Preclinical Evaluation of Amphihevir, a First-in-Class Clinical Hepatitis C Virus NS4B Inhibitor

Amphihevir, a benzofuran derivative, is the first reported hepatitis C virus (HCV) nonstructural protein 4B (NS4B) inhibitor that has advanced to clinical trials (currently in phase Ib trial [CTR20170632]). Here, we report the results of a preclinical study of its potency, toxicity, selectivity, drug metabolism and pharmacokinetics (DMPK), and safety profiles.

ABSTRACT

Amphihevir, a benzofuran derivative, is the first reported hepatitis C virus (HCV) nonstructural protein 4B (NS4B) inhibitor that has advanced to clinical trials (currently in phase Ib trial [CTR20170632]). Here, we report the results of a preclinical study of its potency, toxicity, selectivity, drug metabolism and pharmacokinetics (DMPK), and safety profiles. Amphihevir displayed good antiviral activities against genotype 1a (50% effective concentration [EC₅₀] of 0.34 nM) and genotype 1b (EC₅₀ of 1.97 nM) replicons and no cytotoxicity in 12 cell lines derived from animals and humans. Amphihevir was found to be inactive against other viruses, human kinases, and G protein-coupled receptors (GPCRs), which implies its good selectivity. A 9-day long-term treatment of genotype 1b replicons with amphihevir resulted in a 3.8 Log₁₀ decline of the hepatitis C viral RNA at a concentration of 25× EC₉₀. Drug resistance screening showed that mutations occurred at H94, F98, and V105 of NS4B, which mediated the resistance to amphihevir. This result suggests that NS4B is the main target of amphihevir. There was no cross-resistance between amphihevir and NS5A, NS3/4A, and NS5B inhibitors, suggesting that amphihevir in combination with other anti-hepatitis C virus drugs could be used to treat hepatitis C, as proven by studies of amphihevir and other hepatitis C virus inhibitors. Pharmacokinetic studies demonstrated that amphihevir has good oral bioavailability and appropriate half-life (t_1/2) in rats and dogs, thereby supporting its use once per day. Finally, amphihevir showed good safety profiles in rats and dogs. The results shed light on the use of amphihevir as a potential treatment option for chronic hepatitis C patients.

INTRODUCTION

Chronic hepatitis C (CHC) is an infectious disease caused by the pathogen hepatitis C virus (HCV). It is estimated that 185 million people are infected with HCV worldwide (1), and in 2015, 399,000 deaths caused by diseases associated with HCV infection were reported. Generally, after 20 years of HCV infection, 15% to 30% of patients will develop cirrhosis and 4% of patients will develop hepatocellular carcinoma (HCC) (2). In recent years, several direct-acting antivirals (DAAs), including a nonstructural protein 3/4A (NS3/4A) inhibitor (simeprevir), an NS5A inhibitor (daclatasvir), and an NS5B inhibitor (sofosbuvir), and all-oral combinations, such as ledipasvir-sofosbuvir (Harvoni) or elbasvir-grazoprevir (Zepatier), have been approved. These approvals have resulted in over 90% sustained virologic response (SVR) in CHC patients in the United States. However, the goal proposed in the global health sector strategy (GHSS) of the World Health Organization (WHO), which states that viral hepatitis, including HCV, should be eliminated as a public health threat by 2030 (i.e., 90% reduction in incidence and 65% reduction in mortality), remains distant.

From a functional point of view, HCV proteins can be divided into assembly modules (core, envelope proteins E1 and E2, p7 protein, and NS2) and replication modules (NS3, NS4A, NS4B, NS5A, and NS5B), but some proteins in the replication modules are also involved in HCV assembly, whose mechanism is still unclear (3–7). NS4B, an integrated membrane protein, is a hydrophobic, nonstructural protein containing 261 amino acids with a molecular weight of 27 kDa. Two amphipathic helices (AH) at the N terminus of the protein mediate membrane binding for lipid vesicle formation (8), while a C-terminal arginine-rich motif mediates its binding with negative-stranded RNA. Both activities are essential for the replication of HCV (9). NS4B also harbors a domain that binds to the functional nucleotide GTP. HCV replication is inhibited when GTP is disabled (10). These replication-related functions of NS4B enable its presentation as an attractive drug target for the treatment of CHC.

By analyzing the core structures of several reported compounds (11–16), we designed and synthesized a novel benzofuran derivative and first-in-human NS4B inhibitor, amphihevir (YS_YD_0250). The chemical structure is shown in Fig. 1. By enrolling six groups of healthy volunteers in a phase Ia clinical trial to evaluate its efficacy (single-dose escalation, multidose phase I clinical trial of amphihevir soft capsule [CTR20170085]), we found that amphihevir displays excellent antiviral activity in genotype type 1a (GT1a [HCV-1a]) and GT1b (HCV-1b) replicons, acceptable pharmacokinetic properties, and good safety characteristics in rats and dogs. More importantly, amphihevir in combination with other HCV antiviral drugs exhibited additive or highly synergistic effects, suggesting that it can be used as a component in the development of novel CHC treatment regimens. Amphihevir is currently undergoing a phase I clinical trial in China (i.e., a phase Ib clinical trial) to evaluate the initial effectiveness of soft capsules for its administration in HCV patients.

FIG 1.

Chemical structure of amphihevir. Concentrations for 50% of maximal effect (EC₅₀s): 0.34 nM for GT1a; 1.97 nM for GT1b; 186 nM for GT2a.

Six groups of healthy volunteers were enrolled in a phase Ia clinical trial to evaluate a first-in-human NS4B inhibitor, amphihevir. Recently, a phase Ib clinical trial was launched in China to evaluate the initial effectiveness of amphihevir soft capsules in HCV patients.

