HBV is a small, enveloped DNA virus replicating via an RNA intermediate that specifically infects hepatocytes. After infection, the capsid is transported to the nucleus, where the relaxed circular DNA (rcDNA) is released and converted into a covalently closed circular DNA (cccDNA) persistent form. This cccDNA serves as a transcriptional template for the different viral RNAs. The 3.5-kb pregenomic RNA (pgRNA) is encapsidated and reverse transcribed into new rcDNA. rcDNA-containing nucleocapsids are then either enveloped and released in newly formed virions or redirected toward the nucleus to establish a cccDNA pool. This amplification pathway, together with the long half-life of the cccDNA molecule in the nucleus, contributes to HBV persistence in infected cells. Additionally, HBV DNA integration into the host hepatocyte genome occurs during viral replication.

Chronic HBV infection accounts for 20 to 50% of the HCC cases in the world (3). It is reported that HBV infection increases the risk of liver cancer by up to 100-fold over uninfected individuals, according to different prospective cohort studies (3). It is believed that multiple mechanisms are involved in HBV-induced HCC. In chronic hepatitis B, persistent viral infection and ineffective T cell responses lead to chronic inflammatory liver damage and continuous compensatory proliferation of hepatocytes, which may lead to liver fibrosis or cirrhosis and promote HCC development (4). In addition, HBV may have direct oncogenic potential that contributes to the development of HCC. HBV integration (truncated pre-S, S, or X) into the host hepatocyte genome is a frequent event in HCC (86.4%) and may lead to general genomic instability, including deletions, cis/trans-activation, translocations, generation of fusion transcripts, and general genomic instability (5). However, it is unclear whether HBV integration directly causes hepatocyte transformation or is just an unrelated incidental event during chronic infection. HBx protein has been implicated as a direct oncogenic protein. Studies using HBx transgenic mouse models showed spontaneous HCC development (6). By overexpressing HBx in a human hepatoma cell line or murine liver cells, various studies have reported HBx's effects on several signaling pathways regulating proliferation, differentiation, cell death, oxidative stress, and DNA repair, which may contribute to HCC development (7–10). Additionally, overexpression of wild-type or truncated surface protein can induce endoplasmic reticulum (ER) stress, oxidative stress or DNA damage, or affect cell proliferation and thus potentially promote HCC development (11–14). However, there is no well-accepted mechanism whereby HBV drives carcinogenesis.

The gaps in our knowledge of HBV-induced HCC are partly attributable to the lack of cell culture systems that resemble the physiological status of human hepatocytes and permit highly efficient HBV infection. An HBV-expressing hepatoma cell line, which has HBV genome integration by stable transfection with HBV DNA constructs, has been used for many decades (15). Alternatively, delivery of viral genomes by baculovirus or adenovirus vectors results in productive HBV replication and formation of infectious viruses (16, 17). Although these models have facilitated antiviral development and drug resistance studies, there are significant limitations. Alternatively, HepaRG, a liver progenitor cell line, can be used after 1 month of differentiation, but the low HBV infection efficiency, as well as a mixed cell phenotype (half biliary-like cells and half hepatocyte-like cells), hampers definitive interpretation of hepatocyte-specific interaction with HBV (18, 19). Other models, such as NTCP-overexpressing hepatoma cells, are not optimal due to their transformed nature (19). Thus, findings resulting from these artificial models may not be biologically relevant to HBV-associated HCC (20). Recent studies using hESC or hiPSC differentiated hepatocytes provide a more physiological system, but these hepatocytes are not as completely mature as primary adult hepatocytes (21, 22).

Studying the pathogenic mechanisms of HBV-associated HCC allows us to better understand the critical pathways to HCC development and identify potential targets for preventive or therapeutic intervention. In this study, we took advantage of primary human hepatocytes (PHHs), which have been used as a gold standard for HBV in vitro studies. By optimizing the cell culture conditions, we can reach an HBV infection efficiency close to 100%. Our results demonstrate that HBV infection deregulates cell cycle control to foster an environment with high levels of proviral factors by suppressing the transforming growth factor β (TGF-β) pathway, which is known to be associated with tumorigenesis.

RESULTS

HBV infection alters expression of growth factors.

