Lack of Efficacy of an Inhibitor of PDE4 in Phase 1 and 2 Trials of Patients With Nonalcoholic...

Clinical Gastroenterology and Hepatology 2014;-:-–-

All studies published in Clinical Gastroenterology and Hepatology are embargoed until 3PM ET of the day they are published as correctedproofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.

Lack of Efficacy of an Inhibitor of PDE4 in Phase 1 and 2 Trialsof Patients With Nonalcoholic Steatohepatitis

Vlad Ratziu,* Pierre Bedossa,‡ Sven M. Francque,§ Dominique Larrey,jj Guruprasad P. Aithal,¶

Lawrence Serfaty,# Mihai Voiculescu,** Liliana Preotescu,‡‡ Frederik Nevens,§§

Victor De Lédinghen,jjjj,¶¶ Gabriele I. Kirchner,## Pavel Trunecka,*** Stephen D. Ryder,¶

Christopher P. Day,‡‡‡ Jun Takeda,§§§ and Klaudia Traudtnerjjjjjj

*Université Pierre et Marie Curie, Assistance Publique Hôpitaux de Paris, Hôpital Pitié Salpêtrière, INSERM UMR_S 938, Paris,France; ‡Hôpital Beaujon, INSERM U773, Université Paris Diderot, Paris, France; §Department of Gastroenterology andHepatology, Antwerp University Hospital, Faculty of Medicine and Health Sciences, Antwerp University, Antwerp, Belgium;jjDépartement d’Hépatogastroenterologie et Transplantation, Hôpital Saint Eloi, INSERM1040-IRB, Montpellier, France;¶National Institute for Health Research (NIHR) Biomedical Research Unit in Gastrointestinal and Liver Diseases, NottinghamUniversity Hospitals NHS Trust and The University of Nottingham, Nottingham, United Kingdom; #Assistance PubliqueHôpitaux de Paris, Université Pierre et Marie Curie, Service d’Hépatologie, INSERM UMR_938, Hôpital Saint-Antoine, Paris,France; **Department of Internal Medicine and Nephrology, Fundeni Clinical Institute, Bucharest, Romania; ‡‡Matei BalsNational Institute for Infectious Diseases, Bucharest, Romania; §§Department of Hepatology, University Hospitals KU Leuven,Leuven, Belgium; jjjjService d’Hépato-Gastro-Entérologie, Hôpital Haut Lévêque, Pessac, France; ¶¶INSERM U1053, UniversitéBordeaux Segalen, Bordeaux, France; ##Department of Internal Medicine I, University Hospital of Regensburg, Regensburg,Germany; ***Department of Hepatogastroenterology, Institute for Clinical and Experimental Medicine (IKEM), Prague, CzechRepublic; ‡‡‡Institute of Cellular Medicine, Faculty of Medical Sciences, Newcastle University, United Kingdom; §§§AstellasPharma, Inc, Tokyo, Japan; and jjjjjjAstellas Pharma Europe BV, Leiden, The Netherlands

BACKGROUND & AIMS:

Abbreviations used in this papetransferase; AST, aspartate amidrug concentration time curve;centration time curve over dosinmonophosphate; Cmax, maximucharide; MCD, methionine-cholhepatitis; PDE, phosphodiester

ASP9831 is a phosphodiesterase-4 inhibitor developed to treat nonalcoholic steatohepatitis(NASH); it showed potent anti-inflammatory and antifibrotic effects in preclinical studies. Weevaluated the efficacy and safety of ASP9831 in patients with NASH.

METHODS:
In a phase 1 trial, we determined the optimal therapeutic window of ASP9831 in healthy volun-teers and evaluated 2 doses (50 and 100 mg) in patients with NASH. Based on the positive out-comes of the phase 1 study, we performed a phase 2 trial to compare the biochemical effects ofASP9831 vs placebo. Patients with NASH were assigned randomly to groups given either 50 mg(n[ 33) or 100 mg (n[ 33) ASP9831 twice daily, or placebo (n[ 30), for 12 weeks. The primaryend point was the mean percentage change, from baseline to the end of ASP9831 administration,in serum level of alanine aminotransferase (ALT); secondary outcomes included changes inaspartate aminotransferase (AST) levels, ratio of AST:ALT, and various biomarkers of NASH.
RESULTS:
After 12 weeks of administration, there was no significant change in mean serum levels of ALT(P [ .42) or AST (P [ .20) or other biomarkers in any group, and no significant differenceswere observed among groups. Most adverse events were mild; gastrointestinal disordersoccurred more frequently in the ASP9831 groups than the placebo group.
CONCLUSIONS:
Despite a relevant mechanism of action, ASP9831 did not significantly alter the biochemicalmarkers of NASH, compared with placebo, in a clinical trial. This highlights the difficulties ofdeveloping therapeutics for NASH and the need for more extensive preclinical testing ofmechanisms of potential drug candidates. Clinicaltrialsregister.eu: 2005-001687-31; EudraCTnumbers: 2007-002114-19.
Keywords: Fatty Liver Disease; Therapeutic Targets; Treatment; PDE4.

r: AE, adverse event; ALT, alanine amino-notransferase; AUC, area under the plasmaAUCtau, area under the plasma drug con-g interval tau; cAMP, adenosine 30,50-cyclicm drug concentration; LPS, lipopolysac-ine deficient; NASH, nonalcoholic steato-ase; tmax, time to attain maximum drug

concentration; TNF, tumor necrosis factor; TEAE, treatment-emergentadverse event.