RESULTS

In vitro activity and selectivity of amphihevir.

In the luciferase assay, the average 50% effective concentration (EC₅₀) of amphihevir against the GT1a (H77) replicon was 0.34 ± 0.06 nM (mean ± standard deviation) when cultured in a medium with 10% fetal bovine serum (FBS), while the EC₅₀s of the GT1b (Con1) replicon were 1.97 ± 0.41 nM and 1.12 ± 0.52 nM in the enzyme and reverse transcription-quantitative PCR (qRT-PCR) experiments, respectively. The activity of amphihevir against GT2a (JFH-1) was much more moderate, with an EC₅₀ of 186 nM. The effect of normal human serum on the antiviral activity of amphihevir was tested using HCV-1b replicon cells (cell lines referred to as “replicon cells” are permissive to the replication of HCV replicons) in culture medium containing 10%, 20%, and 50% human serum. Compared to the EC₅₀ for 10% FBS, which was used as the control, the EC₅₀s increased by 1.50-, 2.56-, and 5.30-fold in medium containing 10%, 20%, and 50% normal human serum, respectively. Linear regression analysis estimated that the EC₅₀s of amphihevir to GT1a and GT1b replicons were 3.13 nM and 18.16 nM with 100% human serum.

We further evaluated the anti-HCV activity of long-term treatment with amphihevir. Treating HCV-1b replicon cells with amphihevir for 9 days resulted in a continuous and concentration-dependent decline in replicon RNA. Also, replicon RNA rebound was not observed during the treatment period. At a concentration 25-fold greater than the EC₉₀ (300 nM), amphihevir reduced replicon RNA by nearly 6,300-fold (3.8 log₁₀) (Fig. 2). The compound 1b {3-(1-(3-chloro-6-isopropyl-8-(trifluoromethyl)imidazo[1,2-a]pyridine-2-carbonyl)piperidin-4-yl)oxazolidin-2-one} reported by GSK (15) was chosen as the reference and was found to cause a concentration-dependent reduction in replicon RNA of approximately 1,580-fold (3.2 log₁₀) at a concentration of 2,750 nM. The results indicate that amphihevir causes a stronger continuous inhibition of viral activity.

FIG 2.

Inhibition of HCV replicons after a 9-day continuous treatment. Amphihevir (YS_YD_0250) was used to treat stable HCV GT1b replicon cells for 9 days. The cells were passaged every 3 days, and medium was replaced with fresh medium containing the corresponding concentration of compounds. Cells were collected for cellular RNA extraction, and the RNA was then used to determine the inhibitory activity of the compound against HCV replicons. The amounts of HCV replicon RNA were determined by qRT-PCR.

We proceeded to determine the antiviral selectivity of amphihevir. The cytotoxicity of amphihevir against 10 cell lines derived from different human tissues and 2 animal cell lines was first tested. The 50% cytotoxicity concentrations (CC₅₀s) of amphihevir against these cell lines were greater than 50 μM, implying its good in vitro safety profile. To verify its specificity for HCV, the antiviral activities of amphihevir against different types of viruses and its protein inhibitory activities with respect to kinases and G protein-coupled receptors (GPCRs) were determined. Seven different viruses (influenza virus, respiratory syncytial virus, human rhinovirus, enterovirus 71, hepatitis B virus, herpes simplex virus type 1, and human immunodeficiency virus type 1), including positive- and negative-strand RNA viruses, DNA viruses, retroviruses, and hepadnaviruses, were insensitive to amphihevir, with EC₅₀s greater than 50 μM. Amphihevir also showed no obvious activities against a panel of 24 protein kinases, as well as 21 human GPCRs, except for endothelin receptor Eta (IC₅₀ of 4.7 μM) and vasopressin receptor V1a (IC₅₀ of 9.9 μM). These findings indicate that amphihevir is a highly selective inhibitor of HCV.

Cross-resistance and drug resistance mutations in the NS4B study.

Cross-resistance is an important parameter for predicting the potential of combining a newly developed antiviral with other HCV drugs. A series of representative HCV GT1b replicons containing different mutations in NS3/4A, NS4B, NS5A, and NS5B was selected to evaluate their susceptibility to amphihevir. Only replicons harboring mutations in NS4B, including H94R (a change of H to R at position 94), F98C, and V105M, were resistant to amphihevir, while other mutated replicons whose mutations mapped to NS3/4A, NS5A, and NS5B were as sensitive as the WT counterpart to amphihevir. These results suggest that amphihevir, as an NS4B inhibitor, was not cross-resistant with other HCV drugs, including NS3/4A inhibitors, NS5A inhibitors, and NS5B inhibitors (Table 1). Hence, amphihevir could be used in combination with these inhibitors for the development of new anti-HCV therapies.

TABLE 1.

HCV GT1b genomic mutations identified after selection for amphihevir resistance

Protein	Mutationa	EC50 (nM)	Fold change
NS4B	None (wild type)	0.338	1
	H94R	13.1	38.7
	F98C	19.59	57.9
	V105M	32.36	95.6
NS5A	L31F	0.297	0.9
	Y93H	0.303	0.9
NS3	Q80R	0.274	0.8
	R155K	0.354	1
	A156S	0.296	0.9
NS5B	S282T	0.357	1.1

^aA, alanine; C, cysteine; F, phenylalanine; H, histidine; K, lysine; L, leucine; M, methionine; Q, glutamide; R, arginine; S, serine; T, threonine; V, valine; Y, tyrosine. In NS4B with mutation H94R, the histidine of the 94th residue was mutated to arginine, while the remaining residues were unchanged; the other mutations are annotated similarly.

