Interplay of Amino Acid Residues at Positions 28 and 31 in NS5A Defines Resistance Pathways in Hepatitis C Virus Genotype 2

Hepatitis C virus (HCV) genotype 2 (GT2) represents approximately 9% of all viral infections globally. While treatment outcomes for GT2-infected patients have improved substantially with direct-acting antiviral agents (DAAs) compared to alpha interferon, the presence of polymorphisms in NS5A can impact the efficacy of NS5A inhibitor-containing regimens. Thus, pathways of NS5A resistance were explored in GT2 subtypes using elbasvir, an NS5A inhibitor with broad genotype activity.

ABSTRACT

Hepatitis C virus (HCV) genotype 2 (GT2) represents approximately 9% of all viral infections globally. While treatment outcomes for GT2-infected patients have improved substantially with direct-acting antiviral agents (DAAs) compared to alpha interferon, the presence of polymorphisms in NS5A can impact the efficacy of NS5A inhibitor-containing regimens. Thus, pathways of NS5A resistance were explored in GT2 subtypes using elbasvir, an NS5A inhibitor with broad genotype activity. Resistance selection studies, resistance analysis in NS5A-inhibitor treated virologic failures, and analyses of antiviral activities in replicons bearing a panel of GT2 subtype sequences and amino acid substitutions introduced by site-directed mutagenesis were performed to define determinants of inhibitor susceptibility. Elbasvir showed differential antiviral activities in replicons bearing GT2 sequences. The 50% effective concentration (EC₅₀) values for replicons bearing reference NS5A sequences for GT2a and GT2b were 0.003 and 3.4 nM, respectively. Studies performed with recombinant replicons demonstrated cross talk between amino acid positions 28 and 31. The combination of phenylalanine and methionine at positions 28 and 31, respectively, conferred the highest potency reduction for elbasvir in GT2a and GT2b. This combination was observed in failures seen in the C-SCAPE trial. Addition of grazoprevir, an NS3/4A protease inhibitor, to elbasvir more effectively suppressed the emergence of resistance in GT2 at modest inhibitor concentrations (3× EC₉₀). Ruzasvir, a potent, pan-genotype NS5A inhibitor, successfully inhibited replicons bearing GT2 resistance-associated substitutions (RASs) at positions 28 and 31. The results demonstrate that cross talk between amino acids at positions 28 and 31 in NS5A modulated inhibitor potency and may impact treatment outcomes in some HCV GT2-infected patients.

INTRODUCTION

Hepatitis C virus (HCV) is a highly prevalent human pathogen that is estimated to have infected more than 71 million people globally as of 2015 (1, 2). It is estimated that, worldwide, more than 350,000 lives are lost each year from HCV-related liver disease (3). In recent years, substantial progress has been made in the development of therapies to eradicate the infection. A number of combinations of direct-acting antiviral agents (DAAs) have been approved for treatment of HCV infection (4). These DAAs target nonstructural proteins, including HCV NS3/4A protease, NS5A protein, and NS5B polymerase, that are critical for viral replication (5). Although treatment options now exist for all HCV genotypes (GTs), many infected individuals are not aware of their infection and remain chronically infected or may not have access to therapy and therefore remain at risk for developing liver cirrhosis, end-stage liver disease, and hepatocellular carcinoma.

HCV is genetically diverse, with 7 major GTs recognized, most of which comprise multiple subtypes (6). Each genotype displays a unique geographical distribution (7). The most prevalent and the focus of most studies is GT1. The other major genotypes of broad distribution include GT2 and GT3. Recently, GT3 has been the subject of a number of studies as it is now considered to be a more-difficult-to-treat genotype (8). In contrast, GT2 has not been a major focus of research activity, in part due to its early susceptibility to therapy, including sofosbuvir, an NS5B polymerase inhibitor (9–11). GT2 accounts for about 9% of all HCV infections worldwide, with about 16 million people infected, of whom approximately half live in eastern Asia (12, 13). In sub-Saharan Africa, 23% of HCV infections are attributable to GT2. Among the many GT2 subtypes, the most prevalent are 2a, 2b, and 2c (7). GT2a is found mostly in Asia, Australia, and New Zealand, while GT2b is the most prevalent GT2 subtype in North America (7). GT2c is more prevalent in Europe (7).

