Clinical Resistance to Velpatasvir (GS-5816), a Novel Pan-Genotypic Inhibitor of the Hepatitis C Virus NS5A Protein

Abstract

Velpatasvir (VEL, GS-5816) is a novel pan-genotypic hepatitis C virus (HCV) nonstructural protein 5A (NS5A) inhibitor with activity against genotype 1 (GT1) to GT6 HCV replicons. In a phase 1b 3-day monotherapy study, patients treated with a 150-mg dose of GS-5816 had a mean maximal HCV RNA decline of ≥3.3 log₁₀ IU/ml in GT1a, -1b, -2, -3, and -4. This report characterizes virologic resistance to VEL in these patients. NS5A resistance-associated substitutions (RASs) were detected by deep sequencing (1% cutoff) pretreatment in 22/70 patients, i.e., 10/35 (29%) patients with GT1a, 1/8 (13%) with GT1b, 4/8 (50.0%) with GT2, 5/17 (29.4%) with GT3, and 2/2 (100.0%) with GT4. In GT1a and GT3 patients, pretreatment RASs were associated with a slightly reduced HCV RNA response compared to that of patients without pretreatment RASs; among patients with GT1b, GT2, and GT4, no significant difference in response was observed in those with or without pretreatment RASs. Following treatment, the pattern of emergent RASs was more complex for GT1a than for the other genotypes. In GT1a, substitutions emerged at positions M28, Q30, L31, P32, H58, E92, and Y93, with the most prevalent substitutions at positions Y93, M28, and L31. RASs were observed at two positions in GT1b and GT2 (Y93 and L31), three positions in GT3 (Y93, L31, and E92), and four positions in GT4 (L28, M31, P32L, and Y93). RASs that were present pretreatment persisted through the 48-week follow-up period; however, RASs emerging during treatment were more likely to decline both in prevalence and in frequency within the viral population during follow-up. (This study has been registered at ClinicalTrials.gov under registration no. NCT01740791.)

INTRODUCTION

Hepatitis C virus (HCV) infects an estimated 180 million people worldwide (1). Infection can lead to cirrhosis, hepatocellular carcinoma, and other complications. Novel direct-acting antiviral agents (DAAs) are being developed in combination with pegylated interferon (PegIFN)/ribavirin (RBV) and are also being pursued as components of IFN-free and IFN/RBV-free regimens to improve efficacy and shorten treatment duration. Recently, new DAAs that are well tolerated and highly effective have been approved for the treatment of chronic HCV infection. In December 2013, the FDA approved Sovaldi (sofosbuvir [SOF]; Gilead Sciences), a once-daily pan-genotypic oral nucleotide analog polymerase inhibitor (NI) for the treatment of HCV infection as a component of a combination antiviral treatment regimen for patients with genotype 1 (GT1), GT2, GT3, or GT4 infection (2, 3). The protease inhibitor (PI) Olysio (simeprevir; Janssen Pharmaceutica), in combination with PegIFN/RBV, also received FDA approval in December 2013. Simeprevir has better activity against multiple genotypes than the PIs boceprevir and telaprevir but weak activity against GT2 and no activity against GT3 (4). Another recently approved treatment of GT1 HCV infection in the United States and Europe is Harvoni (Gilead Sciences), a fixed-dose combination of ledipasvir, a nonstructural protein 5A (NS5A) inhibitor, and SOF based on the results of three phase 3 studies (5 – 7). Treatment with Harvoni for durations ranging from 8 to 24 weeks (depending on the HCV GT and patient population) resulted in sustained virologic response rates of 94 to 99%. Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets copackaged with dasabuvir; AbbVie) (8) was approved for the treatment of patients with GT1 infection in December 2014, and on 24 July 2015, Daklinza (daclatasvir; Bristol-Myers Squibb) (9) in combination with SOF was approved by the FDA for the treatment of patients with HCV GT3 infection. Recently, in January 2016, the FDA approved Zepatier (elbasvir and grazoprevir; Merck) with or without ribavirin for the treatment of patients with HCV GT1 and GT4 infections (10).

HCV is characterized by high genetic diversity and is classified into at least six GTs. HCV GT1 to GT6 have been isolated from multiple patients and are further divided into 66 subtypes (11). GT3 infection accounts for 35 to 80% of chronic HCV infections in regions such as the Indian subcontinent (12). In Egypt, where GT4 predominates, the prevalence of HCV is one of the highest in the world, with almost 15% of the population infected (13). To enable effective, well-tolerated, and all-oral therapy across all HCV genotypes, development of pan-genotypic DAAs is needed.

Velpatasvir (VEL, GS-5816) is a pan-genotypic HCV NS5A inhibitor in development to target HCV GT1 to GT6 with broad polymorphism coverage. It is a selective inhibitor of HCV RNA replication with mean 50% effective concentrations (EC₅₀s) against GT1 to GT6 of 6 to 130 pM (14). VEL showed potent activity against some of the clinically significant GT1 NS5A inhibitor resistance-associated substitutions (RASs), as well as clinically prevalent GT2a and GT2b polymorphism M31 and GT3a polymorphism A30K that confer high-level resistance to NS5A inhibitors (e.g., daclatasvir and ledipasvir) (14).

A multiple-ascending-dose study was conducted in which GT1a, GT1b, GT2, GT3, and GT4 HCV chronically infected, treatment-naive patients were treated once daily with VEL as a 3-day monotherapy. This report describes the emergence of NS5A RASs following treatment and the impact of pretreatment RASs detected by population or deep sequencing of NS5A. In addition, the long-term persistence of NS5A RASs was evaluated through week 48 (ClinicalTrials.gov identifier NCT01740791).

(Part of this work was presented at the American Association for the Study of the Liver, Washington, DC, 1 to 5 November 2013.)

MATERIALS AND METHODS

Compounds.

All compounds (VEL and SOF) were synthesized by Gilead Sciences (Foster City, CA).

