Novel NS5B Resistance‐Associated Substitution Emerges Under Failing Sofosbuvir/Ledipasvir Therapy

Watch a video presentation of this article

Watch the interview with the author (English)

Abbreviations

DAA

direct‐acting antiviral

DSV

dasabuvir

FC

fold‐change relative to reference replicon

GT

genotype

HCV

hepatitis C virus

IC₅₀

median inhibitory concentration

LDV

ledipasvir

NS5Ai

NS5A inhibitor

PEG‐IFN

pegylated interferon

PPI

proton pump inhibitor

RAS

resistance‐associated substitution

RBV

ribavirin

RC

replication capacity

SMV

simeprevir

SOF

sofosbuvir

SVR

sustained virological response

SVR12

sustained virological response posttreatment week 12

TND

target not detected

VL

viral load

Case

A 67‐year‐old white man with compensated cirrhosis, hepatitis C virus (HCV) genotype (GT) 1a, IL28B CT, presented for HCV treatment. His Model for End‐Stage Liver Disease score was 6, Child‐Turcotte‐Pugh score 5, transient elastography 26 kPa, and upper endoscopy was negative for varices. Previous HCV treatment included partial/null response to pegylated interferon (PEG‐IFN) and ribavirin (RBV) in 2008. Medical history was significant for moderate aortic stenosis with mild dyspnea on exertion. A stress test was negative for ischemia. The patient had gastroesophageal reflux with symptoms relatively well controlled on the proton pump inhibitor (PPI) omeprazole 20 mg daily.

The patient initiated direct‐acting antiviral (DAA) treatment with sofosbuvir/ledipasvir (SOF/LDV) for an intended treatment duration of 24 weeks. RBV‐based treatment was avoided because of moderate aortic stenosis. The patient was advised to take omeprazole with SOF/LDV in the morning on an empty stomach, 2 hours prior to eating breakfast. The baseline viral load (VL) was 5 million IU/mL and dropped to undetected (<12 IU/mL) by week 4. At week 8, the patient’s VL was 40 IU/mL and when rechecked at week 12, had increased to 70,000 IU/mL. Plasma was sent to Monogram Biosciences (South San Francisco, CA) for confirmation of GT and NS3/4A, NS5A, and NS5B resistance testing.

The patient denied any missed DAA dose, and the only antacid used was the daily PPI taken as recommended. Omeprazole was discontinued at week 12 and substituted with an H2 blocker once daily, 12 hours after administration of SOF/LDV. His cardiologist was consulted for consideration of adding RBV to the regimen. After review of his cardiac catheterization, the cardiologist indicated his aortic stenosis was stable without evidence of coronary artery disease and advised it would be safe to proceed with RBV. The patient started RBV and tolerated weight‐based dosing for the remainder of treatment without significant anemia or other side effects.

At the time of virological breakthrough the following resistance‐associated substitutions (RASs) were identified: NS3/4A Q80K, NS5A Y93H, NS5B L159F, S282T. After adding RBV at week 12, the patient’s HCV VL declined to undetectable levels by week 16. Because the patient was tolerating treatment and was highly motivated to complete 24 weeks of treatment, SOF/LDV/RBV was continued. Given the presence of the Y93H variant, an additional 4 weeks of simeprevir (SMV)/SOF/RBV after the completion of 24 weeks of SOF/LDV/RBV was given for a total treatment duration of 28 weeks. HCV RNA was undetectable at 12 and 24 weeks posttreatment, consistent with a sustained virological response (SVR) (Table 1).

Table 1.

Summary of Treatment Course and RASs

Time Point	HCV VL (IU/mL)	HCV Drug Therapy	Antacids	RAS
Baseline	5,926,990	SOF/LDV	Omeprazole 20 mg once daily on empty stomach with SOF/LDV	NS3/4A: Q80K
				NS5A: Q30H/N/S,Y93H
				NS5B: none
Week 4	<12 TND	SOF/LDV
Week 8	40	SOF/LDV
Week 12	70,000	SOF/LDV/RBV	famotidine 20 mg once daily 12 hours after SOF/LDV
Week 13	30,700	SOF/LDV/RBV		NS3/4A: Q80K
				NS5A: Y93H
				NS5B: L159F, S282T
Week 16	<12 TND	SOF/LDV/RBV
Week 24	<12 TND	SOF/LDV/RBV
Week 28	<12 TND	SOF/SMV/RBV
SVR12	<12 TND	—
SVR24	<12 TND	—

HCV VLs from treatment baseline to posttreatment week 24, revealing initial suppression, virological breakthrough, then SVR. Columns represent HCV DAA therapy completed and use of antacids and resistance‐associated substitutions at baseline and at time of virological breakthrough at week 12.

