Class A capsid assembly modulator apoptotic elimination of hepatocytes with high HBV core antigen level in vivo is dependent on de novo core protein translation

ABSTRACT

Capsid assembly is critical in the hepatitis B virus (HBV) life cycle, mediated by the viral core protein. Capsid assembly is the target for new anti-viral therapeutics known as capsid assembly modulators (CAMs) of which the CAM-aberrant (CAM-A) class induces aberrant shaped core protein structures and leads to hepatocyte cell death. This study aimed to identify the mechanism of action of CAM-A modulators leading to HBV-infected hepatocyte elimination where CAM-A-mediated hepatitis B surface antigen (HBsAg) reduction was evaluated in a stable HBV replicating cell line and in AAV-HBV-transduced C57BL/6, C57BL/6 SCID, and HBV-infected chimeric mice with humanized livers. Results showed that in vivo treatment with CAM-A modulators induced pronounced reductions in hepatitis B e antigen (HBeAg) and HBsAg, associated with a transient alanine amino transferase (ALT) increase. Both HBsAg and HBeAg reductions and ALT increase were delayed in C57BL/6 SCID and chimeric mice, suggesting that adaptive immune responses may indirectly contribute. However, CD8+ T cell depletion in transduced wild-type mice did not impact antigen reduction, indicating that CD8+ T cell responses are not essential. Transient ALT elevation in AAV-HBV-transduced mice coincided with a transient increase in endoplasmic reticulum stress and apoptosis markers, followed by detection of a proliferation marker. Microarray data revealed antigen presentation pathway (major histocompatibility complex class I molecules) upregulation, overlapping with the apoptosis. Combination treatment with HBV-specific siRNA demonstrated that CAM-A-mediated HBsAg reduction is dependent on de novo core protein translation. To conclude, CAM-A treatment eradicates HBV-infected hepatocytes with high core protein levels through the induction of apoptosis, which can be a promising approach as part of a regimen to achieve functional cure.

IMPORTANCE

Treatment with hepatitis B virus (HBV) capsid assembly modulators that induce the formation of aberrant HBV core protein structures (CAM-A) leads to programmed cell death, apoptosis, of HBV-infected hepatocytes and subsequent reduction of HBV antigens, which differentiates CAM-A from other CAMs. The effect is dependent on the de novo synthesis and high levels of core protein.

INTRODUCTION

Hepatitis B is a major global health problem, with approximately 2 billion people estimated to have had an acute infection with the hepatitis B virus (HBV). The World Health Organization have estimated that 296 million people were living with chronic hepatitis B (CHB) infection in 2019, with 1.5 million new infections each year (1). Of those infected as adults with HBV, 1%–5% develop chronic infection, while the remainder clear the infection. Approximately 25%–50% of children infected with HBV at ages 1–5 develop chronic infection, as do >90% of infants infected at birth (2). Approximately 20%–30% of individuals with CHB ultimately develop cirrhosis, liver failure, or hepatocellular carcinoma (1).

Current CHB treatments rarely result in a functional cure, defined as sustained loss of hepatitis B surface antigen (HBsAg) with or without HBsAg seroconversion and undetectable HBV DNA in serum 24 weeks off treatment (3). Finite treatment with an immunomodulator (pegylated interferon-α) can induce loss of HBsAg but only in a limited patient subset (4, 5) and, due to side effects, may be poorly tolerated (6). Treatment with a nucleos(t)ide analog (NA; e.g., tenofovir disoproxil or entecavir) is generally lifelong, well tolerated (6), and associated with a high barrier to resistance. Nucleos(t)ide treatment controls viral replication, eliminates liver inflammation, and prevents cirrhosis but only partially reduces the risk of hepatocellular carcinoma (7–10) and rarely leads to functional cure.

Thus, novel therapeutic approaches with improved efficacy and an acceptable side effect profile are required to provide finite treatment regimens that can achieve high functional cure rates in CHB.

HBV capsid assembly is a critical step in the HBV’s life cycle and, hence, represents a therapeutic target. One of the seven proteins encoded by HBV is core, a protein which self-assembles to form the viral nucleocapsid. Core plays a role in most steps of the HBV life cycle, e.g., subcellular trafficking and transport, release of the HBV genome, capsid assembly, and reverse transcription (11). The HBV life cycle has been well described (11). In brief, the first cytoplasmic step is formation of a capsid containing pre-genomic RNA (pgRNA) and viral polymerase, initiated by polymerase-bound pgRNA (pol-pgRNA) association with three core protein dimers (nucleation). Single dimers are rapidly added to this nucleus to form the nucleic acid containing capsid or nucleocapsid (12). Within the nucleocapsid, reverse transcription of the pgRNA by viral polymerase produces relaxed circular DNA (rcDNA). The rcDNA-containing capsid either shuttles to the nucleus to replenish covalently closed circular DNA (cccDNA) or becomes enveloped, thereby forming an infectious viral particle to be released from the cell (13).

Capsid assembly modulators (CAMs) accelerate the kinetics of capsid assembly, whereby they prevent pol-pgRNA complex encapsidation and block HBV replication (14, 15). CAMs also interfere with cccDNA transcription/de novo formation during the early steps of infection (16–18). Such characteristics differentiate CAMs from NAs, as the latter solely interfere with the reverse transcription and polymerase process, inhibiting HBV replication. There are several types of CAMs which impact pol-pgRNA encapsidation at the nucleation step, and these can be broadly categorized into two classes. CAMs whose mechanism of action (MOA) results in the formation of empty, morphologically intact capsids are referred to as CAM-E (E = empty capsids), and CAMs whose MOA results in the formation of pleiomorphic non-capsid structures (i.e. aberrant particles) are referred to as CAM-A (A = aberrant structure) (19). Examples of the latter class include heteroaryldihydropyrimidine (HAP) chemotypes such as BAY41-4109 (20, 21), GLS-4 (15), JNJ-890 (18), and RO7049389 (22, 23). Interestingly, treatment of adeno-associated virus (AAV)-HBV-transduced mice with RO7049389 not only inhibited HBV DNA replication but also caused a strong decrease of hepatitis B e antigen (HBeAg) and HBsAg (22). Several CAM-Es and CAM-As have entered clinical trials, but meaningful HBsAg reductions have not been observed [for an overview see reference (24)]. In a recent study, Kum et al. presented clearance of HBV-infected hepatocytes through core-dependent hepatocyte death and proliferation by RO7049389, which they studied in different in vitro hepatoma cell lines and primary human hepatocytes as in the AAV-HBV mouse model (25).

In this paper, we studied the underlying biology of CAM-A induced HBsAg loss in vitro and in vivo in AAV-HBV-transduced mice to confirm HAP-induced apoptosis and expanded our studies over that of Kum et al. with a core titration in vivo study, testing CAM-A in an immunodeficient in vivo model and through a translationally relevant siRNA + CAM-A combination in vivo assessment. We also further explored this MOA in HBV-infected chimeric mice with humanized livers. The in vivo dosage for each CAM-A modulator was determined based on public information and previous pharmacokinetic studies.

RESULTS

HAP-1, HAP-2, and HAP-3 have anti-HBV activity in vitro

Three different HAP CAM-A compounds {HAP-1 [close structural analog of RG7907 (22)], HAP-2 [RG7907], and HAP-3 [close structural analog HAP compound]} with favorable pharmacokinetic and absorption, distribution, metabolism, and excretion profiles were used to study the underlying biology of CAM-A-induced HBsAg reduction. HAP-1, HAP-2, and HAP-3 satisfy the general structure as shown in Fig. 1A. The anti-HBV activity of HAP-1, HAP-2, and HAP-3 was evaluated in stable core-inducible HBV-replicating HepG2.117 cells that can be induced for HBV replication (26). The mean 50% effective concentrations (EC₅₀) of HAP-1, HAP-2, and HAP-3 were 86, 151, and 35 nM, respectively (Table 1). HAP-1, HAP-2, and HAP-3 did not show cytotoxicity up to 50 µM, the highest concentration tested, in HepG2 cells (selectivity index >331; Table 1).

