Setanaxib, a first‐in‐class selective NADPH oxidase 1/4 inhibitor for primary biliary cholangitis: A randomized, placebo‐controlled, phase 2 trial

Primary biliary cholangitis (PBC) is a rare liver disease with significant unmet need for second‐line/add‐on treatments. Setanaxib, a NOX1/4 inhibitor, has shown anti‐fibrotic effects in in vitro and animal studies. This phase 2, randomized, multicentre study investigated the efficacy and safety of setanaxib in patients with PBC.


| INTRODUC TI ON
Primary biliary cholangitis (PBC) is a chronic, cholestatic liver disease caused by progressive autoimmune destruction of bile ducts within the liver and the accumulation of bile acids. 1,2 Diagnosis of PBC typically relies on serum tests of liver biochemical parameters, including alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), aspartate transaminase (AST), and bilirubin, which are all typically elevated in PBC. 1 Alanine transaminase (ALT) levels may also be elevated due to inflammation. 3,4 Advanced disease progression leads to the development of liver fibrosis and eventually cirrhosis, which can be assessed through liver biopsy, or non-invasively using transient elastography. 1,4,5 Ursodeoxycholic acid (UDCA) is the current standard of care for patients with PBC. 4 However, up to 40% of patients with PBC have an incomplete response to UDCA and may benefit from add-on or second-line therapies. 6 Biochemical response to UDCA is defined by reduced serum levels of ALP, GGT or bilirubin. 7 Elevated ALP and GGT serum levels have been shown to predict patient outcomes, including liver transplantation and death, 8,9 so incomplete response to UDCA treatment can indicate worsened disease outcomes. 9 High serum GGT levels (≥3.2 × upper level of normal [ULN]) also indicate a need for treatment escalation in patients with low levels of serum ALP (<1.5 × ULN). 9 To date, there is only one second-line treatment (obeticholic acid [OCA]) approved for the treatment of PBC in UDCA incomplete responders, used in combination with UDCA. 6 In both randomized controlled clinical trials and real-world studies, OCA has shown efficacy in lowering ALP and bilirubin in ~50% of patients (according to POISE study response criteria [ALP <1.67 × ULN]). 10,11 Although there is evidence that UDCA may delay histological progression in some patients with PBC, only preliminary data exist to support OCA having an anti-fibrotic effect. 12,13 Fibrates, including bezafibrate, may also be used as an add-on treatment in patients with incomplete response or intolerance to UDCA and have been shown to reduce serum ALP levels over 12 months. 14 However, at this time, bezafibrate is currently unlicensed for use in PBC. 15 Given that liver fibrosis is a major driver of clinical outcomes in liver disease, and strong correlations between advanced fibrosis and liver stiffness have been previously observed, 16 improving or stabilizing liver stiffness is an important treatment target and endpoint.
UDCA has been shown to stabilize or slow liver fibrosis in a subset of patients, i.e., those with a remarkable improvement in ALP, but on the other hand, many patients with incomplete response to UDCA remain at risk of progression to cirrhosis and liver failure, which can ultimately result in liver transplantation or death. 17,18 The most commonly reported symptoms of PBC are fatigue and pruritus (itching), which can be debilitating for patients and impair quality of life (QoL). 1 While treatments such as bezafibrate have shown beneficial effects on pruritus, neither UDCA, OCA nor bezafibrate have shown efficacy in significantly reducing fatigue, despite this symptom affecting >50% of patients. [19][20][21] Setanaxib, a selective inhibitor of the NADPH oxidase (NOX) 1 and 4 isoforms, has demonstrated potential to downregulate multiple fibrogenic and inflammatory pathways and prevent progression to liver fibrosis in in vitro and animal studies. 22,23 The mechanism of action of setanaxib involves both anti-fibrotic and anti-inflammatory effects at the early stages of pathophysiology of some liver disorders, supported by pre-clinical data. 24 These properties make setanaxib a plausible agent to evaluate. In this phase 2 clinical trial, the efficacy and safety of oral setanaxib (400 mg once [OD] or twice [BID] daily), alongside UDCA, were assessed in patients with PBC. PBC-40 fatigue domain scores to Week 24 were +0.3% (24.9%) for setanaxib 400 mg OD, −9.9% (19.8%) for setanaxib 400 mg BID and +2.4% (23.1%) for placebo, p = .027.
Two patients (one placebo, one setanaxib 400 mg BID) experienced serious treatmentemergent adverse events, deemed unrelated to study drug.

