(VX-497) and Interferon-Α in Previously

Antiviral Therapy 10:635–643 A randomized, double-blind, placebo-controlled dose-escalation trial of merimepodib (VX-497) and interferon-α in previously untreated patients with chronic hepatitis C John G McHutchison1,2*, Mitchell L Shiffman3, Ramsey C Cheung4, Stuart C Gordon5, Teresa L Wright 6, John C Pottage Jr 7, Lindsay McNair7, Ene Ette7, Scott Moseley 7 and John Alam7

1Duke Clinical Research Institute and Division of Gastroenterology, Duke University Medical Center, Durham, NC, USA 2Division of Gastroenterology, Scripps Clinic, La Jolla, CA, USA 3Hepatology Section, Virginia Commonwealth University Health System, Richmond, VA, USA 4VAPAHCS, Palo Alto, CA, USA 5William Beaumont Hospital, Royal Oak, MI, USA 6VA Medical Center, San Francisco, CA, USA 7Vertex Pharmaceuticals, Cambridge, MA, USA

*Corresponding author: Tel: +1 919 668 7177; Fax: +1 919 668 7164; E-mail: [email protected]

Inhibition of inosine monophosphate dehydrogenase elevated alanine aminotransferase levels. No pharmaco- (IMPDH) is one of several proposed mechanisms of action kinetic interactions were evident between the two drugs. for ribavirin (RBV), a critical component of the current Analysis of covariance that adjusted for a baseline imbal- treatment for chronic hepatitis C (CHC). ance in HCV-RNA in the intent-to-treat population did This study was a double-blind, placebo-controlled dose- not show any significant differences between the treat- escalation study of a novel, selective, orally active small ment groups, or between MMPD plus IFN-α compared molecule inhibitor of IMPDH, merimepodib (VX-497 or with IFN-α alone. However, the per-protocol primary MMPD) in combination with standard interferon-alpha efficacy analysis based on treatment-compliant patients (IFN-α). Fifty-four treatment-naive patients with genotype- demonstrated a greater reduction in mean HCV-RNA in 1 CHC were randomized to receive IFN-α 3 MIU the combination of 100 mg MMPD plus IFN-α compared subcutaneously three times a week, alone or in combination with IFN-α alone (–1.78 log vs –0.86 log, P=0.037). with 100 mg or 300 mg (every 8 h) of MMPD for 4 weeks. In conclusion, the addition of a selective IMPDH inhibitor At the end of 4 weeks, all patients were offered 48 weeks of to IFN-α was well tolerated. In a low-dose range, the treatment with IFN-α/RBV. The objectives of the study were addition of MMPD may have the potential to add to the to evaluate the tolerability of the IFN-α/MMPD combina- antiviral efficacy of IFN-α. Larger, longer duration trials tion and to evaluate whether MMPD had an on-treatment incorporating pegylated IFN would be required to deter- effect on HCV-RNA, similar to RBV when added to IFN-α. mine whether this combination, alone or with RBV, would The drug combination was generally well tolerated; one increase either early or sustained virological response patient at the higher dose discontinued because of rates.

Introduction

The addition of ribavirin (RBV), a broad-spectrum When combined with either standard recombinant or antiviral agent, to interferon (IFN) therapy, has proven pegylated IFN-α, RBV seems to have two effects: to be a significant advance in the treatment of patients firstly, to increase the percentage of patients who with chronic hepatitis C (CHC) infection. When used achieve HCV-RNA undetectable status as compared alone, RBV decreases alanine aminotransferase (ALT) with IFN-α alone and secondly, of the patients who levels in most patients during treatment. However, liver achieve an undetectable level of HCV-RNA, to enzymes return to baseline values when treatment is decrease the rate of breakthrough/relapse during and discontinued and, when given alone, RBV has little if after treatment. The net effect of the addition of RBV to any effect on serum HCV-RNA concentrations [1]. IFN-α is that it substantially increases the sustained

