| |
A phase 1, randomized, placebo-controlled, 3-day, dose-ranging study of GS-5885, an NS5A inhibitor, in patients with genotype 1 hepatitis C
|
| |
| |
Download the PDF here
Journal of Hepatology July 2012
Eric J. Lawitz1, Daniel Gruener2, John M. Hill3, Thomas Marbury4, Lisa Moorehead5,
Anita Mathias5, Guofeng Cheng5, John O. Link5, Kelly A. Wong5, Hongmei Mo5,
John G. McHutchison5, Diana M. Brainard5
1Alamo Medical Research, San Antonio, TX, USA; 2CRi Worldwide, Philadelphia, PA, USA; 3Avail Clinical Research, DeLand, FL, USA; 4Orlando Clinical Research Center, Orlando, FL, USA; 5Gilead Sciences, inc., Foster City, CA, USA
ABSTRACT
Background & Aims
GS-5885 is an inhibitor of the hepatitis C virus (HCV) NS5A protein and exhibits potent suppression of genotype 1 HCV replicons. The safety, tolerability, pharmacokinetics, antiviral activity, and resistance profile of once-daily GS-5885 doses of 1-90mg were evaluated in patients with chronic genotype 1 HCV.
Methods
Genotype 1 HCV-infected patients were randomized to 3days of once-daily (QD) dosing with placebo (n=12) or GS-5885 1mg (n=10), 3mg (n=10), 10mg (n=20), 30mg (n=10), or 90mg (n=10). Plasma samples for pharmacokinetics, HCV RNA, and NS5A sequencing were collected through day 14.
Results
GS-5885 was well tolerated and resulted in median maximal reductions in HCV RNA ranging from 2.3 log10 iU/ml (1mg QD) to 3.3 log10 iU/ml (10mg QD in genotype 1b and 30mg QD). Emax modeling indicated GS-5885 30mg was associated with>95% of maximal antiviral response to HCV genotype 1a. HCV RNA reductions were generally more sustained among patients with genotype 1b vs. 1a. Three of 60 patients had a reduced response and harbored NS5A-resistant virus at baseline. NS5A sequencing identified residues 30 and 31 in genotype 1a, and 93 in genotype 1b as the predominant sites of mutation following GS-5885 dosing. Plasma pharmacokinetics was consistent with QD dosing.
Conclusions
During 3days of monotherapy, low doses of GS-5885 demonstrated significant antiviral activity in genotype 1a and 1b HCV-infected patients. GS-5885 is currently being evaluated in combination with direct antiviral regimens with and without peginterferon.
introduction
Treatment of genotype 1 chronic hepatitis C virus (HCV) infection with 48weeks of peginterferon alfa-2a (PEG-iFN) and ribavirin (RBV) results in sustained virologic response (SVR) in 40-52% of patients [1], [2], [3], [4]. Addition of telaprevir or boceprevir, both HCV NS3 protease inhibitors, to the PEG-iFN and RBV regimen increases SVR rates to 67-75% [5], [6], [7], [8], [9], [10]. Therapy with PEG-iFN and RBV can cause side effects such as influenza-like fatigue, anemia, and depression, yet PEG-iFN and RBV remain the backbone of HCV therapy because monotherapy with telaprevir or boceprevir leads to the rapid emergence of viral resistance [11]. Development of antiviral agents targeting other HCV proteins to be used in combination may reduce the emergence of resistance and potentially allow for a successful HCV treatment regimen without PEG-iFN.
The NS5A protein plays a role in both viral RNA replication [12] and the assembly of HCV virions [13]. in HCV replicon cells, inhibition of NS5A results in the redistribution of NS5A from the endoplasmic reticulum to lipid droplets and appears to disrupt formation of new replication complexes [14]. Clinically, inhibition of NS5A has been associated with significant reductions in HCV RNA and enhanced SVR rates when combined with PEG-iFN and RBV [15], [16].
GS-5885 is a novel NS5A inhibitor with EC50 (50% effective inhibitory concentration) values of 34pM against genotype 1a and 4pM against genotype 1b replicons [17]. GS-5885 is stable in human liver microsomes and hepatocytes and has no inhibitory effect on the activities of major human CYP enzymes and low liability for induction via activation of xenobiotic receptors such as the aryl hydrocarbon receptor or pregnane X receptor (Gilead Sciences, data on file).
We examined the safety, tolerability, pharmacokinetics, and antiviral activity of once-daily dosing of GS-5885 for 3days in patients with chronic genotype 1 HCV. The GS-5885 doses evaluated ranged from 1 to 90mg daily. Treatment-emergent changes in the NS5A genetic sequence were also assessed.