We proceeded to perform an in vitro resistance selection assay by serial passage of stable replicons in the presence of amphihevir at concentrations ranging from 6 to 30 nM and 40 to 200 nM for the GT1a and GT1b replicons, respectively. Twenty replicon clones selected from each concentration were sequenced, and 26 mutations of the NS4B gene were found in drug resistance screening of HCV-1a, 7 of which occurred more than three times (Table 2). Phe98, including mutations F98C and F98L, had the highest mutation frequency, followed by V105M. A total of 11 mutations of the NS4B gene were found in drug resistance screening of HCV-1b, 3 of which occurred more than three times (Table 3). The mutation with the highest frequency was H94R, followed by V105M and F98L (Table 3).The cross-resistance and drug resistance mutation profiles of amphihevir were highly similar to those of other recently reported NS4B inhibitors, including anguizole, GSK8853, PTC725, and our previously reported imidazo[2,1-b]thiazole NS4B inhibitors (11, 14, 17, 18). This finding suggests that amphihevir mainly acts on the NS4B protein of HCV.

TABLE 2.

Amino acid mutations of the NS4B protein of GT1a replicon clones and their frequencies after screening against amphihevir

GT1a NS4B mutationa	Total no. of mutations	No. of mutations at amphihevir concn (nM) of:
		6	15	30
F98C	8	2	1	5
F98L	8	3	3	2
V105M	7	2	5	0
K52R	5	3	1	1
W50L	4	2	1	1
H31Q	3	1	1	1
F98Y	3	1	1	1

^aC, cysteine; F, phenylalanine; H, histidine; K, lysine; L, leucine; M, methionine; Q, glutamine; R, arginine; V, proline; W, tryptophan; Y, tyrosine.

TABLE 3.

Amino acid mutations in the NS4B protein of GT1b replicon clones and their frequencies after screening against amphihevir

GT1b NS4B mutationa	Total no. of mutations	No. of mutations at amphihevir concn (nM) of:
		40	100	200
H94R	17	5	5	7
V105M	10	2	4	4
F98L	5	2	1	2

^aF, phenylalanine; H, histidine; L, leucine; M, methionine; R, arginine; V, valine.

Combination effects with other anti-HCV drugs.

To prevent the emergence of drug resistance mutations, a combination of drugs is essential for the treatment of CHC. We sought to evaluate the combined use of amphihevir with other direct antiviral drugs, including the marketed drugs interferon (IFN), ribavirin (RBV), an NS3/4A inhibitor (simeprevir), and an NS5B inhibitor (sofosbuvir), as well as an NS5A inhibitor (fopitasvir) that was developed by our group and has entered phase II clinical trials in China. The antiviral activity and selectivity data of fopitasvir are shown in Table 4.

TABLE 4.

Antiviral activity and selectivity data of fopitasvir

Parameter	Results
Anti-HCV activity and cytotoxicity	Mean EC50 ± SD (nM) was 0.139 ± 0.059 for GT1a, 0.0089 ± 0.004 for GT1b, 0.030 ± 0.016 for GT2a, 2.190 ± 0.633 for GT3a, 0.007 ± 0.001 for GT4a, 0.017 ± 0.007 for GT5a, and 0.161 ± 0.071 for GT6a; CC50 was >3,000 nM
Virus specificity	No inhibitory activity against the seven viruses tested (influenza virus, respiratory syncytial virus, human rhinovirus, enterovirus 71, hepatitis B virus, herpes simplex virus type 1, and human immunodeficiency virus type 1); EC50s were >1,000 nM
Protein kinase selectivity	No obvious inhibitory effect on 24 protein kinases at 1 μM and 10 μM, good biosafety selectivity for protein kinases
GPCR selectivity	There were no significant agonistic or antagonistic effects on the 21 GPCR targets tested, and the EC50s (agonistic test) and IC50sa (antagonistic test) were both greater than the highest concn tested (EC50 > 12 μM; IC50 > 10 μM); fopitasvir should have good safety against the off-targets tested

^aIC₅₀, 50% inhibitory concentration.

MacSynergy software was used to estimate the synergy indices for drug combinations. A positive index indicated a synergistic effect, and a negative index indicated an antagonistic effect. The results of combining amphihevir and different HCV drugs are shown in Table 5. The synergy volume for amphihevir and fopitasvir was 117.44 μM²%, indicating a highly synergistic drug-drug interaction. The synergy volumes for amphihevir in combination with IFN, ribavirin, and sofosbuvir were between 28.32 μM²% and 85.67 μM²%, thereby indicating mild to moderate synergistic effect. The test concentrations are shown in Table 6. The combination of amphihevir and simeprevir resulted in an additive interaction, as indicated by a synergy volume of 7.05 μM²%. No obvious cytotoxicity was observed within the range of these drug combinations.

TABLE 5.

Combination effect index of amphihevir in GT1b replicon cell experiment

Compound tested (protein targeted) in combination with amphihevir (targeting NS4B)	Combination effect index (μM2% [95% credibility])a		Comprehensive effect
	Synergistic	Antagonistic
TMC435 (NS3)	7.05	0	Additive
Fopitasvir (NS5A)	117.44	0	Highly synergistic
Sofosbuvir (NS5B)	85.67	−5.3	Moderately synergistic
Ribavirin	56.38	0	Moderately synergistic
IFN-α	28.32	0	Mildly synergistic

^aCombination effect index descriptions: a positive index represents a synergistic effect, and a negative index represents an antagonistic effect; an absolute value of the index that is <25 indicates additive effect; an absolute value of the index that is in the range from 25 to 50 indicates mild synergy or antagonism; an absolute value of the index that is in the range from 50 to 100 indicates a moderate synergistic or antagonistic effect and may have important implications for in vivo effects; an absolute value of the index that is in the range from 100 to 1,000 indicates highly synergistic or antagonistic effects.

TABLE 6.