Fig 1.

Formation of aberrant HBV core protein structures induced by HAP-1, HAP-2, and HAP-3, leading to induction of apoptosis via caspase 3/7 activity and associated with a decreased cell viability. HepG2.117 cells were not induced (A1, +DOX) or induced (A2–A7, −DOX) for core expression or treated with dimethyl sulfoxide (DMSO) as a control (A1, A2), JNJ-0827 (CAM-E) (A3), GLS-4 (A4), HAP-1 (A5), HAP-2 (A6), or HAP-3 (A7). Core protein was immunolabeled with primary core antibody and fluorescently labeled with Alexa 488 secondary antibody (green), while nuclei were co-stained with Hoechst (cyan). Fluorescence was detected by confocal imaging. (A2 and A3) A uniform core distribution in the nucleus and cytoplasm was observed in HepG2.117 cells induced for core expression (−DOX) and treated with DMSO or CAM-E. (A5–A7) Treatment of HepG2.117 with HAP-1, HAP-2, and HAP-3 led to the formation of aberrant core structures in the cytoplasm (represented by “dot-like structures”). The formation of aberrant cytosolic core structures was normalized to a reference HAP compound (GLS-4) used at a concentration of 1 µM (A4) as 100% effect and DMSO treatment as 0% effect. The representative images show the induced effect around the EC₅₀ concentration to induce aberrant core structures. (B) Induction of caspase 3/7 activity and cell viability evaluation of HepG2.117 cells that were not induced (+DOX) or induced (−DOX) for core expression and treated with HAP-1 or HAP-2 for 28 days (HAP-3 was not tested). Caspase activity and cell viability were assessed on days 14, 21, and 28. In each graph, data are shown as mean ± SEM.

TABLE 1.

Antiviral properties and cytotoxicity of HAP-1, HAP-2, and HAP-3 in HepG2.117^a

	Cytosolic aberrant core
Compound	Mean pEC50 (SD)	Converted mean EC50	Mean Emax (%)	N
HAP-1	5.68 M (0.228)	2.08 µM	103%	5
HAP-2	5.54 M (0.049)	2.85 µM	111%	2
HAP-3	6.05 M (0.568)	0.89 µM	88%	4
GLS-4	7.01 M (0.138)	0.097 µM	121%	234
JNJ-0827	>4 M (0.093)	>100 µM	6.9%	244

^a The antiviral properties of HAP-1, HAP-2, and HAP-3 were tested in a dose-response assay in stable HBV-replicating HepG2.117 cells. The potency of HAP-1, HAP-2, and HAP-3 to induce aberrant cytosolic or nuclear core protein structures was assessed in a high-content imaging assay with the EC₅₀ representing the 50% effective concentration and the E_max effect as the maximum effect induced compared to a HAP reference compound (GLS-4, 1 µM). Intracellular HBV DNA was extracted from cell lysates and assessed using qPCR in the standard HBV antiviral assay. The cytotoxicity was assessed in HepG2 cells using an ATP-lite read-out. In each experiment, EC₅₀ values were determined based on the mean core induction/inhibition from two wells per compound concentration. GLS-4 and JNJ-0827 (CAM-E) are included as references. pEC₅₀, negative logarithm of the EC₅₀; EC₅₀, 50% effective concentration; SD, standard deviation; CC₅₀, 50% cytotoxic concentration; E_max, maximal effect to induce core aberrant structures; N, number of experiments; qPCR, quantitative polymerase chain reaction; SI, selectivity index (CC₅₀/HBV DNA EC₅₀).

To confirm that HAP-1, HAP-2, and HAP-3 induced aberrant HBV core protein, the compounds were tested in a high-content imaging assay in HepG2.117 cells designed to identify and quantify the induction of aberrant HBV core proteins (27). HAP-1, HAP-2, and HAP-3 induced aberrant cytosolic and nuclear core protein (represented by “dot-like structures”) formation with a mean 50% effective concentrations of 2.08, 2.85, and 0.89 µM, respectively, for cytosolic aberrant core structure formation and 2.6, 3.27, and 0.904 µM, respectively, for nuclear core structure formation (Table 1), confirming that the compounds possessed a CAM-A MOA (Fig. 1A through E). The aberrant core structure formation is in line with what has been observed for GSL-4, while JNJ-0827, a capsid assembly modulator that induce normal capsids (CAM-E), did not induce any aberrant core structure formation (mean cytosolic and nuclear EC₅₀ >100 µM) (Table 1).

The formation of intracellular aberrant core protein structures could lead to cellular stress responses and, if aberrant core accumulates, to apoptosis. To test if CAM-A-mediated aberrant core protein formation induced apoptosis, HepG2.117 cells were cultured with (suppressed HBV replication) or without (induced HBV replication) doxycycline in presence of HAP-1 and HAP-2 for up to 4 weeks, and a caspase 3/7 assay was performed to detect apoptotic cells (HAP-3 was not included in this study). A dose-dependent caspase 3/7 activity was only detected in HBV replicating but not in HBV replication-suppressed HepG2.117 cells after 2 weeks of treatment and progressed further at 3 and 4 weeks of treatment. The increase in caspase 3/7 activity was associated with a decrease in cell viability, which was evaluated by the remaining cell count present in the well measured via high-content imaging (Fig. 1F). The two distinct readouts, cell viability and caspase activity, are concomitantly affected by HAP treatment and are inversely correlated since the induction of caspase 3/7 reflects an increase in apoptotic or dead cells that are lost in the washing steps of the immunostaining method used for cell viability. The loss of cells treated with high compound concentration during the immunostaining results in a cell count close to 0%. These data indicate that accumulation of aberrant core protein induced apoptosis and that the compound itself in the absence of core did not induce this effect.

CAM-A mediates anti-HBV activity in AAV-HBV-transduced C57BL/6 mice

The anti-HBV activity of HAP-1 was tested in vivo in C57BL/6 mice transduced with 1 × 10¹¹ viral genome (vg) equivalents AAV-HBV. Mice were orally gavaged once daily (QD) with 36 mg/kg HAP-1 or placebo for 28 days. Plasma sampling was performed frequently for the quantification of HBV DNA, HBsAg, and alanine amino transferase (ALT) activity (Fig. 2A). Treatment with HAP-1 induced a dose-related rapid decline of HBV DNA from a pre-treatment 4.2 log₁₀ copies/µL to less than the lower limit of quantification (LLOQ) (2.08 log₁₀ copies/µL) after 7 days of treatment; this suppression was maintained until the end of treatment on day 28. A continuous decline of HBsAg was observed from day 16 onwards in HAP-1-treated mice. The profound HAP-1-induced HBsAg decline was associated with a transient increase in ALT activity (Fig. 2B and C), suggesting that treatment with CAM-A induced a biological effect in the liver, but not chronic toxicity, which is also supported by a lack of body weight loss (Fig. 2D). Liver samples from three mice from the placebo and CAM-A-treated group were analyzed at multiple time points to assess potential HAP-1-induced hepatic effects (Fig. 2A). A time-dependent core and HBsAg reduction was observed in the liver of HAP-1 treated mice, indicating that treatment with HAP-1 induced the elimination of HBV-infected hepatocytes over time (Fig. 2E). No effect on HBV DNA, HBsAg, or ALT was observed in mice treated with placebo (Fig. 2B and C).

Fig 2.

