Targeting the multifunctional HBV core protein as a potential cure for chronic hepatitis B

Abstract

The core (capsid) protein of hepatitis B virus (HBV) is the building block of nucleocapsids where viral DNA reverse transcriptional replication takes place and mediates virus-host cell interaction important for the persistence of HBV infection. The pleiotropic role of core protein (Cp) in HBV replication makes it an attractive target for antiviral therapies of chronic hepatitis B, a disease that affects more than 257 million people worldwide without a cure. Recent clinical studies indicate that core protein allosteric modulators (CpAMs) have a great promise as a key component of hepatitis B curative therapies. Particularly, it has been demonstrated that modulation of Cp dimer-dimer interactions by several chemical series of CpAMs not only inhibit nucleocapsid assembly and viral DNA replication, but also induce the disassembly of double-stranded DNA-containing nucleocapsids to prevent the synthesis of cccDNA. Moreover, the different chemotypes of CpAMs modulate Cp assembly by interaction with distinct amino acid residues at the HAP pocket between Cp dimer-dimer interfaces, which results in the assembly of Cp dimers into either non-capsid Cp polymers (type I CpAMs) or empty capsids with distinct physical property (type II CpAMs). The different CpAMs also differentially modulate Cp metabolism and subcellular distribution, which may impact cccDNA metabolism and host antiviral immune responses, the critical factors for the cure of chronic HBV infection. This review article highlights the recent research progress on the structure and function of core protein in HBV replication cycle, the mode of action of CpAMs, as well as the current status and perspectives on the discovery and development of core protein-targeting antivirals. This article forms part of a symposium in Antiviral Research on “Wide-ranging immune and direct-acting antiviral approaches to curing HBV and HDV infections.”

1. Introduction

Hepatitis B virus (HBV) chronically infects 257 million people worldwide and causes approximately 880 thousand deaths annually due to cirrhosis, hepatocellular carcinoma (HCC) and liver failure (Polaris Observatory, 2018). The current standard of care for chronic HBV infection, that includes six nucleos(t)ide analog viral DNA polymerase inhibitors (NUCs) and pegylated interferon-alpha (pgIFN-α), can significantly reduce viral load and prevent liver disease progression, but rarely induce the loss of HBV surface antigen (HBsAg) or seroconvertion to anti-HBs, the indication of durable immune control of HBV infection or a functional cure (Block et al., 2013; Dienstag, 2009; Lok et al., 2017; Perrillo, 2009). Once started, termination of the NUC-based antiviral therapy usually results in the rebound of viral replication and flares of liver diseases (Alter et al., 2018). The failure to achieve a functional cure by the NUC therapy is attributed to the lack of direct effects on viral covalently closed circular (ccc) DNA in infected hepatocytes and inability to restore a functional antiviral immune response (Chang et al., 2014). Therefore, novel antiviral agents that can induce the loss or inactivation of cccDNA and activate a functional antiviral immune response with a finite therapeutic duration are urgently needed (Liang et al., 2015). Recently, HBV core protein (Cp), the building block of nucleocapsids where viral DNA replication takes place, has emerged as a promising target for the development of novel antiviral therapeutics (Tang et al., 2017; Zlotnick et al., 2015). Several core protein allosteric modulators (CpAMs) are currently under clinical development and have demonstrated a great promise as a key component of curative therapeutics for chronic hepatitis B (Cole, 2016; Yang et al., 2019b). This review highlights the recent research progress in our understanding of the structure and function of core protein in the HBV replication cycle, the mode of CpAM action, the current status and our perspectives on the discovery and development of core protein-targeting antivirals.

2. Structure and function of core protein in HBV life cycle

2.1. Cp is the building block of capsids and plays an essential role during multiple stages of viral replication in the cytoplasm

HBV is a para-retrovirus with a 3.2 kb, partially double-stranded, relaxed circular (rc) DNA genome that encode seven viral proteins, including the small (S), middle (M) and large (L) envelope proteins, core (capsid) protein (Cp), DNA polymerase (Pol), pre-core protein that is proteolytically processed and secreted as soluble e antigen (HBeAg), and X protein (HBx) (Hu et al., 2019). HBV replicates its DNA genome via the reverse transcription of a RNA pregenome in the cytoplasmic nucleocapsids of infected hepatocytes (Mason et al., 1982; Summers and Mason, 1982). As the building block of HBV nucleocapsid, Cp plays distinct roles at multiple steps of the HBV replication cycle. As illustrated in Fig. 1, by binding to its specific receptor, sodium taurocholate cotransporting polypeptide (NTCP), on the plasma membrane, the HBV virion enters hepatocyte by endocytosis. The nucleocapsid is delivered into the cytoplasm after the fusion of the viral envelope with the membrane of the endocytic vesicle (Yan et al., 2012). The Cp on the shell of nucleocapsid binds importin-β and directs the transport of the nucleocapsid to the nuclear pore complex (NPC), where the nucleocapsid disassembles and releases the rcDNA genome into the nucleus for the synthesis of cccDNA (Guo et al., 2010; Guo and Guo, 2015; Schmitz et al., 2010). The cccDNA serves as a template for the transcription of four major species of viral RNA. The 3.5 kb pre-genomic (pg) RNA translates Cp and Pol. Binding of Pol to the stem-loop structure (ε) at the 5’ terminal region of pgRNA initiates the packaging of the Pol-pgRNA complex by 120 copies of Cp dimers to form the nucleocapsid (Bartenschlager and Schaller, 1992; Tavis et al., 1994; Wang and Seeger, 1993). Inside the nucleocapsid, Pol reverse transcribes the pgRNA first to single-stranded (ss) DNA and then to rcDNA (Mason et al., 1982; Summers and Mason, 1982). The deficiency of second-strand DNA synthesis in the nucleocapsids assembled with the W124A mutant or a C-terminally truncated Cp strongly indicates that Cp does not builds an inert container to provide space for viral DNA synthesis, but actively participates in viral DNA synthesis, probably via interaction with Pol, pgRNA/DNA or encapsidated host cellular factors (Nassal, 1992; Tan et al., 2015). Through interaction with envelope proteins studded on the membrane of multi-vesicle bodies (MVBs), Cp directs the envelopment of the rcDNA-containing “mature” nucleocapsids for the assembly and secretion of virion particles (Hu and Liu, 2017). Alternatively, Cp can also mediate the transport of mature nucleocapsids to the nuclear pore complexes (NPCs) for the import of rcDNA into the nuclei to amplify the cccDNA pool (Guo et al., 2007; Guo et al., 2010; Tuttleman et al., 1986). Due to the strong tendency of Cp self-assembly into empty capsids, the nucleocapsids only constitute less than 10% of total capsids in an HBV infected hepatocyte (Ning et al., 2011). Similar to the mature nucleocapsids, the less matured nucleocapsids-containing reverse transcriptional intermediates as well as the empty capsids can also be enveloped and secreted as RNA-containing virions or “empty” virions, respectively. While the RNA-containing virions are 100-fold less abundant than the DNA-containing virions, the empty virions are approximately 100-fold more abundant than the DNA-containing virions in the blood of HBV carriers (Hu and Liu, 2017). It is thus apparent that by assembly of empty capsids and nucleocapsids, Cp involves in almost every steps of viral replication in the cytoplasm, which will be explained in detail below.