Clinical trial population and study design.

This was a phase 1b double-blind, randomized, placebo-controlled, multicenter study of VEL in HCV-infected patients in the United States (ClinicalTrials.gov identifier NCT01740791). Clinical data of this trial have been described previously (15). A total of 87 patients were enrolled and received treatment in 1 of 11 cohorts, each randomized 4:1 to treatment with VEL or a matching placebo for 3 days (except for the GT4 patients, who were not randomized and received VEL). The actual treatments administered are presented in Table 1. One patient discontinued study treatment because of an adverse event, and two patients discontinued the study (one was lost to follow-up, and one withdrew consent) prior to day 17 assessments (these patients were included in the sequencing analyses). VEL was administered once daily as follows: 5, 25, 50, 100, and 150 mg to GT1a patients; 150 mg to GT1b, GT2, and GT4 patients; and 25, 50, and 150 mg to GT3 patients. Eligible patients had plasma HCV RNA levels of >5 log₁₀ IU/ml pretreatment and were treatment naive. Of the 87 patients in this study, 45 had HCV GT1a, 10 had GT1b, 10 had GT2b, 1 had GT3, 19 had GT3a, 1 had GT4, and 1 had GT4a, GT4b, and GT4c. The study was conducted in compliance with the Declaration of Helsinki. The study protocol and informed consent documents were reviewed and approved by the institutional review board of the participating institution, and informed consent was obtained from all patients before any study-specified procedures.

TABLE 1.

Antiviral response to VEL monotherapy

HCV genotype, treatment (n)	Mean max declinea in HCV RNA ± SD (median)	No. of patients sequenced pretreatment	No. of patients/total (%) with pretreatment:		Mean maximal log10 HCV RNA reduction in patients with/without pretreatment RASs
			RASs	RASs by GT
Placebo (17)	0.43 ± 0.24 (0.43)	8	2/8 (25)	2/8 (25)	NAc
GT1a
5 mg VEL (4)	3.68 ± 0.42 (3.85)	4	0/4 (0)	10/35 (28.6)	NA/3.68
25 mg VEL (8)	3.95 ± 0.4 (3.89)	8	2/8 (25.0)		3.76/4.01
50 mg VEL (8)	3.58 ± 1.24 (4.17)	8	3/8 (37.5)		2.32/3.71
100 mg VEL (8)	3.56 ± 0.78 (3.67)	8	3/8 (37.5)		2.87/3.97
150 mg VEL (7)	3.95 ± 0.85 (4.19)	7	2/6 (33.3)		2.9/4.38
GT1b, 150 mg VEL (8)	3.97 ± 1.24 (4.29)	8	1/8 (12.5)	1/8 (13)	4.47/4.39
GT2, 150 mg VEL (8)	4.36 ± 0.53 (4.39)	8b	4/8 (50.0)	4/8 (50)	4.08/4.62
GT3
25 mg VEL (7)	2.84 ± 1.32 (3.25)	7	2/7 (28.6)	5/17 (29.4)	0.99/3.58
50 mg VEL (4)	2.59 ± 1.18 (3.12)	4	1/4 (25.0)		0.84/3.18
150 mg VEL (6)	3.3 ± 0.55 (3.13)	6	2/6 (33.3)		2.8/3.54
GT4, 150 mg VEL (2)	3.47 ± 0.59 (3.17)	2	2/2 (100)	2/2 (100)	3.47/NA
All (87)	NA	78	24 (30.8)	24/78 (30.8)	NA

^aMaximal viral load reduction (log IU/ml) at any time point during the first 17 days.

^bOne patient was determined to have GT1a by NS5A sequencing. This patient with no NS5A RAS pretreatment had selected NS5A RASs posttreatment.

^cNA, not applicable.

Plasma for resistance surveillance was collected pretreatment and on days 2, 4, 5, 7, 10, and 17 and in follow-up weeks 12, 24, and 48. The first samples collected each day during the dosing interval were taken prior to dose administration (predose). The COBAS TaqMan HCV Test Version 2.0 was used for serum HCV RNA assessments. The lower limit of quantitation of the assay was 25 IU/ml. The change from the pretreatment HCV RNA level was determined at each time point.

Viral sequencing.

For every VEL-treated patient and 8 of 17 who received a placebo, samples with HCV RNA levels of >1,000 IU/ml at the pretreatment visit and day 4 and/or day 2, 4, 5, 7, 10, or 17 and in follow-up weeks 12, 24, and 48 were used to amplify the gene for HCV NS5A (DDL Diagnostic Laboratory, Rijswijk, Netherlands), which was deep sequenced with a 1% assay sensitivity cutoff (DDL Diagnostic Laboratory, Rijswijk, Netherlands, or WuxiApptec Co. Ltd., Shanghai, China) with the Illumina MiSeq platform (Illumina, San Diego, CA), except for 1 patient with population sequencing at the pretreatment visit. Population sequencing of the full-length HCV NS5A coding region was performed by Janssen Diagnostics (Beerse, Belgium) by reverse transcription-PCR and standard Sanger sequencing of the bulk PCR product. The sensitivity of detection of resistant variants is approximately 10 to 20% (16). Variants are reported as differences from a genotype-specific reference strain, i.e., GT1b Con1 (AJ238799), GT1a H77 (GenBank accession number NC_004102), GT2 JFH-1 (AB047639), GT3 S52 (GU814263), or GT4 ED43 (GU814265). Deep-sequencing reads were aligned and processed with internally developed software via a multistep method to identify the substitutions present at levels of >1% (17). Sequencing analysis included NS5A class RASs that were summarized by the HCV Drug Resistance Advisory Group (18) and/or recently observed in clinical trials with LDV, VEL, DCV, ABT-267, ABT-530, and MK-8742 (19 – 29) and including positions 24, 28, 30, 31, 32, 38, 58, 92, and 93.