Orthogonally proportioned concentrations of amphihevir and interferon, ribavirin, and sofosbuvir used in testing drug combinations^a

Compound	Concn (nM)
Amphihevir	25	12.5	6.25	3.13	1.56	0.78	0.39	0
Ribavirin or sofosbuvir	25	12.5	6.25	3.13	1.56	0.78	0.39	0
Interferon	50	25	12.5	6.25	3.13	1.56	0.78	0

^aAmphihevir, sofosbuvir, and ribavirin were formulated into a 10 mM mother liquor in 100% DMSO and stored in a nitrogen cabinet. Each compound was first diluted twice. At the time of testing, amphihevir was further orthogonalized with another compound at 8 different concentrations (including a blank control).

We further investigated the effect of amphihevir in combination with fopitasvir on the formation of resistant GT1b clones. When these drugs were used together, the number of clones of drug-resistant replicon cells decreased significantly compared to the number of clones resistant to either drug singly. Amphihevir and fopitasvir were combined at three concentrations of 12 nM, 36 nM, and 108 nM and 0.05 nM, 0.15 nM, and 0.45 nM (about 5×, 15×, and 45× EC₅₀), respectively (Fig. 3). This finding suggests that the combination of amphihevir and fopitasvir could significantly reduce the emergence of resistant replicon clones even in the minimum-concentration combination of 5× EC₅₀ of each compound. Hence, amphihevir-based combination therapies, especially the combination of amphihevir and fopitasvir, could be used to achieve favorable anti-HCV effects.

FIG 3.

Inhibition of drug-resistant HCV replicons by the combination of amphihevir and fopitasvir. Stable HCV GT1b replicon cells were treated with amphihevir or fopitasvir alone or in combination for 21 days in the presence of G418. Replicon cell colonies are indicated by crystal violet staining.

PK in rats and dogs.

To investigate the pharmacokinetics (PK) of amphihevir monotherapy in animals and its effects in combination with fopitasvir, we performed PK studies in rats and dogs.

In the rat experiment, all rats tolerated amphihevir well and did not show any abnormal behavior. There was no significant gender difference (greater than 0.5- or less than 2-fold) in any of the three groups with respect to the systemic exposure (area under the concentration-time curve from time zero to last time of measurement [AUC_0–last] and maximum concentration of drug in serum [C_max]). When the dose increased from 3 mg/kg of body weight to 10 mg/kg (3.33 times) and 30 mg/kg (10 times), the systemic exposure also increased but to a lesser extent (AUC_0–last increased by 2.37-fold and 6.58-fold, respectively). We observed that the half-life of the drug increased significantly and the absolute bioavailability decreased substantially as the dosage increased; this could be explained by the poor solubility of the drug. In fact, at the highest dosage of 30 mg/kg, a substantial portion of the drug could be observed in the gastrointestinal tract in an insoluble particulate state after dissection. This may possibly have led to the slow release of amphihevir and, ultimately, to prolongation of the half-life of the drug. Moreover, after repeated dosing for 7 days, there was no significant drug accumulation in male and female Sprague Dawley (SD) rats, as indicated by the PK parameters in the 10-mg/kg group (Table 7).

TABLE 7.

Rat PK parameters

Parametera	Value for indicated group (dosage and mode of administration)b
	1 (1 mg/kg, single dose, i.v.)		2 (3 mg/kg, single dose, p.o.)		3 (10 mg/kg, p.o.)				4 (30 mg/kg, single dose, p.o.)
					Single dose		7-day repeated dosing (end value)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
C0 or Cmax (ng/ml)	673	171	311	70.6	507	163	345	75.4	899	428
Tmax (h)	—	—	2.00	1.10	1.50	0.548	3.84	3.61	2.67	1.03
t1/2 (h)	1.15	0.318	2.74	0.654	4.16	0.740	4.7	1.47	8.98	0.746
Vss (liters/kg)	1.67	0.173	—	—	—	—	—	—	—	—
CL (ml/min/kg)	17.8	3.03	—	—	—	—	—	—	—	—
AUC0–last (h · ng/ml)	959	197	1,610	339	3,820	1,330	3,205	777	10,600	5,570
AUC0–inf (h · ng/ml)	965	200	1,630	327	3,900	1,370	3,510	545	12,700	6,490
Bioavailability (%)	—	—	65.1	—	40.1	—	—	—	43.1	—

^aC₀, initial blood concentration after injection of the drug for 0.033 h; C_max, maximum concentration of drug in serum; T_max, time to maximum concentration of drug in serum; t_1/2, half-life; V_ss, volume of distribution at steady state; CL, clearance; AUC_0–last, area under the concentration-time curve from time zero to last time of measurement; AUC_0–inf, AUC from time zero to infinity. Bioavailability was calculated from AUC_0–inf and theoretical doses.

^bp.o., oral administration (per os). —, the value for the parameter was not calculated.

With a 1-mg/kg single-dose intravenous (i.v.) injection to beagle dogs, the total clearance rate was 8.14 ± 2.82 ml/min/kg (about 26.3% of hepatic blood flow). The t_1/2 values for clearance were 4.70 ± 2.36 h and 4.20 ± 0.445 h in male and female beagle dogs, respectively. In the fasting state, beagle dogs were orally administered 3, 10, and 30 mg/kg of body weight of amphihevir (prepared as soft capsules). The systemic exposure and C_max increased as the dosage increased, while the absolute bioavailabilities were relative stable, between 12.6% and 18.5% (Table 8), markedly less than those of rats (40 to 65%). The difference in the drug formation between these two species would be a possible explanation for these observations and the differences in drug absorption characteristics between the species and could also contribute to the decline in bioavailability to a certain extent.

TABLE 8.

