Development of anti-HBV agents targeting HBV capsid proteins

Abstract

Hepatitis B is a viral hepatitis, which is caused by infection of hepatitis B virus (HBV). This disease progresses to chronic hepatitis, cirrhosis and liver cancer. To treat hepatitis B, exclusion of virus and covalently closed circular DNA (cccDNA) that is formed in hepatocyte nucleus is necessary. A hepatitis B capsid protein (HBc) is an indispensable protein, which forms the capsid that encapsulates viral DNA. Since HBc is correlated to the transcriptional regulation of cccDNA, this protein would be an attractive target for complete cure of hepatitis B. By in silico screening of a library of compounds, a small compound, Cpd4 (1), which binds to a hydrophobic cavity located in the inner pocket on the tetramer interface of HBc proteins, was identified. In anti-HBV assays, this synthetic compound, Cpd4 (1) decreased the amount of HBV core related antigen (HBcrAg), which has been correlated with the proliferation of HBV, and decreased the amount of HBV surface antigen (HBsAg), which is correlated with the amount of cccDNA. Based on Cpd4 (1) as a lead compound, 20 derivatives of 1 were designed and synthesized and their structure–activity relationships were examined. As a result, specific interactions between each compound and amino acid residues of the target protein appeared to be unimportant but the shape/size of compounds which can bind to the hydrophobic cavity might be important in the expression of high anti-HBV activity, and a more potent derivative, TKB-HBV-CA-001 (3b), was discovered. These results will be useful in the development of novel anti-HBV agents for a complete cure of hepatitis B.

>Design and synthesis of anti-hepatitis B virus (HBV) agents: A small anti-HBV compound, Cpd4 (1), was identified by in silico screening. Its structure–activity relationship studies discovered a more potent derivative, TKB-HBV-CA-001 (3b).

Introduction

Hepatitis B is a viral hepatitis, which is caused by infection by the hepatitis B virus (HBV) of liver cells, and is largely classified into acute hepatitis and chronic hepatitis.^1,2 If chronic hepatitis is developed, the risk of developing cirrhosis and liver cancer, which are the main causes of death, is significantly increased. To date, effective vaccines for HBV have been developed,³ and the incidence rate has been remarkably decreased. The number of HBV-infected patients worldwide however, is still estimated to be in excess of 350 million.⁴ HBV is a virus with a relaxed circular DNA (rcDNA) with a molecular weight of 3.2 kb. A viral genome is contained in its nucleocapsid, which is surrounded by envelop proteins. Hepatitis viruses with the exception of HBV have an RNA genome, thus the replication cycle of HBV is different from those of other viruses.⁵ In general, DNA viruses seldom have gene mutations because they have a calibration mechanism that blocks replication errors. RNA viruses on the other hand often experience genetic mutation. HBV, however does not take advantage of this calibration mechanism and this causes high levels of mutation in spite of its being a DNA virus because its replication cycle might be of the retrovirus type containing a reverse transcription step.⁶ By the difference of base sequences derived from gene mutation, HBV genotypes are classified into 9 genotypes from type A to J (I is a subtype of C).

After invading the human body, HBV enters into liver cells via a Na⁺ taurocholate co-transporting polypeptide (NTCP). After entry of the virus into liver cells, DNA repair causes formation of covalently closed circular DNA (cccDNA) from relaxed circular DNA (rcDNA) in the nucleus. cccDNA is thermodynamically stable and maintained in the nucleus, and becomes a template for viral replication. By host transcription factors, cccDNA is transcribed into 4 viral mRNAs, which are translated into polymerase (Pol), hepatitis B capsid protein (HBc), hepatitis B surface protein (HBs) and hepatitis B X protein (HBx). The longest viral RNA is designated as a pregenomic RNA (pgRNA), which ultimately becomes a viral genome. HBc multimer encapsulates complexes of pgRNA and Pol, and forms nucleocapsid. In the nucleocapsid, by the reverse transcription activity of Pol, rcDNA is synthesized from pgRNA, and the nucleocapsid is then encapsulated by HBs in endoplasmic reticulum and is released extracellularly.^7–10 The part of the nucleocapsid is translocated into the nucleus, and cccDNA is then recycled and remains essentially permanently in the human body.¹¹

