Discovery and preclinical evaluations of GST-HG131, a novel HBV antigen inhibitor for the treatment of chronic hepatitis B infection

Abstract

Chronic hepatitis B (CHB) remains a significant health challenge worldwide. The current treatments for CHB achieve less than 10% cure rates, majority of the patients are on therapy for life. Therefore, cure of CHB is a high unmet medical need. HBV surface antigen (HBsAg) loss and seroconversion are considered as the key for the cure. RG7834 is a novel, orally bioavailable small molecule reported to reduce HBV antigens. Based on RG7834 chemistry, we designed and discovered a series of dihydrobenzopyridooxazepine (DBP) series of HBV antigen inhibitors. Extensive SAR studies led us to GST-HG131 with excellent reduction of HBV antigens (both HBsAg and HBeAg) in vitro and in vivo. GST-HG131 improved safety in rat toxicology studies over RG7834. The promising inhibitory activity, together with animal safety enhancement, merited GST-HG131 progressed into clinical development in 2020 (NCT04499443).

Introduction

There are nearly 250 million people living with chronic hepatitis B virus (HBV) infection. Projected 25 % of all chronic HBV carriers develop serious liver diseases such as chronic hepatitis, cirrhosis, and primary hepatocellular carcinoma. Nearly a million death due to hepatitis B reported by WHO in 2019.¹ Chronic hepatitis B (CHB) therapy targets to improve patient quality of life and overall survival by preventing the progression of the disease to premature death. CHB infection marks a high viral load (HBV DNA) and even higher levels of non-infectious virions containing the tolerogenic viral surface antigen (HBsAg). There are two classes of the approved therapeutic agents, i.e. nucleos(t)ide analogues and interferon-α, for the treatment of CHB.² The current therapy effectively reduces viremia but rarely results in HBsAg loss.³ There is an urgent need to identify novel agents that reduce HBsAg levels and restore HBV virus-specific immunity in patients to achieve HBV cure.⁴.

There are numerous means being explored in clinic for reduction of HBsAg, including HBV-RNAi. However, an oral agent RG7834 attracted our attention. RG7834 is a novel, orally bioavailable small molecule belongs to dihydroquinolizinone (DHQ) chemical class which selectively reduced HBV antigens (both HBsAg and HBeAg) expression both in vitro and in vivo.⁵ RG7834 is clearly differentiated from the existing small molecule therapies like nucleos(t)ides, which have no significant effect on HBV antigen reduction. RG7834 had been taken to clinical trials for CHB, however, the trial was halted in phase I without an official explanation. It is reasonable to believe that RG7834 had safety concerns. Because of importance of HBV surface antigen inhibition in HBV management and RG7834 possesses novel mechanism in lowering HBV antigens. We have decided to initiate medicinal chemistry research for a safer and proprietary HBV antigen inhibitor related to RG7834. In this letter, we disclose a series of dihydrobenzopyridooxazepine (DBP) class of compounds. Structure activity relationship (SAR) study resulted in our clinical candidate GST-HG131. Animal toxicology studies of GST-HG131 side-by-side with RG7834 revealed that GST-HG131 was safer than RG7834 in animals. GST-HG131 moved into clinical development and preliminary PK and safety in clinic have been reported.6, 7.

At the early stage of our research, the chemical structure of RG7834 was not known, compound 1 in Fig. 1 was a starting chemical lead selected from a published Roche patent application (WO2015113990). From the chemical structure of compound 1, proprietary compound 2 and 3a were designed to retain the pharmacophores of the compound 1.⁵ Compound 3a also retained stereochemistry of RG7834 at isopropyl carbon (C-7). The syntheses of compound 2 and 3a are depicted in Scheme 1 and Scheme 2 respectively. The preparation of compound 2 started from alkylation of commercially available compound 4 with 1-bromo-3‑methoxypropane to give compound 5. Ester hydrolysis and chlorination converted compound 5 to benzoyl chloride 6. Compound 6 was transformed into pyrone 8 by reacting with ketoester compound 7. Desired compound 2 was generated from reaction of pyrone 8 with 2-methylaniline in acetic acid, followed by ester hydrolysis.