A fundamental difference between normal and tumor tissues is the regulation of cell growth. It is known that various tumorigenic growth factor signaling pathways are deregulated in human HCC. To study whether HBV infection alters the expression profile of growth factors in hepatocytes, the supernatant of PHHs with or without HBV infection was collected and analyzed with a human growth factor membrane array (Fig. 1A and B). Quantification of spot intensities was performed using ImageJ software, and the levels of growth factors are shown in Fig. 1C and D. The secretion pattern from donor 1192 showed that many growth factors were downregulated by more than 50%, including epidermal growth factor receptor (EGFR), insulin-like growth factor binding protein 3 (IGFBP-3), macrophage colony-stimulating factor (MCSF), neurotrophin-4 (NT-4), platelet-derived growth factor AB (PDGF-AB), TGF-β2, TGF-β3, vascular endothelial growth factor (VEGF), and VEGF receptor 2 (Fig. 1C). The only two upregulated factors were IGFBP-1 and IGFBP-2 (Fig. 1C). Similar downregulated growth factors, including IGFBP-3, IGFBP-4, MCSF, NT-3, NT-4, TGF-β2, TGF-β3, and VEGF receptor 2 (Fig. 1D), were observed by using PHHs from another donor, 1413. Since TGF-βs have been implicated in control of hepatocyte proliferation and development of HCC (23, 24), we focused on the interplay between TGF-βs and HBV infection.

FIG 1.

Human growth factor array. (A) The expression of 41 human growth factors from cell culture supernatant of primary human hepatocytes (donor 1192 and donor 1413) with (+) or without (−) HBV infection was analyzed by a semiquantitative membrane array. (B) Array map of 41 growth factors. (C and D) Quantification of signal dots from donor 1192 (C) and donor 1413 (D) was analyzed using a semiquantitative membrane array. Genes with greater than 50% expression level change are labeled in red. The data are shown as means and standard deviations.

To validate our array result, the expression of TGF-βs was assessed by quantitative real-time (qRT)-PCR in donor-derived PHH cells from commercially available sources that were subsequently infected with HBV. As shown in Fig. 2A, in all three PHHs, HBV infection elicited minor effects on TGFB1 and TGFB3 mRNA levels but caused significantly reduced TGFB2 levels. At the protein level, TGF-β2 was also significantly reduced in the supernatant of HBV-infected cells. (Fig. 2B). To exclude the possibility that components other than infectious virus particles in the virus inoculum induced the alteration of TGF-β2, UV germicidal-irradiation-inactivated virus, a medium-only control, and HBV infection in the presence of the viral entry inhibitor myrcludex B (MyrB) were tested. As expected, only infectious virus, but not inactivated virus or medium control, was able to suppress TGF-β2 expression (Fig. 2C). Furthermore, suppression of TGF-β2 expression by HBV is dependent on the HBV multiplicity of infection (MOI) (Fig. 2D). To analyze the correlation of HBV and TGFB2 mRNA level at the single-cell level, PHHs were infected with HBV at an MOI of 1,000 or mock infected for 8 days and then costained for TGFB2 and HBV RNAs using the RNAscope technique (Fig. 3A) (see Materials and Methods). HBV (Fig. 3B) and TGFB2 RNAs (Fig. 3C) in individual cells were enumerated and plotted. As shown in Fig. 2E, HBV-infected cells clearly had lower TGFB2 RNA signals. In addition, in HBV-infected cultures, HBV low-replication cells (≤200 dots/cell) expressed significantly larger amounts of TGFB2 RNA than HBV high-replication cells (>200 dots/cell) (Fig. 3D), supporting the notion that HBV replication suppresses TGFB2 gene expression. Using PHH-engrafted Alb-uPA/SCID mice (25). we also showed that the expression of TGFB2 is reduced in the livers of HBV-infected mice compared to the uninfected liver (Fig. 3E).

FIG 2.

Expression of TGF-β family members following HBV infection. PHHs from three donors were infected with HBV for 8 days, (A) mRNA levels of TGFB1, TGFB2, and TGFB3 were evaluated by qRT-PCR. (B) The protein level of TGFB2 was measured by ELISA. (C) PHHs (donor 1413) were infected with HBV, UV-inactivated HBV, medium control, or HBV plus 100 nM myrcludex B for 8 days, and expression of TGFB1, TGFB2, and TGFB3 was evaluated by qRT-PCR. (D) PHHs (donor 1832) were infected with HBV at an MOI of 5,000, 2,000, 1,000, 500, or 200. Eight days after infection, expression of TGFB2 was evaluated by qRT-PCR. The data are shown as means and standard deviations. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, not significant.

FIG 3.

RNAscope analysis of HBV and TGFB2. (A) PHHs (donor 1832) were infected with HBV at an MOI of 1,000. Eight days after infection, expression of HBV RNA (green) and TGFB2 (red) was evaluated by RNAscope. (B and C) HBV signals (B) and TGFB2 signals (C) in individual cells were enumerated and plotted. (D) In HBV-infected cultures, TGFB2 signals (number of dots per cell) in HBV low-replication cells (≤200 dots/cell) and HBV high-replication cells (>200 dots/cell) were compared. (E) PHHs from the same donor were transplanted into Alb-uPA/SCID mice, and the mice were infected with HBV (genotype C; 10⁵ genomes). Eight weeks after virus infection, RNAscope staining was performed to detect TGFB2 (red) and HBV-RNA replicates (green) in the liver. Scale bars, 20 μm. The data are shown as means and standard deviations. **, P < 0.01.