Fig. 1. HBV life cycle and steps of HBV replication disrupted by the standard of care medicines (NUCs and IFN-α) and CpAMs.

See text for detailed description on HBV life cycle. NUCs inhibit viral DNA synthesis in the cytoplasmic nucleocapsids. IFN-α inhibits cccDNA transcription and pgRNA encapsidation, in addition to its immunomodulatory activity. Hyperphosphorylated and hypophosphorylated core proteins are illustrated in purple and green, respectively. The positive strand and negative strand of HBV DNA are illustrated in black and red, respectively. DP-rcDNA, deproteinized relaxed circular DNA. MVB, multi-vesicle body.

2.2. Core protein and capsid structures

HBV Cp is a 183-amino acid (aa) polypeptide containing a N-terminal assembly domain (NTD, aa 1–140) and an arginine-rich C-terminal domain (CTD, aa150–183), that are connected by a linker of 9 amino acid residues (Venkatakrishnan and Zlotnick, 2016; Zlotnick et al., 2015) (Fig. 2A). As illustrated in Fig. 2C, the assembly domain has five α helices connected by loops, i.e., the short helix α1 (aa 13–17), longer helices α2 (aa 27–43) and α5 (aa 112–127)), helix bundles α3 (upward) and α4 (downward), an irregular proline-rich loop (aa 128–136) and an extended strand (aa 137–149). The hydrophobic interaction between α3 and α4 helices of two Cp molecules drives the formation of a four-helix bundle at their interface and results in the dimerization of Cp. The dimer is the basic structural and functional unit of Cp.

Fig. 2. Structural features of HBV core protein and HBeAg.

(A) Domain structure of Cp and structural motifs and phosphorylation sites at C-terminal domain are highlighted. (B) Pre-core protein processing and biogenesis of HBeAg and their structure domains are illustrated. (C and D) the secondary structure and dimers of Cp from pdb code 1QGT and HBeAg from pdb code 6CVK are presented with all the five alpha helices color coded to see how the structures differ even though they have the same secondary structures

The structure of HBV capsids assembled in E. coli expressing C-terminally truncated or full-length Cp or HBV infected hepatocytes had been determined by cryo-electron microscopy (Cryo-EM) and X-ray diffraction technologies (Bottcher and Nassal, 2018; Bourne et al., 2006; Wynne et al., 1999; Yu et al., 2013) (Table 1). The predominant capsid species is assembled from 120 copies of Cp dimers arranged in T=4 icosahedral symmetry (Fig. 3). The capsid is approximately 340 Å in diameter with an inner radius of 130 Å and thickness of 20 Å (pdb codes 1QGT, 2G33, 3J2V and 6HTX). The spikes with 25 Å in length and 20 Å in width located on the surface of the capsid are the four-helix bundles of Cp dimers. Cys61 located in the middle of the four-helix bundle can form a disulfide bond to crosslink the Cp dimer. However, such a disulfide bond is not required for capsid assembly (Bourne et al., 2006). In fact, the assembly of capsids is driven by the hydrophobic interaction between the dimer-dimer interfaces. Although each Cp, or subunit, has the same aa sequence, the 240 Cp molecules of a capsid take one of four quasi-equivalent conformations, A, B, C or D (Fig. 3).AB and CD dimers have a nearly two-fold symmetry. Five A subunits surround the five-fold axes and AB dimers extend from the five-fold to quasi-six-fold vertices. CD dimers connect the quasi-six-fold vertices and each quasi-six-fold vertex is surrounded by B–C–D–B–C–D subunits. There are pores around the five-fold, quasi-six-fold, three-fold and local three-fold axes of the capsid (Fig 3). The five-fold and quasi-six-fold axes primarily involve the residues from the irregular proline rich loop (aa 128–136) and the extended strand (aa 137–149). Especially the N137 forms a ring with P134 and P135 providing the structural clamp to hold the ring in place. The three-fold and local three-fold channels involve short helix α1 (aa 13–17) and longer helix α2 (aa 27–43), which are primarily polar residues (D2, K7, R39, E40, E43 and E46).

Table 1.

Structures of HBV capsids and Y132A Cp hexamer with or without CpAMs

No.	PDB ID	Res. (Å)	Structure	Ligand	Method	Year
1	1QGT	3.3	WT (T=4)	NA	X-ray Diffraction	1999
2	2G33	4.0	WT (T=4)	NA	X-ray Diffraction	2006
3	2QIJ	8.9	WT (T=4)	NA	Electron Microscopy	2007
4	3J2V	3.5	WT (T=4) Cp183	NA	Electron Microscopy	2013
5	6HTX	2.7	WT (T=4) Cp183	NA	Electron Microscopy	2018
6	2G34	5.1	WT (T=4)	HAP1	X-ray Diffraction	2006
7	4G93	4.2	WT (T=4)	AT130	X-ray Diffraction	2013
8	5D7Y	3.9	WT (T=4)	HAP18	X-ray Diffraction	2016
9	6BVN	4.0	WT (T=3)	HAP-TAMRA	Electron Microscopy	2018
10	6BVF	4.0	WT (T=4)	HAP-TAMRA	Electron Microscopy	2018
11	6CWT	3.2	WT (T=4)	Fab e21	X-ray Diffraction	2018
12	6CWD	3.3	WT (T=4)	scFv e13	X-ray Diffraction	2018
13	3KXS	2.3	Y132A	NA	X-ray Diffraction	2010
14	4BMG	3.0	Y132A	NA	X-ray Diffraction	2013
15	5WTW	2.6	Y132A	NA	X-ray Diffraction	2017
16	5E0I	2.0	Y132A	NVR10-001E2	X-ray Diffraction	2016
17	5GMZ	1.7	Y132A	4-CH3-HAP	X-ray Diffraction	2016
18	5WRE	1.9	Y132A	HAP_R01	X-ray Diffraction	2017
19	5T2P	1.7	Y132A	SBA_R01	X-ray Diffraction	2017
20	6J10	2.3	Y132A	Ciclopirox	X-ray Diffraction	2019
21	6HU4	2.6	F97L (T=4) Cp183	NA	Electron Microscopy	2018
22	6HU7	2.8	F96L-phos (T=4) Cp183	NA	Electron Microscopy	2018

Fig. 3. Structure of HBV capsids.

Global icosahedral structure of T=4 HBV capsid from pdb code 1QGTwith highlighted sections showing five-fold, three-fold, quasi-six-fold and local three-fold axis. The A, B, C and D monomers are colored green, yellow, red and blue respectively.

2.3. Physical principles of capsid assembly learned from in vitro assembly system

HBV capsids can be assembled in vitro with the CTD-deleted Cp, Cp149. The in vitro assembly reaction occurs as a function of Cp149 dimer concentration, temperature and ionic strength, and results in capsids that are morphologically indistinguishable from capsids purified from HBV expressing cells (Wingfield et al., 1995; Zlotnick et al., 1997). Study of the in vitro Cp assembly reaction kinetics and characterization of the assembly intermediates and products by light scattering, size exclusion chromatography (SEC), charge detection mass spectrometry (CDMS), small angle X-ray scattering (SAXS) technologies has revealed very interesting physical principles of HBV capsid assembly (Perlmutter and Hagan, 2015).