Patients and methods
Patients
Eligible patients were 18-65years of age, with chronic infection with genotype 1a or 1b HCV virus and plasma HCV RNA 5 log10 iU/ml at screening. Patients were HCV treatment naïve and had a body mass index (BMi) of 19-35kg/m2 inclusive, creatinine clearance 70ml/min, and a QTcF interval ≤450ms. Patients with any of the following conditions or characteristics were excluded from participation: known cirrhosis, hepatic decompensation, excessive ongoing alcohol intake, Gilbert's syndrome, evidence of hepatocellular carcinoma, co-infection with HiV or hepatitis B virus, prothrombin time>1.5xULN, albumin<3g/dl, alanine aminotransferase and aspartate aminotransferase levels>5xULN, total bilirubin>ULN, hemoglobin<11g/dl, platelets<90,000/mm3, or absolute neutrophil count<1000 cells/mm3 (<900cells/mm3 for African Americans). Concomitant prescription or non-prescription medications were prohibited during the study unless prior approval was received from the medical monitor. The only exception was the use of hormonal contraception; additional double barrier method contraception was mandated for all women of childbearing potential. All patients provided written informed consent before undertaking any study-related procedures.
Study design
This was a phase 1, multi-center, randomized, double-blind, placebo-controlled, dose-escalation study that included 6 cohorts: 5 cohorts included patients with genotype 1a HCV only, and 1 cohort included patients with genotype 1b HCV only. Patients arrived at a study center the day prior to dosing initiation and were sequestered for approximately 5days. in all cohorts, oral tablets of GS-5885 or matching placebo were administered in a fasted state once-daily for 3days (days 1-3). Doses of GS-5885 in individual cohorts were as follows: 1, 3, 10mg (2 cohorts: genotype 1a and 1b), 30, and 90mg. Each cohort had 12 patients, 10 randomly assigned to active drug and 2 to placebo. Study treatment (GS-5885 or placebo) was assigned to patients according to a centralized randomization schedule generated via computer by the sponsor's Biometrics group. The study protocol was approved by each institution's review board prior to study initiation and was performed in accordance with Good Clinical Practice guidelines outlined by the international Conference on Harmonization.
Safety assessments
From baseline through day 14 follow-up visit, safety was evaluated on the basis of adverse events, physical examinations, clinical laboratory tests, vital signs, and ECG recordings. Concomitant medication intake was also recorded. Treatment-emergent adverse events were summarized by treatment, system organ class, and preferred term using the most current version of the Medical Dictionary for Regulatory Activities (MedDRA®).
Pharmacokinetic and pharmacodynamic assessments
Plasma samples for analysis of GS-5885 concentrations were drawn through 24h after the first dose on day 1 and after dosing on day 3. Additional blood samples were collected on days 4, 5, 6, 7, 8, and 10. Concentrations of GS-5885 were determined in plasma using a validated LC/MS/MS assay with a lower limit of quantification of 1ng/ml. Pharmacokinetic parameters were estimated via non-compartmental methods using Phoenix WinNonlin™ 6.0 (Pharsight Corporation, Sunnyvale, California, USA).
Parameters estimated included area under the plasma concentration-time curve (AUC) of the dosing interval (AUCtau, after administration of the last dose), AUC extrapolated to infinity (AUC∞, after administration of first dose), elimination rate constant (λz), and half-life (T1/2). in addition, maximum observed plasma concentration (Cmax), last quantifiable concentration (Clast), and concentration at the end of dosing interval (Ctau) were also identified.
The pharmacokinetic/pharmacodynamic relationship between GS-5885 doses and HCV RNA levels was explored using a simple Emax exposure-response curve (Phoenix WinNonlin™ 6.0). Appropriateness of the model was assessed visually using diagnostic and predictive check plots. Goodness of fit was assessed by using Akaike's information criterion and residual plots.
Efficacy assessments
HCV RNA
Plasma samples for determining HCV RNA levels were drawn at screening; prior to the first dose on day 1 (baseline); at the following hours after the initial dose: 1, 2, 4, 8, 12, 24, 36, 48, 72, 84; and on the following days: 4, 5, 6, 7, 9, 11, and 14. HCV RNA levels were measured with the Roche COBAS® TaqMan® HCV Test v2.0 for use with the High Pure System, with a lower limit of quantification of 25iU/ml.