© 2014 by the AGA Institute1542-3565/$36.00

http://dx.doi.org/10.1016/j.cgh.2014.01.040

FLA 5.2.0 DTD � YJCGH53706_proof � 19 May 2014 � 8:09 pm � ce


2 Ratziu et al Clinical Gastroenterology and Hepatology Vol. -, No. -

Nonalcoholic steatohepatitis (NASH) is an increas-ingly prevalent liver disease worldwide. Patients

at risk for NASH include those sharing clusters ofcomorbidities intimately related to insulin resistance (eg,obesity/overweight, type 2 diabetes, dyslipidemia).1

Although nonalcoholic fatty liver (ie, steatosis alone orwith mild lobular inflammation) usually is considered abenign, nonevolving condition, NASH is a cause ofcirrhosis, end-stage liver disease, and hepatocellular car-cinoma, which increases liver-related mortality 10-foldand significantly reduces survival.2 No drug therapycurrently is available for NASH; however, in some trials,thiazolidinediones or vitamin E have shown modest ther-apeutic efficacy on histologic parameters.3

NASH differs from steatosis alone by the presence ofhepatic inflammation and cell injury in response to lip-otoxic intermediates that accumulate or are metabolizedin the steatotic liver.4 Several experimental studies inanimal models of steatohepatitis have shown the acti-vation of intrahepatic inflammatory pathways leading tothe activation of inflammatory mediators and recruit-ment of inflammatory cells.5 However, selective phar-macologic blockades of these different pathways did notresult in unambiguous improvement in NASH models,and they have not been proven applicable to humanbeings with this disease.

Adenosine 30,50-cyclic monophosphate (cAMP), a sec-ond messenger, mediates numerous signaling responsesthat result in suppression or inhibition of inflammatory-mediator release, as well as inhibition of fibrosis andsmoothmuscle relaxation. Intracellular levels of cAMP aremodulated by phosphodiesterases (PDEs) that degradecAMP into 50-adenosine monophosphate.6 The super-family of PDE enzymes comprises 11 or more membersdiffering in structure, substrate specificity, inhibitorselectivity, tissue and cell distribution, and regulation bykinases.7 Inflammatory cells crucial to the pathogenesis ofinflammatory diseases preferentially express PDE4,which has led to the development of highly potent andselective PDE inhibitors. These compounds increase theintracellular levels of cAMP and reduce the activation of awide range of inflammatory cells.8 Some of these com-pounds have been approved (ie, roflumilast) or are indevelopment for inflammatory diseases (eg, asthma,chronic obstructive pulmonary disease, arthritis, inflam-matory bowel disease, and skin inflammatory disorders).Pentoxifylline, a nonselective inhibitor of PDEs, reducessome lethal complications of end-stage cirrhosis (eg,encephalopathy, bacterial infection) and increasescomplication-free survival.9 It is not known if these effectsresulted from the inhibition of PDEs or tumor necrosisfactor (TNF)-a synthesis.

ASP9831 is a new PDE4 inhibitor with anti-inflammatory properties mainly directed at activatedmacrophages and Kupffer cells. Upon stimulation bylipopolysaccharide (LPS), ASP9831 inhibits TNF-a pro-duction, in the nanomolar range, from both human andrat peripheral blood mononuclear cells (data on file;

FLA 5.2.0 DTD � YJCGH53706_proo

Astellas Pharma Europe BV, Leiden, The Netherlands).Moreover, ASP9831 significantly reduced alanine ami-notransferase (ALT) activity with a dose effect in 3distinct acute hepatitis models: D-galactosamine andD-galactosamine plus LPS in rats, and concanavalin A inmice (data on file; Astellas Pharma). When administeredconcurrently with a methionine-choline deficient (MCD)diet for 16 weeks, ASP9831 achieved 40% to 50%reduction in aminotransferase levels and improved his-tologic necroinflammation at doses of 1 and 3 mg/kg.There was a significant reduction in hydroxyproline he-patic content and an improvement in fibrosis score,suggesting that ASP9831 also might have antifibroticeffects as a result of its anti-inflammatory activity (dataon file; Astellas Pharma). These preclinical data stronglysupport an anti-inflammatory and antifibrotic potentialfor ASP9831 in chronic models of steatohepatitis andprompted the initiation of phase I and IIa clinical trialsin healthy volunteers and patients with NASH.

Methods

The Supplementary Materials and Methods sectiondescribes in detail the methodology and patient eligi-bility criteria.

In brief, the phase I study first evaluated single- andmultiple-ascending doses of ASP9831 in healthy volun-teers. Outcome variables were frequency and severity ofadverse events (AEs) and pharmacokinetic parameters forplasma and urine. Ex vivo TNF-a production was assessedas a biomarker for inflammation, and safety data werecollected throughout the study. ASP9831 then was eval-uated in patients with NASH to assess whether fibroticliver injury in NASH significantly modifies the pharmaco-kinetic parameters. Based on the results obtained inhealthy volunteers, 2 doses of ASP9831 (50 and 100 mg)were administered sequentially to patients with NASH andadvanced fibrosis for 8 days (Supplementary Figure 1A).Outcome variables were similar to those in healthy vol-unteers (Supplementary Materials and Methods).

The phase II study was a randomized, double-blind,parallel-group, placebo-controlled study investigatingthe effect of 12 weeks of administration of 50 and 100 mgASP9831 twice daily vs placebo on serum level of ALT,various liver injury biomarkers in NASH, and safety(Supplementary Figure 1B). Adult outpatients (age,�18 y)with biopsy-confirmed NASH and increased serum ALTlevel (�1.5 � upper limit of normal: 41 U/L for men and31 U/L for women, but �300 U/L) at screening wereincluded and randomized to one of the treatment groupsin a 1:1:1 ratio (Supplementary Materials and Methods).Each subject signed an institutional review board–approved written informed consent form.

The primary efficacy variable was the percentagechange in serum ALT level after 12 weeks of administrationvs baseline. Because this was a proof-of-concept study toevaluate the efficacy of ASP9831 in NASH, a biochemical

f � 19 May 2014 � 8:09 pm � ce

Figure 1. Inhibition of TNF-a by ASP9831. Young healthymen received single oral doses of ASP0831 (50 mg or 100mg; n ¼ 6 in each dose group) or placebo (n ¼ 20). TNF-aplasma concentration was measured at baseline and up to 24hours after dose intake.