CAM-A anti-HBV activity in AAV-HBV-transduced C57BL/6 wild-type mice. (A) C57BL/6 mice transduced with 1 × 10¹¹ vg equivalents AAV-HBV were treated orally with placebo or HAP-1 (36 mg/kg QD) for 28 days. Time points and number of mice that were used for intermediate liver sample analysis are indicated. (B) Plasma sampling was performed frequently for the quantification of HBV DNA, HBsAg, and ALT. (C) Body weight was monitored during treatment duration. (D) Immunohistochemistry for the detection of HBV core and HBsAg in the liver of mice sacrificed at different time points is shown for placebo and HAP-1-treated mice for days −1, 7, 14, 21, and 28. QD, once daily; LLOQ, lower limit of quantification. In each graph, data are shown as mean ± SEM and represent groups of three mice.

Treatment with CAM-A induces apoptosis of infected hepatocytes in AAV-HBV-transduced mice

HAP-1 induces the formation and accumulation of intracellular aberrant core protein structures. To further understand how the induction of aberrant structures is associated with transient ALT increases and loss of core staining, we explored whether the accumulation of such aberrant subcellular structures may trigger cellular stress responses that result in the induction of apoptosis of HBV-infected hepatocytes in vivo. We applied a multiplex immunofluorescence analysis to liver slices from HAP-1- and placebo-treated mice at multiple time points for the detection of HBsAg, Ki67 (marker for proliferation), and DAPI (DNA stain). A TUNEL immunofluorescence stain for the detection of apoptotic cells that undergo extensive DNA degradation during the late stages of apoptosis was performed. An increase in apoptotic hepatocytes (up to ~3.5%) was observed in HAP-1-treated mice compared to placebo-treated mice (Fig. 3A), which was accompanied by an increase in hepatocyte proliferation (determined by Ki67 measurement) (Fig. 3B.1). At later time points (days 24 and 28), these proliferating hepatocytes were predominantly HBsAg negative (Fig. 3B.2). The detection of apoptotic cells overlapped with the transient increase in ALT activity, suggesting that CAM-A-induced formation and accumulation of aberrant core protein in high level of core-expressing hepatocytes caused this effect. Consistent with the time-dependent HBsAg reduction in plasma, we observed a time-dependent reduction of HBsAg-positive hepatocytes in HAP-1-treated mice (Fig. 3B.3).

Fig 3.

Treatment with CAM-A induces apoptosis and proliferation in the liver of AAV-HBV-transduced mice. (A) Formalin-fixed, paraffin-embedded liver samples were sectioned at 5 µm, mounted on SuperFrost Plus glass slides, and used for TUNEL analysis. The percentages of TUNEL-positive (apoptotic) cells are shown for HAP-1 (36 mg/kg QD)-treated mice compared to placebo-treated mice. (B) Alternative HBsAg/Ki67/DAPI-stained images are shown with the calculation of %Ki67+ cells representing proliferating cells. A subdivision between HBsAg− and HBsAg+ cells of the Ki67+ cells for HAP-1-treated AAV-HBV mice is shown as well. Time-dependent immunofluorescence staining against HBsAg (green) on liver slices with the correlating calculation of %HBsAg+ hepatocytes over time of HAP-1-treated AAV-HBV mice versus placebo treated. For A and B, data are shown as floating bars indicating the range and the mean by black dots. (C) Volcano plots per each time point for log fold change difference expression of genes between placebo and HAP-1-treated AAV-HBV mice, colored by adjusted false discovery rate (fdr) P-value. Pathway ranking plot from gene set enrichment analysis between placebo and HAP-1-treated AAV-HBV mice. Top 5 pathways, with lowest P-value score, are shown. QD, once daily. The data add on to the in vivo study explained in Fig. 2.

Treatment with CAM-A transiently induces immune activation

To assess if additional factors contribute to the CAM-A-mediated clearance of HBV-infected hepatocytes, mouse plasma samples and liver homogenates were analyzed for the presence of immune mediators using the Mouse Luminex Discovery Assay. A transient increase of IP10, MIG, and FAS (increase was less clear for FAS) was observed in plasma and liver homogenates of HAP-1-treated mice (Fig. S1A), but not in mice treated with vehicle only. FASL was only detected in plasma, whereas RANTES and Granzyme B were only detected in liver homogenates (Fig. S1B). No evident HBV antigen-specific T cells were detected in spleen by ex vivo IFN-γ ELISpot on days 7, 16, 21 and 28 post dosing (data not shown).

Effect of CAM-A treatment on liver gene expression

To understand whether changes in RNA expression were observed during the treatment-induced clearance of HBV-infected hepatocytes, liver gene expression was analyzed at different time points. Differentially expressed genes between HAP-1-treated versus vehicle-treated mice were mainly observed between days 12 and 16, with over 500 genes significantly expressed (Fig. 3C; Table S1). Days 14 and 16 showed 1,947 and 1,743 significantly expressed genes, respectively (adjusted P-value threshold of 0.05), indicating profound transcriptional changes. Pathway analysis showed an enrichment in processes involved in antigen processing and presentation and responses to interferon. Increased gene expression for certain interferon-stimulated genes and interferon-related genes (e.g., Cd74, H2-Ab1, H2-Eb1, and H2-Q7) was observed from days 12 to 16, coinciding with the appearance of apoptotic events, which also remained relatively stable until day 28 (Fig. S2).

CAM-A mediates anti-HBV activity in AAV-HBV-transduced C57BL/6 SCID mice lacking T and B cells

The cytokine and microarray data from studies in C57BL/6 mice suggested that adaptive immune responses might contribute to the elimination of HBV-infected hepatocytes and the accompanying reduction in HBsAg. To study this in more detail, C57BL/6 SCID mice that lack mature B and T lymphocytes were transduced with 1 × 10¹¹ viral genome equivalents AAV-HBV and treated orally for 28 days QD with 36 mg/kg HAP-1 or placebo (Fig. 4A). There was no apparent difference in HBV DNA (Fig. 4B), HBeAg (Fig. 4C), and HBsAg (Fig. 4D) levels at the beginning of dosing between C57BL/6 and C57BL/6 SCID AAV-HBV-transduced mice showing that the efficiency of transduction was similar in both mouse strains. Plasma HBV DNA levels decreased to reach the LLOQ after 7 days of dosing in C57BL/6 and C57BL/6 SCID (Fig. 4B), indicating that the compound was equally efficacious in suppressing HBV replication in both mouse strains. However, there was an apparent difference observed in time and extent between C57BL/6 and C57BL/6 SCID transduced mice for both HBsAg decline and transient increase in ALT (Fig. 4D). A delayed and less pronounced decline of HBsAg and a delayed transient increase in ALT were observed in C57BL/6 SCID compared to C57BL/6 mice (Fig. 4D). Also, a delayed and less pronounced HBeAg decline was observed in C57BL/6 SCID mice (Fig. 4C). The reason for this less pronounced transient increase in ALT in this study in the wild-type (wt) mice is unknown. The limited frequency of plasma sampling in mice during longitudinal analyses may preclude timing the blood draw to be coincident with the peak of ALT incidence. However, full target engagement can be assumed since HBV DNA levels dropped to the LLOQ. Both placebo-treated C57BL/6 and C57BL/6 SCID transduced mice did not show a transient ALT increase. Due to the delayed transient ALT increase, a longer follow-up period may have been required to observe the maximum HBsAg effect.

Fig 4.

HBV DNA, HBeAg, and HBsAg decline and transient increase in ALT in C57BL/6 and C57BL/6 SCID mice after HAP-1 treatment. (A) C57BL/6 and C57BL/6 SCID mice that lack mature B and T lymphocytes were transduced with 1 × 10¹¹ vg equivalents AAV-HBV and treated orally for 28 days with HAP-1 (36 mg/kg QD) or placebo. (B and C) Plots represent HBV DNA and HBeAg levels in plasma during treatment duration. (D) Correlation plot is shown between HBsAg levels and ALT increase in plasma during treatment duration. QD, once daily; vg, viral genome; LLOQ, lower limit of quantification. Data are shown as mean ± SEM and represent groups of five mice.