HBV infection deregulates cell cycle pathways.

To study the overall cellular response to HBV infection, the transcriptome of PHHs with or without HBV infection was analyzed by microarray. Anti-HBsAg staining of infected PHHs showed more than 90% HBV infection efficiency (Fig. 4A). Principal-component analysis defined a distinct gene expression pattern of infected PHHs compared to noninfected PHHs (Fig. 4B). Gene set enrichment analysis demonstrated that the pathways involved in cell proliferation and regulating cell cycle progression were affected following HBV infection (Fig. 4C). A heat map and cluster diagram of cyclin D1 pathway genes from cDNA microarray analysis showed that many genes that regulate cell cycle control were altered (Fig. 4D). These pathways are frequently deregulated in cancer and are a biomarker of cancer phenotype and disease progression. To validate our microarray data, a human cell cycle PCR array containing 84 key cell cycle-regulatory genes was utilized (Fig. 4E). Three genes were upregulated, while 21 genes were downregulated (Fig. 4E and Table 1), confirming the microarray data. The biological function of cyclins is to control the cell cycle and growth (26). In HBV-infected cells, cyclin B1 (CCNB1) was greatly reduced (log −3.49); other cyclins, such as CCNA2, CCND3, and CCNE1, and various CDKs were also significantly decreased (Table 1). The suppression of cyclin B1 by HBV infection was further validated in three sets of PHHs (Fig. 4F).

FIG 4.

Gene expression profile of the cell cycle pathway in HBV infection. PHHs (donor 1192) with or without HBV infection were cultured for 8 days. (A) Infection efficiency was determined by HBsAg staining (green). (B) Principal-component analysis of transcriptomic profiles of cells with (blue) or without (red) HBV infection. (C) Pathway analysis of a cDNA microarray was performed, and affected pathways were ranked by P value. (D) Heat map with cluster diagram of cyclin D1 pathway genes from cDNA microarray analysis. (E) Expression of 84 genes key to cell cycle regulation was evaluated by human cell cycle RT² profiler PCR array. (F) mRNA levels of CCNB1 were evaluated by qRT-PCR. The data are shown as means and standard deviations. ***, P < 0.001. (G) PHHs (donor 1192) with or without HBV infection were cultured for 8 days. The cell cycle was analyzed by FACS.

TABLE 1.

Results of RT2 profiler PCR array

^aExpression of 84 genes key to cell cycle regulation was evaluated by human cell cycle RT2 profiler PCR array. Downregulated (boldface) and upregulated (italic) genes are listed.

We next asked whether HBV infection affects cell cycle progression. The cell cycle status of HBV-infected or noninfected PHHs was evaluated by fluorescence-activated cell sorting (FACS) analysis. As expected, most noninfected PHHs were in G₀/G₁ phase (Fig. 4G). Interestingly, more of the HBV-infected PHHs were in the G₂/M phase, suggesting that HBV infection may affect cell cycle control by promoting and then arresting the cells in the G₂/M phase.

TGF-β2 restricts HBV infection.

Since TGFB2 was downregulated by HBV infection, we investigated whether addition of TGFB2 could restrict HBV infection. HBV-infected PHHs were treated with TGFB2, and different HBV infection markers were evaluated. After 7 days of treatment, all HBV infection markers were significantly inhibited (Fig. 5A). Previously, TGFB1 has been reported to inhibit HBV replication in an HBV-expressing hepatoma cell line model (27), probably by suppressing the expression of HNF-4α, which promotes HBV gene transcription (27). To test whether TGFB2 restricts HBV infection through a similar mechanism, the expression of several known HBV transcription factors, including HNF-4α, was measured. As expected, the mRNA levels of HNF4A, PPARA, RXRA, and CEBPB were significantly downregulated upon TGF-β2 treatment (Fig. 5B).

FIG 5.

Expression of proviral host factors in HBV-infected cells. (A and B) PHHs (donor 1192) were infected with HBV and treated with TGF-β2 for 7 days. Cells were collected for RNA extraction. (A) Different HBV cccDNAs and HBV DNAs were evaluated by qPCR, HBV-RNA was evaluated by qRT-PCR, and HBsAg and HBeAg were evaluated by ELISA. (B) Expression levels of different genes were evaluated by qRT-PCR. (C) PHHs from three donors were infected with HBV for 8 days, and mRNA levels of HNF4A, PPARA, RXRA, and CEBPB were evaluated by qRT-PCR. (D) PHHs (donor 1192) were infected with HBV. The expression levels of PPAR and CEBPB were determined by Western blotting. (E) Transcriptomic analysis of PHHs with or without HBV infection were performed and the results are shown as volcano plots. Genes involved in RXR and PPARA activation signaling pathways are shown in red. Fold change (HBV⁺/HBV⁻) is shown on the x axis; a positive fold change represents upregulated genes in HBV-infected cells, and the opposite for a negative fold change. The up- and downregulated genes are detailed in Tables 2 and 3. The data are shown as means and standard deviations. *, P < 0.05; ***, P < 0.001.