- 2014 Pharmacologic PDE4 Inhibition in Human NASH 3

rather than histologic end point was chosen, as recom-mended by experts.10 Secondary variables were as follows:the proportion of patients with a normal ALT level at 12weeks, changes from baseline to the end of administrationin serum aspartate aminotransferase (AST) level, AST:ALTratio, and various biomarkers for NASH (SupplementaryMaterials and Methods). Plasma drug concentration wassparsely measured. Safety was recorded as the frequencyof treatment-emergent AEs (TEAEs).

The safety population comprised patients who took1 or more tablets of study medication. The full-analysispopulation comprised patients who took 1 or more tab-lets of study medication and had 1 or more postbaselineALT values. The per-protocol population comprised pa-tients in the full analysis set who adhered to protocol;principal reasons for exclusion were use of a prohibitedmedication and treatment duration less than 9 weeks.

Statistical Analysis

Outcomes in the phase I study were reported usingdescriptive statistics (mean � SD; coefficient of variation;minimum, median, and maximum; frequencies; andpercentages). Calculations included the evaluation ofdose-proportionality for the area under the plasma drugconcentration time curve (AUC) and the maximum drugconcentration (Cmax) of ASP9831 in plasma (last dose vsfirst dose in the pharmacokinetic study), as well as therelationship between ASP9831 doses and ex vivo TNF-a.

The primary analysis in the phase II study was themean percentage change of serum ALT level vs baseline.Analysis of covariance was used to analyze the percentagechange from baseline in serum ALT level at 12 weeks withthe last observation carried forward (a ¼ .05, 2-sided).Treatment group, ALT level, type 2 diabetes, and bodymass index (<30 or�30 kg/m2) at baseline were includedas fixed effects, and site was included as a random effect.The significance of pairwise comparisons can be statedformally only when the overall comparison was significantto prevent a-error inflation. The primary analysis wasperformed on the full-analysis population. The secondaryanalysis summarized biomarkers using descriptive statis-tics. A sample size of 31 subjects in each arm was based onthe comparison of 2 means; delta, SD, power, and drop-outrate were 20%, 25%, 80%, and 20%, respectively.

All authors had access to the study data and reviewedand approved the final manuscript.

Results

The baseline characteristics of all participants in pha-ses I and II are summarized in Supplementary Table 1.

Phase I

Healthy volunteers. Volunteers were randomized togroups based on ASP9831 dose and demographics (sex


and age). In the single-ascending dose study, the meanCmax and AUC increased proportionally to dose up to 150mg in young men (data not shown). Administration of50 mg ASP9831 twice daily resulted in drug exposuresimilar to that observed with 100 mg once daily. Inyoung men, the mean Cmax was similar on days 1, 9, and10 after 50 mg ASP9831 twice daily but tended to in-crease over time with the 100 mg twice-daily dose. Largeintersubject variability was noted. The change in areaunder the plasma drug concentration time curve overdosing interval tau (AUCtau) between days 9 and 10depended on age and daily dose. Only small fractions ofASP9831 were excreted in the urine after all single- andmultiple-ascending doses in all groups.

Substantial TNF-a suppression was achieved forapproximately 8 to 14 hours with single oral doses be-tween 20 and 400 mg of ASP9831 (SupplementaryFigure 2). A rapid (<1 h) and potent (35% of baseline)inhibition of serum TNF-a was seen after a single dose of50 and 100 mg of ASP9831 (Figure 1). After multipleASP9831 doses, consistent inhibition of TNF-a produc-tion was observed over 24 hours (>50% of baselinevalues) in all groups except for elderly women receiving50 mg twice daily (data not shown). The maximum singledose was set at 300 mg.

ASP9831 was safe and well tolerated. TEAEs weremostly mild. In the single-dose study, treatment-relatedTEAEs were reported in 55.9% and 29.4% of the sub-jects receiving ASP9831 and placebo, respectively, and in65.5% and 46.4% of subjects in the multiple-dose study.In addition, in the single-dose study, doses of 150 mg orhigher of ASP9831 tended to be associated with anincreased incidence of treatment-related TEAEs in youngadults. The most commonly reported TEAEs weregastrointestinal symptoms, dizziness, and headache.There were no deaths or serious AEs. None of the sub-jects withdrew as a result of AEs. No consistent or clin-ically relevant treatment or dose-related trends in mean

f � 19 May 2014 � 8:09 pm � ce


changes in biochemistry (including hepatic parameters)or hematology values were observed. No clinical abnor-malities were noted on routine electrocardiographyassessments.

Patients with nonalcoholic steatohepatitis. Fourteenpatients with biopsy-confirmed NASH and liver fibrosis(Figure 2A) completed the study comparing ASP9831exposure levels with those of healthy volunteers. Themean AUCtau after administration of 50 mg ASP9831twice daily was below the mean AUC after administrationof 100 mg ASP9831 twice daily in healthy volunteers.Therefore, evaluation of 100 mg ASP9831 twice daily inpatients with NASH proceeded as planned. ASP9831 waswell absorbed, with a mean time to attain the maximumdrug concentration (tmax) of 1.26 hours for the 50-mggroup and 1.71 hours for the 100-mg group. The Cmax

and AUCtau increased in a dose-proportional mannerwhen the ASP9831 dose was increased from 50 to 100mg. For most of the other pharmacokinetic parametersassessed, no dose dependency was observed(Supplementary Table 2).

Compared with healthy volunteers, ASP9831 wasabsorbed more slowly in patients with NASH. The meanand median tmax were 2-fold longer, and the mean Cmax

was 2-fold lower. The administration of ASP9831 infasted and fed conditions in healthy volunteers showedthat the presence of food increased tmax and decreasedCmax (data not shown). ASP9831 was given with food inpatients with NASH, which may have affected absorption.

Four of 7 patients given 50 mg ASP9831, and 6 of7 patients receiving 100 mg ASP9831 experienced 1or more TEAEs. All TEAEs were mild or moderatein severity and the majority was categorized as

Figure 2. NASH patient flow diagram for (A) phase I and (BSupplementary Materials and Methods section.


gastrointestinal disorders. No deaths or other seriousAEs occurred and no AE resulted in treatmentdiscontinuation.