Dog PK parameters

Parametera	Value for indicated group (dosage and mode of administration)b
	1 (1 mg/kg, single dose, i.v.)		2 (3 mg/kg, single dose, p.o.)		3 (10 mg/kg, p.o.)				4 (30 mg/kg, single dose, p.o.)
					Single dose		7-day repeated dosing (end value)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
C0 or Cmax (ng/ml)	555	70.3	145	38.0	438	230	371	189	681	285
Tmax (h)			1.25	0.612	1.50	0.548	1.50	0.548	1.33	0.516
t1/2 (h)	4.45	1.54	4.38	1.65	4.72	1.80	4.51	3.22	11.3	8.30
Vss (liters/kg)	2.30	0.350	—a	—	—	—	—	—	—	—
CL (ml/min/kg)	9.30	2.63	—	—	—	—	—	—	—	—
AUC0–last (h · ng/ml)	1,880	516	758	230	3,520	2,410	2,560	1,700	7,150	4,010
AUC0–inf (h · ng/ml)	1,930	587	787	243	4,260	3,070	2,880	2,040	10,400	7,100
MRT0−last (h)	1,880	516	5.33	1.09	7.61	1.98	7.19	1.81	8.17	2.38
MRT0−inf (h)	1,930	587	6.18	1.37	10.7	3.27	9.27	3.28	16.8	10.3
Bioavailability (%)	—	—	13.8	5.47	18.5	13.3	13.3	8.51	12.6	7.16

^aC₀, initial blood concentration after injection of the drug for 0.033 h; C_max, maximum concentration of drug in serum; T_max, time to maximum concentration of drug in serum; t_1/2, half-life; V_ss, volume of distribution at steady state; CL, clearance; AUC_0–last, area under the concentration-time curve from time zero to last time of measurement; AUC_0–inf, AUC from time zero to infinity; MRT_0–last, mean residence time from time zero to last measurement; MRT_0–inf, MRT from time zero to infinity. Bioavailability was calculated from AUC_0–inf and theoretical doses.

^bp.o., oral administration (per os); —, the value for the parameter was not calculated.

By analyzing PK parameters, amphihevir’s active pharmaceutical ingredients (API) and its administration in soft capsules were found to have moderate to good absorption properties in rats and dogs, thus supporting their clinical administration once per day as was determined using data from a phase I clinical trial. As determined by the PK parameters of rats and beagle dogs, oral administration of amphihevir at 288 mg/kg and 300 mg/kg or more once a day in the human body can maintain the blood concentration above the effective concentration. In fact, because amphihevir has a certain cumulative effect in the liver, its effective dose may be lower than the estimated dose.

Safety.

To support early clinical trials and examine the safety of drugs in two animal species, a 4-week toxicity study was conducted with rats and dogs. In the 4-week toxicity study with rats, at the end of administration, the number of white blood cells (lymphocytes) in the 200-mg/kg group was slightly elevated relative to the number in the control group but recovered to normal at the end of the recovery period. There were no obvious abnormalities in rats in the amphihevir group when general status, body weight, food intake, ophthalmology, hematology, blood biochemistry, urine examination, organ weight and the ratio of organ to body weight, gross anatomy, and histopathological examination were assessed. In the dose range of 20 to 200 mg/kg, the exposure of female rats was slightly higher than that of male rats. The proportions of exposure of female and male rats were lower than those of the dose increases. There was no obvious accumulation during 4 weeks of continuous administration. Rats were orally administered 20, 60, and 200 mg/kg of amphihevir for 4 weeks. The no observed adverse effect level (NOAEL) dose was 200 mg/kg. The mean AUC in female rats after the last dose was 11,689.05 h · ng/ml, while that of males was 7,854.32 h · ng/ml.

When beagle dogs were orally administered 20, 60, and 200 mg/kg of amphihevir for 4 weeks, their general status, weight, food intake, electrocardiogram, blood pressure, body temperature, ophthalmology, hematology, blood biochemistry, urine, bone marrow smear, organ weight and the ratio of organ to body weight, gross anatomy observation, and histopathological examination indicated no significant abnormal changes. During the administration period, 2, 3, 5, and 8 dogs in the control group and the 20-, 60-, and 200-mg/kg administration groups, respectively, experienced vomiting 5 to 60 min after administration. (The proportions were 2/10, 3/10, 5/10, 8/10, respectively.) Their vomit consisted of either a white or yellow foam, liquid, or food. The incidence of adverse events increased with increasing doses of amphihevir in each group, which may be drug-induced adverse events.

In the dose range of 20 to 200 mg/kg, a significant gender difference was not found in dogs following amphihevir exposure. The proportions of increased exposure were lower than the rate of dose increase, and we could not find an obvious accumulation after 4 weeks of continuous administration. Beagle dogs were orally administered 20, 60, and 200 mg/kg of amphihevir for 4 weeks, and the NOAEL dose was 200 mg/kg. The mean AUC in female dogs after the last dose was 40,166.26 h · ng/ml, while that for male dogs was 32,850.22 h · ng/ml.

At the 10- and 30-mg/kg PK doses and 200-mg/kg NOAEL doses in rats, drug exposure values were 2.37, 6.58, and 6.07 times higher than for the 3-mg/kg dose, respectively. At the 10- and 30-mg/kg PK doses and the NOAEL dose of 200 mg/kg in beagle dogs, drug exposure showed increases of 4.63, 9.43, and 48.16 times, respectively, compared to that of the 3-mg/kg dose. The possible cause is the difference in animal species and the different degrees of saturation in vivo for the drug. An exposure plateau was observed in rats at doses above 30 mg/kg, which indicates that the exposure will not increase as dose increases. However, there was no plateau in beagle dogs, although the increased proportion of exposure was less than the dose increase ratio.