CD8 T cells are dispensable for CAM-A-mediated anti-HBV activity in AAV-HBV-transduced C57BL/6 mice

Results from a study in AAV-HBV-transduced C57BL/6 SCID mice suggested that adaptive immune responses contribute to clearance of HBV-infected hepatocytes. To study if CD8+ T cells contribute to the observed effect, C57BL/6 AAV-HBV-transduced mice had their CD8+ T cells depleted with an anti-CD8 antibody or were treated with an isotype control antibody and additionally dosed orally (gavage) QD with placebo or 20 mg/kg HAP-2 for 56 days (Fig. 5A). Successful depletion of CD8+ T cells, maintained over time, was confirmed by fluorescence-activated cell sorting (FACS) (Fig. 5B). Treatment with 20 mg/kg HAP-2 leads to a similar response for HBsAg and HBeAg, in both CD8+ depleted and non-depleted mice suggesting that CD8+ T cells are dispensable for the clearance of HBV-infected hepatocytes (Fig. 5C). However, it cannot be excluded that other components of the adaptive immune system are involved.

Fig 5.

CAM-A anti-HBV activity in AAV-HBV-transduced C57BL/6 wt mice with depleted CD8+ T cells. (A) AAV-HBV-transduced C57BL/6 mice with 1 × 10¹¹ vg equivalents were treated orally with placebo or HAP-2 (20 mg/kg QD) for 56 days. CD8+ T cell depletion, intermediate plasma sampling, sacrifices, and number of mice that were sacrificed are indicated on the timeline. (B) FACS analysis gating for CD8+ T cells with representation of CD8+ T cell frequency in blood of HAP-2-treated HBV-AAV mice versus placebo treated. (C) HBsAg and HBeAg response after treatment with HAP-2 versus placebo in both CD8+ depleted and non-depleted mice. QD, once daily; LLOQ, lower limit of quantification. Data are shown as mean ± SEM and represent groups of five mice for the isotype-treated control groups or eight or nine mice for the anti-CD8-treated groups.

Induced elimination of HBV-infected hepatocytes is dependent on high viremia and antigen expression

The AAV-HBV mouse model represents a single-round infection model, as HBV cannot infect mouse hepatocytes (28, 29) nor can mouse hepatocytes support the formation of authentic cccDNA (30), and the transcription of HBV RNAs is dependent on the AAV-HBV template within the transduced hepatocytes. Therefore, the transduction of mice with different AAV-HBV doses results in different levels of viremia and antigen expression in the hepatocytes. To study this core protein dependency in more detail, we transduced mice with four different doses of AAV-HBV (5 × 10⁹, 1 × 10¹⁰, 2.5 × 10¹⁰, and 1 × 10¹¹ AAV-HBV vg equivalents) and treated orally for 56 days QD with 36 mg/kg HAP-1 or placebo (Fig. 6A). An inoculum-dependent increase in HBV DNA and HBeAg (Fig. S3) as well as HBsAg (Fig. 6B) and HBV core-related antigen (HBcrAg) (Fig. 6C) levels was observed in mice transduced with 5 × 10⁹, 1 × 10¹⁰, 2.5 × 10¹⁰, and 1 × 10¹¹ vg equivalents, respectively, indicating that viremia is dependent on the amount of AAV-HBV inoculum used for transduction and not on spread of infection. Treatment with HAP-1 reduced HBV DNA levels to LLOQ (Fig. S3). HBsAg was reduced over time only in mice transduced with 2.5 × 10¹⁰ and 1 × 10¹¹ AAV-HBV vg equivalents (Fig. 6B). An earlier and steeper decline of HBsAg was observed in mice transduced with 1 × 10¹¹, suggesting that higher levels of viremia/expression of core antigen in hepatocytes led to a faster and greater accumulation of aberrant core protein resulting in an earlier and stronger induction of apoptosis upon treatment. In addition, higher levels of HBcrAg and a steeper decline upon treatment were observed in mice transduced with 1 × 10¹¹ vg equivalents (Fig. 6C). No significant effect on HBsAg and HBcrAg was observed in mice transduced with 5 × 10⁹ or 1 × 10¹⁰ vg equivalents. Consistent with the HBsAg observations, a more pronounced and earlier transient increase in ALT activity was observed in mice transduced with 1 × 10¹¹ vg equivalents compared to mice transduced with 2.5 × 10¹⁰ vg equivalents upon treatment (Fig. 6D). No transient increase in ALT activity was observed in mice transduced with 5 × 10⁹ and 1 × 10¹⁰ vg equivalents or in any of the placebo-treated mice (Fig. 6D).

Fig 6.

HBsAg and HBcrAg levels in mice transduced with different amounts of AAV-HBV. (A) C57BL/6 mice transduced with four different amounts of AAV-HBV (5 × 10⁹, 1 × 10¹⁰, 2.5 × 10¹⁰, and 1 × 10¹¹ AAV-HBV vg equivalents) were treated orally with placebo or HAP-1 (36 mg/kg QD) for 28 days. (B–D) Plots represent HBsAg, HBcrAg and ALT levels in plasma during treatment duration. QD, once daily; vg, viral genome; LLOQ, lower limit of quantification. Data are shown as mean ± SEM and represent groups of five mice for the placebo-treated control groups and 10 mice for the HAP-1-treated groups.

CAM-A-induced elimination of HBV-infected hepatocytes is dependent on de novo core protein translation

Given that the CAM-A-induced clearance of HBV-infected hepatocytes is dependent on high viremia and high core protein levels, we next investigated if the induction of apoptosis was dependent on de novo core protein translation. To do this, a GalNac-conjugated HBV siRNA was used in combination with the CAM-A compound in the AAV-HBV mouse model. C57BL/6 mice were transduced with 1 × 10¹¹ vg equivalents AAV-HBV and treated with placebo, HAP-3 (12 mg/kg oral for 12 weeks daily QD), siRNA [3 mg/kg subcutaneously (s.c.) four times every 3 weeks], and a combination of HAP-3 with siRNA, followed by a 12-week washout (Fig. 7A). Plasma samples were collected at multiple time points throughout the study to assess the effect on HBV DNA, HBeAg, HBsAg, HBcrAg, and ALT (Fig. 7B through F). HBV DNA levels declined to and remained at LLOQ during HAP-3 monotherapy and during washout, indicating that HAP-3 monotherapy eliminated HBV-infected hepatocytes as observed previously (Fig. 7B).

Fig 7.

CAM-A-induced elimination of HBV-infected hepatocytes is dependent on de novo core protein translation. (A) C57BL/6 mice were transduced with 1 × 10¹¹ vg equivalents AAV-HBV and treated with placebo, HAP-3 (12 mg/kg oral for 12 weeks daily QD), siRNA (3 mg/kg s.c. four times every 3 weeks), and a combination of HAP-3 with siRNA, followed by a 12-week washout. (B–F) Plasma samples were collected at multiple time points throughout the study to assess the effect on HBV DNA, HBeAg, HBsAg, HBcrAg, and ALT. QD, once daily; vg, viral genome; s.c., subcutaneously; LLOQ, lower limit of quantification. Data are shown as mean ± SEM and represent groups of five mice.