Transient transfection of replicon RNA into Huh7 cells and EC₅₀ determination.

Resistance mutations were introduced into the GT1a (30), GT1b (31), GT2a (32), GT3a (33), and GT4a (34) replicons (backbone GT1a H77, GT1b Con1, GT2a JFH-1, GT3a S52, and GT4a ED43, respectively) by site-directed mutagenesis and tested in transient transfections as previously described (35). Briefly, NS5A mutations were introduced into a plasmid encoding the PI-hRluc replicon with a QuikChange II XL mutagenesis kit in accordance with the manufacturer's instructions (Stratagene, La Jolla, CA). Mutations were confirmed by DNA sequencing. Replicon RNAs were transcribed in vitro from replicon-encoding plasmids with a MEGAscript kit (Ambion, Austin, TX). RNA was transfected into Huh-lunet cells by the method of Lohmann et al. (31). Briefly, cells were trypsinized and washed twice with phosphate-buffered saline (PBS). A suspension of 4 × 10⁶ cells in 400 μl of PBS was mixed with 5 μg of RNA and subjected to electroporation at settings of 960 μF and 270 V. Cells were transferred into 40 ml of prewarmed culture medium and then seeded into 96-well plates (100 μl/well). Compounds were 3-fold serially diluted in 100% dimethyl sulfoxide (DMSO) and added to cells at a 1:200 dilution, achieving a final DMSO concentration of 0.5% in a total volume of 200 μl/well. Cells were treated for 3 days, after which culture media were removed, cells were lysed, and Renilla luciferase activity was quantified with a commercially available assay (Promega, Madison, WI) and a Top Count instrument (PerkinElmer, Waltham, MA). The EC₅₀ was calculated as the compound concentration at which a 50% reduction in the level of Renilla reporter activity was observed compared with control samples with DMSO. Dose-response curves and EC₅₀s were generated with the GraphPad Prism software package (GraphPad Software, La Jolla, CA) by nonlinear regression analysis. The replication level of either reference strains (1b-Con1 and 1a-H77) or chimeric replicons derived transiently from clinical isolates was determined as the ratio of the Renilla luciferase signal at day 4 to that at 4 h postelectroporation to normalize for transfection efficiency. The replication capacity of each replicon was expressed as their normalized replication efficiency compared with that of the reference strain (1b-Con1 or 1a-H77) within the same experiment.

Accession numbers.

All of the NS5A sequences produced in this study were submitted to the GenBank database and assigned accession numbers KT721915 to KT722333.

RESULTS

Antiviral response to VEL in GT1 to GT4 HCV-infected patients.

Three-day monotherapy with VEL produced rapid declines in HCV RNA levels. Among the cohorts dosed with 150 mg of VEL, the median reductions in HCV RNA levels through day 17 were 4.19, 4.29, 4.39, 3.13, and 3.17 log₁₀ HCV RNA IU/ml in the GT1a, GT1b, GT2, GT3, and GT4 groups, respectively (Table 1). GT1a patients dosed with 5, 25, 50, or 100 mg had a median HCV RNA reduction of >3.67 log₁₀, and GT3 patients dosed with 25 or 50 mg had a median reduction of >3.12 log₁₀.

Effect of NS5A RASs present pretreatment.

For all 70 patients who received VEL and 8 placebo-treated patients, sequences were analyzed for the presence of polymorphisms that are known NS5A inhibitor RASs at amino acids 28, 30, 31, 32, 58, 92, and 93. Pretreatment, NS5A RASs (detected at >1%) were present in 24 patients, i.e., 2/8 placebo, 10/35 GT1a, 1/8 GT1b, 4/8 GT2, 5/17 GT3, and 2/2 GT4 patients (Table 1), with some patients having >1 RAS. Seven patients who received VEL had pretreatment RASs at amino acid residue 93 (Y93C/F/H/N) (Table 2). For GT1a, NS5A RASs were observed at positions M28T, Q30H/K/R, L31M/V, H58D, and Y93C/H/F/N. The mean viral load decrease in GT1a patients who were dosed with 150 mg of VEL and had pretreatment NS5A RASs was 2.9 log₁₀ (Table 1), compared with a 4.38-log₁₀ reduction in patients without pretreatment NS5A RASs. The GT1b patient with an NS5A RAS pretreatment (Y93H) had a 4.47-log₁₀ HCV RNA reduction, compared with a mean 4.39-log₁₀ reduction in patients without NS5A RASs (Table 1). In the four GT2 patients with L31M pretreatment, the mean log₁₀ reduction was 4.08 log₁₀, compared with 4.62 log₁₀ in patients without NS5A RASs (Table 1). Of the five GT3 patients with pretreatment NS5A RASs, two were treated with 25 mg of VEL, one was treated with 50 mg of VEL, and two were treated with 150 mg of VEL. The GT3 patients with pretreatment RASs treated with 25 or 50 mg of VEL had a <1-log₁₀ mean HCV RNA reduction, while all of the GT3 patients in these dose groups without pretreatment RASs had >3-log₁₀ reductions. The two patients with RASs treated with 150 mg of VEL had mean HCV RNA reductions of 2.9 log₁₀ and 2.7 log₁₀, compared to 3.54 log₁₀ in patients without NS5A RASs (Table 1; Fig. 1). Both GT4 patients who had variants at positions 30 (one with L30H [45.8%] and L30R [53.7%] and one with L30S [2.4%] and L30H [97.1%]) had a mean HCV RNA reduction of 3.47 log₁₀.

TABLE 2.