Compound 3a was prepared from pyrone 9 as depicted in Scheme 2. Pyrone 9 was obtained using the similar pathway to compound 8 starting from a suitably substituted benzoyl ester (see supplementary material for details). Reaction of pyrone 9 with (R)-valinol in acetic acid provided 4-pyridone 10a. Compound 10a was converted to methanesulfonate, then hydrogenation to remove O-benzyl group, providing a phenol 11a. Cyclization of compound 11a, followed by hydrolysis yielded the desired dihydrobenzopyridooxazepine compound 3a.

Antiviral activity of compound 2 and 3a was first assessed in vitro in HBV transfected HepG2.2.15 cells. Table 1 lists their HBV-DNA and HBsAg inhibitory EC₅₀ data. Compound 2 showed some inhibitory activity for both HBV-DNA and HBsAg, but their EC₅₀s are over 100-fold less than that of compound 1. Compound 3a, however displayed respectable potency for the inhibition of both HBV-DNA and HBsAg. However, the potency is still about 10-fold weaker when compared to the reference compounds (compounds 1 and RG7834). Further chemical optimization of compound 3a is necessary.

Optimization of compound 3a started with C-7 substituents (R¹) SAR exploration. Compound 3b-h were prepared, their SARs were are summarized in Table 2. The syntheses of compounds 3b-g followed the preparation of compound 3a as shown in Scheme 2. Starting from pyrone 9, reaction with commercial (R)- amino alcohols provided corresponding of 4-pyridone derivatives 10b-g. Similar reaction sequence and reaction conditions provided compounds 3b-g. For compound 3 h, the reaction sequence for preparation of compound 3a did not provide the desired compound. The steric bulkiness of tert-butyl group made the ring closure step unfruitful. An alternative pathway was developed as shown in Scheme 3. The synthesis started from commercial compound 12a, which reacted with a commercial cyclic sulfamidate 13 to produce aldehyde 14a with excellent yield. Acidic removal of the BOC group and basification resulted in an intramolecular annulation to generate a dihydrobenzoxazepine 15a. Subsequently, reaction of the dihydrobenzoxaepine 15a with a Danishefsky's diene 16a produced tetrahydrobenzopyridooxazepine 17a. Compound 17a underwent oxidation with p-chloranil to provide dihydrobenzopyridooxazepine compound 18a and hydrolysis generated desired compound 3h.

Antiviral activities of the compounds 3b-h and RG7834 in HBV transfected HepG2.2.15 cells are summarized in Table 2. Comparing to the antiviral potency of compound 3a with compounds 3b-h, the proper size of C-7 substituents (R¹) are important for antiviral activity. Smaller substituents (methyl in 3b, ethyl in 3c) than isopropyl decreased potency. Larger substituents, such as isobutyl (3d) and phenyl (3g), than isopropyl attenuated potency as well. It was puzzling that cyclopropyl 3e, cyclobutyl 3f with somewhat similar size to isopropyl also showed reduced potency. To our delight, tert-butyl substituted compound 3h displayed the most potent activity in both HBV-DNA (EC₅₀: 0.5 nM) and HBsAg (EC₅₀: 0.7 nM) inhibition in this assay. This represented about 10-fold potency improvement from compound 3a.

Because of their excellent in vitro activity, compounds 3a and 3h were chosen for mouse pharmacokinetics studies before in vivo evaluations. Table 3 summarizes mouse PK parameters of both compounds. Compound 3a showed low clearance, and excellent oral exposure (C_max and AUC), but compound 3h showed surprisingly low oral plasma exposure and oral bioavailability at the same dosage given. Although exact reasons are difficult to pinpoint, comparing compound 3h with compound 3a, the ClogP slightly increased because of an extra methyl group in 3h (ClogP: 4.3 vs 3.9 for 3h and 3a respectively). To reduce ClogP and optimize PK properties, C-2 chlorine atom was replaced with acetyl (compound 19b) and methoxy (compound 19c) group to provide compounds with reduced ClogP (Table 4). Both compounds 19b and 19c were prepared similarly to the preparation of compound 3h, starting from different benzaldehydes (12b and 12c) as shown in Scheme 3.