Because this patient was a participant in the HCV‐TARGET study, baseline resistance testing results were available.1 Baseline sequencing revealed NS3/4A Q80K, NS5A Q30H/N/S, Y93H, and no RASs in NS5B.

Given the rare virological breakthrough at week 12 of SOF/LDV in our patient, Monogram Biosciences performed phenotypic drug susceptibility analysis, a procedure not routinely used in clinical practice (Fig. 1 and Table 2). The phenotypic analysis showed that the patient‐derived NS5A region conferred a large reduction in LDV susceptibility with an approximate 40,000‐fold change in the half maximum inhibitory concentration (IC₅₀) and the 95% inhibitory concentration (IC₉₅) values relative to a reference replicon. Consistent with the genotypic data, the patient‐derived NS5B region conferred a reduction in SOF susceptibility with a 12‐ and 15.4‐fold change in IC₅₀ and IC₉₅ values, respectively. In contrast, RBV susceptibility was increased approximately 3‐ to 4‐fold.

Figure 1.

Phenotypic susceptibility analysis. Susceptibility to (A) LDV, (B) SOF, (C) the nonnucleoside polymerase inhibitor DSV, (D) RBV, and (E) an IFN control was evaluated using the GT1b Con1 reference replicon (reference, top panel) and a chimeric HCV replicon containing either a population of NS5A (for LDV) or NS5B regions derived from the patient’s virus (test sample, bottom panel). The vertical dotted line designates the IC₅₀.

Table 2.

Phenotypic Susceptibility Analysis

Inhibitor	Units	Reference Replicon		Test Sample Replicon				Assessment Relative to Reference
		IC50	IC95	IC50	IC95	IC50 FC	IC95 FC
LDV	nM	0.0026	0.0061	112.1	238.1	43116	39025	Reduced susceptibility
SOF	nM	128.4	579.1	1543	8943	12.0	15.4	Reduced susceptibility
DSV	nM	0.82	3.7	3.6	68.8	4.4	18.7	Reduced susceptibility
RBV	µM	19.6	176.6	6.5	41.2	0.33	0.23	Increased susceptibility
IFN	IU/mL	0.96	7.4	0.83	7.0	0.86	0.94	Equivalent susceptibility

Reference replicon: HCV GT1b Con1.

The test sample replicon containing patient‐derived NS5A regions had a large reduction in LDV susceptibility; 43,116‐ and 39,025‐fold change in IC₅₀ and IC₉₅ values, respectively, relative to the reference replicon. The test sample replicon containing patient‐derived NS5B regions had reduced SOF susceptibility with an IC₅₀ FC of 12 and IC₉₅ FC of 15.4 relative to the reference. An IC₅₀ FC of 4.4 and IC₉₅ FC of 18.7 to DSV was also observed. In contrast, RBV susceptibility was increased approximately 3‐ to 4‐fold relative to reference with a 0.33 and 0.23 FC in IC₅₀ and IC₉₅ values, respectively. No significant change in susceptibility to the IFN control was observed, as expected.

From the genotypic resistance analysis, it was also noted that an unusual de novo GT1a amino variant, S226A, emerged as a mixture with the typical wild type amino acid at position 226 in NS5B (S226S/A). In the context of SOF‐resistant (L159F + S282T) patient‐derived replicon clones, the S226A variant increased the fold change by 10 for SOF and dasabuvir (DSV), and had little impact on RBV susceptibility (Fig. 2). S226A also improved replication. In the context of a GT1a replicon, the phenotype of S226A alone was similar to the L159F mutant (J. Reeves, E. Anton, and K. Strommen, unpublished observations).

Figure 2.

Phenotypic susceptibility analysis of clones with and without the NS5B S226A variant. NS5B replicon clones were isolated that contained L159F and S282T with and without S226A for additional phenotypic analysis. The (A) SOF, (B) DSV, (C) RBV, and (D) IFN susceptibility of the GT1b Con1 reference replicon (reference replicon, black bars), a chimeric HCV replicon containing a population of NS5B regions derived from the patient’s virus (test sample population, gray bars), and replicon clones derived from the test sample harboring L159F and S282T RASs without (‐S226A, dark red bars) and with (+ S226A, green bars) the S226A variant were evaluated. Circles with (A, B) solid or (C, D) dashed lines demarcate differences in y axis scale. (E) Replication capacity (RC) expressed as a percentage of the reference replicon was also evaluated. In the context of replicon clones with L159F and S282T, S226A had minimal impact on RBV or IFN susceptibility, but further reduced SOF and DSV susceptibility, and conferred an increase in RC.