Kinetically, the assembly of capsids in the in vitro condition is sigmoidal. The initial step is the nucleation of Cp dimers, i.e., the formation of the trimers of Cp dimers, which is slow or the lag phase (Zlotnick, 2005). The assembly incompetent Cp with a Y132A substitution gives a glimpse of intermediate structures - namely the kinetically trapped trimers of dimers namely AB, CD and EF for the capsid formation (Alexander et al., 2013; Packianathan et al., 2010; Zhou et al., 2017). Cp149_Y132A crystallizes in three forms – open triangle (3KXS), closed triangle (4BMG) and spike-contact packing (5WTW). The assembly becomes fast once more nuclei form, which is followed by an asymptotic phase leading to equilibrium as the free Cp dimer depletes (Zlotnick, 2005). Interestingly, analyses of assembled capsids by CDMS revealed that the in vitro assembly reaction initially yields defective and overgrown particles, which slowly anneal and relax back to form icosahedrons (Lutomski et al., 2017). Moreover, a recent time-resolved small angle X-ray scattering analysis of full-length Cp assembly indicates that HBV capsid is assembled through the formation of two transient, on-path intermediates, a trimer of dimers and a partially closed shell consisting of around 40 dimers (Oliver et al., 2020).

Based on the kinetically-limiting model established by Zlotnick and colleagues, thermodynamic parameters such as nucleus size, nucleation rate, ΔG_contact and elongation rate could be extracted. ΔG_contact, the association energy of dimers, is found to be weak (~3.5 kcal/mol), but it is enough to form stable capsids since each subunit must interact with at least three other subunits (~9–12 kcal/mol as association energies are additive) (Ceres and Zlotnick, 2002). The weak association energy allows thermodynamic editing which helps remove incorrect and weakly associated dimers. If ΔG_contact is too high (or, equivalently, temperature is too low), the system becomes kinetically trapped in intermediates which cannot rearrange anymore due to the strong binding. If ΔG_contact is too low (or, equivalently, temperature is too high), the target structure is not sufficiently stable. For example, a V124W substitution in Cp increases buried hydrophobic surface and in turn increases the rate of capsid assembly while substitutions such as Y132A and W102A decreases buried surface and prevents capsid assembly (Packianathan et al., 2010).

2.4. Capsid assembly in hepatocytes and its regulation by viral and host cellular factors

Assembly of capsids in hepatocytes should follow the general physical principles as observed in the in vitro assembly reaction. However, due to the complex intracellular environment, differences do exist. For instance, capsids can be assembled in hepatocytes expressing Cp142, but not Cp149, two forms of Cp with C-terminal truncation after residues 142 and 149, respectively, suggesting that cellular proteins may interact with the linker region to regulate capsid assembly (Ludgate et al., 2016; Wu et al., 2017). As shown in Fig. 2A, the CTD contains four arginine (R)-rich clusters with nucleic acid binding activity (Birnbaum and Nassal, 1990; Su et al., 2016; Yang et al., 1994) and seven serine (S)/ threonine (T) residues that can be phosphorylated or dephosphorylated during the viral replication cycle (Basagoudanavar et al., 2007; Ning et al., 2017; Perlman et al., 2005; Pugh et al., 1989). Expression of the wild-type Cp or Cp with substitutions at the seven phosphor-acceptor residues in the CTD with aspartic acid to mimic hyperphosphorylated Cp, results in assembled capsids without detectable packaged RNA. In contrast, the expression of Cp with substitution of the seven phosphor-acceptor residues with alanine (A) to mimic the dephosphorylated Cp results in assembled capsids that non-specifically package viral or cellular RNA (Heger-Stevic et al., 2018; Hu and Liu, 2017; Ludgate et al., 2016; Ning et al., 2011; Su et al., 2016). This is because the arginine-rich clusters in the CTD can non-specifically bind RNA through charge interaction. The RNA binding activity of the CTD can be neutralized by the introduction of negatively charged moieties through phosphorylation (Chua et al., 2010).

The selective packaging of Pol-pgRNA complex into a nucleocapsid is a key step of HBV replication and subjected to regulation by many viral and host cellular factors (Hu et al., 2019). Despite its lack of a role in the assembly of empty capsids, the CTD is absolutely required for pgRNA packaging. Sequential C-terminal truncation analyses indicate that deletion of the CTD beyond aa 164 impairs pgRNA encapsidation, whereas residues between aa 164 to 173 are required for primer translocation and positive strand DNA synthesis (Nassal, 1992). The function of the four arginine clusters in the CTD was analyzed by substitution of the individual arginine clusters with either alanine or lysine residues (Lewellyn and Loeb, 2011a). As expected, all the clusters were not required for capsid assembly. However, the alanine (but not lysine) substitution of cluster I to III, but not IV, significantly reduced the amount of encapsidated pgRNA. Moreover, it appears that all four clusters involved in viral DNA synthesis, which is, occasionally, not charge-dependent (Lewellyn and Loeb, 2011a). These results imply that by interaction with pgRNA and viral DNA, the arginine-rich CTD play an essential role in multiple steps of viral genome replication (Su et al., 2016).

HBV core proteins were discovered as phosphoproteins in the late 1980s by using metabolic labeling and phosphatase treatment/immunoblotting assays (Roossinck and Siddiqui, 1987; Yeh and Ou, 1991). Phosphomimic mutagenesis studies indicated that CTD phosphorylation, particularly at S162 and S170, was required for pgRNA packaging (Gazina et al., 2000; Lan et al., 1999; Lewellyn and Loeb, 2011b; Melegari et al., 2005). While several cellular kinases, such as SRPK1, SRPK2 (Daub et al., 2002; Zheng et al., 2005), PKA (Okabe et al., 2006), GAPDH-PK(Duclos-Vallee et al., 1998), PKC (Kann and Gerlich, 1994; Wittkop et al., 2010), CDK2 (Ludgate et al., 2012) and PLK1 (Diab et al., 2017), have been reported to phosphorylate one or multiple phospho-acceptor sites of CTD in vitro, their biological relevance in HBV replication in hepatocytes has not been established. However, our recent studies show that Cp is hyperphosphorylated in free dimers and empty capsids, but hypophosphorylated in pgRNA and DNA-containing nucleocapsids, suggesting that Cp dephosphorylation occurs during pgRNA encapsidation (Zhao et al., 2018). Indeed, further investigation demonstrated that cellular protein phosphatase 1 (PP1) catalyzes Cp dephosphorylation and is required for pgRNA packaging. Interestingly, both PP1 catalytic subunits α and β are co-packaged with Pol-pgRNA complex into nucleocapsids, but not empty capsids (Hu et al., 2020). We believe that the hyperphosphorylation of Cp dimers prevents its non-specific binding of cellular RNA and ensures the specific packaging of pgRNA through recognition of the Pol-pgRNA complex, most likely via Cp-Pol interaction. The sequential dephosphorylation of Cp during nucleocapsid assembly allows the CTD to interact with pgRNA, which may facilitate pgRNA packaging by organizing its folding to ensure its complete packaging. This hypothesis is supported by several independent studies showing that C-terminal truncation or mutation of the last two arginine-rich clusters resulted in failed encapsidation of the 3’ terminal region of pgRNA (Lewellyn and Loeb, 2011a). In addition, the involvement of the CTD in viral DNA synthesis, including template switching and primer translocation as well as DNA chain elongation, suggests that the CTD may be involved in interactions with viral and cellular factors in the viral replication environment (Lewellyn and Loeb, 2011a; Nassal, 1992). One possibility is that the CTD may also work as a nucleic acid chaperone to catalyze the breaking and reforming of base pair interactions. This hypothesis is supported by the fact that the CTD can catalyze nucleic acid strand exchange and hammerhead ribozyme activity in vitro (Chu et al., 2014). The other host cellular factors required for or have a role in regulating nucleocapsid assembly had been extensively reviewed elsewhere (Hu et al., 2019).