Treatment with siRNA alone induced a less pronounced reduction of HBV DNA than HAP-3 and a plateau around week 6 prior to the third siRNA dose. Additional siRNA administration did not reduce the HBV DNA further. Mean HBV DNA levels remained constant during the washout phase, likely due to a long half-life of the siRNA. However, a larger level of variation across mice was observed (increased standard errors on the derived means during the washout compared to the treatment phase). The combination of siRNA with HAP-3 induced a rapid decline of HBV DNA to the LLOQ. A rebound of HBV DNA to levels observed with siRNA monotherapy occurred after 2 weeks washout (week 14), indicating that antiviral activity was mediated by the siRNA and not by HAP-3. Rebounds were also observed for HBeAg and HBcrAg in the HAP-3 + siRNA treatment group during the washout phase, whereas no rebound was observed in the HAP-3 monotherapy group, suggesting that combination treatment with siRNA interfered with the HAP-3-induced clearance of HBV-infected hepatocytes (Fig. 7D and E). This assumption is supported by the observation that a transient increase in ALT activity was only observed in the HAP-3 monotherapy group and not in the HAP-3 plus siRNA combination treatment group (Fig. 7F). Treatment with siRNA alone or in combination with HAP-3 resulted in a faster and steeper HBsAg decline than HAP-3 treatment alone. A rebound was observed during the washout phase (Fig. 7C). Altogether, these data suggest that CAM-A-induced clearance of HBV-infected hepatocytes is dependent on de novo core protein translation.

CAM-A modulation mediates anti-HBV activity in HBV-infected chimeric mice with humanized liver

To study the anti-HBV activity of CAM-A in a model with “authentic” HBV infection, HBV-infected chimeric humanized liver mice were treated orally for 84 days QD with 36 mg/kg HAP-1 or placebo. Approximately four times lower compound concentrations were observed in the liver of chimeric mice compared to C57BL/6 mice treated with the same dose (data not shown).

Repeated serum sampling was performed to quantify HBV DNA, HBeAg, HBsAg, HBcrAg, hAlb and hALT activity. Mice samples were taken at multiple time points during treatment (Fig. 8A) to assess HAP-1-induced effects in the liver. Treatment with HAP-1 induced a rapid decline of HBV DNA, indicating that the compound suppressed HBV replication (Fig. 8B). In contrast to what was observed in AAV-HBV-transduced mice, a clear HBsAg and HBcrAg reduction was only observed from day 35 onward, and the decline was less steep. Interestingly, HBsAg decline was also associated with a transient increase in human ALT activity (Fig. 8B), suggesting that treatment with CAM-A induced a transient biological effect in the human hepatocytes. Stable human albumin levels were observed over the entire treatment duration (Fig. 8C), and no body weight loss (Fig. S4) was observed during HAP-1 treatment, further indicating that the compound itself is unlikely to cause cytotoxic effects. Previous CAM-E study data in chimeric mice have shown that there is no-to-little HBsAg decline during a 2-month period. With only 200 mg/kg QD dosing of JNJ-8320, limited HBsAg effect is observed but linked to HBV RNA stability effect of CAMs (31). A confirmation study showed similar HBsAg decline with associated transient ALT increase after HAP-1 treatment study whereby also a high dose of JNJ-8320 (200 mg/kg QD) (CAM-E) was tested in parallel. Treatment with CAM-E did not result in a transient ALT increase, and the limited HBsAg decline was due to RNA instability (Fig. S5).

Fig 8.

CAM-A anti-HBV activity in HBV-infected chimeric humanized liver mice. (A) HBV-infected chimeric humanized liver mice were treated orally for 84 days QD with 36 mg/kg HAP-1 or placebo. Time points and number of mice that were used for intermediate liver sample analysis are indicated. (B) Serum sampling was performed frequently for the quantification of HBV DNA, HBsAg, HBcrAg, and hALT-1 activity. (C) Detection of human albumin levels in blood during treatment duration. (D) Core and HBsAg detection over time in the liver of mice treated with HAP-1 or placebo with a more detailed image of the core staining on day 35 to visualize the formation of aberrant core structures after HAP-1 treatment versus the normal core detected in the placebo treated condition. QD, once daily. Data are shown as mean ± SEM and represent groups of three mice for the placebo-treated control groups and four mice for the HAP-1-treated groups.

Aberrant core protein, represented by “dot-like structures,” was detected in the liver of mice treated with HAP-1, and a reduction of core and HBsAg positive hepatocytes was observed over time (Fig. 8D). We were not able to detect apoptosis and proliferation of hepatocytes in the liver of HAP-1-treated mice in contrast to the observations in the AAV-HBV mouse model.

DISCUSSION

HBV represents a global health burden that necessitates the development of new regimens for CHB, aimed at achieving increased functional cure rates after finite duration therapy. At a first glance, targets for direct-acting antivirals may be considered exhaustive as the condensed HBV genome only encodes seven proteins (32). However, all these viral proteins fulfill multiple roles in the HBV life cycle, which allows for broader pharmacologic considerations when interfering with their functions. HBV core protein provides an attractive target for the development of new interventions because it acts at multiple key steps during viral replication (11). In this study, we evaluated the anti-HBV activity and mode of action in vitro and in vivo of HAP capsid assembly modulators classified as CAM-A.

HAPs are known to induce heterogeneous forms of capsid in vitro (14). This study is consistent with previous reports showing that treatment with HAP CAM-A modulators induce the aggregation of HBV core protein in an inducible HBV replicating cell line (27). A recent study from Kum et al. showed that in the AAV-HBV model, CAM-A induced aggregation and apoptosis and was also characterized in different in vitro cell line-based and primary human hepatocyte models (25).

We confirmed that accumulation of such aberrant protein structures was shown to lead to apoptosis of the cells in vitro when cultured for an extended period in presence of compound. This was not an artifact since a dose- and time-dependent increase in caspase 3/7 activity was completely absent in cells treated with compound but lacking core expression. The observation that apoptosis is not immediately evident led us to hypothesize that the effects induced by these aggregated core protein structures over time may be the result of an accumulation of the unfolded protein response (UPR). The UPR functions to restore the normal homeostasis of a cell when stressed by the accumulation of misfolded proteins. If this response cannot adequately cope with the insult within a certain time span, the cell undergoes apoptosis (33). Similar observations of aggregate induced effects by HAPs have been described by Huber et al. (34). The authors report on the association of core aggregates with promyelocytic leukemia nuclear bodies (PML-NBs) in infected cells when exposed to CAM-A modulator HAP Bay38-7690. Although the publication does not describe the induction of apoptosis, Huber et al. do suggest that this accumulation of core protein with PML-NBs could potentially disrupt cellular process and impart cell death.

Notably, when transduced mice were dosed with HAPs, not only did they profoundly suppress HBV replication but the treatment also led to the reduction of HBsAg accompanied by a transient increase in ALT in absence of any safety signals. This decline in HBsAg has not been described for compounds belonging to the CAM-E class pointing toward an additional differentiating property of CAM-As. Furthermore, this ALT increase was absent in naïve mice dosed with compound. In examinations of the liver of AAV-HBV-transduced mice receiving compound, it was apparent that compared to placebo-treated animals, mice treated with HAPs showed a time-dependent decline in core- and HBsAg-expressing hepatocytes coinciding with the increase in ALT. To connect the apoptosis observed in vitro with a potential similar event in vivo, we performed a multiplex immunofluorescence on the liver looking at HBsAg, DNA, and Ki67, a known marker for proliferation. These results showed that CAM-A treatment increased the percentage of apoptotic cells along with an enhanced proliferation of HBsAg-negative cells with a subsequent reduction in HBsAg-positive cells and confirmed also the observed core- and CAM-A-dependent cell death via apoptosis published by Kum et al. (25). The accumulation of aberrant core structures within hepatocytes by HAPs resulted in cell death and was represented by a transient ALT increase and repopulation of the liver by non-infected hepatocytes.