As a therapeutic index for hepatitis B, the amount of HBV DNA is useful for the prehension of disease states and evaluation of therapeutic effects. In cases where the amount of HBV DNA is high, the probability of progressing to liver cancer becomes high and thus is an important factor correlated with the prognosis. In addition, the amount of hepatitis B surface antigen (HBsAg) in blood is also used as therapeutic index because it could correspond to the amount of cccDNA in the nucleus of liver cells.¹² The final destination of these therapeutics is complete exclusion of HBV from liver and maintenance of its state with no recurrence of hepatitis. For an example, in addition to negative conversion of HBV DNA and HBsAg in blood, a significant increase of the antibody titer for HBsAg in blood could be a therapeutic index: elimination of HBsAg means exclusion of cccDNA in the nucleus of liver cells, and appearance of antibodies against HBs signifies a sustainable state capable of suppressing reactivation of HBV.¹³ Currently, polyethylene-glycol-interferon-α (PEG-IFN-α) that has immunopotentiative action and reverse transcriptase inhibitors such as entecavir and tenofovir that suppress the replication cycle are used clinically as drugs to combat hepatitis B.⁶ With these drugs, however, a complete cure cannot be achieved. Therefore, development of novel drugs with different mechanisms of action is necessary.

Currently, anti-HBV agents targeting capsid are regarded as new drug candidates.^14,15 Capsid is formed by self-assembly of an HBc dimer. Synthesis of the HBV genome can be inhibited by modification of the HBc self-assembly because reverse transcription of pgRNA is initiated after the formation of nucleocapsid.¹⁶ In addition, HBc is correlated with the regulation of transcription of cccDNA, and some reported agents that target capsid inhibit the entry of viral DNA into nucleus.^14,15 Therefore, compounds targeting HBc have the possibility of inhibiting several replication steps. Amino acid sequences of HBc are highly conserved among several genotypes.¹⁷ Taken together, HBc should be an extremely useful target for complete cure of hepatitis B.

Results and discussion

In silico screening of library compounds for binding to HBc

First, we attempted to identify small molecules which can bind to HBc. The hydrophobic cavity located in the inner pocket of the protein includes Ala54, Ala58, Phe97 and Leu100 on the HBc multimer-interface which was the focus because it has an appropriate size as judged by the crystallographic data of the HBc monomer (PDB ID: 3KXS) (Fig. 1).¹⁸ As compounds used for docking simulation with this hydrophobic cavity, 6 842 684 compounds from a commercially available library of 8 555 483 compounds (eMolecule, San Diego, CA, USA) were selected, as they were expected to have suitable pharmacokinetics in vivo.¹⁹ The structure of each compound was energy-minimized before simulated docking with the above target cavity on the surface of HBc. Binding scores of each compound were simulated using a docking simulation software, Flex-SIS (BioSolveIT GmbH, Sankt Augustin, Germany). This in silico screening identified Cpd4 (1) as a hit compound (Fig. 2). Cpd4 (1) and its derivatives were synthesized and their antiviral activity was evaluated to establish each of their structure–activity relationships.

Fig. 1. The structure of HBc (Adyw strain) (PDB ID: 3KXS). Blue: hydrophobic amino acid region, red (right): Ala54, Ala58, Phe97, Leu100 (multimer-interface site), yellow square (left): the cavity targeted by screened compounds.

Fig. 2. Docking simulation of Cpd4 (1) with HBc (Adyw strain) (PDB ID: 3KXS) by molecular operating environment (MOE), 2020.0901 (Chemical Computing Group, QC, Canada). (a) Docking model of Cpd4 (1) bound to the hydrophobic cavity of HBc. Cpd (1) is represented with purple (carbon), blue (nitrogen), red (oxygen), yellow (sulfur), green (chlorine), and brown (bromine); (b) enlarged structure of Cpd4 (1). Carbon atoms of Ala54, Ala58, Phe97 and Leu100 (multimer-interface site) are shown in red.