2.5. The morphogenesis of complete virions and empty virions depends on distinct structure features of capsids

As mentioned above, both the partially double stranded DNA-containing nucleocapsids and empty capsids can be efficiently enveloped and secreted as complete and empty virion particles, respectively (Blondot et al., 2016; Hu and Liu, 2017). Mutagenesis analyses identified amino acid residues of Cp on the shell of nucleocapsids that are essential for the envelopment of DNA-containing nucleocapsids (Koschel et al., 2000; Orabi et al., 2015; Pairan and Bruss, 2009; Ponsel and Bruss, 2003). While the tip of the spike of capsid appeared not to be directly involved in the morphogenesis of virions, alanine substitution of amino acid residues L60, L95, K96 or I126 abolished virion secretion (Ponsel and Bruss, 2003). These amino acid residues are located on the outer surface of capsid in the depressions between the spikes and possibly bind to envelope proteins and initiate budding of virion particles. Interestingly, F97L mutant Cp causes premature secretion of nucleocapsids containing less mature forms of viral DNA (Yuan et al., 1999), which is most likely due to the amino acid substitution leading to the increased size of an adjacent hydrophobic pocket in the center of the spike (Ceres et al., 2004). In marked contrast, Jianming Hu and colleagues recently demonstrated that the Cp residues essential for the production of complete virions as well as the entire CTD are not required for the envelopment of empty capsids (Ning et al., 2018). However, substitution or deletion mutations of the Cp linker region examined thus far completely blocks the production of empty virions, suggesting a critical role of the linker region in empty virion morphogenesis (Liu et al., 2018).

2.6. Nucleocapsid disassembly and its regulation

In addition to supporting viral DNA replication, another important function of HBV nucleocapsids is to serves as a vehicle to transmit viral genome to susceptible cells of the same or other individuals. On one hand, this latter function requires that the nucleocapsids be stable enough to protect the viral genome from degradation by environmental and host factors, while on the other hand, it also requires that the nucleocapsids disassemble in the proper subcellular compartment of infected cells to release the viral genome for replication. This is particularly a challenge for HBV and related viruses, because progeny mature nucleocapsids can recycle the newly replicated viral DNA back to the nucleus for amplification of the cccDNA pool (Guo and Guo, 2015). Because the uncontrolled cccDNA amplification loop results in hepatocyte death (Summers et al., 1990), the disassembly of mature progeny nucleocapsids must be strictly controlled by viral and host cellular factors (Cui et al., 2015a; Sheraz et al., 2019; Tang et al., 2019). In addition, timely- and spatially-controlled disassembly of nucleocapsids from infecting virions during de novo infection of hepatocytes is also critical for the successful establishment of cccDNA (Guo et al., 2017). Thus far, it is not known where the incoming virion-derived nucleocapsid disassembly occurs. However, it had been identified that several factors regulate the disassembly of intracellular progeny mature nucleocapsids and subsequently, cccDNA synthesis (Chen et al., 2016; Gallucci and Kann, 2017; Guo et al., 2010; Osseman et al., 2017).

The nucleocapsid disassembly is a multistep event that can be revealed as distinct species with different electrophoresis mobility in a native agarose gel electrophoresis (Cui et al., 2013). The physical force driving the nucleocapsid disassembly is most likely the elongation of double-stranded DNA, which is structurally rigid and pushes the nucleocapsid from inside (Dhason et al., 2012; Guo et al., 2007). In addition, interaction with NPC components, such as Nup153, is also important for the uncoating of viral rcDNA (Schmitz et al., 2010). Interestingly, Cp sequence motifs regulate nucleocapsid disassembly are identical or overlapped with that regulating virion particle envelopment (Cui et al., 2015b). This finding implies that the nucleocapsid disassembly (or genome uncoating) and envelopment are mechanistically two mutually excluding events. In addition, several mutagenesis studies indicated that phosphorylation of Cp at the CTD and assembly domain also modulates capsid stability and possibly, nucleocapsid disassembly (Barrasa et al., 2001; Selzer et al., 2015). Particularly, a phosphomimetic mutagenesis analysis of two putative phosphorylation sites in Cp assembly domain, S44 and S49, demonstrated that phosphorylation of the two residues suppressed pgRNA packaging, promoted nucleocapsid disassembly and subsequent cccDNA synthesis. On the contrary, dephosphorylation of the two residues did not affects pgRNA packaging and reverse transcriptional DNA synthesis, but reduced cccDNA formation (Luo et al., 2020). Because residues S44 and S49 locate at the internal surface of capsid, it is possible that phosphorylation of these two residues by encapsidated cellular kinases, such as CDK2, triggers mature nucleocapsid disassembly and cccDNA synthesis. In support of this notion, it was shown that treatment of cells with CDK2 inhibitors reduced cccDNA synthesis (Luo et al., 2020).

3. Core protein: Possible roles in cccDNA function and virus-host cell interaction

In addition to the capsid-related functions, Cp may also regulate cccDNA function as a component of cccDNA minichromosomes and mediate virus-host interaction by interacting with a variety of host cellular proteins. For instance, it was reported that Cp associates with cccDNA and regulates cccDNA transcription (Bock et al., 2001; Chong et al., 2017; Guo et al., 2011). Moreover, it was also demonstrated that Cp recruits IFN-induced cellular proteins APOBEC 3A or 3B to cccDNA minichromosomes for induction of cccDNA cytidine deamination and its subsequent degradation (Lucifora et al., 2014; Xia et al., 2014; Xia et al., 2016). Interestingly, although infection of primary duck hepatocytes with wild-type duck hepatitis B virus (DHBV) or a mutant DHBV deficient in expression of Cp results in the similar amounts of de novo cccDNA synthesis, the cccDNA in mutant DHBV infected cells transcribed significantly less amounts of viral RNA, suggesting an important role of Cp in DHBV cccDNA transcription regulation and/or RNA stability (Schultz et al., 1999). In marked contrast, HepG2-NTCP cells infected with wild-type and Cp-deficient HBV produced similar amounts of cccDNA, viral RNA and HBsAg during a three-week observation period (Qi et al., 2016). These contradictive findings on the role of Cp in DHBV and HBV cccDNA transcription regulation may reflect the distinct structure and unique biological function of the avian and mammalian hepadnaviral Cp in viral replication. Particularly, DHBV Cp is approximately 80 amino acid residues longer than HBV Cp. Because DHBV lacks X protein, it had been speculated that DHBV Cp may have a HBx-like function, which is, at least in part, involved in regulation of cccDNA transcription (Decorsiere et al., 2016; Lucifora et al., 2011).