The treatment for HCV GT2 infection has evolved over the years. Prior to the advent of DAAs, treatment primarily involved interferon injections and use of ribavirin for long treatment durations. While treatment was relatively efficacious and achieved sustained virologic response (SVR) rates of ∼80% after 24 weeks, many debilitating side effects, including flu-like symptoms and anemia, afflicted patients (14). More-successful and safer treatments have become available with the introduction of DAAs, where a >95% SVR is not uncommon. However, susceptibility to therapy can be impacted by the presence of certain polymorphisms, particularly in NS5A (15–18). As NS5A inhibitors have become key components of HCV therapy, we explored differences in GT2 subtypes to obtain a deeper understanding of the underlining factors associated with susceptibility and/or resistance to enable the design of more-potent inhibitors with a higher barrier to resistance. The need for potent inhibitors also stems from the observation that when patients fail NS5A inhibitor-containing therapy, resistance-associated substitutions (RASs) are commonly selected that tend to result in cross-resistance among the members of the inhibitor class. These RASs often persist for a long period (>96 weeks) in patients who fail therapy and therefore may impact future treatment options (19). Thus, we investigated resistance in GT2 with the aid of elbasvir, a potent NS5A inhibitor with broad genotype activity. Elbasvir is approved for the treatment of HCV infection in combination with grazoprevir, an NS3/4A protease inhibitor, in patients infected with either GT1 or GT4, or, for patients infected with GT3, in combination with grazoprevir and sofosbuvir, depending on the jurisdiction.

RESULTS

Elbasvir shows differential replicon activities in GT2 reference sequences.

To understand the impact of polymorphisms in GT2 subtypes, an initial assessment of the potency of elbasvir (formerly MK-8742) was made by comparing its activities in standard reference sequences. The antiviral activities of elbasvir in the key GT2 subtypes were compared to its activities in reference sequences for GT1a and GT1b. As shown in Table 1, elbasvir demonstrated potent activities in the reference sequences for GT1a, GT1b, and GT2a but not in GT2b, which harbors a resistance-conferring methionine residue at position 31. Substitution of methionine to leucine (the alternative polymorphic option) at position 31 (M31L) recovered substantial activity of the compound in the GT2b replicon. While a 10-fold improvement in potency was achieved with the M31L substitution in GT2b, the 50% effective concentration (EC₅₀) and EC₉₀ values represented levels of potency roughly 2 orders of magnitude lower than those seen with GT2a JFH-1 [Table 1].

TABLE 1.

Activity of elbasvir in GT1 and GT2 replicons

Replicon	EC50 ± SD (nM)	Fold shift (relative to GT1a)	EC90 ± SD (nM)	Fold shift (relative to GT1a)
GT1a_NC004102 (H77)	0.007 ± 0.004	1	0.017 ± 0.009	1
GT1b_AJ238799 (Con1)	0.003 ± 0.002	0.4	0.006 ± 0.004	0.4
GT2a_AB047639 (JFH-1)	0.003 ± 0.001	0.4	0.019 ± 0.01	1
GT2b_AB030907	3.4 ± 2.6	486	11.0 ± 4.8	647
GT2b_AB030907 M31L	0.32 ± 0.38	46	1.0± 1.2	59

De novo resistance selection studies with elbasvir in GT2 replicons.

To investigate the breadth of resistance pathways in GT2, two sets of de novo resistance selection studies, in a full-length GT2a (JFH-1) replicon and in a chimeric replicon bearing the GT2b NS5A sequences in the background of GT2a (JFH-1), were conducted with elbasvir. RNA was isolated from the resistant colonies, subjected to reverse transcription, and sequenced to identify potential mutations. The resistant colonies that emerged at each inhibitor concentration tested were pooled and assayed to determine the impact on the antiviral activity of elbasvir. In GT2a, only two amino acid (AA) substitutions, F28S/Y, were detected in the resistant colonies. However, with increasing selective pressure (up to 1,000× EC₉₀ value), all colonies emerged with F28S as the dominant substitution and main driver of resistance in the subtype. The resistant colonies seen at each concentration tested were pooled and assayed, and the results showed that they conferred potency reductions in the range of 3 to 4 orders of magnitude relative to pooled colonies derived from dimethyl sulfoxide (DMSO) treatment controls. In selections conducted in the GT2b replicon bearing the resistant L31M polymorphism in NS5A, the only amino acid substitution detected at all concentrations tested was Y93H. The combination of the Y93H substitution and the preexisting L31M substitution resulted in a replicon cell that was resistant to elbasvir in assays of pooled emergent colonies (at each concentration level) (Table 2).