Both compound 19b and 19c displayed excellent antiviral potency in inhibiting both HBV-DNA and HBsAg as shown in Table 4. With their excellent potency, rodent pharmacokinetics property of compounds 19b and 19c were examined. Their results are shown in Table 5. Both compounds had reasonable plasma clearance, excellent oral exposure in mice, but in rats, compound 19b showed surprisingly poor oral plasma exposure (C_max and AUC) and oral bioavailability. Gratifyingly, compound 19c showed excellent oral exposure in rats. PK properties of compound 19c (GST-HG131)⁸ in Beagle dogs were also assessed. It exhibited low clearance, moderate plasma half-life (T_1/2), high plasma exposure (C_max and AUC) and oral bioavailability.

Further, in vitro ADMET properties of GST-HG131 (19c) were also assessed. Table 6 summarizes the ADMET properties of GST-HG131. GST-HG131 showed moderate permeability in Caco-2 assay, low rat and human plasma protein binding, low clearance in liver microsomal stability assays (all species), no inhibition of CYP enzymes at concentration greater than 50 μM. Furthermore, time-dependent inhibition (TDI) of CYP enzymes or CYP3A4 induction was not detected. In addition, it also did not inhibit hERG channel.

Next, antiviral activity of GST-HG131 in primary human hepatocytes (PHH) expressing HBV was assessed, it showed excellent HBV antigens (both HBsAg and HBeAg) inhibition (Table 7). The inhibitory EC₅₀ for HBsAg was 28.2 nM; HBeAg was 16.0 nM. The peg-interferon (IFN-α) in the same assay displayed inhibitory potency EC₅₀s of 10.7 and 4.4 IU/mL for HBsAg and HBeAg respectively. GST-HG131 did not inhibit HBV-DNA above 50 % in this assay, contrary to the effects in HBV transfected HepG2.2.15 cells. Entecavir is very potent inhibiting HBV DNA in PHH, but it is ineffective in inhibiting either HBsAg or HBeAg. The potency of GST-HG131 was not affected by addition of human serum in the range of 0 ∼ 50 %. Specificity of antiviral effects of GST-HG131 was evaluated against other viruses, such as hepatitis C virus, herpes simplex virus type 1 and influenza virus H1N1 in cell-based assays, and no inhibitory activity observed (see supplementary material for details). This indicates GST-HG131 is an HBV specific inhibitor. GST-HG131 was also not cytotoxic as its CC₅₀s were greater than 50 μM in several cell lines tested (see supplementary material for details).

Mechanism of action (MOA) was elucidated in HepG2.2.15 cells, a stably expressing HBV cell line. GST-HG131 was tested side-by-side with entecavir, the levels of HBV RNAs were measured by Northern Blot as shown in Fig. 2. GST-HG131 dose-dependently reduced the level of 3.5 kb and 2.4/2.1 kb HBV RNAs. The inhibitory effect on 2.4/2.1 kb HBV RNAs was greater than that of 3.5 kb RNA, with significant effect observed at 6.17 nM concentration (panel (a) in Fig. 2). The inhibition of HBV RNAs was also time-dependent, the maximum effect observed at 16 h after treatment initiation (panel (b) in Fig. 2). As expected, entecavir did not show significant effects on HBV RNAs at any time points. In short, GST-HG131 appeared to inhibit HBV antigens through inhibition of HBV RNAs.