Phase II

Eighty-six patients with NASH completed the 12-week study examining the efficacy and safety of pla-cebo and 50 and 100 mg ASP9831 administered twicedaily (Figure 2B). Mean serum ALT values were stablethroughout the study and were similar in all 3 treatmentgroups (placebo and both active treatment groups)(Figure 3A). Findings were similar in the full analysis andper-protocol population. The mean percentage changefrom baseline to end of treatment in serum ALT level was-13.3% with placebo and -1.3% and -9.3% with 50 and100 mg ASP9831, respectively (overall comparison, P ¼.42; estimates with placebo, 9.3 [95% confidence inter-val, -5.9 to 24.5] and 1.4 [95% confidence interval, -13.7to 16.4]).

No significant change from baseline or trend wasobserved in the biomarkers adiponectin, cytokeratin 18,and TNF-a (Figure 3B–D), and no significant differenceswere observed between the ASP9831 and placebogroups. No significant changes were observed for liverinjury biomarkers (Supplementary Figure 3). The meanpercentage change in AST level from baseline to end oftreatment did not differ significantly among treatmentgroups: -8.6% for placebo; and 11.9% and -4.5% for 50and 100 mg ASP9831, respectively (overall comparison,P ¼ .20). Results of the analysis using the per-protocolpopulation were similar to those performed on the

) phase II. *The randomization method is described in the

f � 19 May 2014 � 8:09 pm � ce

Figure 3.Mean values of (A) ALT, (B) adiponectin, (C) cytokeratin 18, and (D) TNF-a. All graphs show the mean � SD in thebiomarker tested in each treatment group. Data from full-analysis set are as follows: placebo, n ¼ 29; 50 mg ASP9831 twicedaily, n ¼ 32; and 100 mg ASP983 twice daily, n ¼ 33. LOCF, last observation carried forward.


full-analysis population. Treatment with ASP9831 didnot induce changes in surrogate markers for insulinresistance.

The incidence of TEAEs was 60.0% in placebo re-cipients and 75.8% and 87.9% in patients receiving 50and 100 mg ASP9831, respectively. The incidence ofTEAEs deemed related to the study drug was 36.7% inthe placebo group and 48.5% and 78.8% in the ASP983150 and 100 mg groups, respectively. Table 1 summarizesthe types of TEAEs reported in 2 or more patients in anygroup. Gastrointestinal disorders were the most commonTEAEs reported overall.

Table 1. TEAEs Reported in 2 or More Patients in Any Group in

System organ class, n (%)

Gastrointestinal disordersNervous system disordersGeneral disorders and administration site conditionsMusculoskeletal and connective tissue disordersInfections and infestationsRespiratory, thoracic, and mediastinal disordersPsychiatric disordersSkin and subcutaneous tissue disordersEar and labyrinth disordersRenal and urinary disordersImmune system disorders


All except 3 TEAEs were categorized as mild or mod-erate in severity. The exceptions were 2 events in theplacebo group (1 case of severe headache and 1 case ofanaphylactic shock resulting in death) and 1 event in theASP9831 50-mg group (a case of severe muscle spasmsleading to study drug withdrawal). No consistent, signifi-cant changes in biochemical or hematologic values werenoted.

One death occurred in the study (placebo group), as aresult of anaphylactic shock after a wasp sting; the deathwas judged unrelated to the study drug. Two otherserious AEs also occurred in the placebo group. One,

Phase II

Placebo(n ¼ 30)

ASP9831

50 mg(n ¼ 33)

100 mg(n ¼ 33)

9 (30.0) 14 (42.4) 22 (66.7)6 (20.0) 10 (30.3) 7 (21.2)

10 (33.3) 9 (27.3) 4 (12.1)3 (10.0) 5 (15.2) 7 (21.2)6 (20.0) 7 (21.2) 3 (9.1)2 (6.7) 6 (18.2) 1 (3.0)2 (6.7) 3 (9.1) 3 (9.1)3 (10.0) 4 (12.1) 0

0 1 (3.0) 2 (6.1)1 (3.3) 0 2 (6.1)2 (6.7) 1 (3.0) 0

f � 19 May 2014 � 8:09 pm � ce


moderate diarrhea, prompted study discontinuation andwas described as recovering/resolving. The other event,hospitalization for mild angina pectoris, led to temporarystudy drug withdrawal and was documented as resolved.One other event, intermittent mild dyspepsia in a patientreceiving 100-mg ASP9831, also resulted in study drugwithdrawal.

Discussion

Despite a strong preclinical rationale, ASP9831 failedto show any therapeutic efficacy on biochemical param-eters in patients with histologically confirmed NASH andincreased aminotransferase levels. Neither 50 nor 100mg ASP9831 twice daily reduced ALT values vs placeboover a 12-week treatment period. The study includedonly patients with persistently increased ALT levels andmore than 1.5 � the upper limit of normal on inclusion.The ALT values remained high on therapy with no sig-nificant spontaneous reduction in either group; there-fore, the lack of biochemical effect was not caused byfluctuating ALT levels or regression to the mean. Also,the duration of therapy probably could not account forfailure to lower ALT level because several active mole-cules significantly reduced ALT level over a comparabletime period in other trials.11–14 It is possible that,because of the weak association between ALT and thehistologic severity of NASH, a therapeutic effect ofASP9831 may have been missed. However, to be feasible,pilot trials need to rely on measurable surrogate out-comes, and ALT level is considered a suitable surrogatemarker for liver injuries. In addition, most potential drugcandidates for NASH that usually improved histologyalso reduced ALT levels.3,11,15 Finally, insufficient dosingis another possible but unlikely explanation for the lackof biochemical effect because the dosing used in thishuman study was calculated based on effective regimensin rodent models of liver disease.

A negative pilot trial with a new pharmacologic agentshould raise questions about the validity of the preclinicaldata that provide the rationale for its testing in humanbeings. Clinical trials are costly and involve significantcommitment fromboth the patients and physicians aswellas risk management by the sponsor. Therefore, lessonsfrom negative trials are important, and future attemptsshould be based on a sound assessment of the drug can-didate’s potential. The current proof-of-concept trial rai-ses the following questions: (1) was there sufficientrationale for the mechanism of action of the drug; (2) wasthe proof of efficacy in preclinical models convincing; and(3) was there sufficient evidence to extrapolate themechanism of action to a specific organ disease?