virological response (SVR) rate [2–5]. Although the herpes simplex virus type-1 are 12.4 µM, 1.1 µM and side-effect profiles of these drugs preclude their use in 6.3 µM, respectively [6]. These values compare with some patients, the combination of pegylated IFN-α and 44.6 µM, 20.9 µM and 162 µM, respectively, for RBV. RBV is currently the most effective therapy for patients The combination of MMPD and IFN-α shows additive with CHC. activity against encephalomyocarditis virus, a single- The major limitation of RBV therapy is the develop- stranded RNA virus. The in vitro effects have been ment of drug-induced haemolytic anaemia, with the further explored in a hepatitis C viral subgenomic majority of patients demonstrating a 2–3 g/dl decrease replicon system based on Bartenschlager’s method [10]. in haemoglobin during treatment. Decreases in haemo- In this assay, the IC50 values for MMPD, RBV and globin concentrations to less than 10 g/dl, necessitating IFN-α are 0.5 µM, 41.8 µM and 3.1 IU/ml, respectively. a reduction in the dose of RBV, have also been In the same assay, MMPD has demonstrated additive observed in approximately 8% of patients receiving activity when combined with either RBV or IFN-α [11]. combination therapy [3]. Given both the benefits and The clinical activity of MMPD monotherapy has limitations of RBV therapy, a drug that captures the been explored over the dose range 100–400 mg when clinical antiviral effects of RBV without the toxicity administered for 28 days to patients with hepatitis C would be an advance in the treatment of hepatitis C. who had previously failed treatment with IFN-α [12]. RBV was originally described as an inhibitor of the MMPD was well tolerated with only generally mild human enzyme inosine monophosphate dehydrogenase digestive system adverse events. A statistically signifi- (IMPDH), which catalyses the rate-limiting step in the cant reduction in end-of-treatment serum ALT de novo biosynthesis of guanine nucleotides. Inhibition compared with baseline was observed in patients of IMPDH reduces the intracellular levels of guanine receiving 200 mg and 400 mg MMPD. No demon- nucleotides, which are required for RNA and DNA strable change in HCV-RNA was noted. These results synthesis. Depletion of guanine nucleotides subse- are similar to those seen with RBV monotherapy. quently leads to antiproliferative and antiviral proper- The primary objective of the study was to evaluate ties. In various in vitro antiviral assays, the activity of the safety and tolerability of the IFN-α/MMPD combi- RBV is partially reversed when exogenous guanosine is nation; the secondary objective was to examine added, indicating at least some of the antiviral effect of whether MMPD had an on-treatment effect similar to RBV may be attributed to IMPDH inhibition [6]. that of RBV, that is, whether the combination of IFN-α However, the lack of full reversal indicates other mech- and MMPD enhanced the on-treatment antiviral anisms are involved and in vivo mechanisms may be activity when compared with IFN-α alone. Standard distinct from what is seen in vitro. Other proposed therapy was deferred for 4 weeks; therefore, the poten- mechanisms include inhibition of viral polymerase, tial for on-treatment effects was evaluated as change in activity as a viral mutagen and activity as a modulator HCV-RNA over 4 weeks of treatment. of TH1/TH2 (antiviral/inflammatory cytokines) responses [6–9]. The relative contribution of IMPDH Patients and methods inhibition versus other mechanisms to the clinical activity of RBV is unknown. Male and female patients aged 18–70 years with geno- Merimepodib (VX-497 or MMPD) is a novel, selec- type-1 CHC infection, not previously treated with tive and reversible inhibitor of IMPDH. MMPD antiviral therapy and candidates for IFN-α/RBV uncompetitively inhibits both isoforms of IMPDH with therapy, were enrolled in this study. All patients had

Ki values that are approximately 25-fold lower than detectable serum HCV-RNA as determined by reverse those observed with RBV (7–10 nM versus 250 nM). transcription PCR, compensated liver disease (as The rationale for the development of MMPD in evidenced clinically, and by normal albumin, hepatitis C includes the possibility that a selective prothrombin time and bilirubin levels), and had no IMPDH inhibitor could lead to the clinical antiviral histological evidence of cirrhosis on liver biopsy activity of RBV without its associated toxicity, and that obtained within 3 years of study entry. Exclusion a more potent IMPDH inhibitor might have a greater criteria included a white blood cell (WBC) count <1500 antiviral effect. cells/mm3, a platelet count <100 000/mm3 and haemo- The antiviral activity of MMPD has been evaluated globin <9.5 g/dl. In addition, patients with a history of in vitro against a host of viruses including a flavivirus haemoglobinopathies or coinfection with either related to hepatitis C virus [6]. Data from these studies hepatitis B or HIV were excluded, as were those with indicate that MMPD possesses more potent broad- other causes of liver disease. In order to determine spectrum antiviral activity than RBV under the experi- whether or not the inhibition of IMPDH could add to α mental conditions. The IC50 for bovine viral diarrhoea the antiviral activity of IFN- , a high degree of study virus (a flavivirus), respiratory syncytial virus and medication compliance was required.