Viral sequencing
Population sequencing of the HCV NS5A gene was performed for all patients at baseline (day 1 prior to dosing), on day 4, and on day 14 by Virco BVBA (Beerse, Belgium). in addition, deep sequencing was performed at baseline for a few selected patients. For deep sequencing, seven independent PCR reactions were performed to amplify the NS5A gene from one sample. The pool of PCR products was fragmented into smaller fragments (150-550 base pairs in length) and sequenced by 454 pyrosequencing technology developed by Virco BVBA.
End points and statistical analyses
The primary efficacy end point was antiviral activity of GS-5885. Reduction in HCV RNA was summarized as continuous change from baseline in log10 HCV RNA. All statistical summaries and analyses were performed using SAS® software (SAS institute, Cary, North Carolina, USA).
Results
Study population
Between August 2010 and January 2011, a total of 72 patients were randomized at nine study centers in the United States. All patients except one, who was lost to follow-up after day 8, completed the study through the day 14 follow-up visit. Overall, 72% (52/72) of patients were male, and the mean (SD) age was 48 (8.3) years-old (Table 1). Mean pre-treatment HCV RNA levels were similar between dosing groups, with a range of 6.3 (30mg GS-5885) to 6.8 log10 iU/ml (90mg GS-5885 and placebo). Two patients (1 assigned to 10mg GS-5885 and 1 assigned to placebo) were excluded from analyses because of study drug administration error.
Safety assessments
GS-5885 was well tolerated at all doses evaluated. No serious adverse events were reported, and no patients interrupted or discontinued dosing because of adverse events. From the 70 patients included in the safety analysis, 22 reported a total of 41 clinical adverse events (Table 2). Headache was the most common adverse event considered related to study drug (n=4/70, 6%). There was no dose-dependent trend apparent for any adverse events. All adverse events were mild or moderate in severity (2 moderate adverse events in the 1-mg and 30-mg groups and four moderate adverse events in the placebo group). in the 1-mg cohort, the two moderate adverse events were both rash; in the 30-mg cohort, the two moderate adverse events were headache and frequent urination. There were no clinically significant abnormalities as judged by physical examination, ECG, or clinical laboratory assessments.
Pharmacokinetic assessments
GS-5885 exhibited time-independent, near-linear pharmacokinetics (Fig. 1 and Table 3). Maximal concentrations were achieved 4-6h (median Tmax) after dosing across all dose groups. At 1-mg dose, all patients had quantifiable concentrations up to 24h after dosing on day 3. The median apparent plasma half-life across cohorts from 3 to 90mg ranged between 22 and 50h, supporting once-daily dosing. Pharmacokinetic parameters of 10mg GS-5885 were similar between HCV genotype 1a or 1b infection. Except at 1-mg dose, plasma concentrations 24h following dosing well exceeded the protein-adjusted mean EC90 for genotype 1 [18]. Specifically, mean plasma GS-5885 Ctau on day 3 for the 3-, 10-, 30-, and 90-mg doses were approximately 2.6-, 10.6-, 51.1- and 127.4-fold, respectively, higher than the protein-binding-adjusted mean EC90 value against HCV genotype 1a (0.91ng/ml).
Efficacy assessments
Antiviral response
Administration of GS-5885 resulted in rapid reductions in HCV RNA concentrations at all doses tested. By 8h post 1st dose, median reductions in HCV RNA were 2 log10 iU/ml in cohorts dosed at 3mg and above. in all cohorts receiving 3mg GS-5885 daily, median HCV RNA reductions were 2.5 log10 iU/ml at 12h and 3 log10 iU/ml by 36h. Overall, median maximal reductions from baseline were>3 log10 iU/ml for all dosing cohorts except the 1-mg group, which had a median (range) maximal reduction of 2.3 (0.95, 3.57) log10 iU/ml. Median (range) maximal HCV RNA reductions for the other groups were as follows: 3.1 (2.72, 3.66) for 3mg, 3.2 (1.60, 3.67) for genotype 1a patients receiving 10mg, 3.3 (2.28, 4.13) for genotype 1b patients receiving 10mg, 3.3 (0.88, 4.17) for 30mg, and 3.1 (1.48, 4.03) for 90mg. For placebo, the median (range) maximal HCV RNA reduction was 0.3 (0.03, 0.66) log10 iU/ml. Median maximal HCV RNA reductions were similar for genotype 1a and 1b patients receiving 10mg GS-5885; however, HCV RNA suppression was more sustained in genotype 1b patients (Fig. 2). No patient in any cohort experienced viral rebound, defined as HCV RNA increase by>1 log10 from nadir during the dosing period. iL28B status (CC vs. non-CC) did not appear to influence HCV RNA decline (data not shown).