TABLE 2.

Summary of de novo resistance selections in GT2 subtypes^a

Genotype	Concn of elbasvir (fold over EC90 value)	Variant(s) observed at all concns	Elbasvir potency in pooled surviving colonies (EC90, nM)	Fold shift
GT2a	DMSO	nd	0.019	1
	1×	F28S, F28Y	47	2,474
	10×	F28S, F28Y	627	33,000
	100×	F28S	827	43,526
	1,000×	F28S	911	47,947
GT2b_31M	DMSO	nd	11	1
	1×	Y93H	>>1,000	>100
	10×	Y93H	>>1,000	>100
	100×	Y93H	>>1,000	>100
	1,000×	Y93H	>>1,000	>100

^aFold shift values are expressed relative to the DMSO value (set to 1) and rounded to whole numbers. nd, none detected.

To confirm that the detected amino acid substitutions were required and responsible for the resistant colonies, they were reengineered into the parental replicons and characterized. A GT2a (JFH-1) replicon bearing either the F28S or F28Y substitution conferred a 99,000-fold or 93-fold potency reduction (based on EC₅₀ values), respectively. This was consistent with the selection of only the F28S substitution at higher concentrations of the inhibitor. An introduction of the Y93H substitution in addition to the L31M substitution in GT2b resulted in a replicon that further reduced the potency of elbasvir by a factor of 20 (see Table S1 in the supplemental material).

Activity of elbasvir in replicons bearing synthetized sequences of GT2 clinical isolates.

Given the potential for single amino acid substitutions to elicit resistance, elbasvir was profiled against a broad set of GT2 clinical isolates. The NS5A sequences were evaluated using the transient replicon assay to circumvent fitness issues that might prevent generation of stable replicons. In the transient assay, luciferase activity was monitored as a function of replication (see Materials and Methods). Fourteen sequences were investigated that represented subtypes 2a, 2b, and 2c, with 6, 5, and 3 NS5A sequences, respectively. One of the 5 GT2b sequences failed to replicate adequately even as a transient replicon.

Elbasvir exhibited substantial potency differences among the GT2 subtypes (Table 3). Apart from the reference GT2a (JFH-1) sequence, the GT2a sequences were less susceptible than the GT2b and GT2c sequences. Although the GT2b sequences were more susceptible than the GT2a sequences, the levels of potency differed among the 4 sequences and ranged from 0.01 to 1.5 nM. The extent of potency variation among GT2c sequences was comparable to that seen among the GT2b sequences. The data are summarized in Table 3.

TABLE 3.

Elbasvir inhibitory potency in transient HCV replicons bearing synthesized GT2 NS5A sequences from clinical isolates^a

GT	Replicon accession no.	EC50 (nM) ± SD	EC90 (nM) ± SD
2a	AB047639 (JFH-1)	0.039 ± 0.018	0.088 ± 0.03
	D00944	29.7 ± 3.4	36 ± 2.0
	AY746460	24.5 ± 6.8	42 ± 12
	AF238481	14.8 ± 2.6	21 ± 5.5
	AF169003	15.7 ± 3.2	24 ± 9
	AF169002	7.7 ± 2.2	15 ± 9
2b	AF238486	1.22 ± 0.372	4.23 ± 2.75
	AY232730	Unfit	Unfit
	AY232742	1.54 ± 0.41	3.09 ± 1.25
	AY232749	0.010 ± 0.001	0.031 ± 0.012
	DQ430815	0.605 ± 0.220	1.67 ± 0.593
2c	KC197227	0.024 ± 0.006	0.107 ± 0.032
	KC197228	4.94 ± 0.024	9.17 ± 1.50
	D50409	0.135 ± 0.063	0.388 ± 0.245

^aThe GT2 NS5A sequences were evaluated in the JFH-1 Gluc replicon background. Potency was estimated by changes in Gaussia luciferase levels. Values represent averages of results from ≥3 experiments. Unfit, replicon gene did not replicate in culture.

As the reference GT2a (JFH-1) sequence behaved differently from other GT2a sequences, we conducted a phylogenetic analysis to investigate its relatedness to the other sequences tested. The sequences for each subtype clustered together, except for the reference GT2a (JFH-1) sequence, which emerged as an outlier (see Fig. S1 in the supplemental material).