The study protocol was reviewed and approved by Assay methods institutional review boards at each of the respective Serum HCV-RNA concentrations were performed study sites, and informed consent in writing was using the quantitative NGI Superquant® RT-PCR assay obtained from each patient prior to screening and enroll- with a lower limit of detection of less than 100 ment into the study. The study was conducted in the US copies/ml (National Genetics Institute, Los Angeles, prior to market approval of pegylated interferons. CA, USA) [13]. HCV genotyping was performed with Innolipa HCV (Innogenetics, Zwijnaarde, Belgium) as Study design described previously [14]. Serum IFN-α levels were The study was divided into two parts: the core study determined by using a commercially available ELISA and the rollover phase. During the core study, patients test kit validated for detection of recombinant IFN-α were assigned to receive either 100 mg MMPD (Vertex (IntronA). Plasma MMPD levels were assessed by Pharmaceuticals, Cambridge, MA, USA) orally every Inveresk Research (Tranent, Scotland, UK) using a vali- 8 h, 300 mg MMPD orally every 8 h or an identical dated, specific, sensitive and reproducible HPLC assay placebo, each in combination with recombinant IFN-α with UV detection. The limit of quantification of the (Intron®A, Schering-Plough, Kenilworth, NJ, USA) assay was 50 ng/ml. 3 MIU three times a week subcutaneously for 4 weeks. Following completion of treatment in the core study, all Sample size and statistical methods patients were offered entry and treatment in a rollover The study was primarily designed to assess the safety phase in which they could receive open-label Rebetron® and tolerability of the MMPD/IFN-α combination (IFN-α2b + RBV; Schering-Plough) for up to 48 weeks. therapy and it was determined that 16 patients per The core study was designed as a randomized, treatment (24 patients per treatment group) would double-blind, placebo-controlled dose-escalation study. provide a sufficient sample size to achieve this objec- Patients were enrolled into one of two dosing groups of tive. In addition, based on the assessment of available 27 patients each: group I patients received IFN-α and data [15–17] and the use of multiple baseline and end- either placebo or 100 mg MMPD; group II patients of-treatment HCV-RNA samples, a sample size of 16 received IFN-α and either placebo or 300 mg MMPD. patients per treatment was determined to provide Patients in group I began treatment in parallel, while approximately 80% power to detect a 10% difference patients in group II began treatment once 14 patients in in the relative change from baseline to week 4 in serum group I had completed 4 weeks of treatment. HCV-RNA with an alpha of 0.05 using a 1-sided test In order to minimize differences in outcome (towards higher response in the combination arms than response variables (that is, baseline HCV-RNA) the placebo arm). It was specifically assumed that the between the two dosing groups, the initial randomiza- IFN-α alone arm would show a 70% (0.5 log) reduction scheme was such that no patient was to be tion in HCV-RNA levels at 4 weeks, whereas the high randomized and start treatment until a sufficient dose of MMPD would show an 80% (0.7 log) reduc- number of patients had passed screening and a ‘pool’ of tion. The assumed standard deviation was 11% in each 54 eligible study participants had been established. arm. Because this was an early Phase II study and a Once the pool was established, patients were to be relevant question was whether MMPD treatment randomized to their respective dosing groups and added to the antiviral effects of IFN-α, a 1-sided test blinded treatment assignments. Patients in group I were was pre-specified in the protocol. In addition, the per- to start treatment immediately, while group II patients protocol analysis only included those patients taking at were to have their treatment deferred until the requisite least 95% of the prescribed MMPD/placebo doses and number of group I patients had completed treatment. 100% of IFN-α doses. To increase the likelihood that However, because of an initially slow rate of patient 48 patients completed the study with the required high screening, it became apparent that the group II patients level of compliance with the study schedule, 54 patients would have their treatment deferred for several months. (27 patients per dosing group or 18 per treatment arm) Therefore, randomization was initiated after the first were enrolled. The study protocol pre-specified that 17 patients were screened, and patients were subse- data from patients receiving placebo in each dose group quently randomized and treated in their respective dose would be ‘pooled’ for analysis purposes. groups as screening was completed. The randomization The antiviral activity assessment was based on the process was stratified according to the investigational percentage change in serum HCV-RNA from baseline site and median baseline viral load (≤2×106 vs >2×106 to week 4. Since analysis of percentage change and copies HCV-RNA per ml). Patients were assigned to change in log HCV-RNA are equivalent, and provide group I or II in a 1:1 ratio. Within each dose group, information that can be compared with various patients were randomized to treatment in a 2:1 ratio published data, the results in this report are expressed as (that is, IFN-α plus MMPD vs IFN-α plus placebo). change in log HCV-RNA. Five repeated measurements

Antiviral Therapy 10:5 637 JG McHutchinson et al.

(screen, baseline 1, 2, 3 and predose on study day 1) treatment, resulting in 53 patients (ITT analysis) who were taken at least 1 day apart and within 12 weeks entered the treatment phase of the study. The majority prior to day 1 (for screen and baseline) for quantifica- of patients were male and Caucasian. Within each tion of serum HCV-RNA. Four separate samples at group, 17–22% were African-American, higher least 1 day apart (scheduled on study days 25, 28, 29 percentages than reported in most published hepatitis and day 30, the day following the last dose of study C treatment studies. The three treatment arms were medication) were taken at week 4 for determination of well-balanced in terms of race, age and weight (Table serum HCV-RNA. Medians for each patient were 1). Serum HCV-RNA levels were statistically signifi- calculated from the baseline and week 4 measurements, cantly higher at baseline in the 100 mg treatment arm respectively. The two medians were used to calculate compared with either the IFN-α alone or the 300 mg the change in log HCV-RNA from baseline to week 4. dose group (Table 1). In addition, a higher proportion Because a statistically significant difference in baseline of male patients were enrolled in the 100 mg group HCV-RNA levels between the IFN-α alone and the 100 (83%) than in the 300 mg (41%) and placebo group mg group was evident, and since baseline HCV-RNA (56%). has been well described as an important prognostic Of the 53 patients treated, 50 patients completed the factor in treatment outcome, an analysis of covariance 4-week study treatment period. One patient in the (ANCOVA), with baseline HCV-RNA as a covariate, 300 mg treatment group discontinued at day 25 of the was used to evaluate the HCV-RNA results. ANCOVA study due to an increase in ALT levels. This patient is analysis including the 300 mg group was attempted, further discussed in the adverse events section below. but could not be performed because the assumption of One patient in the 100 mg dose group was withdrawn parallel response slopes was not met. at day 25 due to syncopal episodes, and one patient in During the course of the core study, treatment the IFN-α alone group withdrew consent on day 8. compliance was assessed by questioning each patient at Compliance with treatment was generally good, with every visit; pill and bottle counts that allowed determi- mean compliance with MMPD/placebo reported to be nation of adherence to the therapy were performed. approximately 98%, and all but one patient in the Compliance was recorded in the source documents and 100 mg dose group reporting taking 100% of their case record form. All patients randomized who IFN-α injections. Seven patients (one in IFN-α alone, received at least one dose of study medication are three each in the two MMPD groups) did not meet the included in an intent-to-treat (ITT) analysis, which is pre-specified criteria of taking at least 95% of their also presented. drug doses to be included in the per-protocol analysis.