Exposure-response modeling
The simple Emax model adequately described the relationship between GS-5885 exposure (assessed as AUC over the dosing interval) and maximum reductions in HCV RNA concentrations. Emax values predicted by the model were close to the observed values. As seen in Fig. 3, Emax modeling indicated that exposures achieved following administration of 30mg GS-5885 provide>95% of maximal antiviral response in genotype 1a infected subjects. inter-subject variability in exposure/response varied across cohorts, with the highest variability observed at the 90-mg dose. However, fairly robust responses were observed through the range of exposures achieved.
NS5A sequence analysis
Population sequencing performed for all patients at baseline revealed 5 patients with detectable pre-existing resistance-associated NS5A mutations. Four of these 5 were identified in GT1a patients, and 2 of the 4 had maximal HCV RNA reductions ≤1.6 log10 iU/ml (Fig. 4). NS5A mutation Q30E/Q was identified in one of these patients (Fig. 4A) and L31M in the other (Fig. 4B). A third patient, dosed at 10mg GS-5885, had a maximal HCV RNA reduction of -1.6 log10 iU/ml with no detectable resistance mutations by population sequencing, which has a limit of detection of ~25%. in this third patient (Fig. 4C), 454 pyrosequencing of the baseline sample revealed that 12% of the viral species harbored an NS5A Y93C mutation. For all 3 of these patients, the mutations observed at baseline were predominant at days 4 and 14. Plasma exposures of GS-5885 in these 3 patients were similar to others within their groups.
Treatment-emergent mutations in NS5A were assessed by population sequencing at day 4 (approximately 24h after the last dose of GS-5885) and day 14. in patients with genotype 1a HCV, the spectrum of detected resistance mutations was complex (Table 4). Of the 10 HCV genotype 1a patients dosed with 1-mg GS-5885, 3 had resistance mutations detected at NS5A positions 30 or 31 by population sequencing at either day 4 or day 14. No changes from baseline at known resistance amino acid positions (NS5A amino acids 28, 30, 31, 93) were detected for the remaining 7 patients at day 4 and day 14. irrespective of racial background or ethnicity, all patients dosed at 3-mg GS-5885 had resistance mutations detected by population sequencing. The fold shift in EC50 of GS-5885 vs. these mutants suggests M28T had the lowest resistance and Y93H had the highest (Table 4). M28T and Q30H were only detected in the 1-, 3-, and 10-mg cohorts with HCV genotype 1a. At the 30- and 90-mg doses, these two mutations were not detected, which is consistent with suppression of lower level resistant variants at higher doses. Conversely, the highest level resistant variants among patients with genotype 1a virus, Y93C and Y93H, were selected more frequently in patients who received 90-mg GS-5885. The two resistance mutations with the broadest representation across dosing cohorts with genotype 1a HCV were Q30R and L31M, with Q30R having the highest frequency. All 10 patients with genotype 1b HCV, who received 10-mg GS-5885, harbored Y93H following treatment, consistent with Y93H being the primary GS-5885 resistance mutation selected in the genotype 1b replicon (data not shown).
Discussion
This is the first report on the safety and antiviral activity of GS-5885 in HCV-infected patients. in this study of patients with genotype 1 HCV, 3days of once-daily dosing with GS-5885 were well tolerated and led to substantial declines in HCV RNA during the dosing interval. The median maximal reduction in HCV RNA ranged from 2.3 log10 iU/ml (1mg QD) to 3.3 log10 iU/ml (10mg QD in genotype 1b and 30mg QD). Median maximal reductions were>3 log10 iU/ml for all groups receiving 3mg or higher GS-5885 daily, and the magnitude of viral suppression was similar between them. GS-5885 produced a rapid reduction in HCV RNA over the initial 12h of therapy, and the response was greater than has been reported for telaprevir or boceprevir monotherapies [19], [20]. The kinetics of HCV RNA reductions over the first 24h after initiating GS-5885 was similar to that seen with a single dose of the NS5A inhibitor BMS-790052 [15]. Theoretically, a more rapid and profound reduction of HCV RNA may result in better suppression of wild type virus and reduce the likelihood of de novo resistant mutants during treatment. Whether GS-5885 provides an advantage in clinical response over other classes of HCV inhibitors warrants further study.
in vitro replicon data demonstrate that the antiviral activity of GS-5885 is less against genotype 1a than 1b (mean EC50 vs. 1a=34pM; mean EC50 vs. 1b=4pM). Although the mean maximal HCV RNA reductions with 10-mg GS-5885 in genotype 1a and 1b patients were not vastly different (3.2 vs. 3.3 log10 iU/ml), viral suppression lasted longer after treatment cessation in the 1b patients. Furthermore, the pattern of resistance mutations was significantly different between 1a and 1b patients, suggesting that antiviral activity differs in these two patient populations [21].