Patients with pretreatment NS5A RASs by deep sequencing

HCV genotype, treatment (n) or parameter	No. (%) of patients with NS5A RAS(s):
	M28T	L30H/R	Q30H/K	A30K	L30S	Q30R	L31M/V	H58D	Y93C/F/H/N
Placebo (8)			1				1		1
GT1a
5 mg VEL (4)
25 mg VEL (8)			1				2
50 mg VEL (8)	2		1			1	3	1	1
100 mg VEL (8)			2			3	2	1	1
150 mg VEL (7)	1						2	1
GT1b, 150 mg VEL (8)									1
GT2, 150 mg VEL (8)							4
GT3
25 mg VEL (7)									2
50 mg VEL (4)									1
150 mg VEL (6)				1					1
GT4, 150 mg VEL (2)		2			1
Total	3 (3.8)	2 (2.6)	5 (6.4)	1 (1.3)	1 (1.3)	4 (5.1)	14 (17.9)	3 (3.8)	8 (10.3)

FIG 1.

Maximum changes from pretreatment HCV RNA loads. The mean ± the standard deviation of the maximum viral load reductions in patients dosed with GS-5816 for 3 days are shown. (A) GT1a patients. (B) GT3 patients. (C) GT1 to GT4 patients dosed with 150 mg of GS-5816. All patients with pretreatment RASs are indicated with the pretreatment percentage of each RAS shown in parentheses, except for one GT1a patient with only population sequencing available (pretreatment RAS L31L/M).

Substitutions selected in HCV of patients following VEL treatment through day 17.

To identify the HCV NS5A variants that are potentially associated with virologic resistance to VEL, the full-length NS5A coding region was analyzed during treatment and posttreatment by deep sequencing with a 1% cutoff. Samples were obtained during treatment (days 2 to 3) or posttreatment (days 4 to 10) and on day 17, respectively, and were analyzed when the viral load was ≥1,000 IU/ml. Of the 46/70 VEL-treated patients without pretreatment NS5A RASs, sequences were available from 40 and 46 on days 2 to 10 and day 17, respectively. All patients (40/40; 100%) with available sequences on days 2 to 10 had emergent NS5A RASs, and 80.4% (37/46) still had RASs on day 17 (Table 3). Emergent NS5A RASs were not detected during the posttreatment period in the two placebo-treated patients whose samples were sequenced. NS5A RASs emerged on treatment at more positions in patients with GT1a than in patients with other GTs and included substitutions at positions M28, Q30, L31, P32, H58, E92, and Y93. RASs at positions Y93, M28, and L31 were the most prevalent in GT1a patients (Table 4). RASs were observed at two NS5A positions in GT1b and GT2 patients (Y93, L31) and three positions in GT3 patients (Y93, L31, E92). In two GT4 patients, NS5A RASs emerged at positions L28, M31, P32L, and Y93 (Table 5). L31M/V and Y93H were the most commonly observed RASs emerging on treatment in GT1b and GT2 patients, and E92K and Y93H/N were the most prevalent RASs emerging in GT3 patients (Table 4).

TABLE 3.

Persistence of emergent NS5A RASs by deep sequencing in patients without pretreatment RASs

HCV genotype, treatment (n)	No. (%) of patients with RASs at posttreatment time point:					No. of patients with RASs at wk 48 by GT/total (%)
	Days 2–10	Day 17	Wk 12	Wk 24	Wk 48
GT1a
5 mg VEL (4)	4/4 (100)	1/4 (25.0)	0/4 (0)	1/4 (25.0)	0/4 (0)	6/21(28.6)
25 mg VEL (8)	6/6 (100)	5/6 (83.3)	3/6 (50.0)	2/6 (33.3)	0/5 (0)
50 mg VEL (8)	3/3 (100)	3/4 (75.0)	1/5 (20.0)	1/4 (25.0)	1/5 (20.0)
100 mg VEL (8)	5/5 (100)	5/6 (83.3)	3/5 (60.0)	2/4 (50.0)	1/3 (33.3)
150 mg VEL (7)	4/4 (100)	5/5 (100)	4/4 (100)	4/4 (100)	4/4 (100)
GT1b, 150 mg VEL (8)	4/4 (100)	4/5 (80.0)	4/5 (80.0)	3/4 (75.0)	2/4 (50.0)	2/4 (50.0)
GT2, 150 mg VEL (8)	1/1 (100)	3/4 (75.0)	0/1 (0)	0/1 (0)	0/2 (0)	0/2 (0)
GT3						3/6 (50.0)
25 mg VEL (7)	6/6 (100)	4/5 (80.0)	NAa	NA	1/4 (25.0)
50 mg VEL (4)	3/3 (100)	3/3 (100)	NA	NA	NA
150 mg VEL (6)	4/4 (100)	4/4 (100)	NA	NA	2/2 (100)
GT4, 150 mg VEL (2)	0/0	0/0	NA	NA	0/0	0/0
All	40/40 (100)	37/46 (80.4)	15/30 (50.0)	13/23 (56.5)	11/33 (33.3)	11/33 (33.3)

^aNA, not applicable because patient missed study visit.

TABLE 4.

NS5A RASs detected through week 48 by deep sequencing in GT1a, GT1b, GT2, and GT3 patients without pretreatment RASs