Results Pharmacokinetics and pharmacodynamics MMPD Cmaxss and week 2 steady-state AUCss for Baseline characteristics and study conduct MMPD and IFN-α are shown in Table 2. Comparison Fifty-four patients were randomized. One patient in the of, for example, 1 and steady-state AUC indicated no 300 mg group was withdrawn from the study prior to accumulation of VX-497 over the dosing interval. The

Table 1. Mean baseline characteristics and demographics of randomized patients

IFN-α + placebo IFN-α + 100 mg MMPD IFN-α + 300 mg MMPD

Gender, M:F 10:8 15:3 7:10 Age, years 48.4 ±9.7 45.4 ±3.0 44.8 ±5.3 Weight, kg 77.7 ±13.8 86.8 ±12.6 83.7 ±23.1 Race Caucasian 10 14 11 African-American 4 3 3 Hispanic 4 1 3 log HCV-RNA, copies/ml 6.72 ±0.6 *7.1 ±0.32 6.55 ±0.68 ALT, IU/ml 88.1 ±38.7 131.2 ±103.8 78.7 ±45.8 Albumin, g/l 40.7 ±2.2 39.6 ±3.6 40.5 ±2.2 Total bilirubin, µmol/l 7.7 ±4.2 6.7 ±3.1 6.7 ±3.1 Prothrombin time, s 12.3 ±0.7 12.3 ±0.5 12.4 ±0.6

Except gender, race and log HCV-RNA, values represent mean (±SD) values for the group. For log HCV-RNA, the value is the median of screening, baseline 1, 2, 3 and day 1 (pre-dose) results. *Baseline difference in HCV-RNA levels; 100 mg versus placebo (P=0.04), 100 mg versus 300 mg (P=0.004). ALT, alanine aminotransferase; IFN, interferon; MMPD, merimepodib.

MMPD pharmacokinetic results are similar to the results Analysis of the correlation between change in log reported when it was given as monotherapy in patients HCV-RNA and baseline HCV-RNA within the group with hepatitis C [12]. IFN-α levels were similar across the of patients receiving IFN-α alone revealed a statistically treatment groups, indicating that MMPD does not affect significant inverse correlation [that is, patients with the pharmacokinetic behaviour of IFN-α. IFN-α-induced high HCV-RNA at baseline demonstrated lower serum neopterin levels were similar across the three changes in log HCV-RNA (Kendall Tau-β correlation groups, indicating that, at least by this measure, MMPD coefficient=0.47, P=0.039)]. Since baseline viral load is does not affect the pharmacodynamics of IFN-α. an important prognostic factor in treatment outcome, and differences in baseline viral load were observed Antiviral effects between the 100 mg MMPD group and the IFN-α Median changes in log HCV-RNA over the 4 weeks of alone group, the most appropriate analysis of differ- the core study are shown in Table 3. The median reduc- ence in change in HCV-RNA between these two groups tion in log HCV-RNA was greater in the 100 mg treat- was an ANCOVA with baseline HCV-RNA as a ment group as compared with the IFN-α alone patients covariate [18]. The outcome of this analysis for the at all timepoints in the per-protocol population, and at core study is shown in Figure 1. Since the 300 mg 2 and 4 weeks in the ITT population. In the 300 mg MMPD group did not show any difference in baseline treatment group, the median reduction in log HCV- HCV-RNA compared with placebo, the ANCOVA RNA tended to be smaller than that observed in both analysis was not performed as it would not have been of the other treatment groups. Without any adjust- statistically meaningful for this comparison. ments for baseline differences in HCV-RNA, statistical Since the secondary objective of the study was to comparison between the MMPD combination treated explore whether there were additive antiviral effects in groups and the IFN-α alone treated patients did not the combination of IFN-α and MMPD as compared reveal any significant differences. with IFN-α alone, the primary analysis population

Table 2. Median pharmacokinetic and pharmacodynamic parameters

IFN-α + placebo IFN-α + 100 mg MMPD IFN-α + 300 mg MMPD

MMPD

Cmaxss ,ng/ml — 1635 (220–3620) n=18 3435 (410–6250) n=17

AUCss , ng/ml•h — 3940 (2010–8320) n=18 8360 (3090–17 790) n=17 IFN-α

Cmaxss , pg/ml 64 (43–87) n=10 68 (29–91) n=7 67 (50–91) n=6

AUCss , pg/ml•h 252 (126–403) n=10 228 (79–373) n=7 335 (230–405) n=6

Serum neopterin AUCss , ng/ml•h 88 (39–119) n=8 92 (44–199) n=11 90 (44–171) n=15

α Cmaxss and AUCss for MMPD and neopterin are based on week 2 pharmacokinetic data. Cmaxss and AUCss for IFN- are based on week 4 pharmacokinetic data when data were available. In cases where week 4 IFN-α concentrations were not available, week 2 pharmacokinetic data were used. Values in parentheses represent minimum to maximum range. AUC, area under the curve; IFN, interferon. ss, steady-state; MMPD, merimepodib.