Design of small molecules derived from Cpd4 (1)

Based on the carbamothioylbenzamide scaffold in Cpd4 (1), 20 derivatives were designed. Substituents on the pyridine ring or the phenyl ring of Cpd4 (1) were replaced by other atoms or groups (Fig. 3). Compounds were designed with a halogen atom at position 5 of the pyridine ring (2a–2d), or with a CF₃ group at the same position (3a–3k). In addition, compounds with no substituent, a Me group, or an NO₂ group at position 5 of the pyridine ring (4–6), a compound with an isoquinoline ring in place of the pyridine ring (7), and a compound having Ph and CF₃ groups at the positions 4 and 5, respectively, of the pyridine ring (8) were also designed.

Fig. 3. Design of Cpd4 (1) derivatives.

Synthesis of Cpd4 (1) derivatives: retrosynthetic analysis

A possible synthesis of derivatives of Cpd4 (1) is outlined in Scheme 1. For the synthesis, Cpd4 (1) derivatives were divided into three segments I, II and III based on the retrosynthetic analysis (Scheme 1). The thiourea scaffold of target compounds can be synthesized by the nucleophilic addition of 2-aminopyridine moieties, segment II, to isothiocyanates (segment I), which could be synthesized by treatment of the benzoyl chloride moieties (segment III), with ammonium thiocyanate.

Scheme 1. Retrosynthetic analysis of Cpd4 (1) derivatives.

Synthesis of an intermediate (12) to produce compound 8

For the synthesis of compound 8, an intermediate (12), corresponding to segment II, was prepared. Suzuki–Miyaura coupling of a phenyl group to 2-chloro-4-iodo-5-(trifluoromethyl)pyridine (9), and subsequent replacement of the Cl atom in 10 by a Boc-protected amino group produced compound 11. This was followed by deprotection of the Boc group to yield the 2-aminopyridine (12) (Scheme 2). Other 2-aminopyridine moieties, which correspond to segment II, are commercially available.

Scheme 2. Synthesis of an intermediate (12) to produce compound 8.

Synthesis of derivatives of Cpd4 (1)

Treatment of benzoyl chloride derivatives (13) with ammonium thiocyanate yielded the corresponding isothiocyanates which reacted with an amine (14) to produce derivatives (2–8) of Cpd4 (1) (Table 1). Some compounds such as TKB-HBV-CA-001 (3b) and compound 7 showed relatively low yields because of low recovery after recrystallization and reprecipitation of target compounds. Compound 3i was synthesized by removal of the methyl group of compound 3j (Scheme 3). This step also gave a low yield because of the low recovery following reprecipitation.

Synthesis of Cpd4 (1) derivatives.

Entry	Compounds	R1	R2	R3	Yield (%)
1	2a	Br	Br	—	36
2	2b	CF3	Br	—	82
3	2c	Cl	F	—	58
4	2d	Cl	Cl	—	58
5	3a	F	CF3	—	35
6	TKB-HBV-CA-001 (3b)	Cl	CF3	—	12
7	3c	Br	CF3	—	31
8	3d	I	CF3	—	35
9	3e	CF3	CF3	—	75
10	3f	Me	CF3	—	35
11	3g	i Pr	CF3	—	40
12	3h	t Bu	CF3	—	57
13	3j	OMe	CF3	—	32
14	3k	NO2	CF3	—	29
15	4	Cl	H	—	49
16	5	Cl	Me	—	48
17	6	Cl	NO2	—	61
18	7	Cl	—		9
19	8	Cl	—		27

Scheme 3. Synthesis of compound 3i from compound 3j.

Evaluation of anti-HBV activity and cytotoxicity

PXB cells are liver cells, isolated from PXB mice.²⁰ To evaluate the anti-HBV activity of the synthesized compounds, amounts of two antigens, HBcrAg and HBsAg, were measured by an enzyme-linked immunosorbent assay (ELISA) in PXB cells, which have been infected by the HBV genotype C (HBV/C). The amount of HBcrAg corresponds to the total amount of HBV DNA and therefore to the proliferation of HBV, and the amount of HBsAg is correlated to the amount of cccDNA in the nuclei of liver cells, which shows levels of exclusion of HBV.¹² Both of these two antigens have important therapeutic indices. The cytotoxicity of these compounds was determined based on reduction of the viability of HepG2 cells,²¹ derived from human liver cancer, by a methyl thiazolyltetrazolium (MTT) assay.²²