HCV genotype, variant	No. (%) of patients with variant at:
	Day 5	Day 7	Day 10	Day 17	Wk 12	Wk 24	Wk 48
GT1a	9	16	22	24	24	22	21
K24R		1 (6.3)
M28T	3 (33.3)	5 (31.3)	9 (40.9)	9 (37.5)	4 (16.7)	2 (9.1)
Q30E				1 (4.2)
Q30H	1 (11.1)	1 (6.3)	5 (22.7)	5 (20.8)	3 (12.5)	1 (4.5)	1 (4.8)
Q30K				1 (4.2)	1 (4.2)	1 (4.5)
Q30L					1 (4.2)
Q30R		1 (6.3)	5 (22.7)	6 (25.0)	7 (29.2)	5 (22.7)	2 (9.5)
L31I				1 (4.2)
L31M	2 (22.2)	2 (12.5)	3 (13.6)	10 (41.7)	8 (33.3)	8 (36.3)	5 (23.8)
L31V	2 (22.2)	8 (50.0)	14 (63.6)	15 (62.5)	11 (45.8)	7 (31.8)	2 (9.5)
P32L		1 (6.3)	2 (9.1)
H58D		1 (6.3)		4 (16.7)	3 (12.5)	2 (9.1)	2 (9.5)
Y93C		3 (18.8)	6 (27.3)	7 (29.2)	3 (12.5)
Y93H	6 (66.7)	11 (68.8)	14 (63.6)	11 (45.8)	4 (16.7)	1 (4.5)	1 (4.8)
Y93N	5 (55.6)	9 (56.3)	13 (59.1)	11 (45.8)	4 (16.4)	3 (13.6)
Y93W	1 (11.1)				1 (4.2)	1 (4.5)	1 (4.8)
Y93R		1 (6.3)	1 (4.5)
Y93S		1 (6.3)
GT1b	0	0	4	5	5	4	4
L31M			2 (50.0)	4 (80.0)	4 (80.0)	3 (75.0)	2 (50.0)
L31V			2 (50.0)	4 (80.0)	4 (80.0)	1 (25.0)
Y93C				2 (40.0)	1 (20.0)
Y93H				4 (80.0)	4 (80.0)	2 (50.0)	1 (25.0)
Y93N				2 (40.0)
Y93S				1 (20.0)
GT2	0	1	1	4	1	1	6
L31I				1 (25.0)
L31M				1 (25.0)
L31V			1 (100)	2 (50.0)
Y93H		1 (25.0)	1 (100)	3 (75.0)
GT3	5	8	17	12	0	0	6
L31V				1 (8.3)
L31P	1 (20.0)
E92K	3 (60.0)	2 (25.0)	4 (23.5)	2 (16.7)
Y93H	5 (100)	8 (100)	12 (70.6)	11 (91.7)			3 (50.0)
Y93N		2 (25.0)	3 (17.6)
Y93R			1 (5.9)

TABLE 5.

Persistence of NS5A RASs by deep sequencing in patients with pretreatment RASs

Patient no.	Genotype	No. of NS5A RASs at each visit						No. of new or additional RASs that developed and persisted through wk 48
		Pretreatment (n = 22)	Days 2–10 (n = 19)	Day 17 (n = 20)	Wk 12 (n = 10)	Wk 24 (n = 10)	Wk 48 (n = 16)
1a	GT1a	4	5	4	4	4	4	2
2	GT1a	5	7	4	4	4	4	0
3	GT1a	2	3	3	2	2	2	0
4	GT1a	1	10	9	NAc	NA	NA	NA
5	GT1a	1	9	7	6	2	3	2
6	GT1a	3	10	9	8	8	7	4
7b	GT1a	8	8	8	8	9	8	1
8	GT1a	2	4	3	3	3	3	1
9	GT1a	1	NA	NA	3	2	2	1
10	GT1a	3	8	8	5	5	5	2
11	GT1b	1	3	3	3	3	3	2
12	GT2b	1	3	4	NA	NA	1	0
13	GT2b	1	1	1	NA	NA	1	0
14	GT2b	1	3	3	NA	NA	1	0
15	GT2b	1	NA	3	NA	NA	1	0
16	GT3a	1	1	1	NA	NA	NA	NA
17	GT3a	1	NA	NA	NA	NA	NA	NA
18	GT3a	1	1	1	NA	NA	NA	NA
19	GT3a	1	4	3	NA	NA	NA	NA
20	GT3a	1	1	1	NA	NA	NA	NA
21	GT4a	2	8	5	NA	NA	2	0
22	GT4a	2	7	5	NA	NA	2	0

^aPatient 1 had Q30K, Q30R, Y93F, and Y93H RASs at the screening visit. The Q30R and Y93F RASs did not persist. Q30H and Q30N RASs developed and persisted through week 48.

^bPatient 7 had a K24R RAS at the screening visit that was not observed at week 48. The H58D RAS developed and persisted through week 48.

^cNA, not applicable because patient missed study visit.

Of the 22 VEL-treated patients with pretreatment NS5A RASs, sequences were available on days 2 to 10 and day 17 from 19 and 20 patients, respectively (Table 5). Fifteen (79.0%) of 19 and 14 (70.0%) of 20 patients had additional NS5A variants that emerged on days 2 to 10 and day 17, respectively (Table 5; see Table S1 in the supplemental material). The GT1a patients developed additional variants at positions K24R, M28T, Q30N/H/K/R/E, L31I/M/V, H58D, A92T, and Y93H/C. The GT1b patient developed variants at position L31M/V, GT2b patients developed variants at positions L31V and Y93H/N (data not shown), and GT4 patients developed variants at positions L28T, M31V, P32L, and Y93H/C/N/W (Table 6; see Table S1 in the supplemental material).

TABLE 6.

Frequencies of pretreatment and emergent NS5A RASs in two GT4a patients

Patient no.	Frequency (%) of NS5A RAS(s)
	Pretreatment	Day 4	Day 5	Day 7	Day 10	Day 17	Wk 48
1	L30H (45.83), L30R (53.72)	NAa	NA	NA	L28T (3.03), L30H (91.00), L30R (8.92), M31V (3.49), P32L (1.49), Y93H (21.57), Y93W (2.69), Y93C (4.24)	L30H (96.43), L30R (3.48), M31V (3.63), Y93H (10.36), Y93C (3.81)	L30H (20.42), L30R (79.50)
2	L30H (97.07), L30S (2.36)	L30H (>99), Y93H (20.23)	L30H (>99), M31V (12.31), Y93H (48.68), Y93N (3.15), Y93C (4.65)	L30H (98.82), M31V (20.62), P32L (1.35), Y93H (37.06), Y93S (1.38), Y93N (1.59), Y93C (2.06)	L30H (98.44), L30S (1.04), M31V (26.13), P32L (1.18), Y93H (46.11), Y93N (1.00), Y93C (3.48)	L30H (97.97), L30S (1.43), M31V (28.51), Y93H (33.11), Y93C (4.00)	L30H (80.93), L30S (18.62)

^aNA, not applicable because patient missed study visit.