The biological effects of ASP9831 observed in preclini-cal studies were deemed sufficient to initiate testing of thismolecule in humanNASH. Themajor effect of ASP9831, andthe basis for its testing in this inflammatory hepatic con-dition, is a potent inhibition of TNF-a production from


peripheral blood mononuclear cells stimulated by LPS(data on file; Astellas Pharma). Of note, the magnitude ofTNF-a inhibition in response to LPS (50%)was comparablewith that observed under similar experimental conditionswith roflumilast, a PDE4 inhibitor approved for severechronic obstructive pulmonary disease exacerbations.16

TNF-a may be central to the pathogenesis of steatohepa-titis and, as such, its inhibition is assumed to be an attrac-tive pharmacologic target. However, this trial failed to showimprovement in ALT values, a surrogate of hepaticinflammation in NASH. Interestingly, a more than 50% in-hibition of TNF-a expression was observed in healthy in-dividuals treated with 100 or 200 mg ASP9831 daily butnot in patients with NASH. Possibly, the reduction of TNF-asignaling through cAMP induction could trigger compen-satory inflammatory pathways, resulting in TNF-a induc-tion. These results are reminiscent of the failure ofmetformin to improve human NASH through TNF-a inhi-bition.17,18 However, pentoxifylline, which also reducesTNF-a expression in animal models of NASH, improvedhistology in a small, 1-year, randomized, controlled trial15;however, the unexpected improvement in fibrosis but notin hepatocyte ballooning observed in this trial calls forconfirmatory clinical studies.

The decision to proceed with a therapeutic trial ofASP9831 in patients with NASH was based on the confir-mation of a range of perceived beneficial anti-inflammatory and antifibrotic effects of the drug in theMCDmodel. These results includeda60%reduction inALTlevels and a significant reduction in the histologic scores ofinflammation (fromamedian of 2 in the untreated group toamedian of 0 in the highest dose group) and fibrosis (froma median of 2 to a median of 1). There were no, ornonsignificant, reductions in steatosis and hepatic tri-glyceride levels, suggesting that ASP9831 has anti-inflammatory rather than antisteatogenic properties. Themagnitude of these anti-inflammatory and antifibrotic ef-fects in the MCD model appears similar to that of piogli-tazone in in vivo models of liver fibrosis.19 Pioglitazonewaseffective in reversing steatohepatitis and reducingALTlevels in human NASH.3 However, the MCD model isnotoriously imperfect as a rendition of the human disease.Therefore, we believe the results of this trial emphasize theimportance that all futuredrug candidates shouldbe testedin several animal models besides the MCD model. More-over, a full assessment of the potential of a drug shouldinclude testing in 2 animal models of fibrosis unrelated tononalcoholic fatty liver disease because some drugs mightbe particularly attractive as antifibrotics in NASH.

The relevance of the PDE family in the pathogenesis ofNASH is incompletely established. Specific expression ofeach PDE isoform in steatohepatitis is unknown, and thereare no data on the contribution of PDE4 inhibition in NASHmodels. Moreover, there are 4 different subtypes of PDE4displaying variable expression, activity, and effects indifferent inflammatory cells, such as macrophages, mono-cytes, and neutrophils.8 Current PDE4 inhibitors in devel-opment (includingASP9831)arenot subtypeselective.Their

f � 19 May 2014 � 8:09 pm � ce


efficacy in certain pathologic models might be reducedbecause it is dependent on the subtype distribution and ac-tions of specific inflammatory cells. Further studies areneeded to understand if other PDE isoforms should be tar-geted to obtain a full anti-inflammatory action in NASH.

Finally, administration of ASP9831 for 12 weeksseemed tolerable in patients with NASH. Most of thereported TEAEs were nausea, vomiting, or headache;they were mild and did not lead to patient discontinua-tion. These side effects were shown to be dose-limitingwith PDE4 inhibitors such as roflumilast.8

In conclusion, despite an attractive rationale based onits anti-inflammatory mechanism of action, the selectivePDE4 inhibitor ASP9831 was safe but inefficient inreducing ALT levels and other surrogates of liver nec-roinflammation in patients with NASH. This study high-lights the difficulties of translational science and theneed for a much more thorough assessment and largerbody of preclinical data before considering a drugcandidate ready for human testing in NASH.

Supplementary Material

Note: To access the supplementary material accom-panying this article, visit the online version of ClinicalGastroenterology and Hepatology at www.cghjournal.org,and at http://dx.doi.org/10.1016/j.cgh.2014.01.040.

References
1. Chalasani N, Younossi Z, Lavine JE, et al. The diagnosis and
management of non-alcoholic fatty liver disease: Practice Guide-line by the American Association for the Study of Liver Diseases,American College of Gastroenterology, and the American Gastro-enterological Association. Hepatology 2012;55:2005–2023.

2. Ratziu V, Bellentani S, Cortez-Pinto H, et al. A position paper onNAFLD/NASH based on the EASL 2009 Special Conference.J Hepatol 2010;53:372–384.

3. Sanyal AJ, Chalasani N, Kowdley KV, et al. Pioglitazone, vitaminE, or placebo for nonalcoholic steatohepatitis. N Engl J Med2010;362:1675–1685.

4. Neuschwander-Tetri BA. Hepatic lipotoxicity and the patho-genesis of nonalcoholic steatohepatitis: the central role ofnontriglyceride fatty acid metabolites. Hepatology 2010;52:774–788.

5. Cusi K. Role of obesity and lipotoxicity in the development ofnonalcoholic steatohepatitis: pathophysiology and clinical im-plications. Gastroenterology 2012;142:711–725.

6. Dastidar SG, Rajagopal D, Ray A. Therapeutic benefit of PDE4inhibitors in inflammatory diseases. Curr Opin Investig Drugs2007;8:364–372.