Table 3. Median change in log HCV-RNA from baseline

IFN-α + placebo IFN-α + 100 mg MMPD IFN-α + 300 mg MMPD

Per-protocol population End of week 1 –0.52 (–2.63, 0.48) n=16 –0.67 (–1.97, 0.64) n=14 –0.27 (–1.68, 0.07) n=13 End of week 2 –0.58 (–3.67, 0.07) n=16 –1.07 (–2.49, 0.49) n=14 –0.53 (–1.43, 0.34) n=13 End of week 4 –0.62 (–3.67, 0.28) n=16 –1.62 (–3.91, 0.29) n=14 –0.56 (–2.49, 0.17) n=14

ITT population End of week 1 –0.69 (–2.63, 0.48) n=17 –0.66 (–1.97, 0.64) n=18 –0.14 (–1.68, 0.19) n=16 End of week 2 –0.53 (–3.67, 0.07) n=17 –0.97 (–2.49, 0.49) n=18 –0.31 (–1.43, 0.34) n=16 End of week 4 –0.66 (–3.67, 0.28) n=17 –0.90 (–3.91, 0.29) n=17 –0.51 (–2.49, 0.17) n=17

Numbers in parentheses represent maximum and minimum change in HCV-RNA within each treatment group at the specified timepoint. The ‘baseline’ value is calculated as the median of the screening, baseline 1, baseline 2, baseline 3 and day 1 (pre-dose) results. The ‘week 4’ value is the median of day 25, 28, 29 and 30 results. IFN, interferon; ITT, intent-to-treat; MMPD, merimepodib.

Antiviral Therapy 10:5 639 JG McHutchinson et al.

Figure 1. Mean change in log HCV-RNA from baseline to HCV-RNA. In the 300 mg every 8 h MMPD group, week 4 only one patient had this early response (1/17, 5.9%). Similarly, in the per-protocol population, an early viro- n=17 n=14 n=17 n=16 logical response was observed in 6/17 patients (35.3%) 0 in the IFN-α alone group, 7/14 (50%) in the 100 mg –0.2 every 8 h MMPD group and only 1/14 (7.1%) in the 300 mg every 8 h MMPD group. –0.4 All patients who completed the initial 4-week study –0.6 period entered the rollover phase. The outcomes in this –0.8 group, who subsequently received ongoing treatment with IFN-α/RBV combination treatment, are shown in –1 Table 4. The majority of patients (21/34) discontinued –1.2 treatment prior to completion of the 12-month course –1.4 because of failure to achieve a virological response at or after 6 months of therapy. Twelve patients achieved –1.6 an SVR at 6 months post-treatment follow-up, with –1.8 equal distribution across the treatment arms. ITT Evaluable for efficacy –2 Adverse events IFN IFN + 100 mg MMPD In general the most common adverse events were consistent with the administration of IFN-α-based Adjusted mean change in log HCV-RNA from ANCOVA analysis with baseline as therapy (Table 5). There was no apparent worsening or a covariate for IFN-α alone group and group receiving IFN-α combined with 100 mg (every 8 h) MMPD. ANCOVA analysis including the 300 mg group was appearance of new adverse events observed in the attempted, but could not be performed because the assumption of parallel MMPD-treated patients. Of the adverse events response slopes was not met. ITT, intent-to-treat; MMPD, merimepodib. reported, nausea was most frequently observed with MMPD treatment: 53% of the 300 mg treatment group reported nausea versus 33% of the 100 mg group and included the per-protocol population, defined as those 28% of the IFN-α alone patients. subjects who were treatment-compliant during the 28- Haemoglobin concentrations and WBC counts in day study period. We found a significant (P=0.037; the three treatment arms over the 4-week study treat- mean change: –0.86 log for IFN-α and –1.78 log for ment showed a general downward trend, but the IFN-α/MMPD) effect favouring the combination treat- changes were similar in the three treatment groups ment, while a trend (P=0.15; mean change: –0.93 log (Figure 2). Thus, MMPD did not appear to add to the for IFN-α and –1.43 log for IFN-α/MMPD) was seen haematological toxicity of IFN-α. in the ITT analysis. One patient in the 300 mg group developed signifi- An additional analysis examined the proportion of cant elevation in ALT levels that required discontinua- patients who achieved a ≥2 log drop in HCV-RNA at tion of study drugs. This patient’s baseline ALT was the end of the core study (day 30) as compared with 40 IU/ml (normal range 6–34), and increased to baseline. In the analysis of the ITT population, these 319 IU/ml on day 25 of treatment. Therapy was discon- proportions were similar in the IFN-α alone and tinued on day 28. Maximal levels of ALT and aspartate 100 mg every 8 h MMPD groups; in each group, 7/18 aminotransferase (ASP) (739 and 482 IU/ml, respec- patients (38.9%) demonstrated a ≥2 log drop in tively) were noted on day 29. The alkaline phosphatase,

Table 4. Study outcomes (all patients) following treatment with IFN-α/RBV in the rollover phase

IFN-α + placebo IFN-α + 100 mg MMPD IFN-α + 300 mg MMPD

Enrolled into rollover phase 17 17 17* Discontinued prematurely due to virological failure 7 5 9 Completed 12 months’ treatment 6 5 5 HCV-RNA negative at end of 12 months’ treatment 6 5 5 HCV-RNA negative at 6 months’ follow-up (SVR) 4 4 4

*Includes one patient randomized to 300 mg MMPD in the core study that did not receive study medication. The patient was enrolled directly into the rollover phase to receive standard therapy with IFN-α and RBV. IFN, interferon; MMPD, merimepodib; RBV, ribavirin; SVR, sustained virological response.