Persistence of NS5A RASs in patients during long-term follow-up through week 48.

The persistence of NS5A RASs in patients who developed NS5A RASs was evaluated during long-term follow-up at weeks 12, 24, and 48. Of the 48 patients who did not have pretreatment NS5A RASs, sequences were available from 30, 23, and 33 patients at weeks 12, 24, and 48, respectively (Table 3). NS5A RASs were observed in 50.0% (15/30), 56.5% (13/23), and 33.3% (11/33) of the patients at weeks 12, 24, and 48, respectively. Eleven patients still had detectable NS5A RASs at week 48, i.e., 6/21 (28.6%) GT1a, 2/4 (50.0%) GT1b, 0/2 GT2, and 3/6 (50.0%) GT3 patients (Table 3). Most of the newly emerged RASs during treatment did not persist through week 48, except Y93H in patients with GT1a, GT1b, and GT3 HCV infections; L31M in patients with GT1a and GT1b HCV infections; and Q30H/R, L31V, Y93W, and H58D in patients with GT1a HCV infections (Table 4). The most prevalent RAS detected at day 17 for each GT (GT1a, L31V; GT1b, L31M/V Y93H; GT2, Y93H; GT3, Y93H) showed a gradual decline in prevalence at posttreatment follow-up, but L31V for GT1a and Y93H for GT1b and GT3 were still observed in a few patients at week 48 (Table 4). A gradual decline in the number of patients with RASs was observed in all of the treatment groups except the highest-dose group, 150 mg of VEL (Table 3). A decline in the frequencies of most NS5A RASs was observed through week 48 (Fig. 2B).

FIG 2.

Changes in the frequency of NS5A RASs among patients with or without pretreatment RASs through 48 weeks posttreatment. (A) Changes in frequencies among patients with pretreatment RASs included Y93H (n = 3 [GT1a, n = 2; GT1b, n = 1]), L31M (n = 11 [GT1a, n = 7; GT2b, n = 4]), H58D (GT3a, n = 3), Q30H (GT1a, n = 2), M28T (GT1a, n = 3), and Q30R (GT1a, n = 2). (B) Changes in frequencies among patients without pretreatment RASs included Y93H (n = 38 [GT1a, n = 18; GT1b, n = 6; GT2b, n = 2; GT3, n = 12]), L31M (n = 16 [GT1a, n = 11; GT1b, n = 4; GT2b, n = 1]), L31V (n = 25 [GT1a, n = 18; GT1b, n = 5; GT2b, n = 1; GT3, n = 1]), H58D (GT1a, n = 5), M28T (GT1a, n = 17), Q30H (GT1a, n = 6), Q30R (GT1a, n = 10), Y93N (n = 21 [GT1a, n = 14; GT1b, n = 2; GT3a, n = 5]). Means ± standard deviations of frequencies of NS5A variants in patients with RASs and available samples at each visit are shown.

Among the 22 patients with NS5A RASs pretreatment, sequences were available from 10, 10, and 16 at posttreatment weeks 12, 24, and 48, respectively (Table 5). In the 16 patients with available sequences at week 48, all of the NS5A RASs that were observed pretreatment were detected at week 48 and had very similar frequencies, except for four RASs in 3 GT1a patients who had low frequencies pretreatment that were not observed at week 12, 24, or 48 (1 patient with Y93F [1.22%] and Q30R [3.23%], 1 patient with Q30H [3.04%], and 1 patient with K24R [1.55%]; see Table S1 in the supplemental material). In the two patients with GT4, only the pretreatment RAS at position L30 persisted through week 48, whereas RASs that developed during treatment did not persist (Table 6). The frequencies of most of the NS5A RASs (Y93H, Q30H/R, H58D) that were observed pretreatment increased and persisted through week 48 (Fig. 2A). The prevalence of RASs persisting through week 48 was higher in patients with RASs detected pretreatment than in patients without pretreatment RASs (Fig. 2B).

Phenotypic and cross-resistance analyses.

The predominant RASs detected pretreatment or selected following VEL treatment in each GT were introduced into wild-type GT1a, GT1b, GT2a, GT3a, and GT4a replicons (backbone GT1a H77, GT1b Con1, GT2a JFH-1, GT3a S52, and GT4a ED43, respectively) by site-directed mutagenesis and phenotyped. In GT1a replicons, Y93H/N/W/R resulted in >400-fold reduced susceptibility to VEL (Table 7). The L31V and P32L variants conferred 67.5- and 28.4-fold changes in susceptibility to VEL, respectively. On the other hand, the M28T, Q30H/K/R, L31M, H58D, and Y93C substitutions conferred <20-fold reduced susceptibility to VEL in GT1a replicons. In GT1b replicons, Y93H/N/C/S and L31M/V resulted in a <3.3-fold reduced susceptibility to VEL (Table 7). In GT2, L31 variants had ≤10.1-fold reduced susceptibility. In GT3a, A30K and L31M variants had ≤70.8-fold reduced susceptibility to VEL. However, Y93H had a 723.5-fold change in the VEL EC₅₀ (Table 7). In GT4a, variants at position 30, which were detected in two patients pretreatment, had ≤2.4-fold reduced susceptibility to VEL (Table 7), suggesting little to no effect of these mutations on the EC₅₀s. No reduced susceptibility to the nucleoside inhibitor SOF was observed in any NS5A mutant (Table 7).

TABLE 7.