Interestingly, it was reported recently that the phosphorylation status of residues S44 and S49 at Cp assembly domain affects the production and conversion of minus-strand covalently closed rcDNA, an intermediate of cccDNA synthesis (Luo et al., 2017; Tang et al., 2019; Zhao and Guo, 2020), to cccDNA, strongly suggesting a possible role of Cp in cccDNA synthesis in the nucleus (Luo et al., 2020). In addition, a recent proteomic analysis identified 60 Cp-interacting proteins in the nuclei of human hepatocytes. The majority of Cp-interacting host factors are RNA binding proteins that are involved in various aspects of mRNA metabolism. Particularly, SRSF10, a RNA binding protein that regulates RNA splicing, appears to modulate HBV RNA abundance, suggesting that Cp may regulate viral and cellular RNA metabolism via interaction with SRSF10 and other host cellular factors (Helene et al., 2020).

Because CpAM treatment usually induces the nuclear or cytoplasmic redistribution of Cp (Corcuera et al., 2018; Huber et al., 2018), it can be anticipated that CpAMs may modulate Cp interaction with its cytoplasmic and/or nuclear partners and consequentially modify HBV-hepatocyte interaction. Concerning the effects of CpAMs on the nuclear function of Cp, only one published study showed that prolonged treatment of HBV-infected HepaRG cells with JNJ-827 and JNJ-890 can reduce HBV RNA and HBsAg production, presumably by inhibition of cccDNA transcription or promotion of viral RNA decay (Lahlali et al., 2018). The effect of CpAM on Cp recruitment of APOBEC 3A and 3B for cccDNA decay as well as viral and cellular RNA metabolism remain to be investigated.

4. Core protein allosteric modulators: Antiviral activities and mode of action

4.1. Multiple chemotypes of CpAMs and their antiviral activity

Since the late 1990s, several chemotypes of compounds had been discovered to inhibit HBV replication via reduction of the amounts of capsids, such as heteroaryldihydropyrimidine (HAP) derivatives (Deres et al., 2003), or pgRNA-containing nucleocapsids, such as phenylpropenamides (PPA) and sulfamoylbenzamides (SBA) derivatives (Campagna et al., 2013b; King et al., 1998a). Mechanistic studies in the in vitro capsid assembly system indicated that those compounds accelerated, but not inhibited, Cp assembly into aberrant or normal capsids and were thus designated as core protein allosteric modulators (CpAMs), but not capsid assembly inhibitors (Zlotnick et al., 2015). As illustrated in Fig. 4, many structural distinct CpAMs had been discovered (Nijampatnam and Liotta, 2019). Based on the structures and metabolism of the induced Cp assembly products in vitro and in hepatocytes, CpAMs can be categorized into two distinct phenotypes. While type I CpAMs, including HAPs and possibly also Ciclopirox (Kang et al., 2019), induce the assembly of aberrant capsids or non-capsid Cp polymers that form aggregates and are subsequently degraded in hepatocytes (Bourne et al., 2006; Kang et al., 2019; Stray et al., 2005a; Stray and Zlotnick, 2006; Wu et al., 2013), type II CpAMs, such as PPA, SBA, etc., induce the assembly of morphologically “normal” capsids devoid of pgRNA and Pol (Amblard et al., 2020; Campagna et al., 2013a; Corcuera et al., 2018; King et al., 1998b; Wang et al., 2015; Yang et al., 2016). As will be discussed in detail below, all the CpAMs examined so far disrupt Cp assembly by binding to a hydrophobic pocket between the interface of Cp dimers to enhance dimer-dimer interaction. In agreement with the pleiotropic function of Cp in the HBV life cycle, CpAMs have also been demonstrated to disrupt multiple steps of HBV replication.

Fig. 4.

Chemical Structures of representative CpAMs.

4.2. Biophysics of CpAM disruption of capsid assembly

Altering the conformational state of Cp by changing the thermodynamics (pair-wise contact energies) or kinetics (rate of nucleation) can result in the misassembly of capsids and potentially disrupt the pregenome encapsidation process (Zlotnick and Mukhopadhyay, 2011). HAP compounds enhance the rate and extent of Cp assembly and act as allosteric effectors to induce an assembly-active state. While HAP1 misdirects assembly and leads to aberrant particles (Bourne et al., 2006; Stray et al., 2005b), HAP 12 leads to irregular non-capsid like polymers by increasing the association energy (~2 kcal/mol) and the rate of assembly (Venkatakrishnan et al., 2016). Type II CpAMs, such as PPAs, act as assembly accelerators by exclusively changing the reaction kinetics (Katen et al., 2010). Unlike HAPs, they do not misdirect assembly, but just speed it up and result in the production of ‘empty’ capsids due to the imbalance created between the rate of assembly of capsids and packing of Pol-pgRNA complex.

CpAMs differentially strengthen the association between Cp dimers and accelerate the kinetics of assembly. Thermodynamics and kinetics play a crucial role in binding energy and mode of CpAM action as well as the acquisition of CpAM resistance (Tan et al., 2013). While the thermodynamics and kinetics of in vitro Cp assembly have a strong and predictable effects on the Cp assembly product morphology, only the kinetics of in vitro assembly had a strong correlation with the activity of CpAM inhibition of HBV replication in human hepatocytes (Bourne et al., 2008). In addition, study of the wild-type and the V124 mutant Cp assembly kinetics revealed a nice correlation between the area of hydrophobic surface at the dimer-dimer interface and ΔG_contact (Tan et al., 2015). The resistance of T109 mutant Cp to GLS4, a HAP derivative, is also correlated with ΔG_contact. The most stable inter-dimer interface has the highest resistance to the CpAMs (Ruan et al., 2018).

4.3. Structure biology of CpAM-misdirected Cp assembly

The structures of Cp149 capsids incubated with CpAMs have been investigated using X-ray crystallography (pdb code 2G34, 4G93 and 5D7Y) (Table 1). All the assembly modulating compounds examined thus far bind to the “HAP” pocket, which is at the inter-dimer interface and consists of a concave interface made up of residues from one chain - F23, P25 of loop 2, D29, L30, T33, L37 from helix2, W102, I105, S106, T109, F110 of helix4, Y118 and F122 of helix5, I139, L140, S141 of loop6) and cap from the other chain – V124, W125, R127 from helix5 and T128, P129, Y132, R133, P134 of loop6 as shown in Fig. 2C.

CpAMs do not have a uniform allosteric effect on capsid structure. A small difference in structure leads to a large (mostly unpredictable) difference in the T=4 capsid after binding of the CpAM. HAP 1 (pdb code 2G34) causes the five-fold vertices to protrude, three-fold vertices to open, and quasi-six-fold vertices to flatten. It changes the quaternary but not the tertiary structure (Bourne et al., 2006). AT-130 (pdb code 4G93), a PPA derivative, causes quaternary changes like HAPs but also induces compensatory tertiary structural changes that lead to the formation of normal capsids (Katen et al., 2013). The quaternary changes might result in assembly misdirection. Binding of Cp by HAP1 and AT-130 expanded the capsid diameter by up to 10Å and caused the capsids to crystallize with different unit cell parameters from apo capsids. HAP18 (pdb code 5D7Y) caused only minor changes in quaternary structure and decreased the capsid diameter by 3Å (Venkatakrishnan et al., 2016).