7. Lugnier C. Cyclic nucleotide phosphodiesterase (PDE) super-family: a new target for the development of specific therapeuticagents. Pharmacol Ther 2006;109:366–398.

8. Spina D. PDE4 inhibitors: current status. Br J Pharmacol 2008;155:308–315.

9. Lebrec D, Thabut D, Oberti F, et al. Pentoxifylline does notdecrease short-term mortality but does reduce complications inpatients with advanced cirrhosis. Gastroenterology 2010;138:1755–1762.


10. Sanyal AJ, Brunt EM, Kleiner DE, et al. Endpoints and clinicaltrial design for nonalcoholic steatohepatitis. Hepatology 2011;54:344–353.

11. Belfort R, Harrison SA, Brown K, et al. A placebo-controlled trialof pioglitazone in subjects with nonalcoholic steatohepatitis.N Engl J Med 2006;355:2297–2307.

12. Ratziu V, Sheikh MY, Sanyal AJ, et al. A phase 2, randomized,double-blind, placebo-controlled study of GS-9450 in subjectswith nonalcoholic steatohepatitis. Hepatology 2012;55:419–428.

13. Ratziu V, de Ledinghen V, Oberti F, et al. A randomizedcontrolled trial of high-dose ursodeoxycholic acid for nonalco-holic steatohepatitis. J Hepatol 2011;54:1011–1019.

14. Ratziu V, Giral P, Jacqueminet S, et al. Rosiglitazone fornonalcoholic steatohepatitis: one-year results of the randomizedplacebo-controlled Fatty Liver Improvement with RosiglitazoneTherapy (FLIRT) trial. Gastroenterology 2008;135:100–110.

15. Zein CO, Yerian LM, Gogate P, et al. Pentoxifylline improvesnonalcoholic steatohepatitis: a randomized placebo-controlledtrial. Hepatology 2011;54:1610–1619.

16. Timmer W, Leclerc V, Birraux G, et al. The new phosphodies-terase 4 inhibitor roflumilast is efficacious in exercise-inducedasthma and leads to suppression of LPS-stimulated TNF-alpha ex vivo. J Clin Pharmacol 2002;42:297–303.

17. Lin HZ, Yang SQ, Chuckaree C, et al. Metformin reverses fattyliver disease in obese, leptin-deficient mice. Nat Med 2000;6:998–1003.

18. Musso G, Gambino R, Cassader M, et al. A meta-analysis ofrandomized trials for the treatment of nonalcoholic fatty liverdisease. Hepatology 2010;52:79–104.

19. Galli A, Crabb DW, Ceni E, et al. Antidiabetic thiazolidinedionesinhibit collagen synthesis and hepatic stellate cell activationin vivo and in vitro. Gastroenterology 2002;122:1924–1940.

Reprint requestsAddress requests for reprints to: Vlad Ratziu, MD, PhD, Hôpital Pitié-Salpê-trière, 47-83 Boulevard de l’Hôpital, Paris 75013, France. e-mail: [email protected]; fax: (33) 142161049.

AcknowledgmentsThe authors are indebted to the following investigators who contributed to theASP9831 trial: Professor C. Beglinger, Dr T. Berg, Dr N. Forestier, ProfessorP. Galle, Professor L. Georghe, Professor G. Gerken, Dr J. Henrion, ProfessorY. Horsmans, Dr L. Husova, Dr C. Ionescu, Professor P. Michielsen, DrD. Mirela, Dr B. Muellhaupt, Dr P. Newsome, Professor F. Oberti, Dr I. Pirvu-lescu, Dr V. Serfati, Professor I. Sporea, Professor C. Stanciu, Dr S. Stewart, DrP. Urbanek, Professor H. Van Vlierberghe, and Professor S. Zeuzem. The au-thors also would like to thank D. Arnet, M. Beer, S. Bromley, M. Crabos,H. Darlington, I. Devogel, A. Dirix, I. Goergen, E. Groeneveld, M. Guilbert,S. Harris, A. Hasting, K. Lamriben, E. Lavenne, P. Machacek, A. Morshuis,K. Pechandova, B. Schmidt, B. Tanevska, S. van Os, J. Veeneman, N. Verbist,R. Wentz, and E. Zhalovaga from Astellas Pharma Europe; and L. Lungu fromFarmaceutica Remedia for their contribution to the study.

Conflicts of interestThese authors disclose the following: Vlad Ratziu is a consultant for Astellas,Genfit, Gilead, Enterome, GalMed, Phenex, and Roche-Genentech; SvenFrancque is consultant for Genfit, Gilead, Janssen, and Actelion; GuruprasadAithal is a consultant for Basilea Pharmaceutica, Aegerion Pharmaceuticals,Abbott, and Dr Falk Pharma; Pavel Trunecka is a consultant for Astellas;Stephen Ryder is a consultant for Astellas; Christopher Day is a consultant forAbbott and Astellas; Jun Takeda is an employee of and owns stock in Astellas;and Klaudia Traudtner is an employee of Astellas. The remaining authorsdisclose no conflicts.

FundingAll clinical trials were sponsored by Astellas Pharma Europe BV. The authorsreceived editorial and writing assistance from Excerpta Medica, throughfinancial support by Astellas Pharma Europe.

f � 19 May 2014 � 8:09 pm � ce

http://www.cghjournal.org


mailto:[email protected]

mailto:[email protected]

Supplementary Materials and Methods

Phase I

Healthy volunteers. This double-blind, randomized,placebo-controlled study, conducted at a single site inThe Netherlands, evaluated single and multipleascending doses of ASP9831 and matching placebo insequential groups. Healthy younger and older volunteers(aged 18‒55 y and 65‒80 y, respectively) with a bodymass index of 18 to 30 kg/m2 were eligible to partici-pate. Primary outcome variables were frequency andseverity of AEs (including safety laboratory tests), Cmax,and AUC from time zero extrapolated to infinity for theopen-label, food-effect part.