Contributions of specific amino acid substitutions in GT2 to inhibitor susceptibility.

In order to obtain further insights into the differences in subtype susceptibilities, we compared amino acid residues at key positions associated with resistance to NS5A inhibitors across the clinical isolate sequences (20). An alignment of the amino acids detected at each of the positions within the sequences is shown in Table 4. The residues observed in the reference sequences for GT1a, GT1b, and GT2b are also provided for comparison. The alignment of the GT2 clinical isolate sequences is available in Fig. S2. The amino acids observed at four positions (i.e., at positions 32, 38, 92, and 93) were conserved across the sequences and were therefore not expected to contribute substantially to potency differences. The single substitution at position 30 (in GT2c D50409) and position 58 (in GT2a AF169002) compared to other sequences was not likely to be a major factor given the profile of the compound across the three subtypes. While the amino acid residue at position 24 was conserved within sequences of a subtype and differed between GT2a and GT2b/GT2c, the potency of elbasvir in the reference GT2a JFH-1 isolate compared to its potency in the other clinical isolates did not support an interpretation of a meaningful contribution from that position. The process of elimination resulted in two potentially impactful amino acid positions: 28 and 31. The most susceptible GT2a, GT2b, and GT2c sequences, represented by AB047639 (JFH-1), AY232749, and KC197227, respectively, all retained a leucine (L) at position 31 instead of a methionine (M), potentially implicating the position in the susceptibility results. While most of the less susceptible clinical isolates retained a phenylalanine (F) at position 28, there appeared to be exceptions for GT2a JFH-1. Position 28 was also linked to reduced susceptibility based on de novo resistance selection studies (Table 2). Hence, we investigated the contributions of amino acid changes at positions 28 and 31 to elbasvir activity in GT2 replicons.

TABLE 4.

Analysis of NS5A sequences from GT2 clinical isolates

Subtype	Clinical isolate	Amino acid at indicated position associated with inhibitor resistance in NS5A
		24	28	30	31	32	38	58	92	93
GT2a	AB047639	T	F	K	L	P	S	P	C	Y
	D00944	T	F	K	M	P	S	P	C	Y
	AY746460	T	F	K	M	P	S	P	C	Y
	AF238481	T	F	K	M	P	S	P	C	Y
	AF169003	T	F	K	M	P	S	P	C	Y
	AF169002	T	F	K	M	P	S	S	C	Y
GT2b	AF238486	S	L	K	M	P	S	P	C	Y
	AY232730	S	L	K	M	P	S	P	C	Y
	AY232742	S	L	K	M	P	S	P	C	Y
	AY232749	S	L	K	L	P	S	P	C	Y
	DQ430815	S	L	K	M	P	S	P	C	Y
GT2c	KC197227	S	F	K	L	P	S	P	C	Y
	KC197228	S	C	K	L	P	S	P	C	Y
	D50409	S	F	R	L	P	S	P	C	Y
GT1aa	NC004102	K	M	Q	L	P	S	H	A	Y
GT1ba	AJ238799	Q	L	R	L	P	S	P	A	Y
GT2ba	AB030907	S	L	K	M	P	S	S	C	Y

^aThe amino acid residues for the subtype reference sequences are included for reference.

We focused on the 2 major global subtypes, GT2a and GT2b, in order to address the contributions from the two positions in a rational manner. Four reciprocal constructs with specific substitutions in NS5A based on their presence in the clinical isolates were engineered into the parental subtype reference sequences. As summarized in Table 5, in the GT2a replicon, the introduction of methionine at position 31 in place of leucine (L31M) resulted in a replicon that conferred a 564-fold potency loss with respect to elbasvir. This L31M-induced resistance was reversed when the amino acid residue at position 28 was switched from a phenylalanine to leucine (F28L). Interestingly, the reversal of the resistance in GT2a occurred despite the presence of the methionine at position 31 (M31). The reciprocal substitutions were similarly investigated in GT2b; the introduction of a L28F substitution resulted in a replicon that was 13-fold less susceptible to elbasvir. In line with observations in GT2a, an M31L substitution restored susceptibility to the compound; this was also in spite of the presence of the resistance-associated phenylalanine at position 28 (F28). Amino acid substitutions at positions 28 and 31 had similar effects on the potency of daclatasvir, another NS5A inhibitor, in GT2 replicons (see Table S2). Although the combination of L28 and L31 was not specifically tested, we note that isolate AY232749 had that combination and was susceptible to elbasvir. Thus, the specific residues at positions 28 and 31 appear to have a cooperative influence on inhibitor susceptibility in the GT2 subtypes.