The Y132 residue near the C-terminal end of the assembly domain plays an important role in interdimer interaction, which facilitates capsid assembly. The Y132A substitution causes the protein to become assembly deficient (Bourne et al., 2009; Packianathan et al., 2010). Four structures with CpAMs in the assembly inactive Y132 Cp hexamer are available till now (pdb codes 5E0I (Klumpp et al., 2015), 5GMZ (Qiu et al., 2016), 5WRE and 5T2P (Zhou et al., 2017)). These structures are at higher resolution compared to the X-ray structures of whole capsid and hence we can visualize water molecules and see some water-mediated hydrogen bonding of the ligand to the dimer interface. Fig. 5 shows the ligand-interaction diagram of the compounds HAP18, Ciclopirox, AT-130 and SBA_R01, respectively, with the core protein observed in the X-ray crystal structures. We can see that although the chemical structures are varied, there are some interactions such as hydrogen bond with W102 and aromatic ring (often with halogens on them) on the hydrophobic pocket formed by L30, L37, I105 and V124 has been observed in most instances.

Fig. 5. Illustration of the interaction of representative CpAMs with Cp at the dimer-dimer interface of the capsid and Y132A Cp hexamer X-ray crystal structures.

The ligand interaction diagram with negatively charged residues in red, positively charged residues in blue, hydrophobic residues in green and polar residues in blue. The purple line shows hydrogen bonding and orange line represents halogen bond interaction with the compounds. The green line denotes a stacking interaction. See text for detailed description.

4.4. Effects of CpAMs on assembled capsids and nucleocapsids

The dual effects of capsid-targeting antiviral agents on capsid assembly and disassembly have been reported in dengue virus and human immunodeficiency virus (Klumpp and Crepin, 2014; Scaturro et al., 2014). It is thus very interesting to note that CpAMs also disrupt the integrity of assembled HBV capsids. In fact, it was reported a decade ago that treatment with higher concentrations of HAPs can induce the disassembly of in vitro assembled capsids (Stray and Zlotnick, 2006). Using an engineered addressable thiol (mutating L30, I105, S106 and V124 to cysteine) in the HAP pocket, it was shown recently that ligands bound to the pocket triggers capsid disassembly in a dose-dependent manner (Qazi et al., 2018). Using a fluorophore [tetramethylrhodamine (TAMRA)]-labeled HAP, it was further demonstrated that binding of HAP compound induces capsid-wide asymmetric deformations (outbreaks, flat regions and sharp angles) (Schlicksup et al., 2018). These in vitro studies imply that the HAP pocket may transiently adopt destabilizing conformations and play a key role in both capsid assembly and disassembly (Hadden et al., 2018). However, CpAM treatment apparently does not induce the disassembly of capsids purified from the cytoplasm of hepatocytes supporting HBV replication. Instead, some HAP derivatives, such as HAP18 and GLS4, can induce the electrophoresis mobility shift of assembled empty capsids and DNA-containing nucleocapsids in native agarose gel electrophoresis (Ko et al., 2019), indicating the induction of global structural alteration of capsids (Schlicksup et al., 2018; Wu et al., 2018). Interestingly, we found recently that CpAMs can preferentially induce the disassembly of mature nucleocapsids and differentially modulate cccDNA synthesis from de novo and intracellular amplification pathways (Guo et al., 2017). Apparently, the mode of CpAM action on capsid structure alteration, mature nucleocapsid disassembly and de novo cccDNA synthesis remains to be further investigated.

Pharmacologically, inhibition of de novo cccDNA synthesis by CpAMs can only be achieved at the concentrations approximately 10 to 34 fold higher than that required to inhibit pgRNA assembly and viral DNA replication in primary human hepatocytes (Berke et al., 2020; Ko et al., 2019). While the finding is consistent with the observation that that disruption of capsid integrity requires the coordinative binding of CpAMs at many HAP pockets of a capsid (Stray and Zlotnick, 2006), it also argues whether the effective CpAM concentrations can be achieved in vivo in humans to inhibit de novo cccDNA synthesis. However, potent suppression of HBV pgRNA encapsidation will result in significant reduction of infectious virion production and consequentially reduce de novo infection and new cccDNA synthesis. Interestingly, a recent study showed that NUC therapy not only efficiently reduced HBV load, the residual virions produced by NUC-treated hepatocytes are not infectious, due to the incorporation of NUC, that irreversibly terminates viral DNA chain elongation (Liu et al., 2020). Therefore, it is conceivable that combination therapy of CpAM and NUC can achieve a more profound inhibition of HBV replication by targeting both pgRNA encapsidation and viral DNA synthesis, which should further reduce the production of infectious virion and thus prevent de novo cccDNA synthesis. In agreement with this notion, a recently reported clinical trial demonstrated that in the majority of patients, 6 months of combination therapy using entecavir and ABI-H0731 abolished residual HBV DNA replication and significantly reduced the level of HBsAg, indicative of a significant reduction of cccDNA levels in the liver (Mark S. Sulkowski, 2019).

5. CpAMs also disrupt the metabolism of pre-core protein

Pre-core ORF encodes a 25 kDa polypeptide (p25) that contains 29 amino acid residues extension at the N-terminus of core protein. As illustrated in Fig. 2B, p25 is co-translationally targeted by the signal peptide located in the N-terminal leader sequence to the endoplasmic reticulum (ER), where the first 19 residues of the leader sequence are cleaved by cellular signalase, producing a 22-kDa protein (p22). The p22 form is further processed by a furin-like protease in a post-ER compartment to remove C-terminal 29 residues to yield a 17 kD protein (p17), which is subsequently secreted out of the hepatocyte as HBeAg (Ito et al., 2009). Therefore, except for the 10 remaining leader residues at its N-terminal, HBeAg (p17) shares the entire assembly domain and linker region with Cp. Due to the formation of an intramolecular disulfide bond between the leader peptide C(−7) and C61 in the middle of assembly domain, HBeAg dimer has an entirely different quaternary structure in comparison with Cp dimer and is incompetent to assemble capsid (DiMattia et al., 2013). However, in a reducing condition, disruption of the intramolecular disulfide bond of p22 or HBeAg confers their ability to assemble capsids with reduced stability, due to the existence of the N-terminal leader peptide. Interestingly, a small fraction of p22 can be retro-transported from the ER lumen into the cytoplasm, where p22 may form heterodimers with Cp and assemble chimeric capsids and nucleocapsids. The observed inhibition of HBV replication by pre-core gene products in hepatocytes in vitro and in transgenic mice in vivo suggests an important regulatory role of p22 and/or HBeAg in HBV replication via assembly of unstable chimeric nucleocapsids with Cp (Guidotti et al., 1996). It had also reported that p22 attenuates interferon response by inhibiting STAT1 nuclear translocation (Mitra et al., 2019).