DISCUSSION

NS4B is a nonstructural protein that induces the formation of a membrane network which acts as a scaffold for HCV replication complexes and can interact with other nonstructural proteins to bind viral RNA (19, 20). NS4B also has nucleoside triphosphatase (NTPase) activity and plays an important role in viral assembly (6, 21). These important roles in HCV replication make it an attractive target in the development of new HCV inhibitors. Since the first small-molecule inhibitor against NS4B was reported in 2008, many pharmaceutical companies have developed multiple NS4B inhibitors with different backbones. These inhibitors can be roughly classified into two types according to their mechanism of action. The first type is a set of compounds represented by clemizole, which inhibits viral replication by inhibiting the attachment of NS4B to viral negative-stranded RNA. The other type is a panel of 4BAH2 inhibitors represented by anguizole and GSK8853, which inhibit viral replication by inhibiting dimerization of the NS4B protein and inducing its intracellular rearrangements. Only clemizole entered clinical trials, but the trial was terminated in phase I due to the low activity and narrow safety window of the compound. In this study, we report the results of a preclinical study with amphihevir, a first-in-class clinical 4BAH2 inhibitor.

Amphihevir exhibited excellent inhibitory activity against GT1a and GT1b replicons but poor activity against GT2a replicons; this is highly similar to properties of other previously reported NS4B inhibitors, such as GSK8853 and PTC725. These results suggest that NS4B inhibitors that are currently being developed may be suitable for patients with hepatitis C caused by HCV GT1a and GT1b, which account for most CHC patients in most countries. By investigating the cytotoxicity of amphihevir against 12 cell lines and the inhibitory activities of various human GPCRs and kinases, we confirmed that this compound is a highly specific HCV inhibitor, ultimately suggesting its high selectivity and safety.

Considering the structural similarity of amphihevir to our previously reported imidazole [2,1-b]-thiazole compounds and GSK8853, we speculated that they share similar mechanisms of action and resistance profiles. Therefore, cross-resistance studies and drug-induced drug resistance mutation screening studies were conducted. The cross-resistance studies showed that replicons containing mutations at the H94, F98, and V105 sites on the NS4B protein were resistant to amphihevir, GSK8853, PTC724, and imidazo[2,1-b]-thiazole compounds, suggesting that they may target the NS4B protein. Further drug-induced mutation screening studies confirmed that amphihevir-induced resistance mutations of HCV GT1a and GT1b replicons were mainly located at H94, F98, and V105 of NS4B, which is similar to findings for GSK8853 and PTC724. Because the amino acids in these sites in the GT2a replicon are replaced by other amino acids (T94, L98, and L105), this implies inherent resistance to current NS4B inhibitors and explains the poor HCV GT2a replicon activity of these NS4B inhibitors. Mutations at sites such as K52, W50, and H31 also caused resistance to amphihevir, which was not observed in GSK8853 and PTC724, suggesting that the resistance spectrum of NS4B inhibitors and their resistance barriers could be optimized via structural changes.

Combination therapy is essential to prevent the emergence of viral resistance mutations when treating hepatitis C. In this study, we examined the combined effects of amphihevir and other representative HCV drugs available on the market. The results showed that these combinations had additive or synergistic effects, thereby indicating the feasibility of developing a combination therapy for hepatitis C based on our NS4B inhibitor. It is worth noting that amphihevir exhibits a highly synergistic effect when combined with our NS5A inhibitor, fopitasvir, and this combination can effectively inhibit the formation of resistant clones at different concentrations. This finding suggests that amphihevir and fopitasvir as a combined therapy could result in high sustained virologic response when used in the clinical treatment of CHC. Due to novel mechanisms of action, new hepatitis C treatment regimens based on NS4B inhibitors, such as amphihevir, are expected to be used to treat hepatitis C patients who fail or relapse after treatment with other regimens. In addition, there is a new all-oral treatment in preclinical development that is based on these drugs.

PK studies in mice and beagle dogs revealed the excellent oral absorption and acceptable bioavailability of amphihevir. Based on the concentration-time curves, the t_1/2 values derived at different doses, and the effect of serum on amphihevir’s EC₅₀, we hypothesized that the possible starting dose in humans is 1 to 5 mg/kg of body weight. At the presumed efficacious dose in humans, the predicted AUC exposure would be 2,000 to 4,000 h · ng/ml and the C_max would be 200 to 450 ng/ml. The safety results also support the use of this dose as the initial dose before escalation. These results provide a basis for implementing a dose for our phase I clinical trial. However, a patient’s clinical dose should be estimated based on the AUC, C_max, and blood drug concentration at 24 h (C₂₄) indicators of single- and multiple-dose studies performed with healthy people. The phase Ia pharmacokinetic study in healthy volunteers showed that the oral dose of 300 mg of amphihevir could maintain the plasma concentration above EC₉₀ within 24 h, which supports the starting dose of 300 mg in patients in the coming phase Ib trial.

MATERIALS AND METHODS

Compounds.

Amphihevir {5-cyclopropyl-2-[4-(2-oxooxazolidin-3-yl) piperazine-1-carbonyl]-7-(trifluoromethyl) benzofuran-3-carbonitrile} and its soft capsules were prepared by Changzhou Yinsheng Pharmaceutical Company, Changzhou, Jiangsu Province, China, the sponsor of the clinical trials. The company also provided fopitasvir and its tablets for use as the novel anti-HCV NS5A clinical compound. Daclatasvir, entecavir, and BMS-433771 were synthesized by WuXi AppTec (Shanghai) Co., Ltd., Shanghai, China. Telaprevir and sofosbuvir were bought from Shanghai Haoyuan Biotech Co., Ltd. Alpha interferon (IFN-α), ribavirin, paclitaxel, pleconaril, and acyclovir were purchased from Sigma, Shanghai, China. AG7088 was purchased from Axon Medchem BV, Shanghai, China. Oseltamivir was purchased from TRC, Massachusetts, USA. Reference materials, such as histamine, pyrilamine maleate salt, (±)-SKF-38393, etc., for testing GPCR selection were purchased from Sigma. Staurosporine was purchased from LC Laboratories, Massachusetts, USA. Positive reference compounds for GPCR selectivity studies were purchased from Sigma.