The in vivo activity of GST-HG131 against HBV was evaluated in an AAV/HBV mouse model. In this model, HBV DNA is encapsidated in adeno-associated virus (AAV) capsid, and is delivered into mouse hepatocytes via tail vein injection. HBV DNA in mouse serum usually reaches a plateau around 4 weeks post AAV/HBV inoculation and persists for at least several months, hence establishment of a persistent HBV viremia. In this experiment, the compound treatment was initiated 4 weeks after AAV/HBV infection. GST-HG131 was orally dosed with dosages of 3, 10, and 30 mg/kg, once-a-day for 4 weeks. RG7834 and tenofovir disoproxil fumarate (TDF) were used as positive controls, dosed at 10 and 1 mg/kg respectively. A combination group of GST-HG131 and TDF was also included. Mouse serum was collected for determination of the levels of HBV DNA, HBsAg, HBeAg by PCR or ELISA. The results are illustrated in Fig. 3. GST-HG131 demonstrated dose-dependent reduction of HBsAg (Fig. 3(a)) and HBeAg (Fig. 3 (b)), while did not reduce HBV DNA (Fig. 3(c)) in this model. The magnitude of reduction of both antigens was similar to RG7834 at the same dose level. Addition of TDF to GST-HG131 appeared to deliver marginal effects in lowering HBV-DNA when compared to TDF alone, but failed to provide significant effects in lowering both antigens when compared to GST-HG131 alone. There was no significant body weight changes in all treatment groups when compared to vehicle control, indicating that the mice had a good tolerance to all doses of GST-HG131.

GST-HG131 was evaluated in vitro for its selectivity against a panel of kinases and receptors (see supplementary materials for details). The results showed GST-HG131 neither inhibit any kinases, nor bind to receptors. Other safety assessments of GST-HG131 were also done. It did not form adducts with glucothione in GSH trapping experiment. It was negative in Ames and micronucleus assays indicating low genotoxicity potential.

For in vivo safety, GST-HG131 and RG7834 were evaluated side by side in a 14-day rat preliminary toxicology study. In this study, three doses (100, 300, and 1000 mpk) of both compounds were orally administered to rats once a day for 14 days. Clinical observations, blood chemistry and histopathology were examined. GST-HG131 and RG7834 tolerated well at low dose (100 mpk) and medium dose (300 mpk). At high dose (1000 mpk), RG7834 showed more organ abnormality, such as mild myocardial degeneration of heart, mild central lobular hepatocyte degeneration of liver, mild white myeloid lymphocytopenia of spleen, moderate acute mucosal inflammation of duodenum; on the other hand, GST-HG131 at the same dose level showed only mild white myeloid lymphocytopenia in the spleen. Next, functional observational battery (FOB) tests of both compounds were investigated at medium (300 mpk) and high dose (1000 mpk). The major findings were mydriasis and lower than normal rectal temperature. RG7834 exerted mydriasis at both doses (1/5 rats at 300 mpk and 5/5 rats at 1000 mpk) and lower than normal rectal temperature (36.5 °C in 2/5 rats at 300 mpk and 35.2 °C in 5/5 rats at 1000 mpk), but GST-HG131 showed recoverable lower than normal rectal temperature only at both doses 1 h after dosing in less numbers of rats, and recovered quickly afterward. Both toxicology tests in rats demonstrated GST-HG131 might exhibit better tolerability than RG7834 at the same dose level in animals.

In summary, based on dihydroquinolizinone (DHQ) chemical series related to RG7834, we designed a series of dihydrobenzopyridooxazepine (DBP) series of HBV surface antigen inhibitors. Extensive SAR studies led us to our clinical candidate GST-HG131. It demonstrated excellent HBsAg and HBeAg inhibitory potency in PHH cell lines in vitro. In a mouse AAV-HBV model, it displayed dose-dependent inhibition of HBsAg, HBeAg, and appeared to provide marginal additive inhibitory effect of HBV-DNA when combination with TDF. In a head-to-head comparative toxicology study in rats, GST-HG131 appeared to be safer than RG7834 at the same dose level. The improved safety profile over RG7834, and its efficacious HBV antigen lowering properties in vitro and in vivo, prompted us to move GST-HG131 into clinical development for the treatment of chronic hepatitis B infection in 2020 (NCT04499443). Preliminary human PK and safety in clinic have been reported.⁷.