In addition to the intracellular effects of p22 on HBV replication and innate antiviral immune response, the secreted pre-core gene product HBeAg had been speculated to suppress host antiviral immune response and favor the establishment of persistent HBV infection (Milich, 2019). Pre-core gene is preserved in all the hepadnaviruses. While studies with duck hepatitis B virus (DHBV) and woodchuck hepatitis virus (WHV) showed that the deficiency of pre-core gene expression does not significantly impair the replication fitness of the viruses in ducks (Zhang and Summers, 1999) and woodchucks (Chen et al., 1992), infection of neonate woodchucks with WHV deficient in pre-core protein expression results in persistent infection, but wild-type WHV infection is transient (Chen et al., 1992). A recent study in mouse models revealed that the maternal-derived HBeAg potentiates persistent HBV infection by altering the function of intrahepatic macrophages in the offspring (Tian et al., 2016). A clinical study indicates that HBeAg induces the expansion of monocytic myeloid-derived suppressor cells to inhibit T cell function (Yang et al., 2019a).

Recently, a study showed that treatment of HBV replicating hepatocytes with HAP_R10, but not AT-130, efficiently inhibited HBeAg secretion. Mechanistic studies support a notion that HAP_10 induces p22 assembly into non-capsid aggregates, which is subsequently degraded (Yan et al., 2019). Interestingly, HAP_10 failed to inhibit HBeAg secretion from hepatocytes supporting the replication of a HBV variant encoding HAP-resistant core protein, suggesting that HAP_R01-mediated HBeAg and core protein reductions are mediated via the same mechanism. Considering the function of p22 in suppression of IFN response in hepatocytes as well as the potential role of HBeAg in regulation of host antiviral immune response, it is reasonable to speculate that HAP derivatives may be able to restore the host antiviral immune response via reducing intracellular and secreted pre-core gene products. It will be interesting to investigate whether the CpAMs that can inhibit HBeAg secretion, such as HAP_10 and JNJ-827 (Lahlali et al., 2018; Yan et al., 2019), are more efficacious in achieving the functional cure of chronic HBV infection than the CpAMs that cannot inhibit HBeAg secretion in clinical trials.

6. CpAMs in clinical development

A number of structurally diverse CpAMs, representing both type I and type II CpAMs, are in clinical trials and constitute one of the most actively investigated mechanisms for chronic hepatitis B (CHB) cure. NVR 3–778 (type II) was the first CpAM to demonstrate clinical proof of concept (POC), showing HBV DNA and RNA reductions in patients when administered as a monotherapy for 28 days and in combination with PegIFNα−2a (Yuen et al., 2019b). Since then, newer generations of CpAMs have demonstrated improved efficacy in POC studies, as monotherapy and in combinations with standard of care (SOC) drugs and investigational agents (Table 2).

Table 2.

Most advanced CpAMs in clinical trials for the treatment of CHB patients

CpAM (MoA Type)	Chemical Class	CpAM Dose and dosing frequency	Reduction from BL (mean or median log10 IU/mL or copies/mL) in Serum						Clinical Status	Sponsor	Ongoing or Completed Clinical Trial^
			Clinical Phase	HBV DNA	HBV RNA	HBeAg#	HBsAg	HBc rAg
NVR 3-778 (Type II)	SBA	100 mg QD 200 mg QD 400 mg QD 600 mg BID 1000 mg BID 600 mg BID + PegIFNα2a (SC)/PegIFNα2a alone	Ph 1b (28d/EOT)	0.19 0.20 0.49 1.43 1.35 1.97/1.06	nd nd nd 1.42 1.27 2.10/0.89	0.02 0.04 0.04 0.09 0.07 0.21/0.29	0.03 0.03 0.03 0.02 1.00 0.08/0.02	nd nd nd 0.01 0.13 0.10/0.28	D/C	Novira/ JNJ	NCT02112799; NCT02401737;
JNJ-6379 (Type II)	SPA	25 mg QD* 75 mg QD 150 mg QD 250 mg QD	Ph 1b (28d/EOT)	2.16 2.89 2.70 2.70	2.30 1.85 1.83 1.43	NS NS NS NS	NS NS NS NS	≥0.5 in a subset of patients mainly with ALT>1x and <2.5x ULN	Ph 2	Janssen	NCT02662712, NCT04208399, NCT03864601, NCT03945539, NCT03111511, NCT02933580, NCT03982186, NCT04288310, NCT04129554, NCT03361956
JNJ-0440 (Type II)	SBA	750 mg QD 750 mg BID	Ph 1 (28d/EOT)	3.2 3.3	2.65 2.94	0.2 0.2	NS NS	0.85 0.62	Ph 1		NCT03439488
ABI-H0731 (Type II)	DBT	100 mg QD 200 mg QD 300 mg QD 400 mg QD	Ph 1b (28d/EOT)	1.3/2.2** 1.9/2.4** 2.9/2.5**$ NR/3.9**	1.2/NR** 1.7/NR** 2.3/NR**$ NA/NR**	NS NS NS NS	NS NS NS NS	NS NS NS NS	Ph 2	Assembly	NCT02908191, NCT03780543, NCT03577171, NCT03576066, NCT03109730
		300 mg QD + ETV/ETV alone (HBeAg+ Rx naïve) 300 mg QD + NRTI/NRTI alone (HBeAg+/- NUC suppressed)	Ph 2a Study 202 (24 week) Phase 2a Study 201 (24 week)	5.30/4.19 TND in 22 of 27 pts/ 0 out of 12 pts	2.34/0.61& 1.74/0.09	-	-	-
ABI-H2158 (Type II)		100 mg QD	Ph 1b (14d/EOT)	2.3	2.1&	NR	NR	NR	Ph 1		NCT03714152, NCT04083716, NCT04142762
AB-506 (Type II)	AI	160 mg QD 400 mg QD	Ph 1b (28d/EOT)	2.10 2.80	2.37 2.40	NR NR	0.021 0.113	NR NR	D/C	Arbutus
GLS4JHS (Type I)	HAP	120 mg QD (+ RTV)	Ph 1b (28d/EOT)	1.42	NR	0.25	0.06	NR	Ph 2	HEC Pharma	NCT04147208, NCT03638076, NCT03662568
		240 mg QD (+ RTV)		2.13	NR	0.30	0.14	NR
		120 mg BID (+ RTV)	Ph 2 (48-week study/TW24 data	3.28¥	~3.0¥	0.57¥	0.20¥	~1.9¥
		120 mg TID (+RTV)	reported)	4.40¥	~3.0¥	1.06¥	0.44¥	~1.9¥
RO9389 (Type I)	HAP	200 mg BID 400 mg BID 200 mg QD 600 mg QD 1000 mg QD	Ph 1b (28d/EOT)	2.66@ 3.20@ 3.00@ 2.92@ 3.17@	2.09@ 2.55@ 2.55@ 2.54@ 2.43@	NS NS NS NS NS	NS NS NS NS NS	NR NR NR NR NR	Ph 2	Roche	NCT03570658, NCT03717064, NCT02952924, NCT04225715

Notes: D/C = discontinued; SC = subcutaneous; QD = once daily; BID = twice daily; NR = not reported;

^*100 mg loading dose followed by 25 mg QD dosing; BL = baseline; EOT = end of treatment; NS = no significant changes; SBA = sulfamoylbenzamide; SPA = Sulphamoylpyrrolamide; AI = aminoindane; DBT -dibenzothiazepinecarboxamide; HAP = heteroaryldihydropyrimidine;

^**HBeAg positive/HBeAg Negative CHB patients; ETV = entecavir;

^&= pgRNA; TND = DNA target not detected in a semiquantitative PCR assay; JNJ-6379 = JNJ-56136379; JNJ-0440 = JNJ-64530440; RO9389 = RO7049389; RTV = ritonavir;

^@median decline log IU/mL; NRTI = approved nucleoside reverse transcriptase inhibitor therapy;

^{^}only trials registered at US clinicaltrials.gov shown; Ph = phase;

^$= HBV DNA and RNA data of one patient with Cp T109M variant in the 300 mg QD cohort was excluded from analyses.