HBV infection increases the expression of host proviral factors.

Since TGFB2 suppresses the expression of different viral transcription factors, we hypothesize that HBV should augment the expression of proviral transcription factors in infected cells. In three sets of PHHs, HBV infection significantly induced expression of HNF4A, PPARA, RXRA, and CEBPB (Fig. 5C). The upregulation of PPAR and CEBPB was also confirmed by Western blotting (Fig. 5D). Similarly, many genes involved in RXR or PPAR activation signaling were induced upon HBV infection (Fig. 5E and Table 2).

TABLE 2.

RXR and PPAR activation signaling genes

GeneID	P value (positive vs. negative)	Fold change (positive vs. negative)
RXR activation signaling genes
ABCB11	0.0911644	1.24303
ABCB4	0.000553729	4.81421
ABCG1	0.157468	−1.38681
ABCG5	4.88E−05	2.19116
ABCG8	4.99E−05	2.63559
AHSG	0.000679832	2.09384
AKT3	0.687272	1.03852
ALB	0.0362356	1.05105
APOA2	0.000108545	2.68045
APOA5	0.000116081	2.88756
APOC2	0.00579067	1.07703
APOE	0.000919102	3.07845
APOF	0.000989539	2.70936
APOM	0.000123263	2.44931
C9	1.02E−05	5.80559
CCL2	0.000715844	−3.00396
CD14	3.14E−05	2.63888
CYP7A1	4.25E−06	9.16339
FASN	0.000132539	3.92209
FBP1	7.66E−05	2.82723
FETUB	3.67E−05	5.14559
FGFR4	0.00022089	2.04813
G6PC	9.12E−05	2.73953
GC	0.000171316	3.67231
HADH	0.000277282	1.96858
HPR	2.20E−05	2.55136
IL-18	8.42E−05	−2.39413
ITIH4	7.76E−05	2.05537
KNG1	1.99E−05	6.43038
LPA	0.000308573	5.85663
LY96	0.00221458	−2.01444
PCK2	5.32E−05	2.94396
PKLR	1.57E−05	5.51542
PON1	0.000142437	2.59116
SCD	0.000223152	2.29584
SERPINF2	0.000148847	3.45863
SLC22A7	0.0341398	1.97954
SLC27A5	4.22E−05	6.42102
SLCO1B1	1.58E−05	2.09569
SREBF1	0.000144665	2.78553
TF	0.0032699	5.00837
TLR4	0.00206251	−2.31052
PPAR activation signaling genes
GLIPR1	0.000439501	−1.44459
TEAD2	0.493776	−1.07708
TEAD4	0.00378904	−2.6338
SREBF1	0.000144665	2.78553
ABCB1	0.00829129	−1.157
ACSL1	0.000201855	1.58296
ABCB4	0.000553729	4.81421
CYP4A11	0.0285436	1.64966
APOA2	0.000108545	2.68045
APOA5	0.000116081	2.88756
FABP1	3.86E−05	3.52895
FADS2	0.00564342	1.85606
FADS1	1.63507	1.63507
NR1D1	2.53739	2.53739
CYP4A11	3.92158	3.92158
CYP1A2	0.100419	1.20837
HMGCS2	0.000359626	9.97594
ANGPTL4	0.00533475	−1.90632
TNFRSF21	0.000927324	−1.74483
CYP7A1	4.25E−06	9.16339
CTGF	0.00146618	−2.88962
ANKRD1	0.000232916	−3.06628

To assess the in vivo relevance of our proposed mechanism, we exploited the microarray gene expression data generated previously from liver biopsies of HBV-infected chimpanzees (28). Not all of the genes we identified here were detectable due to the low sensitivity of the microarray chip that was used. Consistent with our in vitro data, HBV infection resulted in an elevation of hepatic CEBPB expression in the chimpanzees (Fig. 6A). This induction correlated well with the HBV DNA levels in the blood, which likely reflected HBV replication in the livers of the infected chimpanzees.

FIG 6.

Proviral host factors and cell cycle alteration. (A) Hepatic expression of CEBPB and viremias of three HBV-infected chimpanzees at different time points of infection. (B and C) HBV-infected PHHs at different cell cycle stages as sorted by flow cytometry. HBV RNA (B) and the expression of different proviral host factors (C) were evaluated by qRT-PCR. The data are shown as means and standard deviations. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Increased expression of host proviral factors is linked to HBV-associated cell cycle alteration.