TABLE 5.

Activity of elbasvir in transiently expressed GT2 replicons bearing reciprocal substitutions at positions 28 and 31 in NS5A

Genotype (accession no.)	Amino acid residue at indicated key position		EC50 ± SD (nM)	Fold shift (relative to the parental replicon)	Fold shift (relative to GT1a)
	28	31
GT2a (AB047639)	F	L	0.039± 0.018	1	5.6
	F	M	22 ± 8	564	3,143
	L	M	0.15 ± 0.03	3.9	21.4
GT2b (AB030907)	L	M	4.8 ± 0.83	1	686
	F	M	63 ± 6.9	13	9,000
	F	L	3.5 ± 0.6	0.7	100

Overcoming the emergence of NS5A resistance.

Having determined and demonstrated the contribution of NS5A resistance pathways in GT2, we investigated strategies to overcome the effects. We began by investigating the combination of elbasvir and grazoprevir, which is an HCV NS3/4A protease inhibitor with a nonoverlapping resistance mechanism. The combination substantially inhibited the emergence of resistance in GT2. Figure S3 shows a matrix of pictures of representative dishes from the inhibitor combination studies. Table S3 and Table S4 show the number of resistant colonies and the substitutions detected and the potency shifts that they confer with respect to elbasvir, respectively. The substitutions detected occurred at positions 28, 31, and 93.

An alternative strategy to increase the barrier to resistance with a potential to overcome emergence of resistance was to develop a more potent, pan-genotype NS5A inhibitor. The chemical approaches undertaken that resulted in the discovery of ruzasvir have been reported recently (21). We tested ruzasvir against the same replicons bearing the reciprocal amino acid substitutions in GT2a and GT2b subtypes. As reported in Table 6, ruzasvir was more potent and the fold shifts relative to the parental replicons were lower than what had been seen for elbasvir. Interestingly, the trends seen were the same as those observed in the case of elbasvir; the combination of phenylalanine and methionine at positions 28 and 31, respectively, resulted in the least susceptible replicons among all substitutions. The effect of the resistance-conferring substitutions was reversible with the introduction of the more permissive and nullifying substitution established with elbasvir (Table 5).

TABLE 6.

Activity of ruzasvir in replicons bearing reciprocal substitutions at positions 28 and 31 in NS5A from GT2a and GT2b subtypes

Genotype (accession no.)	Amino acid residue at indicated key position		EC50 ± SD (nM)	Fold shift (relative to the parental replicon)	Fold shift (relative to GT1a)
	28	31
GT2a (AB047639)	F	L	0.008 ± 0.003	1	1
	F	M	0.172 ± 0.037	21.5	22
	L	M	0.027 ± 0.024	3.4	3.4
GT2b (AB030907)	L	M	0.038 ± 0.013	1	4.8
	F	M	2.7 ± 0.35	71	338
	F	L	0.015 ± 0.005	0.4	1.9