Cell lines.

The hepatocellular carcinoma cells, Huh-7 and Huh-7.5.1, were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) containing 10% fetal bovine serum (FBS; Corning), 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. The replicons of HCV used in this paper are all subtype replicons, which contain nonstructural (NS) protein NS3, NS4A, NS4B, NS5A, and NS5B genes and luciferase genes (22, 23) and were established by WuXi AppTec (Shanghai) Co., Ltd. Replicators used to prepare stably transfected cells contain the resistance selection neo gene, while replicators used for transient experiments do not contain this gene. In addition, stable transfection of HCV replicon cells, including wild-type GT1a (HCV-1a [H77]) and GT1b (HCV-1b [Con1]) cell lines, was established by WuXi AppTec. An infectious HCV cell culture system (HCVcc, GT2a, JFH-1) was constructed and provided by Wuhan Virus Research Institute, Chinese Academy of Sciences. Cells used to determine cytotoxicity were from 5 different companies or institutes: Huh-7 cells were from AppTec in the United States; HL-7702 and 16HBE cells were from Shanghai Zhengchuang Industrial Co., Ltd., Shanghai, China; Hep2, A375, MDBK, MDCK, Hela, Colo-205, and Caki-1 cells were from ATCC; MT-4 cells were from Ottawa University, Canada; and CCRF-CEM cells were purchased from ECACC. Twenty-one GPCR targets for GPCR screening were used for an early in vitro study for safety assessment (24).

Reagents and instruments.

Reagents included Bright-Glo luciferase assay reagent (catalog no. E2650; Promega), CellTiter-Fluor cell viability assay reagent (catalog no. G6082; Promega), T7 RNA polymerase (catalog no. P1320; Promega), quantitative PCR detection reagent (TaqMan EZ RT-PCR core reagents, catalog no. N8080236-403028; Applied Biosystems), CellTiter-Glo cell viability assay reagent (catalog no. G7573; Promega), crystal violet (catalog no. C3886; Sigma), RNA extraction kit (catalog no. 74104; Qiagen), cDNA reverse transcriptase (catalog no. 4368814; Applied Biosystems), DMEM (catalog no. 1960; Invitrogen), F-12 medium (catalog no. 11765; Invitrogen), FBS (catalog no. 10099; Invitrogen), G418 (catalog no. 10131-027; Invitrogen), blasticidin (catalog no. A11139; Invitrogen), and a Fluo-4 direct kit (catalog no. F10471; Invitrogen). The 24 protein kinases were purchased from Thermo Fisher (under the Invitrogen brand name).

The main instruments used in these projects were an automated liquid workstation (Echo; Labcyte), an electroporator (Xcell; Bio-Rad), a fluorescence quantitative PCR instrument (model number 7900; Applied Biosystems), a scanner (ScanJet G4050; HP), an automated liquid handling robot (Bravo; Agilent), a plate reader (EnVision 2104; PerkinElmer), a reagent dispenser (Multidrop; Thermo Fisher), 384-microwell black plates (product no. 6007279; PerkinElmer), 384-microwell plates (item no. 781280; Greiner), a cellular screening system (FLIPR; Molecular Devices), a plate assembler (POD 810; Labcyte), and a cell viability analyzer (Vi-Cell XR; Beckman Coulter).

Animals.

Rats for the PK study were purchased from Shanghai Slaccas Company, while dogs were purchased from Beijing Marshall Biotechnology Co., Ltd., Beijing, China. Rats for the safety study were purchased from Charles River Laboratories, Shanghai, China. Animals were handled according to relevant animal welfare guidelines. The experimental animal license was approved by the Science and Technology Department of Jiangsu Province [no. SYXK(SU)2014-0016].

Human subjects.

The use of human subjects was approved by the Ethics Committee of the First Hospital of Jilin University. Clinical trials must follow the Helsinki Declaration (2013 version), the NMPA's Drug Clinical Trial Management Practices (GCP), and related regulations.

In vitro activity and specificity studies.

The anti-HCV activity of amphihevir was determined with a method reported previously (16, 25, 26). Human serum shift effect (27) was used to assess the extent of decline in drug activity in human serum.

All 21 GPCR targets were derived from the GPCR targets for early in vitro safety assessment. To study the 24 kinases belonging to different types, the Z′-Lyte method (Life Technologies) was used to measure the enzyme activities and inhibition rates of the compound to determine whether the compound had inhibitory activity against the 24 kinases.

Nine-day inhibitory activity of amphihevir on HCV replicons.

The inhibitory effect of amphihevir on HCV replicon RNA replication over a 9-day continuous treatment period was measured using HCV-1b replicon cells (28). HCV-1b cells were seeded in culture flasks using DMEM without G418 and with amphihevir (12 nM, 60 nM, or 300 nM) at a final concentration that was approximately 1-fold, 5-fold, or 25-fold greater than the EC₉₀. Cells were passaged on days 3 and 6. Each passage was performed with fresh medium containing the corresponding concentration of compound, and a sample of cells was collected. The culture was terminated on the 9th day, and cells were collected to determine HCV replicon RNA content. After collection of the cell sample, cellular RNA was extracted. The content of HCV replicon RNA in each sample was determined by qRT-PCR. In this experiment, the neo gene in the replicon was used as a target to quantify viral RNA, the cyclophilin A gene in the cell was detected as an internal reference, and the RNA content was homogenized. The experimental method was the same as the qRT-PCR method described above. The ratio of HCV RNA after the action of the compound to the HCV RNA in the compound-free control was calculated to determine the extent of inhibition.