Since HBV infection altered cell cycle progression and promoted expression of host proviral factors, as we demonstrated earlier, we next asked whether HBV-infected cells in distinct cell cycle phases might exhibit different HBV replication levels or host proviral factor expression levels. HBV-infected PHHs were sorted into G₀/G₁, S, and G₂/M phases by FACS, and HBV transcriptional activity (defined as HBV RNA/cccDNA) was evaluated in each phase (Fig. 6B). Cells in the G₂/M phase fraction contained a larger amount of viral RNA than those in the G₀/G₁ and S phases, suggesting that HBV replicates more favorably in this cell cycle phase. Interestingly, we noted that cells in G₂/M phase also expressed higher levels of HNF4A, RXRA, and CEBPB but a lower level of TGFB2 (Fig. 6C), indicating that hepatocytes in this phase provide a more permissive environment for HBV replication.

To determine which viral proteins may be responsible for the alteration of host gene expression, we infected PHHs with adenovirus vectors expressing individual HBV genes (17). HBV infection was used as a positive control, while adenovirus expressing green fluorescent protein (GFP) was used as a negative control. The efficiency of adenovirus transduction was more than 90%, as determined by the proportion of GFP-positive cells (Fig. 7A). Interestingly, adenoviruses expressing surface antigen (HBs), core protein (HBc), X protein (HBx), and e antigen (HBe) did not induce higher levels of HBV transcription factors than the negative-control group, suggesting that components of active viral replication, either the replication intermediates or other by-products, and not individual viral proteins are responsible for the induction of proviral host factors associated with cell cycle alteration (Fig. 7B).

FIG 7.

Overexpression of different viral proteins by an adenovirus vector. (A) PHHs were transduced with adenoviral vectors expressing individual viral proteins or a control adenoviral vector. The infection efficiency was determined by GFP expression (green). (B) The expression levels of different proviral genes were evaluated by qRT-PCR. HBV infection was used as a positive control. PHHs (donor 1832) were infected with HBV and treated with 100 nM myrcludex B or 5 μM entecavir. (C and D) Eight days after infection, HBV-DNA and HBeAg from cell culture supernatants were evaluated by qPCR and ELISA, respectively (C), and expression of HNF4A, RXR, and PPAR was evaluated by qRT-PCR (D). The data are shown as means and standard deviations. **, P < 0.01; ***, P < 0.001; ns, not significant.

To further investigate the mechanism, HBV-infected PHHs were treated with either myrcludex B or entecavir. As expected, while the entry inhibitor myrcludex B sufficiently blocked virus infection, a reverse transcriptase inhibitor strongly suppressed viral DNA but showed no effect on viral protein production (Fig. 7C). Interestingly, entecavir treatment significantly blocked the effects of HBV infection, supporting the potential role of HBV replication in stimulating expression of these proviral factors (Fig. 7D).

DISCUSSION

While the mechanisms by which chronic HBV infection causes HCC are largely undefined, HBV likely causes HCC via both indirect and direct pathways. Antiviral therapy against HBV can reduce the risk of HCC in adult patients, but in individuals on effective antiviral therapy, the risk is not eliminated (29, 30). Early diagnosis, prevention, and treatment are important strategies to control hepatitis B-associated HCC.

The TGF-β superfamily proteins are multifunctional profibrotic cytokines that play a critical role in the regulation of cell growth, differentiation, and development in a wide range of biological systems (31). TGF-β signaling is initiated with ligand-induced oligomerization of serine/threonine receptor kinases and phosphorylation of the cytoplasmic signaling molecules SMADs (32). Activated SMADs regulate diverse biological effects by association with transcription factors, resulting in cell-state-specific modulation of transcription. TGF-β signaling can also affect SMAD-independent pathways, including ERK, SAPK/JNK, and p38 MAPK pathways (33, 34). TGF-β has a dual role in carcinogenesis. It acts early as a tumor suppressor by inhibiting cell proliferation but later stimulates cancer progression by its action on the tumor environment (33). Regulation of TGF-β expression and function is complicated, involving transcriptional, posttranslational, and protein-protein interactions. Deregulation of the TGF-β pathway has also been shown to be a key genetic event leading to HCC (24). The role of TGF-β in HBV-induced HCC is not fully understood.

TGF-βs are well-known growth inhibitors in a variety of cells and key regulators in checking hepatocyte proliferation during liver regeneration (23, 35). TGF-βs are known to cause cell cycle arrest at the cell cycle restriction point between the G₁ and S phases (36). Here, we show that HBV infection specifically downregulates the transcription of TGF-β2. Among the TGF-β family members, TGF-β2 appears to be uniquely regulated at the transcriptional level by epigenetic modifications (37, 38). The question of whether HBV infection results in an epigenetic reprogramming impacting growth-related genes like TGF-β2 is intriguing and needs further study.