Table 5. Adverse events for ITT population in the core study

IFN-α + placebo (n=18) IFN-α + 100 mg MMPD (n=18) IFN-α + 300 mg MMPD (n=17)

Asthenia 13 (72%) 11 (61%) 7 (41%) Headache 8 (44%) 8 (44%) 5 (29%) Nausea 5 (28%) 6 (33%) 9 (53%) Chills 7 (39%) 6 (33%) 6 (35%) Myalgia 5 (28%) 8 (44%) 6 (35%) Fever 6 (33%) 3 (17%) 0 Thinking abnormal 4 (22%) 4 (22%) 1 ( 6%) Diarrhoea 3 (17%) 3 (17%) 2 (12%) Abdominal pain 2 (11%) 2 (11%) 4 (24%)

Number of patients (percentages) reporting adverse events. Events listed include those observed in more than 15% of the total study patients. ITT, intent-to-treat. gamma-glutamyl transferase (GGT) and total bilirubin Discussion levels remained within the normal range and the patient was asymptomatic. The transaminase increases fully The objectives of the study were to evaluate the safety resolved 5 weeks after treatment discontinuation. and tolerability of the IFN-α/MMPD combination, and Another patient in the 300 mg group had a transient to evaluate whether the combination of IFN-α and ALT elevation to greater than five times the upper limit MMPD enhanced on-treatment antiviral activity of normal (baseline: 105 IU/l, day 15: 237 IU/l) that compared with IFN-α alone over 4 weeks of treatment. resolved during continued study drug administration. At the end of the 4-week study period, all patients ALT levels otherwise showed a modest decline in all received treatment with a standard course of IFN-α three treatment groups over 4 weeks, with no significant and RBV. The SVR rate of approximately 25% in the differences between the treatment groups. genotype-1 patient population suggests that the effects of IFN-α and RBV were not compromised by the initial 4-week study treatment. A significant challenge for this study was that Figure 2. Haematologicol parameters: (A) haemoglobin levels MMPD was expected to produce antiviral effects and (B) WBC counts similar to RBV, an effect size that is small relative to A both the variability in HCV-RNA assays and the inherent patient-to-patient variability in terms of IFN- 160 α response. In order to maximize the statistical power 155 150 to determine if MMPD had an additive antiviral effect, 145 certain statistical approaches were pre-specified. This 140 135 included the use of a 1-sided statistical test and the defi- 130 125 nition of a per-protocol population that only included 120 patients who had been compliant with therapy. In addi- Haemoglobin levels, g/l 01529tion, the use of multiple HCV-RNA assessments both at Study day baseline and at the end of the 4 weeks helped mitigate Placebo 100 mg MMPD 300 mg MMPD the effects of assay and within patient variability in B HCV-RNA levels. However, despite these approaches the ITT analysis did not demonstrate a statistically

/ml 8 significant difference, and the per-protocol population 9 7 6 was modestly statistically significant. A significant 5 issue in the statistical analysis was that the inter-subject 4 3 variability was significantly greater than the sample 2 size calculations had assumed: the assumed standard 1 0 deviation in change in log HCV-RNA was approxi- WBC counts, 1x10 01529mately 0.2 log prior to therapy, while the observed Study day value was approximately 1.0 log. One potential reason Placebo 100 mg MMPD 300 mg MMPD for the high variability observed in this study was the high variability in baseline HCV-RNA. There were no MMPD, merimepodib; WBC, white blood cells. limits on serum HCV-RNA concentrations prior to

Antiviral Therapy 10:5 641 JG McHutchinson et al.