Cpd4 (1), showed significant anti-HBV activity (EC₅₀ = 44 nM for HBcrAg, EC₅₀ = 690 nM for HBsAg), and showed no significant cytotoxicity below 100 μM. These results are shown in Table 2. The anti-HBV activity of Cpd4 (1) is remarkably higher than that of tenofovir, which is used clinically with patients with hepatitis B.⁶ Although the reverse transcriptase inhibitor, tenofovir, did not show significant activity against HBsAg, Cpd4 (1) showed potent activity against HBsAg, suggesting that Cpd4 (1) blocks one or more steps of the replication cycle. To search for potent compounds, compounds 2a or 2b, which retain a bromine atom at position 5 of the pyridine ring and have a bromine atom or a trifluoromethyl group at position 4 of the phenyl ring, respectively, were evaluated. Compound 2b showed comparable anti-HBV activities to those of Cpd4 (1), but Compound 2a showed lower activities than Cpd4 (1) in terms of HBcrAg and HBsAg. This suggests that a trifluoromethyl group or a chlorine atom on the phenyl ring is suitable but a bromine atom is not. Cpd4 (1) derivatives 2c and 2d, which both maintain a chlorine atom at position 4 of the phenyl ring and have fluorine and chlorine atoms at position 5 of the pyridine ring, respectively, were evaluated. Compounds 2c and 2d showed significantly lower anti-HBV activities than Cpd4 (1) in terms of HBcrAg and HBsAg, suggesting that a fluorine or chlorine atom in the pyridine ring is disadvantageous. Next, Cpd4 (1) derivatives 3a–3k, which maintain a trifluoromethyl group at position 5 of the pyridine ring and have various moieties at position 4 of the phenyl ring, were evaluated. Only compound TKB-HBV-CA-001 (3b) with a chlorine atom at position 4 of the phenyl ring showed remarkably higher anti-HBV activities than Cpd4 (1) among compounds 3a–3d which have any of four halogen atoms at position 4 of the phenyl ring, and compound 3c with a bromine atom at the same position 4 exhibited moderate activities. This indicates that at position 4 of the phenyl ring, a chlorine atom is the best choice and a bromine atom is the second best but fluorine or iodine are not suitable, in derivatives that maintain a trifluoromethyl group at position 5 of the pyridine ring. Compounds 3e–3h with various alkyl groups at position 4 of the phenyl ring have no significant anti-HBV activity. Compound 3j with a methoxy group also was inactive. As a result, any alkyl group at position 4 of the phenyl ring is unsuitable, in derivatives that maintain a trifluoromethyl group at position 5 of the pyridine ring. Compounds (3i and 3k) with hydroxy and nitro groups at position 4 of the phenyl ring, respectively, exhibited significant anti-HBV activity, indicating that hydrophilic substituents are suitable. Since compound TKB-HBV-CA-001 (3b) with a chlorine atom at position 4 of the phenyl ring showed by far and away the highest anti-HBV activities among compounds Cpd4 (1), 2a–2d and 3a–3k, compounds 4–8 with a chlorine atom at position 4 of the phenyl ring were evaluated. Compounds 4–6 with a hydrogen atom, methyl or nitro group at position 5 of the pyridine ring showed lower anti-HBV activities than Cpd4 (1) and TKB-HBV-CA-001 (3b). Compounds 7 and 8, with additional phenyl rings fused or attached to the pyridine ring exhibited lower anti-HBV activities than Cpd4 (1) and TKB-HBV-CA-001 (3b). Even compared to compounds 2c and 2d, compound TKB-HBV-CA-001 (3b) with a chlorine atom at position 4 of the phenyl ring, has the highest anti-HBV activities among the synthesized derivatives. In terms of cytotoxicity based on reduction of the viability of HepG2 cells, some of the synthesized compounds exhibited significant cytotoxicity (CC₅₀: 20–90 μM) although the cause of cytotoxicity is not clear, but compounds 2b, 3b, 3c and 3k which have remarkable anti-HBV activity failed to show any cytotoxicity at concentrations below 100 μM. The cytotoxicity of compounds 2b and 3b was evaluated in other cell lines (Table 3), but compounds 2b and 3b did not show strong cytotoxicity (CC₅₀ >60 μM) even in cell lines derived from the bile duct.