Reduced TGF-β expression by HBV appears to promote the progression of infected cells into the G₂/M phase. The highly decreased cyclin B1 level observed in HBV-infected cells may also impede G₂/M transition and thus arrest the cells in the G₂/M phase. Since PHHs in culture barely proliferate (39), it is not clear whether HBV can indeed cause G₂/M arrest in actively dividing cells. Expression of other cyclins or CDKs are also affected by HBV, consistent with deregulated cell cycle control in HBV-infected cells. Furthermore, TGF-βs can induce apoptosis in hepatocytes (40, 41). HBV has been reported to inhibit apoptosis in host cells (42). As a strategy for persistence in the infected hepatocyte, HBV may inhibit cell death by disrupting the TGF-β signaling pathway, thereby promoting malignant progression.

Progression through the cell cycle consists of an ordered and tightly regulated cascade of events that are controlled by different factors. A common strategy of viruses is to create in the infected host cell an environment favorable for efficient replication and spreading. Many viruses disrupt the cell cycle by targeting specific steps of the cell cycle to enhance viral replication. For instance, some DNA viruses are able to induce a G₁/S phase arrest in order to replicate their genomes at the same time as cellular DNA synthesis (43). The myxoma virus M-T5 gene regulates cell cycle progression at the G₀/G₁ checkpoint, thereby protecting infected cells from diverse innate host antiviral responses normally triggered by G₀/G₁ cell cycle arrest (44). Influenza A virus induces favorable conditions for viral protein accumulation and virus production by inducing G₀/G₁ cell cycle arrest in infected cells (45). Enterovirus 71 mediates cell cycle arrest in S phase through its nonstructural protein 3D (46). HIV-1 Vif protein promotes the G₁/S transition to promote infection and latency (47). Herpes simplex virus 1 Vmw110 protein inhibits the progression of cells through mitosis and from G₁ into S phase of the cell cycle (48). After cytomegalovirus infection, the host cell is blocked prior to S phase to provide a favorable environment for viral replication (49). Epstein-Barr virus causes G₀/G₁ cell cycle arrest through induction of cyclin-dependent kinase inhibitors (50).

Here, we demonstrated that HBV infection promotes progression to the G₂/M phase and possibly causes G₂/M arrest of infected cells. Similarly, a number of viruses, including DNA viruses, RNA viruses, and retroviruses, cause cell cycle arrest between the G₂ and M phases, allowing viruses to replicate their genomes before the cells enter mitosis. Human papillomavirus induces G₂/M phase arrest through the maintenance of Cdk1 phosphorylation by E4 proteins (51). G₂/M arrest of hepatocytes caused by hepatitis C virus via antioxidative stress and TGF-β signaling has also been reported (52, 53). HIV Vpr protein directly inactivates the Cdk1/cyclin B1 complex through p21 (54, 55) and disrupts cell cycle progression by various mechanisms (56, 57).

HBV replication was shown previously to be independent of the cell cycle in an HBV transgenic-mouse model (58). HBV DNA and core protein in the livers of HBV transgenic mice were transiently reduced by cytokines induced during massive liver regeneration following partial hepatectomy, while viral RNA was unchanged (58). Despite massive hepatocyte turnover, this cytokine-mediated antiviral effect is independent of cell proliferation, suggesting that HBV can replicate efficiently in resting and dividing hepatocytes. However, in an HBV-infected humanized-mouse model, proliferation of HBV-infected hepatocytes provoked a dramatic decrease of HBV infection markers, including cccDNA (59, 60), suggesting that HBV cccDNA may be depleted during mitosis (61) Thus, HBV replication may be highly dependent on cell cycle phases. Previous studies also demonstrated that HBV can manipulate cell cycle regulation machinery in order to replicate more efficiently (62).

HBx protein has been implicated in deregulation of the cell cycle and, further, promotes carcinogenesis. An HBx-expressing murine liver cell line presented a deregulated G₂/M checkpoint, thus rescuing cells from apoptosis (63). By adenoviral vector overexpression, HBx has been shown to accelerate the transition through checkpoint controls at G₀/G₁ and G₂/M in human hepatoma Chang cells (64) but to arrest HepG2 cells at G₂/M through activation of cyclin B1-CDK1 kinase (65). Furthermore, overexpression of HBx in Chang cells induces G₂/M arrest through the ATM-Chk2 pathway (66). These divergent findings may reflect the nonphysiological nature of the hepatoma cells and the transfection used for these studies. In this study, we did not observe an overt effect of HBx on cell cycle regulation. To our knowledge, the present study is the first to report that in HBV-infected primary human hepatocytes, viral replication possibly promotes progression to and then arrest in the G₂/M phase, which likely creates an optimal cellular environment for viral replication. Additionally, blockade of mitosis is beneficial for maintenance of the cccDNA pool (59, 60).