DISCUSSION

We employed resistance selection and reverse pharmacology to characterize the activity of elbasvir in NS5A from HCV GT2. An initial evaluation of the compound’s activity in the FDA-recommended reference sequences for GT2 revealed approximately 1,000-fold-higher potency in GT2a than in GT2b. Unexpectedly, reversion of the naturally occurring resistant M31 residue to the susceptible M31L substitution restored elbasvir potency in the GT2b replicon only partially. To obtain a deeper understanding of the pathways of resistance for elbasvir in GT2, colony formation assays were conducted that enabled the identification of amino acid substitutions (at position 28 in GT2a and positions 31 and 93 in GT2b) that conferred potency reductions to the compound when reengineered into the parental replicon. The positions associated with resistance in the colony formation assays predicted the activity of elbasvir in a panel of replicons bearing NS5A sequences derived from clinical GT2 isolates of the major subtypes. The higher potency of elbasvir in GT2b and GT2c replicons than in GT2a (except isolate JFH-1) was demonstrated to be a result of amino acid substitutions at positions 28 and 31. With respect to GT2a JFH-1, our data suggest that the sequence may not be a representative reference for GT2a. Phylogenic analysis demonstrated that it does not cluster with the other GT2a sequences tested, and, indeed, the response to elbasvir that it exhibited differed from the responses seen with the other GT2a sequences. These differences were instructive and were exploited to make deductive analyses of the potential impact of amino acid positions on elbasvir potency. While it has been reported that the presence of a methionine residue at position 31 in GT2 may render sequences resistant to NS5A inhibitors, the influence of the residue at position 28 had not previously been demonstrated (20). A comparison of the sequences of GT2a and GT2b revealed that amino acid residues at positions 28 and 31 may account for the differential activities of elbasvir in the subtypes. Further analysis demonstrated that the amino acids at these two positions act in concert to influence inhibitor potency. The combination of phenylalanine (F) and methionine (M) at positions 28 and 31, respectively, conferred the greatest potency reductions for the inhibitor. Interestingly, one residue may act to nullify the effect of the other with respect to susceptibility. For elbasvir, the presence of L31 not only abolished the resistance-enhancing effect of F28 but also conferred susceptibility; this effect was more pronounced with ruzasvir (Table 6). Clearly, the nature of the manner in which a compound interacts with NS5A influences the contribution of the two positions to inhibitory activity. Unfortunately, the available crystal structures of domain 1 of NS5A do not enable us to address the issue of the structural basis of the interplay between the residues at positions 28 and 31 as the constructs used for crystallization either do not span the positions or are not clearly resolved (22–25). One may speculate that the positioning of the two residues occupying positions 28 and 31 potentially creates a binding locus that critically influences how the compound interacts with NS5A. Productive binding interactions may be favored in one configuration while being disfavored in a different configuration. Availability of data representing the tertiary structure of the entirety of domain 1 (preferably in complex with an inhibitor) should enable a better understanding of the underlying mechanism of the impact of the two positions on inhibitor potency.

In an analysis of 178 GT2 sequences available in public databases, the proportions of sequences bearing the more-difficult-to-inhibit combination of F28 plus M31 were ∼80% in GT2a and <10% in GT2b. The number of GT2c sequences available was too low (only 4 sequences) to conclude much from it; however, other sequences with unknown subtypes also appeared to have a not insignificant prevalence of the more resistant combination (Table S5). The existence of the phenomenon in other HCV genotypes will require further studies.

Given the tremendous diversity of HCV sequences across genotypes, the potential for preexisting RASs (as observed in some GT2 clinical isolates) is high. One strategy to overcome the emergence of resistance is that of using combinations of available inhibitors with nonoverlapping resistance pathways directed at distinct viral proteins. The combination of elbasvir, an NS5A inhibitor, and grazoprevir, an NS3/4A protease inhibitor, created a higher barrier to resistance and suppressed the emergence of resistant colonies at lower multiples of their respective EC₉₀ values than were seen with each inhibitor alone. The pathways of resistance in NS5A were not altered from those observed when elbasvir was used alone; however, the combination was additive (at a minimum) or synergistic.

An alternative and a more challenging strategy to improve activity against nonsusceptible genotypes is that of designing and synthesizing a more potent compound. The chemistry strategies that resulted in the development of ruzasvir were reported recently (21). Profiled against the same panel of replicon cells, ruzasvir was more potent than elbasvir, although the effect of the interplay between the amino acid residues at positions 28 and 31 on inhibitor potency remained. Both elbasvir and ruzasvir have been evaluated in HCV patients infected with GT2 in clinical trials. Comparisons of the clinical studies are complicated because the combination partners are different and because there are differences in patient dispositions, among other factors. Nonetheless, there are observations from the individual trials that are important to consider. In the phase 2 C-SCAPE trial, the combination of elbasvir (50 mg) and grazoprevir (100 mg) administered once daily for 12 weeks with ribavirin resulted in 80% of GT2-infected patients achieving a SVR (26). GT2 patients harboring the L31M substitution achieved a lower SVR rate (67%) than those without the substitution (93%). Interestingly, all four patients who failed to achieve a SVR harbored the more resistant combination of F28 plus M31. In part A of the phase 2 C-CREST-1 and C-CREST-2 trials that investigated a 3-drug combination over an 8-week treatment duration, the combination of grazoprevir (100 mg), uprifosbuvir (450 mg), and ruzasvir (60 mg) was selected for further testing over the combination of grazoprevir, uprifosbuvir, and elbasvir (50 mg), as the former consistently resulted in higher 12-week SVR (SVR12) rates in HCV genotypes, including GT2 (27). In the phase 2 C-BREEZE trial, the combination of ruzasvir (60 mg) and uprifosbuvir, an NS5B polymerase prodrug (450 mg), administered once daily for 12 weeks resulted in a 100% SVR among GT2-infected patients (28). It seems reasonable to conclude that the success of the trial was, in part, due to the activity of ruzasvir in GT2.