treatment resulting in wide ranges in baseline HCV- Since MMPD has both antiviral and immunosup- RNA, with nearly half the patients having more than pressive properties, the adoption of the maximal toler- 10 000 000 copies/ml at baseline. In future studies, the ated dose principle for IMPDH inhibitors may not be use of a narrower range of HCV-RNA concentrations warranted in anti-HCV therapy in combination with at baseline may be recommended. IFN-α. Therefore, the expectation of antiviral activity as The limited antiviral effects seen in this study need to a function of dose with IMPDH inhibitors may not be be interpreted cautiously. One inherent limitation is the applicable. Systemic immunosuppression would be short duration of the study. Whether a full course of expected to offset any antiviral effects since other treatment would increase the percentage of patients immunosuppressants, such as prednisone, have been who achieve HCV-RNA undetectable status at end-of- shown to increase HCV-RNA levels [21]. This finding is treatment and/or increase SVR is unknown and cannot consistent with the fact that RBV is a low-potency be predicted given the design and objectives of the IMPDH inhibitor and as such would not be expected to current study. In this study, the ITT analysis did not produce the immunosuppressive properties demon- reveal a statistically significant effect. However, several strated by more potent agents such as MMPD and points should be considered. The per-protocol popula- mycophenolate mofetil. tion was pre-specified in the protocol as the primary Recent data have suggested that potent IMPDH efficacy population for analysis, as was the use of a inhibitors such as MMPD may in fact enhance the 1-sided test. Since baseline viral load is recognized as antiviral activity of RBV. Depletion of intracellular an important prognostic factor in treatment outcome, guanine nucleotide pools may result in increased the use of ANCOVA was the most appropriate analysis mismatch pairing by RBV and thus increase the trans- given the statistically significant difference between lational mutation rate and contribute to error cata- groups in baseline HCV-RNA levels. In addition, the strophe [9,22]. This concept is further supported by the magnitude of HCV-RNA change demonstrated in the findings of Zhou et al. [23], in which MMPD greatly IFN-α group was within the expected range and similar enhanced the antiviral activity of RBV in the subge- to prior reports [15,16]. nomic replicon. In this respect, rather than MMPD The apparently lower activity of the 300 mg being a potential replacement for RBV, the best MMPD/IFN-α combination relative to IFN-α alone, in approach to using MMPD may be as an additional retrospect, is perhaps consistent with expected outcomes therapy to IFN-α and RBV. Since the initiation of this at high levels of IMPDH inhibition. Mycophenolate clinical study, pegylated interferons (PEG-IFNs) have mofetil (Cellcept®; Roche Pharmaceuticals, Nutley, NJ, replaced standard interferons in the treatment of USA), the only marketed selective uncompetitive hepatitis C. In a study of viral kinetics in patients inhibitor of IMPDH, clearly has immunosuppressive treated with PEG-IFN [24], the decrease in viral load properties as it is approved for post-transplantation over 28 days of PEG-IFN monotherapy was 0.86 log, management and has demonstrated clinical activity in the same change that was seen with interferon a number of autoimmune disorders [19]. Studies to monotherapy in this study. Therefore, further studies date have indicated that the addition of transplant with PEG-IFN combination therapy and with MMPD doses of mycophenolate mofetil (2000 mg/day) to are required. IFN-α fails to enhance response rates in patients with In conclusion, the results of this study indicate hepatitis C who have failed to respond to antiviral MMPD at low doses has a potentially similar effect to therapy [20]. For mycophenolate mofetil and MMPD, RBV in increasing on-treatment antiviral efficacy when the mechanism of immunosuppression is thought to be added to IFN-α therapy. Larger, longer duration treat- via a systemic effect on lymphocyte proliferation and ment trials incorporating PEG-IFN, and using MMPD activation. The Cmaxss levels achieved in the current either alone or in combination with RBV, are required study would be expected to have an effect on lympho- to further explore the findings from this study. A small cyte proliferation, but the median trough concentra- Phase II study, which evaluated the safety of the triple tion for the 100 mg dose, which was 168 ng/ml as combination (PEG-IFN, RBV and MMPD), has been opposed to 250 ng/ml for the 300 mg dose, would not completed [25] and a study designed to evaluate the be expected to have an effect. Moreover, the liver effect of this regimen on SVR rates is in progress. concentrations of MMPD in a C14-ADME study in rats were approximately four- to fivefold higher than Support plasma concentrations. Thus, it may be that at lower doses the concentrations achieved in the liver are suffi- Supported in part by research grants from Vertex cient for activity, while at higher doses, systemic Pharmaceuticals, Cambridge, MA, USA and clinical concentrations leading to immunosuppressive effects research centre grants from Scripps Clinic, La Jolla, are also achieved. CA, USA (MO1-RR00833).