As an obligate intracellular pathogen, a virus co-opts an infected cell to provide an environment favorable to its replication. Collectively, our results support the hypothesis that HBV hijacks cell cycle machinery to foster an environment with high levels of proviral factors (Fig. 8). HBV infection results in reduced TGF-β signaling, possibly leading to cell cycle deregulation and G₂/M cell cycle arrest. In this state, hepatocytes express more HBV transcription factors, including HNF4A, RXR, and CEBPB, which enhance HBV replication. These transcriptional factors are not known to be the direct targets of the TGF-β signaling pathway (67), suggesting that the observed TGF-β effect may be indirect. It is also possible that TGF-β downregulation may not be the mechanism that deregulates cell cycle progression in HBV-infected cells. Regardless, the resulting deregulated cell cycle signaling coincidently creates a premalignant condition and promote carcinogenesis.

FIG 8.

HBV hijacks cell cycle machinery to upregulate proviral factors. In this proposed mechanism, HBV infection results in deregulation of TGF-β2 expression, affects cell cycle regulation, and arrests cells in the G₂/M phase. In this phase, hepatocytes produce more nuclear factors, including HNF4A, RXR, PPAR, and CEBPB, to enhance HBV replication. On the other hand, deregulated cell cycle signaling, which affects the control of correct entry and progression through the cell cycle, is known to promote carcinogenesis.

MATERIALS AND METHODS

Hepatocyte cultures.

Cryopreserved PHHs from four different donors (Hu1192, Hu1413, Hu1457, and Hu1832) were purchased from Life Technologies Corp. (Carlsbad, CA, USA). HMC493 was purchased from BD Biosciences (Woburn, MA, USA). Cells were recovered in hepatocyte thaw medium (Thermo Scientific, Waltham, MA, USA) and plated in William's E medium (WEM) plus plating supplements (Thermo Scientific, Waltham, MA, USA). One day after plating, the PHHs were maintained in WEM.

HBV infection.

The HBV inoculum was produced from the supernatant of HepG2.2.15 cells, concentrated using centrifugal filter devices (Centricon Plus-70 and Biomax 100.000; Millipore Corp., Bedford, MA), and quantified by HBV-DNA quantitative PCR (qPCR). Immediately after collection, the virus stock was divided into aliquots and stored at −80°C until use. The medium control was prepared in the same way from the supernatant of HepG2 cells. For infection, inoculation of cells was performed at an MOI of 5,000 in WEM containing 5% polyethylene glycol (PEG) 8000 (Sigma-Aldrich, St. Louis, MO, USA) for 16 h. At the end of the incubation period, the cells were washed three times with PBS and cultured in WEM.

Human growth factor antibody array.

A human growth factor antibody array (ab134002; Abcam, Cambridge, United Kingdom) was used to investigate growth factor secretion profiles in cell culture supernatants. One milliliter of undiluted supernatant was used for each membrane, with sample incubation overnight at 4°C according to the manufacturer's instructions, followed by large-volume wash and incubation with biotin-conjugated anticytokine mixture overnight at 4°C. Densitometry data were obtained using ImageJ (National Institutes of Health, Bethesda, MD, USA [http://imagej.nih.gov/ij/]) and used to compare different samples after background subtraction and normalization to positive-control spots.

ELISA.

TGF-β2 from cell culture supernatant was quantified using the Human TGF-beta 2 Quantikine enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems). Secreted HBeAg and HBsAg were determined using a human HBeAg ELISA kit (CD BioScience, Shirley, NY, USA) and a human HBsAg ELISA kit (CD BioScience, Shirley, NY, USA), respectively.

Microarray.

PHHs (Hu1192) were mock infected or infected with HBV at an MOI of 5,000. At 8 days after infection, total cellular RNA was extracted and purified with TRIzol RNA isolation reagent (Thermo Scientific, Waltham, MA, USA). RNA quality and quantity were analyzed using an Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA). Cellular RNA was then amplified using an Agilent Enzo kit. Amplified RNA was hybridized to an Affymetrix Human 133 Plus 2.0 microarray chip containing 54,675 gene transcripts. For microarray analysis, a >1.5-fold change in expression combined with a >95% probability of being expressed differentially (P < 0.05) was considered to be biologically significant. Bioinformatics and statistical analyses of microarray data were performed at the National Institute of Diabetes and Digestive and Kidney Diseases Genomics Core Facility.

Bioinformatics analysis.

Prior to analysis with Ingenuity Pathway Analysis software, genes were filtered by a P value of less than 0.05 and a fold change of greater than 2 or less than −2. Data from the microarray were used.