In summary, our studies in GT2 revealed resistance pathways and uncovered cross talk between amino acid residues at positions 28 and 31 influencing inhibitor potency and may be useful to gauge susceptibility of NS5A sequences in the clinic. The studies helped inform discovery of a next-generation NS5A inhibitor that, used together with an NS5B prodrug, resulted in effective treatment of GT2-infected patients in a phase 2 trial.

MATERIALS AND METHODS

Compounds.

Elbasvir (NS5A inhibitor), N,N′-[[(6S)-6-phenyl-6×H-indolo[1,2-c][1,3]benzoxazine-3,10-diyl]bis[1H-imidazole-5,2-diyl-(2S)-2,1-pyrrolidinediyl[(1S)-1-(1-methylethyl)-2-oxo-2,1-ethanediyl]]]bis[carbamic acid] C,C′-dimethyl ester (Fig. 1A), was prepared as reported previously (29).

FIG 1.

Chemical structures of (A) elbasvir (NS5A inhibitor), (B) grazoprevir (NS3/4A protease inhibitor), and (C) ruzasvir (NS5A inhibitor).

Grazoprevir (NS3/4A protease inhibitor), N-[[[(1R,2R)-2-[5-(3-hydroxy-6-methoxy-2-quinoxalinyl)pentyl]cyclopropyl] oxy]carbonyl]-3-methyl-l-valyl-(4R)-4-hydroxy-l-prolyl-(1R,2S)-1-amino-N-(cyclopropylsulfonyl)-2-ethenylcyclopropanecarboxamide cyclic (1→2) ether (Fig. 1B), was prepared as reported previously (30, 31).

Ruzasvir (NS5A inhibitor), N,N’-[[(6S)-6-(2-cyclopropyl-5-thiazolyl)-1-fluoro-6×H-indolo[1,2-c][1,3] benzoxazine-3,10-diyl]bis[1H-imidazole-5,2-diyl-(2S)-2,1-pyrrolidinediyl[(1S)-1-(1-methylethyl)-2-oxo-2,1-ethanediyl]]]bis[carbamic acid] C,C’-dimethyl ester (Fig. 1C), was prepared as reported previously (21).

Cell culture.

Human hepatoma cell line Huh7 or Huh7.5 (32) was cultured in Dulbecco’s minimal essential medium (DMEM) supplemented with 2 mM glutamine, nonessential amino acids (NEAA), 0.075% sodium bicarbonate, 100 U/ml penicillin, 100 μg/ml streptomycin, 10% fetal bovine serum (FBS), and 10 mM HEPES at a pH of 7.5. Stable replicons generated in Huh7 or Huh7.5 cells were cultured in 0.5 mg/ml G418 (Cellgro, VA). The generation and establishment of HCV replicons in Huh cell lines were described previously (33).

Generation of replicons. (i) Stable HCV replicon system.

Full-length HCV subgenomic replicon cDNAs for GT1a (H77) (34), GT1b (Con1) (35), and GT2a (JFH-1) (36) have been described elsewhere. The GT2a (JFH-1) cDNA was used as a model genome background to receive NS5A gene sequences from GT2b (JPUT971017) to create chimeric replicon genomes. Chimeric cDNAs were designed and synthesized (Genewiz, South Plainfield, NJ) as cassettes for cloning into replicon background vectors (Table 1). Replicons bearing the NS5A genes to be tested carried the complete NS5A sequence (Table 1). RNA was transcribed from XbaI-linearized plasmids using a T7 MEGAscript kit (Ambion/Life Technologies) per the manufacturer’s protocol and was used to transfect Huh7 cells for stable cell line generation, as described previously (33).

(ii) NS5A genes synthesized from clinical isolates in a GT2a (JFH-1)-based transient replicon system.