Anti-HBV activities and cytotoxicity of the synthesized Cpd4 (1) derivatives.

Entry	Compounds	R1	R2	R3	HBcrAg	HBsAg	CC50b (μM)
					EC50 ratio to Cpd4 (1)a	EC50 ratio to Cpd4 (1)a
1	Cpd4 (1)	Cl	Br	—	(44 nM)	(690 nM)	>100
2	2a	Br	Br	—	>12	>4.8	41.8
3	2b	CF3	Br	—	1.7	0.63	>100
4	2c	Cl	F	—	6.2	8.1	>100
5	2d	Cl	Cl	—	7.5	6.9	66.3
6	3a	F	CF3	—	>12	>4.8	42.2
7	TKB-HBV-CA-001 (3b)	Cl	CF3	—	0.032	0.005	>100
8	3c	Br	CF3	—	3.8	2.7	>100
9	3d	I	CF3	—	>12	>4.8	>100
10	3e	CF3	CF3	—	16	>2.9	>100
11	3f	Me	CF3	—	>12	>4.8	89.8
12	3g	i Pr	CF3	—	>12	>4.8	61.9
13	3h	t Bu	CF3	—	>12	>4.8	>100
14	3i	OH	CF3	—	0.86	>4.8	23.7
15	3j	OMe	CF3	—	>12	>4.8	75.9
16	3k	NO2	CF3	—	2.7	0.37	>100
17	4	Cl	H	—	7.6	5.5	26.8
18	5	Cl	Me	—	3.7	3.8	43.8
19	6	Cl	NO2	—	20	12	>100
20	7	Cl	—		2.9	31	79.1
21	8	Cl	—		22	11	>100
	Tenofovir				(360 nM)	(>2 μM)

^aEC₅₀ values are the concentrations for 50% reduction of amounts of two antigens, HBcrAg and HBsAg, based on ELISA results in the culture supernatants of PXB cells infected with HBV/C. Ratios to the EC₅₀ values are shown.

^bCC₅₀ values are the concentrations for 50% reduction of the viability of HepG2 cells based on an MTT assay. The data are a mean of at least two independent experiments.

Cytotoxicity of compounds 2b and TKB-HBV-CA-001 (3b) in several cell lines.

Tissue	Cell line	CC50a (μM)
		2b	TKB-HBV-CA-001 (3b)
Liver	HuH-7	92.7	>100
	Li-7	>100	>100
	HLE	>100	>100
	Li90	62.1	>100
Bile duct	HuCCT1	>100	>100
	RBE	81.1	>100
	SSP-25	76.3	61.3
Gallbladder	G-415	>100	>100

^aCC₅₀ values were determined by an MTT assay after culture of each cell line with serially diluted solutions using a microplate reader.

Conclusion

Hepatitis B is a viral hepatitis, which can progress to cirrhosis and liver cancer. To date, PEG-IFN-α and reverse transcriptase inhibitors such as entecavir and tenofovir have been clinically used as therapeutic drugs for hepatitis B.⁶ However, since complete cure with these existing drugs cannot be achieved, development of novel drugs with different mechanisms of action is necessary. HBc, which is correlated with the transcriptional regulation of cccDNA, could be an attractive target because the exclusion of cccDNA is critical for complete cure of hepatitis B. In this study, in silico screening of library compounds identified Cpd4 (1) as binding to a hydrophobic cavity located in the inner pocket on the tetramer interface of HBc proteins. Two different assays disclosed the significant anti-HBV activities of the synthetic compound Cpd4 (1). Cpd4 (1) remarkably decreases the amount of two types of HBV antigens: core-related (HBcrAg) and surface antigens (HBsAg), which might reflect proliferation of HBV and the amount of cccDNA, respectively. The subsequent structure–activity relationship studies of Cpd4 (1) found a more potent derivative, TKB-HBV-CA-001 (3b), which has prominently higher anti-HBV activity than tenofovir. Although tenofovir showed no significant activity against the antigen HBsAg, TKB-HBV-CA-001 (3b) showed potent activity. Tenofovir inhibits a reverse transcriptase, but TKB-HBV-CA-001 (3b) might block multiple steps of the replication cycle including the recycling of cccDNA. Tenofovir and TKB-HBV-CA-001 (3b) might show a synergistic effect when both of these agents are administrated at the same time because the action mechanisms of these agents are different. In addition, compounds 2b, 3c and 3k showed significant activity against HBcrAg and HBsAg. These structure–activity relationship studies, suggest that the carbamothioylbenzamide scaffold should be a key structure. There are however, no clear rules governing the nature and sites of substituents, and the shape and size of compounds rather than any specific interaction between compounds and the target protein HBc might be important to express high anti-HBV activity. An ability to occupy the hydrophobic cavity appears to be important to express high anti-HBV activity. The present results could lead to the development of novel anti-HBV agents.