RT² profiler PCR array human cell cycle.

The mRNA samples from the microarray were converted into cDNA using the RT² first-strand kit (SA Biosciences, Frederick, MD, USA). The cDNA was then added to the RT² SYBR green qPCR master mix (SA Biosciences, Frederick, MD, USA). Next, each sample was aliquoted on 24 human common cytokine PCR arrays. All the steps were performed according to the manufacturer's protocol for the Roche Lightcycler 480 system. The PCR array data were analyzed with the manufacturer's online software.

Flow cytometry analysis of the cell cycle.

PHHs were infected with HBV for 7 days and then harvested with trypsin. After suspension, Vybrant Dye Cycle Green (Thermo Scientific, Waltham, MA, USA) was added to each tube to reach a final concentration of 10 μM. The cells were incubated at 37°C for 30 min in the dark. After washing, the cells were suspended in cold phosphate-buffered saline and analyzed by flow cytometry (BD Biosciences, San Jose, CA) using 488-nm excitation and 520-nm emission.

Quantitative real-time PCR.

For intracellular HBV, total DNA, and cccDNA quantification, total cellular DNA was purified from infected cells using a NucleoSpin tissue kit (Macherey-Nagel, Düren, Germany). HBV total DNA qPCR was performed using the rcDNA1745 fw/rcDNA1844 rev primer pair. DNA samples for cccDNA qPCR were treated with 500 U/ml T5 exonuclease (NEB, Ipswich, MA, USA) at 37°C for 30 min. A cccDNA-specific primer pair, cccDNA 92 fw/cccDNA 2251 rev, was used for qPCR. Extracellular HBV DNA quantification was performed with DNA extraction using a DNeasy blood and tissue kit (SA Biosciences, Frederick, MD, USA). The rcDNA1745 fw/rcDNA1844 rev primer pair was used for qPCR. For qRT-PCR, RNA was extracted using an RNeasy kit (SA Biosciences, Frederick, MD, USA) and transcribed into cDNA with a Maxima first-strand cDNA synthesis kit (Thermo Scientific, Waltham, MA, USA) according to the manufacturer's instructions. qPCR was performed using the LightCycler 480 system with SYBR green Master (Roche, Manheim, Germany) or Probe Master (Roche, Manheim, Germany). All the primers and probes are listed in Table 3. The expression ratio of a target gene against the reference housekeeping gene (ACTB for RNA and PRNP for DNA) is presented.

TABLE 3.

Primers and probes

^aFAM, 6-carboxyfluorescein; 3IABkFQ, 3′ Iowa Black FQ quencher.

RNAscope assay.

In situ detection of HBV nucleic acid was performed using the RNAscope assay (Advanced Cell Diagnostics, Newark, CA, USA). TGFB2- and HBV-specific probe sets were provided by the manufacturer. Hybridizations were performed for 2 h with target probes at 40°C. Following signal amplification, probes conjugated to Atto 550 (Advanced Cell Diagnostics, Newark, CA, USA) were imaged using a Zeiss LSM 700 confocal microscope.

Western blotting.

Lysates from PHHs were obtained by adding 100 μl M-PER mammalian protein extraction reagent (Thermo Scientific, Rockford, IL, USA) per well to the cells and incubating them at 37°C for 5 min. Fifty microliters of 3× SDS sample buffer (Aviva Systems Biology Corp., San Diego, CA, USA) was added, and samples were incubated at 99°C for 5 min. Protein lysates were separated on a NuPAGE 4 to 12% Bis-Tris gel (Life Technologies Corp., Carlsbad, CA, USA) and then blotted onto a nitrocellulose membrane (iBlot 2 NC mini stacks; Thermo Scientific, Waltham, MA, USA) using iBlot (Thermo Scientific, Waltham, MA, USA). The membrane was blocked with Odyssey blocking buffer (Li-Cor Biosciences, Lincoln, NE, USA) for 1 h at room temperature, followed by overnight incubation with primary antibody at 4°C. Anti-rabbit or anti-mouse secondary antibody conjugated to horseradish peroxidase (Sigma, St. Louis, MO, USA) was used to visualize the stained bands with an enhanced chemiluminescence visualization kit (Thermo Scientific, Waltham, MA, USA). The primary antibodies used were rabbit anti-CEBPB (3087S; Cell Signaling, Danvers, MA, USA), rabbit anti-RXRA (5388S; Cell Signaling, Danvers, MA, USA), and mouse anti-β-tubulin (T8328; Sigma, St. Louis, MO, USA).

Statistical analysis.

Student's unpaired two-tailed t tests were performed with GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). The data are means and standard deviations. Two-sided P values of <0.05 were considered significant.

Accession number(s).

Gene expression data were deposited in the Gene Expression Omnibus under accession number GSE118295.