In exploiting the single-round infection characteristics of the AAV-HBV mouse model, we also demonstrated that the size of the inoculum determines the level of antigen expressed in the transduced hepatocytes and, in turn, the extent of decline in viral antigen and ALT activity upon treatment with CAM-A. No CAM-A-mediated antigen reduction and transient increases in ALT were observed in mice with low(er) levels of viremia, implying that a steady state between CAM-A-induced formation of aberrant core protein structures and their degradation in hepatocytes of those mice could be maintained. In contrast, CAM-A treatment induced a faster and steeper decline of antigen in mice with the highest level of viremia suggesting that faster accumulation of aberrant core protein led to an earlier induction of apoptosis. Our titration studies in transduced mice suggest that HBcrAg levels above 8 log₁₀ IU/mL would be required to observe a CAM-A-mediated HBsAg decline. High HBcrAg levels above 8 log₁₀ U/mL are mainly observed in naïve HBeAg+ patients, and HBcrAg exhibits good correlation with intrahepatic cccDNA (35). Therefore, it is likely that HBsAg reductions during CAM-A treatment would only be observed in treatment-naïve, HBeAg+, and especially immunotolerant patients who have the highest HBcrAg levels in the periphery and high core protein levels in hepatocytes. In patient populations with low levels of replication and low intrahepatic core antigen levels, a CAM-A is unlikely to significantly contribute to HBsAg reduction and functional cure.

This is corroborated by the CAM-A/siRNA combination study data that showed the CAM-A-mediated effect of HBV antigen reduction is dependent on the de novo synthesis of core protein and that preexisting core protein levels are insufficient to induce apoptosis.

Given that transduced mice do not recapitulate the full replication cycle of HBV and to exclude that this model, therefore, harbors a bias toward this pharmacodynamic effect of a CAM-A HAP modulator, we also conducted a study in chimeric mice with a humanized liver. These mice can be infected with HBV, and the level of viremia is dependent on de novo infection and viral spread. Consistent with the observations in vitro, the formation of aberrant core protein structures into “dot-like structures” was also apparent in hepatocytes of chimeric mice treated with CAM-A, demonstrating that similar pharmacological effects occur in human hepatocytes in vivo. This model also exhibited an HBsAg decline and transient increase in human ALT albeit delayed and less pronounced compared to transduced mice suggesting that a similar biological effect was induced. However, we were not able to observe apoptosis and proliferation, likely due to too early and infrequent sampling in the study. The difference between the two mouse models might be attributed to an approximately lower compound exposure in chimeric mice compared to AAV-HBV mice at the same dose. In addition, chimeric mice do not possess an adaptive immune system, and we cannot exclude that the delays and reduced extent of HBsAg decline and ALT are due to an immunocompromised background.

To elaborate further on the potential role of adaptive immunity in response to CAM-A treatment, we compared the antiviral effects in both transduced wild-type and SCID mice and evaluated the response of HAP treatment in transduced mice depleted in CD8+ T cells. These studies collectively showed that CD8+ T cells are not required for CAM-A-mediated HBsAg declines. No HBV-specific T cells were detected in the spleen suggesting that CAM-A-mediated induction of HBV-infected hepatocyte cell death is unlikely to generate potent immunogenic responses to control the virus and eliminate infected cells. However, it cannot be excluded that other factors beyond the induction of apoptosis contribute to the observed effects as we observed differentially expressed genes in the AAV-HBV mouse model before the ALT flare leading up to the HBsAg decline. Gene set enrichment showed that the upregulated genes were involved in antigen presentation and major histocompatibility complex class II. Interestingly, CD74 was upregulated and is known as a receptor for MIF-mediated signaling in tissue repair (32). In the liver, this same MIF-CD74 signaling has shown a protective role in nonalcoholic fatty liver disease (33). All these observations at the RNA level suggest that some component of the adaptive immune responses might play a role in the clearance of HBV-infected hepatocytes, although it is not clear what it is and to what extent it is directly contributing or collateral to the induced cell death. This is also supported by the transient induction of cytokines that can be linked to adaptive immune responses (e.g., RANTES and Granzyme B). However, it also cannot be excluded that the detection of certain cytokines is a response toward the induction of apoptosis and the clearance of apoptotic cells.

The immunostimulatory effect of a CAM-A (BAY41-4109) had been evaluated in vitro (36). Direct treatment of HBV-infected primary human hepatocytes (PHH) did not stimulate the production of type I and type III interferon (IFN). However, treatment enhanced the antiviral interferon-stimulated gene (ISG) expression induced by IFN-α in HBV-infected PHHs. Our AAV-HBV mouse data showed a transient induction of inflammatory response in the liver which was associated with a transient ALT increase, HBsAg decline, and elimination of infected hepatocytes. However, there was no induction of type I IFN in the liver by CAM-A treatment. These data suggest that treatment with CAM-A alone is not sufficient to stimulate potent immune response to control HBV infection in AAV-HBV-transduced mice, and additional immunological studies are required to investigate the role of the adaptive immune responses in CAM-A-induced HBsAg clearance.

In summary, we demonstrated that CAM-A treatment eradicates some HBV-infected hepatocytes in preclinical animal models through selective induction of apoptotic processes and that this effect is dependent on de novo translation and high levels of core antigen. HBV core protein synthesis is dependent on the presence of cccDNA (13), and as a result, CAM-A treatment would only lead to the specific eradication of HBV-infected hepatocytes with transcriptionally active cccDNA, leaving hepatocytes with silenced cccDNA and/or integrated HBV DNA untouched. Meaningful HBsAg reductions have not been observed in clinical trials with CAM-A so far (24). However, increases in ALT activity in subjects with high HBcrAg levels treated with RO7049389 were reported (22). Since subjects were only treated for up to 4 weeks it cannot be excluded that HBsAg reductions would have occurred during longer treatment as observed in our chimeric mouse study.

MATERIALS AND METHODS

Compounds

HAP-1, HAP-2, and HAP-3 were synthesized at WuXi AppTech, Shanghai, People’s Republic of China, with a purity of >99%. GLS-4 was used as the positive high control for the induction of aberrant core structures and was synthesized in house. JNJ-61030827 (JNJ-0827) was used as a CAM-E-negative control for in vitro characterization and was synthesized in-house. JNJ-61758320 (JNJ-8320) was used as a CAM-E-negative control in the chimeric mice study and was synthesized at WuXi AppTech, Shanghai, People’s Republic of China. The GalNac-conjugated HBV siRNA was synthetized in a contract research organization (CRO). All compounds were prepared for research purposes only.

Antiviral assay with HepG2.117 cells and cytotoxicity assay

Details of the antiviral and cytotoxicity (HepG2) assays have been described previously (17). In brief, compounds at several concentrations were added to cultured HepG2.117 cells (kindly provided by Dr. M. Nassal, University Hospital Freiburg, Germany). On day 4, HBV DNA was extracted from the cells and quantified by a quantitative polymerase chain reaction (qPCR) assay (17). Percent inhibition [EC₅₀ and 90% effective concentration (EC₉₀) values) was calculated using the difference in threshold cycle number between treated and control wells. Cultured HepG2 cells were incubated for 4 days with several concentrations of HAP-1, HAP-2, and HAP-3. Cytotoxicity was evaluated by resazurin read-out, and 50% cytotoxic concentration (CC₅₀) and 90% cytotoxic concentration (CC₉₀) were calculated. Selectivity index was defined as the ratio of the mean CC₅₀ or CC₉₀ value of HAP-1, HAP-2, and HAP-3 in HepG2 cells to the mean EC₅₀ value of HAP-1, HAP-2, and HAP-3 assessed in HepG2.117 cells.