Secondary variables in the single-dose study werepharmacokinetic parameters measured in plasma (AUCfrom time zero extrapolated to infinity, AUC from timezero to the last quantifiable sample, Cmax, tmax, terminalelimination half-life, apparent volume of distribution,and apparent total body clearance) and in urine (amountexcreted in urine unchanged from time zero to the lastquantifiable sample, amount excreted unchanged in urinefrom time zero extrapolated to infinity, percentage ofamount excreted in urine unchanged from time zero to thelast quantifiable sample, percentage of amount excretedunchanged in urine from time zero extrapolated to infinity,and renal clearance). Secondary variables in the multiple-dose study were the same pharmacokinetic parameters asin the single-dose study but were measured after the firstdose (day 1) and the last dose (evening of day 9/morningof day 10), including AUCtau. Ex vivo LPS-induced TNF-aproductionwas assessed inperipheral bloodmononuclearcells as a biomarker for inflammation after single andmultiple doses. It was analyzed in relation to baseline andexpressed as a percentage of the baseline value. Safetydata were collected throughout the study.

Patients with nonalcoholic steatohepatitis. This 2-part, open-label study evaluated ASP9831 in patientswith NASH to assess whether fibrotic liver injury inNASH significantly modifies the pharmacokinetic pa-rameters, necessitating dose adjustment (SupplementaryFigure 1A). Adult outpatients (age, �18 y) with histo-logically confirmed NASH on a biopsy performed within1 year before first study dose and with increased serumALT levels on 2 or more occasions within 1 year beforescreening and less than 300 U/L at screening wereincluded. The histologic diagnosis of NASH was estab-lished by a central pathologist (P.B.) based on the asso-ciation of steatosis (>10%), hepatocyte ballooning, andlobular inflammation.10 Outcome variables were as fol-lows: AUCtau, Cmax, tmax, terminal elimination half-life,apparent total body clearance, apparent volume of dis-tribution, and trough plasma concentration (Ctrough)based on plasma sampling on day 8. The study was

conducted at 2 study sites in France and 1 site in theUnited Kingdom.

In the first part of this study, 50 mg of ASP9831 wasadministered twice daily to patients with NASH andadvanced fibrosis (ie, bridging fibrosis, Kleiner stage 3).If the mean area under the curve from time zero to 24 hafter administration of the drug (AUC0-24h) in patientswith NASH receiving 50 mg of ASP9831 twice daily didnot exceed the mean AUC0-24h in healthy volunteersreceiving 100 mg of ASP9831 twice daily, the secondpart of this study was initiated. In this second part, 100mg ASP9831 twice daily was administered to patientswith noncirrhotic NASH. If mean AUC0–24h after 100 mgASP9831 twice daily did not exceed the mean AUC0–24hin healthy volunteers receiving the same dose, the ran-domized part of the phase II study was initiated.

Phase II

This randomized, double-blind, parallel-group,placebo-controlled study investigated the effect of 12weeks of administration of 50 and 100 mg ASP9831twice daily vs placebo on serum ALT, safety, and variousliver injury biomarkers in NASH. It was conducted at 22sites across 7 countries (Belgium, Czech Republic,France, Germany, Romania, Switzerland, and the UnitedKingdom) (Supplementary Figure 1B).

Adult outpatients (age, �18 y) with noncirrhoticNASH documented by a liver biopsy performed within 1year before randomization were eligible. The histologicdiagnosis of NASH was established as described earlier.Increased serum ALT level (�1.5 � upper limit ofnormal: 41 U/L for men and 31 U/L for women, but�300 U/L) was required at screening in a central lab-oratory (BARC nv, Industriepark, Ghent, Belgium), withhistorical local serum ALT levels greater than 1.5 �upper limit of normal or greater than 60 U/L on 1 ormore occasions in the prior 6 months and no docu-mented normal value in the previous year. Primaryexclusion criteria were as follows: hepatic cirrhosis;uncontrolled type 2 diabetes (hemoglobin A1C level,>8.5%); liver disease other than NASH; excessivealcohol use within 5 years of screening; current averagealcohol intake of more than 20 g/day for women ormore than 30 g/day for men; and weight change morethan 5% in the prior 6 months. The randomization wasperformed by fixed permutation blocks and patientswere allocated to 1 of 3 treatment groups in a 1:1:1ratio. The appearance and packaging of the ASP9831tablets and placebo were identical to ensure that theparticipants were blinded to the received treatment; inaddition, neither the investigator, nor the pharmacist,nor the sponsor was aware of the treatment group. Therandomization codes were created by an external orga-nization and had been concealed until the end of study.


7.e1 Ratziu et al Clinical Gastroenterology and Hepatology Vol. -, No. -

The primary efficacy variable was the percentagechange in serum ALT level after 12 weeks of treatmentvs baseline. Because this was a proof-of-concept study toevaluate the efficacy of ASP9831 in NASH, a biochemicalrather than histologic end point was chosen—if negative,this study was not expected to be followed by subse-quent, larger trials with histologic end points. Thisapproach has been recommended by experts.10 Second-ary variables were as follows: the proportion of patientswith normal ALT level at 12 weeks; and changes frombaseline in AST, AST/ALT ratio, biomarkers of liverfibrosis (FibroTest; Biopredictive S.A., Paris, France),steatosis (SteatoTest; Biopredictive S.A.), inflammation(TNF-a), apoptosis (cytokeratin 18), and other bio-markers (NASHTest and adiponectin; Biopredictive S.A.).Other serum measurements included inflammatory cy-tokines/markers such as interleukin-1b, interleukin-2R,interleukin-6, interleukin-10, C-reactive protein, andfibrosis markers (eg, hyaluronic acid, type III procollagenN-peptide, transforming growth factor b, type IVcollagen 7S domain). Plasma drug concentration wassparsely measured. Safety was recorded as the frequencyof TEAEs. To safeguard the blinding of the treatment andavoid any potential biases, assessment of all biomarkerswas performed by a central laboratory; the participatingsites received specific instructions on sampling, handling,shipping, and storage. The results of the analysis of bio-markers for inflammation, fibrosis, steatosis/NASH, andapoptosis, as well as insulin resistance, were not com-municated to the investigators during the study, unlessthe biomarker was part of the safety assessment.