The transiently expressed HCV replicon system was based on a subgenomic GT2a (JFH-1) sequence (36), modified with the Gaussia luciferase (G-luc) gene integrated in-frame in the 5′ cistron (as a MluI-PmeI insertion) to replace the NPTII gene, and was used to analyze the resistance phenotypes of HCV NS5A genes from GT2 subtypes. Sequences representing these GT2 subtypes were identified from searches of GenBank (NCBI, U.S. National Library of Medicine) and the European HCV database (euHCVdb; EMBL); accession number identifiers are listed in the relevant tables. The NS5A gene of interest was introduced by cloning a cassette of the modified target gene (made by gene synthesis; Genewiz, South Plainfield, NJ) into the GT2a-G-luc background.

(iii) Inhibition studies in replicon cells (TaqMan assay).

Stable replicon cell lines were tested for compound susceptibility. Briefly, Huh 7.5 replicon-bearing cells (referred to here as replicon cells) were seeded in 384-well plates in DMEM containing 0.5 mg/ml G-418 (33). Compounds were serially diluted in dimethyl sulfoxide (DMSO) and added to cells (in the presence of 5% fetal calf serum [FCS]) at a 1:200 dilution 24 h after cells were seeded. The final DMSO concentration in the medium was 0.5% (vol/vol). After 72 h of incubation, cells were harvested and subjected to real-time PCR (RT-PCR) analysis as reported previously (33, 37–39). The primer/probe sets used for the PCR analysis on an ABI Prism 7900HTS sequence detection system (Perkin Elmer) were reported previously (39). The threshold cycle (C_T) values were plotted against the log of compound concentrations and fitted to the sigmoid dose-response model using DataAnalyzer 4.0.30639.0 (Merck Canada Inc., Kirkland, QC, Canada) to obtain the EC₉₀—the drug concentration that was needed to reduce the response by 90% compared to that seen with a no-treatment control plate. The EC₉₀ value was computed as the drug concentration required for an increase of 3.2 ΔC_T over the baseline. The 50% effective concentration (EC₅₀) was the concentration needed to achieve a ΔC_T of 1 over the baseline. Each replicon cell was tested at least 3 times to compute the mean EC₅₀ and EC₉₀ values.

(iv) Inhibition studies in replicon cells (luciferase assay).

Transfection of G-luc-bearing replicon RNA, transcribed from cDNA using T7 MEGAscript (Ambion/Life Technologies), into Huh 7.5 cells results in replication of the HCV RNA and expression of the G-luc protein (40). The levels of expressed G-luc protein directly correlate with HCV RNA copy number and viral protein translation. The transiently expressed replicon system allows growth and characterization of viruses bearing mutations that may have detrimental effects on viral fitness and that may result in an inability to establish stable replicons. In this protocol, 5 × 10⁶ Huh 7.5 cells were transfected by electroporation with 0.5 μg replicon RNA on a Bio-Rad Gene Pulser Xcell system using exponential protocol values of 270 V, 950 μF capacitance, and resistance of 100 Ω. Culture supernatant was collected and refreshed after 6 h and each day thereafter.

To measure the effect of changes in genomes on compound potency, transfected cells were seeded in 96-well plates (3 days after RNA transfection). Compounds were serially diluted in DMSO and added to cells (in the presence of 10% FCS) at a 1:200 dilution 24 h after the cells were seeded. The final DMSO concentration in the medium was 0.5% (vol/vol). After 48 h, the medium with compound was refreshed. After 72 h of total compound incubation, supernatants were harvested and subjected to measurement of Gaussia luciferase activity using a BioLux Gaussia Luciferase assay kit (New England Biolabs) per the manufacturer’s protocol (40). Luciferase activity was measured on an Envision plate reader (Perkin-Elmer). EC₅₀s were computed by calculating the percentage of inhibition of luciferase activity at each concentration, with EC₅₀s derived using nonlinear regression (curve fit)–sigmoidal dose-response (variable slope) data and GraphPad Prism software (GraphPad Software, Inc.).

De novo resistance selection studies.

To select replicon cells bearing genomes resistant to inhibitors, subconfluent monolayers of replicon cells were cultured with concentrations of the compounds at multiples of the EC₉₀ values (39). All cells were passaged at a 1:10 ratio when they were ∼95% confluent. The colonies that survived selection were pooled and expanded for further analysis. Total cellular RNA was isolated from pooled colonies and amplified by RT-PCR. The RT-PCR products were purified with a QIAquick PCR purification kit (Qiagen), and the NS5A region was sequenced. Alternatively, the RT-PCR products were cloned into TOPO TA vector (Invitrogen), and the plasmid DNA from 12 bacterial colonies was sequenced.