EC₅₀ fold change and replication capacities of NS5A mutants

Genotype (EC50 [nM]) and mutation(s)	Mean replication capacity (%) ± SDb	EC50 fold change vs wild typea
		VEL	SOF
GT1a (0.014)
M28T	39.5 ± 6.3	7.5 ± 4.2	0.91 ± 0.1
Q30E	83.7 ± 20.1	17.7 ± 2.4	0.95 ± 0.1
Q30H	66.7 ± 10.5	2.3 ± 0.9	0.94 ± 0.1
Q30K	91.1 ± 22.8	10.4 ± 2.0	1.05 ± 0.1
Q30R	82.5 ± 36.8	2.2 ± 0.2	1.12 ± 0.3
Q30L	28.4 ± 7.4	0.5 ± 0.2	0.6 ± 0.1
L31M	67.1 ± 6.5	16.0 ± 6.2	1.1 ± 0.2
L31V	110.7 ± 32.6	67.5 ± 27.5	0.98 ± 0.2
L31I	128 ± 86.4	4.4 ± 0.8	0.94 ± 0.1
P32L	108.6 ± 26.5	28.4 ± 3.7	1.1 ± 0.3
H58D	85.0 ± 33.9	7.3 ± 1.6	1.1 ± 0.14
Y93C	13.2 ± 6.3	3.8 ± 2.3	0.58 ± 0.1
Y93H	45.5 ± 16.2	609.1 ± 92.4	0.8 ± 0.1
Y93N	44.9 ± 13.7	2,758 ± 533.6	0.8 ± 0.2
Y93W	60.2 ± 36.0	99.8.9 ± 383.6	NDd
Y93R	56.3 ± 11.9	497.2 ± 235.2	ND
Y93S	12.4 ± 4.5	63.9 ± 7.3	ND
Y93F	144.3 ± 176.4	2.45 ± 0.2	1.1 ± 0.1
Q30H L31M	91.1 ± 34.7	105.7 ± 78.1	ND
Q30H Y93H	48.9 ± 33.6	2,835 ± 1,758	ND
Q30R L31M	132.0 ± 8.3	198.4 ± 53.9	1.2c
Q30R Y93H	24.0 ± 9.1	>419	0.8 ± 0.2
GT1b (0.016)
Y93H	54.5 ± 10.5	3.3 ± 1.3	1.1 ± 0.2
Y93N	21.3 ± 16.7	2.8 ± 0.8	1.3c
Y93C	72.5 ± 29.0	0.2 ± 0.02	1.8c
Y93S	40.4 ± 0.1	0.6 ± 0.2	ND
L31M	273.9c	1.7c	1.8c
L31V	113.0 ± 3.5	2.3 ± 0.1	ND
GT2a (0.008)
L31M	130.1 ± 42.8	1.5 ± 0.7	0.9 ± 0.2
L31V	37.2 ± 0.1	10.1 ± 3.7	0.9 ± 0.1
L31I	0.4 ± 0.1	No replication	No replication
Y93H	53.0 ± 2.7	45.7 ± 8.7	No replication
GT3a (0.004)
Y93H	30.5 ± 2.7	723.5 ± 129.7	0.9 ± 0.04
A30K	133.0 ± 30.0	50.0 ± 7.3	1.1 ± 0.1
L31M	95.2 ± 43.6	70.8 ± 12.4	1.0 ± 0.1
L31V	204.7c	0.3c	1.0 ± 0.2
L31P	0.2 ± 0.1	No replication	No replication
E92K	1.5 ± 0.1	0.7 ± 0.1	ND
Y93N	0.4 ± 0.1	No replication	No replication
Y93R	0.6 ± 0.1	No replication	ND
GT4a (0.009)
L30H	17.3 ± 0.1	1. 5 ± 1.0	0.8 ± 0.6
L30S	52.0 ± 0.1	1.1 ± 0.2	1.3 ± 0.6
L30R	67.9 ± 0.1	2.4 ± 0.7	0.9 ± 0.3
Y93H	55.1 ± 26.7	2.9 ± 0.01	0.41 ± 0.1
Y93C	2.9 ± 1.7	1.2c	0.88c
P32L	No replication	No replication	No replication
Y93N	3.0 ± 0.3	12.7 ± 5.0	0.5 ± 0.2

^aEC₅₀ fold change compared with GT1a PI RLuc NS5A shuttle vector replicon control.

^bReplication capacity calculated as a percentage of that of the GT1a protease inhibitor RLuc NS5A shuttle vector replicon control.

^cData from one experiment.

^dND, not done.

DISCUSSION

VEL (GS-5816) is a potent pan-genotypic HCV NS5A inhibitor with activity across HCV GT1 to GT6 that has better in vitro activity against GT2 and GT3 than the NS5A inhibitors LDV and DCV (14, 36). A phase 1b monotherapy study was conducted in which GT1a, GT1b, GT2, GT3, and GT4 HCV chronically infected, treatment-naive patients were treated once daily with VEL for 3 days. Here, the impact of pretreatment NS5A RASs, emergence of RASs following treatment, and long-term persistence of RASs were evaluated in patients enrolled in that study. Consistent with in vitro data, treatment with VEL resulted in a significant antiviral effect in GT1 to GT4 HCV-infected patients (37).