The safety population comprised patients who took 1or more tablets of study medication. The full-analysis pop-ulation comprised patients who took 1 or more tabletsof study medication and had 1 or more postbaseline ALTvalues recorded. The per-protocol population comprisedpatients in the full-analysis set who adhered to protocol;principal reasons for exclusion were use of a prohibitedmedication and treatment duration shorter than 9 weeks.

Ethics

Clinical study protocols conformed to the ethicalguidelines of the 1975 Declaration of Helsinki.Studies were conducted in accordance with GoodClinical Practices. Each subject signed an institu-tional review board–approved written informedconsent form.

Statistical Analysis

Outcomes in the phase I study were reported usingdescriptive statistics (mean � SD; coefficient of varia-tion; minimum, median, and maximum; frequencies; andpercentages). Calculations included evaluation of dose-proportionality for AUC and Cmax of ASP9831 inplasma (last dose vs first dose in the pharmacokineticstudy), as well as the relationship between ASP9831doses and ex vivo TNF-a.

The primary analysis in the phase II study was themean percentage change of serum ALT level vs baseline.Analysis of covariance was used to analyze percentagechange from baseline in serum ALT level at 12 weekswith the last observation carried forward (a ¼ .05, 2-sided). Treatment group, ALT level, type 2 diabetes, andbody mass index (<30 or �30 kg/m2) at baseline wereincluded as fixed effects, and site was included as arandom effect. Significance of pairwise comparisons canbe stated formally only when the overall comparisonwas significant to prevent a-error inflation. The primaryanalysis was performed on the full-analysis population.The secondary analysis summarized biomarkers usingdescriptive statistics. A sample size of 31 subjects in eacharm was based on comparison of 2 means; delta, SD,power, and drop-out rate were 20%, 25%, 80%, and20%, respectively.

All authors had access to the study data and reviewedand approved the final manuscript.


- 2014 Pharmacologic PDE4 Inhibition in Human NASH 7.e2

Supplementary Table 1. Baseline Characteristics of Healthy Volunteers and Patients With NASH Enrolled in Phase I andPhase II

Phase I Phase II

Healthy volunteers NASH patients NASH patients

Single increasingdoses

Multiple increasingdoses ASP9831 ASP9831

ASP9831(n ¼ 102)

Placebo(n ¼ 34)

ASP9831(n ¼ 55)

Placebo(n ¼ 28)

50 mg(n ¼ 7)

100 mg(n ¼ 7)

Placebo(n ¼ 30)

50 mg(n ¼ 33)

100 mg(n ¼ 33)

Men, n (%) 72 (70.6) 24 (70.6) 32 (58.2) 16 (57.1) 6 (85.7) 5 (71.4) 22 (73.3) 24 (72.7) 24 (72.7)Mean age � SD, y 36.3 � 20.2 36.7 � 20.4 39.8 � 20.6 48.3 � 18.1 53.9 � 11.9 50.9 � 9.8 43.1 � 13.6 45.9 � 12.7 46.2 � 11.5Mean BMI � SD,

kg/m223.8 � 2.8 25.1 � 2.7 24.6 � 2.6 24.5 � 2.2 32.3 � 3.4 30.8 � 3.7 31.8 � 5.6 31.2 � 4.9 29.5 � 5.2

Fibrosis stagegroup, n (%)

NA NA NA NA

No fibrosis 0 0 4 (13.3) 1 (3.0) 6 (18.2)Stage 1 0 2 (28.6) 12 (40.0) 9 (27.3) 12 (36.4)Stage 2 0 2 (28.6) 7 (23.3) 16 (48.5) 8 (24.2)Stage 3 7 (100) 3 (42.9) 7 (23.3) 7 (21.2) 7 (21.2)Stage 4 0 0 0 0 0

BMI, body mass index; NA, not applicable.

Supplementary Table 2. Pharmacokinetics of ASP9831 in Patients With NASH in Phase I With Noncompartment Analysis

Parameters Statistics

ASP9831

50 mg 100 mg

tmax, h n 7 7Mean (SD, %CV) 1.26 (1.23, 98) 1.71 (1.58, 92)

Cmax, ng/mL n 7 7Mean (SD, %CV) 1465 (503, 34) 3320 (1098, 33)

AUCtau, h*ng/mL n 7 7Mean (SD, %CV) 4622 (1831, 40) 10825 (4105, 38)

t1/2, h n 7 5Mean (SD, %CV) 6.47 (1.71, 26) 7.01 (1.77, 25)

CLSS/F, L/h n 7 7Mean (SD, %CV) 12.2 (4.25, 35) 10.3 (3.52, 34)

Vz/F, L n 7 5Mean (SD, %CV) 113 (47.1, 42) 114 (47.2, 42)

Ctrough, ng/mL n 7 7Mean (SD, %CV) 97.1 (56.7, 58) 197 (122, 62)

CLSS/F, apparent body clearance after extravascular dosing; Ctrough, trough plasma concentration; %CV, coefficient of variation; t1/2, apparent terminal eliminationhalf-life; Vz/F, apparent volume of distribution.



SupplementaryFigure 1. (A) Phase I and(B) phase II study designsevaluating ASP9831 in pa-tients with NASH. BID,twice daily.

Supplementary Figure 2. Inhibition of TNF-a by ASP9831.Young healthy men received single oral doses of ASP0831(5–400 mg; n ¼ 6 in each dose group) or placebo (n ¼ 20).TNF-a plasma concentration was measured at baseline andup to 24 hours after dose intake.


- 2014 Pharmacologic PDE4 Inhibition in Human NASH 7.e4

SupplementaryFigure 3.Mean � stan-dard deviation values ofliver injury biomarkers: (A)NASHTest score, (B)fibrosis, and (C) steatosisin each treatment group.LOCF, last observationcarried forward.



Lack of Efficacy of an Inhibitor of PDE4 in Phase 1 and 2 Trials of Patients With Nonalcoholic...

Documents

Transcript of Lack of Efficacy of an Inhibitor of PDE4 in Phase 1 and 2 Trials of Patients With Nonalcoholic...