^#= applies to HBeAg-positive patients; TW = treatment week.

^¥= mean max log₁₀ IU/ml reduction from BL.

In a 28-day POC study in treatment-naïve CHB patients, JNJ-6379 (type II) administration showed reductions in HBV DNA and RNA from baseline (BL) without significant reductions in HBeAg or HBsAg levels, although a subset of patients showed serum HBV core-related antigen (HBcrAg) reduction presumably due to direct inhibition of core protein release (Lenz et al., 2019; Zoulim et al., 2018). The antiviral activity plateaued at ~3 log reduction in HBV DNA, at doses exceeding 75 mg QD (Zoulim et al., 2018). Interestingly, JNJ-6379 showed a long plasma half-life of 120–148 h (Yogaratnam et al., 2019). JNJ-6379 is being evaluated in a triple combination study with an RNAi inhibitor and a NUC where interim results have shown robust reductions in HBsAg, HBV DNA and RNA after 3 months on therapy (Yuen et al., 2019c). JNJ-0440, a SBA derivative (Table 2), has also demonstrated clinical POC in a 28-day study, with 750 mg QD and BID dosing showing equivalent HBV DNA reductions in plasma and modest effects on HBcrAg levels (Gane et al., 2019a). In contrast to JNJ-6379, JNJ-0440 has a shorter plasma half-life of 9–16 h. The BID dosing of JNJ-0440 resulted in higher plasma trough concentrations (C_tau:plasma-adjusted EC₉₀ ratio) in comparison to QD dosing, presumably increasing the likelihood of engaging the secondary mechanism of CpAMs (Gane et al., 2019a).

ABI-H0731 (type II CpAM) has also demonstrated antiviral activity in CHB patients in a 28-day POC study (Yuen et al., 2018; Yuen et al., 2020). One patient in the 300 mg cohort who exhibited a reduced level of HBV DNA decline (1 log vs mean 2.9 log₁₀ IU/mL for the cohort), showed a 70% to 100% enrichment of Cp T109M BL variant over the course of treatment. ABI-H0731+NUC combination is currently being evaluated in NUC-suppressed HBeAg positive and HBeAg negative patients, as well as NUC-naïve HBeAg positive subjects (Sulkowski et al., 2019). In the NUC-naïve group, ABI-H0731 and entecavir combination demonstrated faster and greater declines in serum HBV DNA and pgRNA from baseline versus entecavir alone, following 24 weeks of administration. These patients enrolled for an additional 52 weeks of combination therapy showed continuous decrease in serum HBV DNA to week 60. Similarly, in NUC-suppressed HBeAg positive and HBeAg negative patients, ABI-0731 and NUC combination showed a greater reduction in HBV pgRNA levels from baseline in comparison to entecavir alone, with 22/27 patients exhibiting undetectable HBV DNA vs 0/12 patients in the NUC alone arm using an inhouse higher sensitivity assay. Interestingly, HBV pgRNA declines also correlated with reductions in HBeAg, HBcrAg and HBsAg levels in patients that had undergone treatment for 16–60 weeks with the combination. A second-generation CpAM, ABI-H2158, is also undergoing POC study in a 14-day monotherapy trial, with a 100 mg QD dose demonstrating reductions in serum HBV DNA and pgRNA in HBeAg positive patients. Higher doses of ABI-H2158 are under evaluation (Yuen et al., 2019a).

AB-506 (type II) has also demonstrated POC with a 400 mg QD dose resulting in mean HBV DNA and RNA reductions of 2.8 and 2.4 log₁₀ IU/mL by day 29 (Sims et al., 2019). One subject (160 mg dose) showed non-response, attributed to the presence of Cp I105T mutation at baseline, which is known to reduce in vitro susceptibility to AB-506 by ~19 fold.

Two type I CpAMs, GLS4 and RO9389, are currently in Phase 2 clinical trials. Phase 1 studies showed that administration of GLS4 did not result in sufficient plasma trough levels, which was attributed to metabolic instability. A co-dosing strategy with the known CYP450 3A4 inhibitor ritonavir (RTV) was used to boost GLS4 exposures (Ding et al., 2017; Zhang et al., 2019; Zhao et al., 2019). RO9389 treatment has also demonstrated robust declines in serum HBV DNA (Gane et al., 2019b).

ALT elevations have been reported in some patients with CpAMs (Gane et al., 2019b; Lenz et al., 2019; Sims et al., 2019; Sulkowski et al., 2019; Vandenbossche et al., 2019). In the case of AB-506 a 28-day healthy volunteer study directly attributed ALT flares to AB-506, leading to its discontinuation. Whether ALT elevations observed with other CpAMs are related to an immunological reawakening of an anti-HBV response or liver toxicity require further studies.

It is clear that both types I and II CpAMs could potentially become part of future treatment regimens for CHB patients. While the above examples are limited to the most advanced CpAMs, a number of others in preclinical/Phase 1a studies are poised to shape a highly competitive landscape.

7. Perspectives of Cp-targeting antiviral therapy

CpAMs are an emerging new mechanistic class of anti-HBV agents for chronic hepatitis B patients. The clinical experience to date has shown that a combination of CpAM and NUC is a viable therapeutic regimen that leads to more pronounced viral suppression than a NUC alone and confirms the hypothesis that NUCs are inherently “leaky” in inhibiting viral replication. However, important questions still remain such as: What mechanistic combination(s) will ultimately increase the rate of functional cure? Is there a role of the secondary mechanism of CpAMs, such as inhibition of cccDNA synthesis and/or transcription and suppression of HBeAg secretion, on antiviral efficacy or rates of functional cure? Do type I vs type II CpAMs have differential effects on efficacy endpoints and/or safety? What is the implication of baseline Cp variants on clinical efficacy? What is the significance of ALT elevations observed with some CpAMs; are they good (i.e. due to immune reactivation) or bad (toxicity) flares? As more CpAM combinations with other investigational agents enter clinical testing, the field is set for an exciting future that will hopefully address these important questions.

Highlights.

Core protein is a building block of nucleocapsids and involved in multiple steps of the HBV replication.

CpAMs bind to a hydrophobic pocket between core protein dimers to misdirect nucleocapsid assembly and disassembly.

CpAMs not only block viral DNA replication, but also inhibit de novo cccDNA formation.

Targeting the multi-functional core protein is an effective antiviral approach against HBV infection.