Exposure-response analyses indicate that GS-5885 doses 30mg provide>95% of maximal antiviral responses in patients with genotype 1a HCV. The plasma Ctrough concentrations of GS-5885 at 30mg and 90mg were 230-fold and 572-fold higher than the protein-adjusted genotype 1a EC50, respectively. The results suggest that GS-5885 dosing beyond 90mg is unlikely to cause further meaningful reductions in HCV RNA.
Three patients in the current study had substantially reduced HCV RNA reductions as compared with median values in their respective cohorts. These patients were subsequently found to have variants in their baseline NS5A sequence that are known to confer resistance to GS-5885. Whether these findings are clinically relevant depends in part on the frequency of these mutations in HCV-infected patients. Studies assessing the prevalence of known HCV drug-resistance mutations in untreated HCV-infected patients have focused on NS3 protease inhibitors or NS5B non-nucleoside inhibitors, and their results indicate dominant resistant mutations are common [22], [23]. Few data are available regarding the prevalence of NS5A resistance mutations in genotype 1a and 1b infected patients. in the European HCV database of over 500 genotype 1a sequences, the reported prevalence of mutations Y93C and L31M is 0.2% and Q30E is 0.0% [24], [25]. Mutants observed in this study retained full susceptibility to other classes of antivirals, including direct acting anti-HCV agents and interferon (data not shown).
The resistance profile for GS-5885 mapped primarily to residues 28, 30, 31, and 93 of HCV NS5A. in genotype 1a patients, resistance mutations were spread across these residues, whereas in genotype 1b patients, Y93H was the sole resistance variant noted by population sequencing. For genotype 1a patients, NS5A, M28T, and Q30H resistance mutations were not detected in patients receiving GS-5885 30mg. This is consistent with a lower fold resistance observed with these two mutations against the genotype 1a replicon when compared with other variants (Table 4). Plasma Ctrough concentrations of GS-5885 at 30 or 90mg were higher than protein-adjusted EC50 values for these two mutants (Table 3, Table 4). While the replication capacity of the genotype 1a M28T mutant is 31%, the replication level of the Q30H mutant is substantially higher at 75% of the wild-type 1a replicon [15]. Suppression of these two mutants with GS-5885 at 30mg or higher potentially increases the resistance barrier of GS-5885-containing combination regimens.
The picomolar potency, pharmacokinetic properties supporting once daily dosing, and current safety profile of NS5A inhibitors make them an attractive target for drug development [26]. in addition, GS-5885 demonstrated additive to synergistic antiviral activity in vitro when combined with iFN, RBV, and multiple other classes of direct-acting antivirals including GS-9451 (protease inhibitor), GS-9669 (non-nucleoside polymerase inhibitor), GS-9190 (non-nucleoside polymerase inhibitor), and GS-6620 (nucleoside analog) (Gilead Sciences, data on file). Therefore, GS-5885 in combination with PEG/RBV or direct-acting antivirals is currently being evaluated in phase 2 studies. The antiviral activity, including subtype differences, observed with GS-5885 is similar to BMS-790052, the only other NS5A inhibitor also in phase 2 development [15]. The pattern of resistance mutations selected by GS-5885 is also similar to that of BMS-790052 [15], where the Y93H mutation dominated with genotype 1b virus, and mutations at residues 28, 30, 31, and 93 were predominant with genotype 1a virus. Furthermore, the levels of resistance against the Q30R, Y93C, and Y93H single mutations in genotype 1a were similar between GS-5885 and BMS-790052. However, the in vitro activity as measured by EC50 against other resistant mutations varies for the two compounds. in genotype 1b virus, BMS-790052 is significantly more active against the Y93H mutation (0.05 nM for BMS-790052 vs. 5 nM for GS-5885) [24]. in contrast, in genotype 1a, GS-5885 is 4-5 times more active against the M28T and Q30H mutations than BMS-790052, while the latter is approximately 2 times more potent against the L31M mutation.
in conclusion, data from the current study demonstrate that GS-5885 is a potent, once-daily inhibitor of HCV replication in vivo and is well tolerated in treatment-naïve genotype 1 HCV infected patients. These results confirm NS5A as an important target for direct acting anti-HCV agents and support further clinical development of GS-5885. GS-5885 is currently being evaluated in multiple combination therapies with and without PEG-iFN alfa-2a in phase 2 studies in treatment-naive and treatment-experienced HCV-infected patients.
|
|
| |
| |
|
|
|