High-content imaging HBV core protein aggregation assay

Details of the high-content imaging core protein aggregation assay have been described previously (26). In brief, compounds were tested in dose range in presence of 1% dimethyl sulfoxide (DMSO) using a 25 µM or 100 µM start concentration and a serial dilution factor of ¼ in core expressing HepG2.117 cells. Compound plates with seeded cells were incubated for 72 hours at 37°C and 5% CO₂. After 72 hours, cells were fixated and permeabilized, and core protein was detected with primary monoclonal mouse core antibody (Abcam Ab8637). After overnight incubation at 4°C, primary antibody was detected with Alexa Fluor-488 goat anti-mouse secondary antibody. Nuclei and cytoplasm were co-stained using Hoechst 33258 (nuclei) and HCS CellMask Deep Red (entire cells) solution. Images were captured via high-content imaging using the Opera Phenix high-content screening system (Perkin Elmer), whereby image analysis via an Acapella script identified nuclei based on the Hoechst 33258 channel, cytoplasm segmentation was based on the HCS CellMask Deep Red channel, and the texture feature was normalized to “% effect” such that the average over low control wells (i.e., non-compound-treated core-expressing HepG2.117 cells) corresponds to 0% effect and the average over the high controls (i.e., 1 µM GLS-4-treated core-expressing HepG2.117 cells) corresponds to 100% effect. Non-linear log-logistic curve fitting was chosen for calculation of EC₅₀ values using GraphPad Prism or Phaedra software which is an open source platform for data capture and analysis of high-content screening data (37).

Core-dependent apoptosis induction study in HepG2.117 cells

After passaging, HepG2.117 cells were resuspended either in core-expressing medium, containing DMEM supplemented with 2% Tet-system-approved fetal calf serum (FCS; Clontech), 2 mM alanyl-glutamine, and 1× non-essential amino acids (NEAA; Sigma-Aldrich), or in core-suppressing medium, containing DMEM supplemented with 2% Tet-system-approved FCS, 2 mM alanyl-glutamine, 1× NEAA, and 100 ng/mL doxycycline. A total of 2,000 cells per well were seeded in compound-containing 384-well black poly-D-lysine PhenoPlate tissue culture-treated plates with an optically clear bottom (Perkin Elmer).

HAP-1 and HAP-2 were tested in dose range in the presence of 1.5% DMSO using 50 µM start concentration and a serial dilution factor of 1/3 in core-expressing HepG2.117 cells or core-suppressed HepG2.117 cells. Compound plates with seeded cells were incubated at 37°C and 5% CO₂ with media and compound refreshes performed every 3 to 4 days. Plates were incubated until day 28 with intermediate end points at days 14 and 21 for parallel assay plates. Caspase 3/7 activity was measured via the CellEvent Caspase-3/7 Green detection reagent (Thermo Fisher Scientific). The CellEvent Caspase-3/7 green detection reagent was added to the cells in a final dilution of 1/400 and incubated for 1 hour at 37°C without CO₂. After incubation, images were captured using the CellVoyager CV8000 (Yokogawa) whereby image analysis using a custom Acapella (Perkin Elmer) script calculated the green fluorescent signal of the CellEvent Caspase-3/7 green detection reagent. Cell viability was measured on parallel plates stained according to the high-content imaging HBV core protein aggregation assay using the cell count feature as a measure for cytotoxicity.

Studies in AAV-HBV-transduced and HBV-infected chimeric mice with humanized liver

Male C57BL/6 and C57BL/6 SCID mice (age 5 weeks upon arrival) were injected with AAV-HBV diluted in phosphate buffered saline (PBS) to 1.0 × 10¹¹ viral genome in 200-µL volume/mouse via tail veins. To study core protein dependency, mice were transduced with four different doses of AAV-HBV (5 × 10⁹, 1 × 10¹⁰, 2.5 × 10¹⁰, and 1 × 10¹¹ AAV-HBV vg equivalents). A vein blood sample was taken regularly (generally once per week) prior to treatment to evaluate HBV transduction efficiency by measurement of plasma HBV DNA, HBsAg, HBeAg, HBcrAg, and ALT levels. Mice were randomly allocated into groups according to the plasma HBV DNA, HBsAg, HBeAg level, and body weights to achieve balanced distribution across groups. The animal welfare and euthanasia were conducted according to the standard operating procedures approved by Institutional Animal Care and Use Committee.

Quantification of HBV DNA in plasma of AAV-HBV-transduced mice

HBV DNA in mouse plasma was isolated with the QIAamp 96 DNA Blood Kit according to

manufacturer’s instructions (Qiagen, 51162). The isolated HBV DNA was quantified by qPCR using FastStart Universal Probe Master (Roche, 04914058001). The standard curve DNA of 10⁷ copies/μL was prepared by 60-fold dilution of 5 ng/µL pAAV2-HBV1.3 plasmid DNA, followed by a 10-fold serial dilution in AE buffer to generate standards ranging from 10⁷ to 10 copies/μL. HBV DNA levels were interpolated according to the standard curve. The LLOQ of the HBV DNA assay is determined as 120 copies/μL plasma.

HBsAg and HBeAg measurement in plasma of AAV-HBV-transduced mice

For the detection of HBsAg, mouse plasma samples were first diluted 50-fold and then further diluted another 24-fold. Fifty microliters of the 1,200-fold diluted plasma were used for determination of HBsAg levels using the HBsAg ELISA kit (Autobio, CL 0310) following the supplier’s manual.

For the detection of HBeAg, mouse plasma samples were diluted 50-fold (4-µL plasma + 196-µL PBS). Fifty microliters of the 50-fold diluted plasma were used for determination of HBeAg using the HBeAg ELISA kit (Autobio, CL 0312) following the supplier’s manual.

Measurement of HBcrAg in plasma of AAV-HBV-transduced and chimeric mice

Quantitative levels of HBcrAg were determined using the Lumipulse G HBcrAg assay, which measures simultaneously denatured HBeAg, HBcAg, and the precore protein p22cr (aa −28 to aa 150). Samples were handled according to the manufacturer’s instructions. HBcrAg levels are quantified in units per milliliter.

ALT activity assay in plasma of AAV-HBV-transduced mice

ALT levels in mouse plasma were detected using the ALT Activity Assay Kit (Sigma, MAK052) following the supplier’s manual.

Liver immunohistochemistry with liver samples from AAV-HBV-transduced mice

Formalin-fixed, paraffin-embedded liver samples were cut and labeled for immunohistochemistry with a polyclonal rabbit anti-HBcAg antibody (Abcam, ab115992, dilution factor 1:250) or a polyclonal horse anti-HBsAg antibody (Abcam, ab9193, dilution factor 1:250). Isotype-matched antibodies were used as control primary antibodies, and HBV-negative liver tissue was used as control tissue. Positive staining was observed under a microscope, and representative photos were captured under 200-fold view.

Liver immunofluorescence with liver samples from AAV-HBV-transduced mice

Formalin-fixed, paraffin-embedded liver samples were sectioned at 5 µm, mounted on SuperFrost Plus glass slides (VWR, 631-9483), and immunofluorescently stained using a VENTANA Discovery Ultra (Roche) for HBsAg (Abcam, Ab859, clone 3E7, diluted 1:1)/Ki67 (Abcam, Ab15580, diluted 1:500)/DAPI. Alternatively, TUNEL assay (Invitrogen, C10617) was performed manually following the vendor’s guidelines. All immunofluorescence-stained sections were covered with mounting media (DAKO, S3023) and scanned on a NanoZoomer S60 (Hamamatsu). Isotype-matched antibodies were used as control for primary antibody, and HBV-negative liver tissue was used as control tissue and for biomarker positivity interpretation.

Liver immunofluorescence cell segmentation and quantification

Immunofluorescent images were subjected to cell segmentation using HALO image analyses software (Indica Labs) and CytoNuclear FL module to allow the identification of single cells and biomarker expression. Briefly, tissue boundaries and individual cells were identified by the presence of nuclear dye (DAPI) positivity. Afterward, starting from the localization of a nucleus, the software gradually expands the area to identify the cell boundary (cytoplasm), where the biomarker intensity is calculated. Based on the overall intensity distribution or sample controls, cell positivity for HBsAg, TUNEL, and Ki67 was determined to assign cellular phenotypes. Fraction of biomarker phenotypes was calculated over the total number of cells identified (number of DAPI positive nuclei in the liver section).