Experimental

In silico screening of antiviral candidates

To perform the in silico screening, we first obtained the structure of HBc (PDB ID: 3KXS) from the Protein Data Bank (https://www.rcsb.org/) and attempted to identify small compounds which are capable of binding to HBc. We chose a hydrophobic cavity located in the inner pocket near to the residues of Ala54, Ala58, Phe97 and Leu100 on the HBc multimer-interface with an appropriate size judged by crystallographic data of the HBc monomer (PDB ID: 3KXS). We then selected 6 842 684 compounds from a commercially available library of 8 555 483 compounds (eMolecules, San Diego, CA, USA), which were expected to have relevant pharmacokinetics in vivo as judged by the following factors; (i) numbers of H-bond donors (≦5) and acceptors (≦10), (ii) molecular weight (200 ≦ M.W. ≦ 600), (iii) calculated log P (≦5), and (iv) number of rotatable bonds (≦10), because these factors are correlated to ADME of compounds that can be administered orally. The structure of each compound was energy-minimized with the MMFF94x force field as implemented in MOE (Chemical Computing Group, QC, Canada) before applying them to a docking simulation. Next, we simulated binding scores of each compound with the putative target cavity on the surface of HBc using a docking simulation software, Flex-SIS (BioSolveIT GmbH, Sankt Augustin, Germany).

Synthesis of Cpd4 (1) and its derivatives

The synthetic methods for Cpd4 (1) and its derivatives are described in Schemes 2 and 3 and Table 1. The purity of all of the final compounds, measured by NMR is >95%. Experimental procedures including characterization data are provided in the ESI.†

Evaluation of anti-HBV activity and cytotoxicity

PXB cells are human hepatocytes, which are isolated from the PXB-mouse, a chimeric mouse model with a humanized liver.²⁰ To evaluate the anti-HBV activity of the synthesized compounds, amounts of two antigens, HBcrAg and HBsAg, were measured by an enzyme-linked immunosorbent assay (ELISA/Lumipulse f, Fujirebio Inc., Tokyo, Japan) in the culture supernatants of PXB cells, which were infected with HBV genotype C (HBV/C) in the presence or absence of various concentrations of compounds. In brief, fresh medium containing serially diluted compounds were added to 96-well plates seeded with PXB cells on day 2. On day 0, HBV/C were added to each well and plates were stored in a CO₂ incubator. After culturing the cells for 20–28 h, culture medium was removed, the cells were washed with PBS, then fresh medium and compounds were replenished. The culture supernatants were harvested on days 12–14 and were subjected to further analysis.

The amount of HBcrAg corresponds to the total amount of HBV DNA and therefore proliferation of HBV, and the amount of HBsAg is correlated to the amount of cccDNA in the nucleus of liver cells, which shows the levels of exclusion of HBV.¹² Both of two antigens serve as important therapeutic indices. The cytotoxicity of these compounds was determined based on reduction of the viability of HepG2 cells,²¹ derived from human hepatocellular carcinoma, by an MTT assay,²² and the optical density was measured by a kinetic microplate reader (Vmax; Molecular Devices, Sunnyvale, CA).

Author contributions

T. K., M. A., M. N., K. T., T. I., Y. M., K. S., K. Y.: investigation; T. K., M. N.: analysis; M. M., H. T.: supervision; T. K., M. A., M. N., K. T., M. M., H. T.: writing.

All authors have given approval to the final version of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.