Screening of cross-resistance and drug resistance mutations for NS4B.

Cross-resistance (29) and drug resistance mutations in NS4B (11) were screened as follows. Resistance characteristics for mutant replicons resistant to amphihevir were studied by comparison to the known compounds NS3 protease (simeprevir), NS5A protein inhibitor (dacaltasvir), and NS5B RNA polymerase inhibitor (sofosbuvir). All resistant-mutant replicons were based on the GT1b replicon. Resistance mutations were introduced by point mutation and confirmed by nucleotide assay. In vitro-transcribed replicon RNA was transfected into Huh-7 cells by electroporation (30). After cells were cultured, the luciferase reporter gene expression level was determined using Bright-Glo to calculate the inhibition rate of the compound against the HCV replicon. Nonlinear fit analysis of compound inhibition data using GraphPad Prism (version 5) software yielded EC₅₀s for the compounds.

In vitro drug resistance screening of amphihevir was performed in stable GT1b and GT1a replicon cells with HCV-1b and HCV-1a, respectively. cDNA was sequenced using the Sanger method with NS4B-specific primers. The NS4B mutant gene was identified by comparing the amphihevir-screened resistant-replicon’s NS4B sequence to the parallel-passaged amphihevir-free replicon sequence.

Combinations of amphihevir and other HCV inhibitors.

The combined effects of amphihevir and other types of HCV inhibitors were determined in HCV-1b replicon cells (31). Each compound was diluted twice, and then amphihevir and the other compounds were orthogonally proportioned at 7 different concentrations and added to 96-well plates in sets of three duplicate wells. HCV-1b cells were seeded at a density of 8,000 cells per well of 96-well plates containing compounds with the concentration of 0.5% dimethyl sulfoxide (DMSO). HCV-1b cells were cultured for 3 days at 5% CO₂ and 37°C. The cell growth assay reagent, CellTiter-Fluor, was added to the first plate of each combination, and the plate cultured for 1 h at 37°C in a 5% CO₂ incubator. Fluorescence signal values were measured using the EnVision plate reader, and cell viability percentages were calculated to determine the cytotoxicities of the compound combinations. With or without cytotoxicity, the luciferase luminescence substrate, Bright-Glo, was then added to each well and the luminescence signal’s value was detected with the EnVision plate reader within 5 min. The experimental data were processed using MacSynergy software to analyze the effects of amphihevir combined with different HCV inhibitors (32).

To determine the antiviral activity of amphihevir in combination with fopitasvir, the method described above was used. Replicon cells appeared as cell colonies after 3 weeks of treatment with the compound and were stained with 1% crystal violet.

In vivo PK.

Twelve male and 12 female rats were randomly divided into four groups, with three males and three females per group. Rats in the i.v. group were injected via the tail vein, and blood samples were collected via a jugular vein catheter. For this group, the solution was prepared using DMSO/polyethylene glycol 400 (PEG 400)/Solutol at a ratio of 15:25:60 and administered at 1 mg/kg daily. The remaining three groups were orally administered a single dose of 3, 10, or 30 mg/kg/day, respectively. The plasma samples were taken before and after the 1st and 7th day, respectively. Volumes of 0.25 ml of blood samples were placed in labeled centrifugal tubes containing K₂-EDTA (0.5 M) for centrifugation at 4°C for 15 min. The supernatant (plasma) was rapidly frozen on dry ice and stored in a refrigerator at −70 ± 10°C until liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis. The concentrations of amphihevir in rat plasma samples were analyzed by LC-MS/MS that had a verified lower limit of quantification (LLOQ) of 1.00 ng/ml and upper limit of quantification (ULOQ) of 1,000 ng/ml. WinNonlin version 6.3.0 (Pharsight, Mountain View, CA) pharmacokinetic (PK) software was employed to analyze the PK parameters using a noncompartment model. The PK parameters were calculated by the logarithmic linear trapezoidal method.

Twelve male beagle dogs and twelve female dogs were divided into four groups with three male and three female dogs per group. The amphihevir solution for the i.v. group was prepared using DMSO/PEG 400/H₂O at a ratio of 20:60:20. This group was administered 1 mg/kg/day, while the remaining three groups were fed amphihevir soft capsules containing 3, 10, or 30 mg/kg/day (specifications of soft capsules were 15 mg and 60 mg). Blood samples were collected from the peripheral veins of all animals without anesthesia. Animals were treated according to the method described above.

Four-week toxicity studies.

The pretest results showed that no significant toxicity was observed in rats and dogs when orally administered 400 mg/kg and 200 mg/kg amphihevir for 28 consecutive days. Therefore, we selected 20, 60, and 200 mg/kg for rats and 3, 100, and 200 mg/kg for dogs as the low, medium, and high doses of amphihevir, respectively. The control group was administered the same volume of corn oil via oral administration.

In the 4-week toxicity test with SD rats, 42 rats were in the control group (12 per group for the toxicity study), and 54 rats were included in each amphihevir dosing group (24 per group for the toxicity study); each group had an equal number of males and females. Ten beagle dogs were tested in each dosing group (equal number of males and females) in the 4-week toxicity test.

Dosing frequency was 1 dose per day, with continuous administration for 4 weeks and withdrawal for 4 weeks. Each group of rats was orally administered corn oil, while dogs were administered soft capsules.

The amount administered to each animal was adjusted based on weight obtained from the most recent measurement. General condition, body weight, food intake, hematology, blood biochemical examination, gross anatomy, and toxicokinetic parameters were assessed for the animals.