Cytokine detection

Mouse plasma samples were used unprocessed, and mouse liver samples were homogenized by bead beating on a TissueLyser twice during 3 min at 25 Hz/s after addition of a 5-mm stainless steel bead (Qiagen, 69989) and 450-µL homogenate buffer [0.1% Triton X-100 and 2.5 mM EDTA in DPBS, supplemented with protease inhibitor (Roche, 05056489-001)]. Homogenized samples were centrifuged for 15 min at 12,000 × g after which supernatant is collected and normalized based on total protein content.

Detection of cytokine and chemokine biomarkers was done with Luminex xMAP bead-based assay platform, in combination with Milliplex kits (Merck-Millipore, MCD8MAG48K-PX15 and MCYTOMAG-70K) following the manufacturer’s manual. In brief, samples are brought in contact with magnetic beads which are each coated with a specific capture antibody. After an analyte is captured by the beads, a biotinylated detection antibody is introduced, and the reaction mixture is then incubated with Streptavidin-PE conjugate, the reporter molecule, to complete the reaction on the surface of each bead. Each individual bead is identified by its “bead signature” and quantified based on fluorescent reporter signals using a Biorad Bio-Plex 200 system. Standard curves were generated with standards supplied with the kits; a best-fit line was determined by regression analysis using 4PL logistic curve-fit.

Microarray

Mouse liver (10 mg) was homogenized by bead beating on a TissueLyser twice during 60 s at 20 Hz/s after addition of two 3.5-mm stainless steel beads (Qiagen, 69989) and 450-µL RLT Plus lysis buffer (Qiagen, 1053393). After homogenization, a total of 400-µL lysate was used for RNA extraction on a QIASymphony liquid handling system using the QIAsymphony RNA extraction kit (Qiagen, RNA_CT400), according to the manufacturer’s instructions.

Transcriptome analysis was done using the Mouse Clariom S Assay (Thermo Fisher Scientific, 902930), in combination with the GeneChip WT PLUS Reagent kit (Thermo Fisher Scientific, 902280). A total of 100-ng RNA was used as input, and downstream process was done according to the manufacturer’s guidelines. In brief, double-stranded cDNA is first synthesized using random primers and then used as template for in vitro transcription. The cRNA is used as input for a second round of first-strand cDNA synthesis, producing single-stranded sense cDNA. After fragmentation and end-labeling, the targets are hybridized to a multi-sample Clariom S plate array on a GeneTitan Multi-Channel Instrument (Thermo Fisher Scientific).

Gene expression values were normalized by robust multiarray average normalization on the microarray probe-level data, and downstream analysis was done in R version 3.4.2.

Unsupervised analysis using spectral maps was performed to remove technical outliers. A supervised analysis was performed using the limma package (38) for comparisons between treatment groups across each time point, and corrected P-values for multiple testing across genes ≤ 0.05 were considered significant. The affected pathways were analyzed with mean log P (MLP) analysis with GO Biological Process and Ingenuity Pathway Analysis (Qiagen). The considered cutoffs for MLP were lower (5) and upper (250) threshold for gene set size where 7,100 pathways from the Biological Process and 667 pathways from IPA were used.

In vivo depletion of CD8+ T cells

Male AAV-HBV mice were intraperitoneally injected with anti-CD8β (BioXcell, 53.5.8) or IgG1 (BioXcell, HRPN) antibodies at 100 µg/mouse on day −4 and day −3 before compound treatment, followed with QW administration until day 39. Antibody injections were strengthened to Q2W (biweekly) from days 39 to 56 to maintain efficient elimination of CD8+ T cells. Concurrently, AAV-HBV mice were orally administered with CAM-A (HAP-2, 20 mg/kg, QD) compound from day 0 until day 56. CD8+ T cell frequency in CD3+ T cells were monitored in blood at weekly basis by flow cytometry using anti-mouse CD8α antibody (BD, 53.6.7).

Studies in HBV-infected chimeric mice with humanized liver

Male HBV-genotype-C-infected uPA/SCID chimeric mice (PXB-Mouse; 21–25 weeks of age; weight, 17.8–23.0 g) were from PhoenixBio (Higashi-Hiroshima, Japan). All mice had blood human albumin levels above 7.8 mg/mL and serum HBV DNA levels above 10⁶ copies/mL. Median (range) HBV DNA levels across all randomized mice were 8.0 (6.6–8.8) log₁₀ copies/mL. Mice received food and water ad libitum.

The use of the animals for this study was approved by the Animal Ethics Committee of PhoenixBio (resolution no. 1495). All the experimental procedures used to treat live animals in this study were approved by the Animal Ethics Committee of PhoenixBio.

Determination of blood human albumin

At all blood samplings, human albumin levels were determined using the clinical chemistry analyzer (BioMajesty Series JCA-BM6050, JEOL) on saline-diluted blood and latex agglutination immunonephelometry (LZ Test “Eiken” U-ALB, Eiken Chemical).

Serum HBV DNA quantification

HBV DNA was extracted from serum using the SMITEST EX-R&D Nucleic Acid Extraction Kit (Medical & Biological Laboratories) and dissolved in nuclease-free water, after which real-time polymerase chain reaction was performed using the TaqMan Fast Advanced Master Mix (Applied Biosystems, Thermo Fisher Scientific) and ABI Prism 7500 sequence detector system (Applied Biosystems).

Serum HBsAg and HBeAg quantification

Serum HBsAg and HBeAg concentrations were determined by SRL (Tokyo, Japan) based on the ChemiLuinescence ImmunoAssay (CLIA) developed by Abbott (Abbott Park, IL; ARCHITECT SYSTEM).

Immunofluorescence stain for the detection of core and HBsAg in cryosections of mouse liver

Fresh frozen liver samples were fixated for 10 min in PFA 4%. After a PBS wash, sections of 10 µm were cut, mounted on SuperFrost Plus glass slides (VWR, 631-9483), and immunofluorescently stained using a VENTANA Discovery Ultra (Roche) for HBsAg (Abcam, Ab859, clone 3E7, diluted 1:1)/DAPI or Core (DAKO, B0586, diluted 1:2,000)/DAPI. All immunofluorescence-stained sections were covered with mounting media (DAKO, S3023) and scanned on a NanoZoomer S60 (Hamamatsu). Isotype-matched antibodies were used as control for primary antibody, and HBV-negative liver tissue was used as control tissue and for biomarker positivity interpretation.

DATA AVAILABILITY

The processed microarray and raw data files have been deposited in the Gene Expression Omnibus (GEO) repository under the accession number GSE246563.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.01502-23.

Text S1. jvi.01502-23-s0001.docx.

Abbreviation list.

Figure S1. jvi.01502-23-s0002.tif.

Treatment with CAM-A induces apoptosis and proliferation in the liver of AAV-HBV transduced mice.

Figure S2. jvi.01502-23-s0003.tif.

Changes in interferon-stimulated and interferon-related genes after HAP-1 treatment.

Figure S3. jvi.01502-23-s0004.tif.

HBV DNA and HBeAg levels in mice treated with different doses of AAV-HBV.

Figure S4. jvi.01502-23-s0005.tif.

Body weight of HBV-infected chimeric mice treated with HAP-1 or placebo.

Figure S5. jvi.01502-23-s0006.tif.

HBV DNA, HbsAg, and hALT-1 levels in serum after treatment with HAP-1 or JNJ-8320 in HBV-infected chimeric humanized liver mice.

Supplemental legends. jvi.01502-23-s0007.docx.

Legends of the supplementary figures.

Table S1. jvi.01502-23-s0008.txt.

Significant genes from the differential expression analysis.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.