HCV NS5A has emerged as an attractive viral target for small-molecule inhibition. Although there is no known enzymatic activity for NS5A, it is essential for viral replication (38). DCV has a low barrier to resistance, and RASs at amino acids positions 28, 30, 31, and 93 are selected following DCV treatment (39). In addition, some preexisting polymorphisms in the gene for NS5A are associated with a reduced antiviral response in patients during monotherapy with NS5A inhibitors, including L31M, Y93C, and Y93H polymorphisms in GT1 (36, 37); the highly prevalent (>50%) M31 polymorphism in GT2 (40); and A30K in GT3 (associated with relapse during DCV-SOF combination therapy) (41). The impact of pretreatment NS5A RASs on the antiviral response was investigated in patients enrolled in this VEL phase 1 monotherapy study. Pretreatment RASs were detected in 22/70 (31.4%) patients by deep sequencing, i.e., 28.6% of GT1a patients, 12.5% of GT1b patients, 50.0% of GT2 patients, 29.4% of GT3 patients, and 100.0% of GT4 patients. The pretreatment frequencies of RASs ranged from 1.2 to >99%, and the majority of these RASs persisted through week 48. No association with specific frequencies of RASs was found that could predict a higher likelihood of persistence. In GT1b, GT2, and GT4 patients, the HCV RNA reduction was high, independently of the presence of pretreatment NS5A RASs; however, in GT1a and GT3 patients, the presence of pretreatment RASs was associated with reduced viral response in some patients. The HCV RNA decline was consistent with in vitro HCV replicon analyses for GT1a, in which replicons harboring Y93H/N or double mutants resulted in a >100-fold reduced susceptibility to VEL. A similar finding was obtained for Y93H in GT3 replicons. However, the reduced viral response was not the same for all GT1a and GT3 patients with pretreatment RASs, possibly because of different frequencies of pretreatment RASs and other individual genetic variation within the HCV quasispecies affecting susceptibility to VEL. Thus, the in vitro EC₅₀s of specific RASs may not always be predictive of in vivo viral responses, given the diversity of the viral quasispecies. Interestingly, among the 10 GT1a patients with pretreatment RASs, 7 had multiple RASs before treatment was initiated. The natural presence of such RASs is uncommon in treatment-naive patients (42); thus, the reason for this is unclear and we cannot exclude possible prior NS5A treatment in these patients or the possibility that GT1a viruses are more prone to naturally occurring variation and acquired RASs to NS5A inhibitors. Of the two GT4 patients enrolled in this study, both had NS5A RASs at positions L30/H/S/R pretreatment. These variants did not show reduced susceptibility to VEL in vitro, and patients had >3-log₁₀ reductions in HCV RNA, indicating that the presence of these RASs prior to treatment did not preclude a viral response to VEL.

Following treatment with VEL, NS5A RASs emerged on treatment at more positions in the GT1a-infected patients than in those infected with the other genotypes; emergent RASs were observed at only two NS5A positions in GT1b and GT2 (L31, Y93) and three positions in GT3 (Y93, L31, E92). In GT4, emergent NS5A RASs were observed at positions L28, M31, P32L, and Y93. GT1a-infected patients showed RASs at NS5A positions M28, Q30, L31, P32, H58, and Y93, with Y93, M28, and L31 being the most prevalent. In GT1b and GT2b, L31M/V and Y93H were the most commonly observed RASs, and in GT3, Y93H and E92K were the most frequently observed RASs following treatment with VEL. On day 17, emerging NS5A RASs were observed in only one of four patients who received 5 mg of VEL but in the majority of the patients who received larger doses. This might be explained by the degree of antiviral pressure at different doses. More substantial suppression of wild-type virus by larger doses resulted in more frequent detection of resistant variants, although RASs may be generated during treatment because of the high error rate of HCV RNA replication. This is consistent with other phase 1 HCV DAA monotherapy studies (28, 43, 44) and suggests that RASs preexist at low levels (<1% of viral population) and become detectable once the wild-type population is sufficiently suppressed.

The possible persistence of NS5A RASs selected during VEL monotherapy was investigated in longitudinal samples posttreatment of the patients enrolled in this study. The frequencies of most of the NS5A RASs that were present prior to treatment increased and persisted through the 48-week follow-up period. Additional NS5A RASs that developed during treatment were more likely to become undetectable during the follow-up period. Indeed, 22 (66.7%) of 33 patients with available sequences and without pretreatment NS5A RASs did not have detectable NS5A RASs at week 48 (with a 1% cutoff). RASs were detected at low frequencies at week 48 in 29% of GT1a patients and 50% of GT1b and GT3 patients. In contrast, the treatment-emergent NS5A RASs were no longer detected above the 1% deep-sequencing assay cutoff in any of the GT2 or GT4 patients at week 48. In GT1a patients, L31M was the most frequent RAS persisting at week 48. However, through week 12, L31V was more common, which correlates with a greater reduction in the susceptibility of NS5A-L31V viruses to VEL, compared with L31M in the GT1a replicon. The prevalence of Y93H and H58D gradually declined over time, but they were still detectable at low levels in a few patients at week 48. In contrast, M28T, Q30E/K/L, L31I, and Y93C/N were no longer detected at week 48. In GT1b, L31M and Y93H were still detected at week 48; however, L31V and Y93C/N/S were not observed at week 48. In GT3 patients, only Y93H persisted through week 48. Overall, the gradual decline in the number of patients with NS5A RASs and the frequencies of newly emergent RASs through week 48 suggests that viruses with these RASs are less fit than the wild-type virus; however, the persistence of specific variants at the same positions differed from patient to patient.

Phenotypic analyses of the RASs detected pretreatment or selected following VEL treatment showed that most of the RASs in the GT1a replicon demonstrated no (≤2.5-fold EC₅₀ change, Q30L/R/H, Y93F) or low- to midlevel resistance (2.5- to 100-fold EC₅₀ change, Q30K/E, L31I/M/V, P32L, H58D, Y93C/S). High levels of resistance (>100-fold EC₅₀ change) were observed in Y93H/N/R/W and double mutants. All of the NS5A RASs in the GT1b replicon conferred <3.3-fold resistance to VEL. The majority of the GT2a, GT3a, and GT4a single mutants displayed no or low levels of resistance to VEL; however, Y93H in GT3a conferred 723.5-fold reduced susceptibility to VEL.

In summary, VEL demonstrated broad genotypic activity and improved activity against preexisting resistant variants in a phase 1b 3-day monotherapy study. Although pretreatment RASs were associated with a slightly reduced HCV RNA response in GT1a and GT3 patients, current HCV treatment strategies are based on the combined use of DAAs and should overcome any minor impact of pretreatment RASs. Combination treatment with VEL and SOF may provide an effective treatment option for patients infected with HCV GT1 to GT6.

Funding Statement

This work was supported by Gilead Sciences.