References hepatitis C patients during treatment with ribavirin. Journal of Gastroenterology & Hepatology 1994; 1. Di Bisceglie AM, Conjeevaram HS, Fried MW, Sallie R, 9:587–591. Park Y, Yurdaydin C, Swain M, Kleiner DE, Mahaney K & 14. Stuyver L, Rossau R, Wyseur A, Duhamel M, Hoofnagle JH. Ribavirin as therapy for chronic hepatitis C. Vanderborght B, Van Heuverswyn H & Maertens G. A randomized, double-blind, placebo-controlled trial. Typing of hepatitis C virus isolates and characterization of Annals of Internal Medicine 1995; 123:897–903. new subtypes using a line probe assay. Journal of General 2. Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Virology 1993; 74:1093–1102. Ideo G, Bain V, Heathcote J, Zeuzem S, Trepo C & 15. Khakoo S, Glue P, Grellier L, Wells B, Bell A, Dash C, Albrecht J. Randomised trial of interferon alpha 2b plus Murray-Lyon I, Lypnyj D, Flannery B, Walters K & ribavirin for 48 weeks or for 24 weeks versus interferon Dusheiko GM. Ribavirin and interferon alfa-2b in chronic alpha 2b plus placebo for 48 weeks for treatment of hepatitis C: assessment of possible pharmacokinetic and chronic infection with hepatitis C virus. Lancet 1998; pharmacodynamic interactions. British Journal of Clinical 352:1426–1432. Pharmacology 1998; 46:563–570. 3. McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, 16. Neumann AU, Dahari H, Conrad A, Pianko S & Lee WM, Rustgi VK, Goodman ZD, Ling MH, Cort S & McHutchison JG. Early predication and mechanism of the Albrecht JK. Interferon α-2b alone or in combination with ribavirin/IFN-α dual therapy effect on chronic hepatitis C ribavirin as initial treatment for chronic hepatitis C. New virus (HCV) infection. Hepatology 1999; 30:309A. England Journal of Medicine 1998; 339:1485–1492. 17. Nyberg L, Albrecht J, Glue P, Gianelli G, Zambas D, Elliot 4. Manns MP, McHutchison JG, Gordon SC, Rustgi VK, M, Conrad A & McHutchison J. Changes in serum Shiffman M, Reindollar R, Goodman ZD, Koury K, Ling hepatitis C virus RNA in interferon nonresponders M & Albrecht JK. Peginterferon alfa-2b plus ribavirin retreated with interferon plus ribavirin: a preliminary compared with interferon alfa-2b plus ribavirin for initial report. Journal of Clinical Gastroenterology 1999; treatment of chronic hepatitis C: a randomised trial. Lancet 28:313–316. 2001; 358:958–965. 18. Senn S. Statistical Issues in Drug Development, 1997. New 5. Fried MW, Shiffman ML, Reddy KR, Smith C, Marinos G, York: John Wiley. Gonçales FL, Häussinger D, Diago M, Carosi G, 19. Kitchin JE, Pomeranz MK, Washenik K & Shupack JL. Dhumeaux D, Craxi A, Lin A, Hoffman J & Yu J. Rediscovering mycophenolic acid: a review of its mecha- Peginterferon alfa-2a plus ribavirin for chronic hepatitis C nism, side effects, and potential uses. Journal of the virus infection. New England Journal of Medicine 2002; American Academy of Dermatology 1997; 37:445–449. 347:975–982. 20. Cornberg M, Hinrichsen H, Teuber G, Berg T, Naumann 6. Markland W, McQuaid TJ, Jain J & Kwong AD. Broad U, Falkenberg C, Zeuzem S & Manns MP. Mycophenolate spectrum antiviral activity of the IMP dehydrogenase mofetil in combination with recombinant interferon alfa-2a inhibitor VX-497: a comparison with ribavirin and demon- in interferon-nonresponder patients with chronic stration of antiviral additivity with alpha interferon. hepatitis C. Journal of Hepatology 2002; 37:843–847. Antimicrobial Agents & Chemotherapy 2000; 44:859–866. 21. McHutchison JG, Ponnudurai R, Bylund DL, Anguiano A, 7. Lau JY, Tam RC, Liang TJ & Hong Z. Mechanism of Pockros PJ, Mondala T & Wilkes LB. Prednisone with- action of ribavirin in the combination treatment of chronic drawal followed by interferon alpha for treatment of HCV infection. Hepatology 2002; 35:1002–1009. chronic hepatitis C infection: results of a randomized 8. Thomas HC, Torok ME, Forton DM & Taylor-Robinson controlled trial. Journal of Clinical Gastroenterology 2001; SD. Possible mechanisms of action and reasons for failure 32:133–137. of antiviral therapy in chronic hepatitis C. Journal of 22. Contreras AM, Hiasa Y, He W, Terella A, Schmidt EV & Hepatology 1999; 31(Suppl 1):152–159. Chung RT. Viral RNA mutations are region specific and 9. Crotty S, Cameron C & Andino R. Ribavirins antiviral increased by ribavirin in a full-length hepatitis C virus mechanism of action: lethal mutagenesis? Journal of replication system. Journal of Virology 2002; Molecular Medicine 2002; 80:86–95. 76:8505–8517. 10. Bartenschlager R & Lohmann V. Novel cell culture systems 23. Zhou S, Liu R, Baroudy BM, Malcolm BA & Reyes GR. for the hepatitis C virus. Antiviral Research 2001; 52:1–17. The effect of ribavirin and IMPDH inhibitors on hepatitis C virus subgenomic replicon RNA. Virology 2003; 11. Lin C, Lin K & Kwong AD. Antiviral activity of rationally 310:333–342. designed IMPDH inhibitors in a cell-based hepatitis C virus replication system. 8th International Symposium on 24. Hermann E, Lee J-H, Marinos G, Modi M & Zeuzem S. Hepatitis C Virus & Related Viruses. 2–5 September 2001, Effect of ribavirin on hepatitis C viral kinetics in patients Paris, France. Abstract O-73. treated with pegylated interferon. Hepatology 2003; 37:1351–1358. 12. Wright T, Shiffman ML, Knox S, Ette E, Kauffman RS & Alam J. Dose ranging study of VX-497, a novel oral 25. Marcellin P, Horsmans Y, Nevens F, Grangé JD, IMPDH inhibitor, in patients with hepatitis C. Hepatology Bronowicki JP, Vetter D et al. A phase II, placebo- 1999; 30(Suppl.):990A. controlled study of merimepodib (VX497), in combination with pegylated interferon-alfa, and ribavirin in patients 13. Tong MJ, Hwang SJ, Lefkowitz M, Lee SD, Co RL, with chronic hepatitis C non-responsive to previous Conrad A, Schmid P & Lo KJ. Correlation of serum HCV therapy with interferon-alfa and ribavirin. Journal of RNA and alanine aminotransferase levels in chronic Hepatology 2004; 40(Suppl. 1):145.

Received 20 August 2004, accepted 9 June 2005

Antiviral Therapy 10:5 643