Discussion

To our knowledge, this is the first known case report of virological breakthrough with the emergence of dual NS5B L159F and S282T RASs in a patient naive to a SOF‐containing regimen. SOF is an NS5B nucleotide polymerase inhibitor that blocks viral replication through chain termination and has a high barrier to resistance. SOF RASs are detected in only 1% of SOF‐based treatment failures, and SOF RASs are transient given poor fitness and rapid reversion with wild type.2, 3 Post hoc sequencing analyses of 978 treatment failures in more than 12,000 patients enrolled in phase 2/3 clinical trials found that treatment‐emergent SOF resistance with the S282T substitution occurred in 10 patients (1%), 4 with GT1a virus who did not respond successfully to SOF/LDV. One participant developed L159F with S282T during retreatment with SOF/LDV.3 Likewise, NS5B L159F and S282T RASs were detected in only 1% and 1.7% of GT1a samples submitted to Monogram Biosciences for routine NS5B inhibitor resistance testing, respectively.4 These studies show that L159F, S282T, and L159F+S282T confer a 1.6‐, 17‐, and 24‐fold reduction in SOF susceptibility and reduce replication to 11%, 3%, and 3%, respectively, in a GT1a replicon.3 S282T has not been detected at treatment baseline, and baseline SOF NS5B resistance testing is not recommended prior to treatment or retreatment.2, 5, 6

In contrast with SOF, the NS5A inhibitor (NS5Ai) LDV has a low barrier to resistance, and baseline NS5A RASs are present in up to approximately 16% of GT1 viruses.7 The impact of these RASs on treatment outcomes varies according to the NS5Ai drug used, the type and number of RASs, treatment regimen, and patient characteristics.8 Our patient, despite being naive to NS5Ais, had significant baseline substitutions at two positions, NS5A Q30H/N/S and Y93H. In a GT1a replicon, the Y93H RAS confers an approximate 1000‐fold reduction in susceptibility to LDV, and in combination with the Q30H confers up to a 35,000‐fold reduction.6

In addition to the presence of double NS5A RASs at baseline, other patient factors may have contributed to absence of sufficient drug pressure, thereby allowing the emergence of NS5B substitutions leading to virological breakthrough at treatment week 12. The concomitant use of the PPI omeprazole, known to decrease LDV absorption, has been identified as a negative predictor of SVR with SOF/LDV in real‐world studies9 and may have contributed to a less favorable response. Moreover, substituting the PPI with an H2 blocker at week 12 may have relatively favored higher LDV serum levels. To a greater degree, patient factors such as the presence of cirrhosis, previous failure of PEG‐IFN/RBV, and GT1a infection are characteristics that increase the propensity to treatment failure along with baseline NS5A RASs. In retrospect, had we been aware of the presence of dual baseline NS5A RASs, along with the other known negative predictors of SVR, we would have more vigorously pursued the addition of additional drug pressure with RBV from treatment onset.

In clinical practice, RBV has been used with DAAs in patients with multiple negative predictors of SVR to prevent relapse. NS5B S282T has previously been associated with an approximate 3‐ to 10‐fold increased RBV susceptibility in vitro, which may translate into a clinical benefit.10, 11, 12 This would be supported in our patient’s case given the increased susceptibility in phenotypic analysis and the rapid decline in VL after addition of RBV to the failing SOF/LDV regimen. RBV is a nucleotide analogue that can incorporate in the RNA‐dependent polymerase, although the precise mechanism of action on viral replication remains elusive.13

It is important to note that SOF/LDV for 24‐week duration is no longer a preferred regimen for GT1a PEG‐IFN/RBV–experienced patients with compensated cirrhosis per American Association for the Study of Liver Diseases/Infectious Diseases Society of America and European Association for the Study of the Liver HCV treatment guidelines. This revised recommendation is supported by post hoc analyses revealing that baseline NS5A RASs adversely impact response to SOF/LDV without the use of RBV.7 The availability of next‐generation DAA regimens has led to additional revisions in recommendations.5

This case highlights how a detailed evaluation of patients who experience virological breakthrough or treatment failure contribute to our understanding of resistance to DAAs and will continue to inform HCV treatment guidelines. The unique role of RBV in the treatment of patients with novel DAA‐associated resistance is also notable.