Conclusions:
The primary endpoint was not met. However, the secondary endpoints provide preliminary evidence for potential anti-cholestatic and anti-fibrotic effects in PBC, supporting the further evaluation of setanaxib in a future phase 2b/3 trial.

| Randomization and masking
Eligible patients were randomized 1:1:1 to setanaxib 400 mg OD, setanaxib 400 mg BID or placebo, alongside UDCA, for 24 weeks ( Figure 1). Patients were stratified at study entry by disease severity (defined as baseline serum GGT <2.5 × ULN or ≥2.5 × ULN). To ensure balanced treatment groups, an interactive web-based randomization system was used to assign a unique randomization number in ascending, sequential order (with associated treatment arm) to each patient, based on pre-determined blocks of randomization numbers for each stratification level. The investigator was responsible for enrolling patients and assigning a randomized number (entered in their electronic case report form).
The sponsor, patients, investigators, the investigator's staff, persons performing assessments, data reviewers, and statisticians all remained blinded to the identity of the study treatments, concealed using investigational medicinal products which were identical F I G U R E 1 Study design. Liver stiffness assessments were performed at baseline and Week 24 using vibration-controlled transient elastography at study sites with available equipment. PBC-40 questionnaires were self-completed at baseline and Weeks 12 and 24 by patients. All patients received setanaxib 400 mg BID or OD, or placebo, in addition to UDCA. ALP, alkaline phosphatase; BID: twice daily; GGT, gamma-glutamyl transferase; OD, once daily; UDCA, ursodeoxycholic acid. in packaging, labelling, administration schedule, appearance, and odour. Patients self-administered twice daily: four active capsules and four placebo capsules for the setanaxib 400 mg OD group and eight active or eight placebo capsules for the setanaxib 400 mg BID and placebo groups, respectively. Any unblinding procedures were reported by a principal investigator in the electronic case report system; none were reported. Randomization data were kept strictly confidential and were only accessible to authorized personnel until the unblinding of the trial.

| Procedures
Patients self-administered setanaxib 400 mg OD or BID or placebo orally for a total of 24 weeks. Baseline assessments were performed on Day 1; patients were then assessed on Weeks 2, 6, 12, 18, and 24.
An interim analysis was conducted when 80%-90% of the planned number of randomized patients had completed Week 6; these analyses did not impact future study conduct and the results were not disclosed to trial investigators. Patients were followed up to 28 days after the end of treatment. If treatment was discontinued early, an early termination visit was scheduled, ensuring patients had six post-

| Outcomes
The primary efficacy endpoint, assessed by a central laboratory, was percentage change from baseline to Week 24 in serum GGT. Serum GGT was selected as the primary endpoint for this study due to its use as a marker of inflammation and oxidative stress, and the proposed anti-inflammatory and anti-fibrotic mechanism of setanaxib.
We also report the following main secondary efficacy endpoints: percentage change in serum ALP, high-sensitivity C-reactive protein (hsCRP), fibrinogen, ELF, FIB-4 and APRI score, total and conjugated bilirubin, liver stiffness, serum levels of collagen fragments, and percentage change in fatigue score (based on the PBC-40 questionnaire) to Week 24. 24,25 In addition, changes in serum GGT to Week 6 are also summarized. A full list of all study outcomes, including all 17 secondary endpoints, is available in Appendix 3. To measure fatigue, patients answered 11 fatigue-based questions on the PBC-40 questionnaire using a 5-point Likert scale (highest score = highest impact) and points were summed to provide a total score. 25 The PBC-40 questionnaire was chosen to measure fatigue because it represents a patient-derived, disease-specific method of tracking fatigue symptoms, relevant to patients from varying geographical areas. 25 Safety outcomes were monitored from baseline to the end of the follow-up period (28 days after last treatment administration).
Adverse events (AEs) were coded using MedDRA™ version 21.1 and graded according to the common terminology criteria for AEs (CTCAE) Version 4. We report treatment-emergent AEs (TEAEs; defined as any AE that started on/after the date of first treatment dose), related TEAEs, serious TEAEs, severe TEAEs (defined as any AE grade 3 or above using the CTCAE), TEAEs leading to discontinuation or interruption of study treatment, and AEs leading to death.

| Statistical analysis
A sample size of 34 patients within each treatment group (a total of 102 patients) was required to have 80% power to detect a 28% difference in mean percentage changes from baseline in serum GGT between treatment groups. A Wilcoxon Mann-Whitney test was used for this sample size estimate; this was the most conservative approach for statistical power. Standard deviations (SD) of 30 for the setanaxib groups and 40 for the placebo group were assumed based on a recent phase 3 clinical trial of OCA. 26 An overall two-sided Type I error of 5% was assumed and the Hochberg method was used to adjust the alpha level for the two-dose comparisons of percentage change in serum GGT and ALP at Week 24; setanaxib 400 mg BID and setanaxib 400 mg OD dose levels were tested against alpha levels of 0.047 and 0.023, respectively. This step-down procedure, after accounting for correlation, ensured that the overall alpha level was no more than 5%. No further multiplicity adjustments were made for the comparison of other efficacy endpoints. Demographic data were compared for setanaxib 400 mg OD and setanaxib 400 mg BID versus placebo using an Analysis of Variance (ANOVA) or a Mann-Whitney test for continuous variables as appropriate, and Fisher's exact test for categorical variables.
The primary endpoint was analysed using the intention-to-treat population (all randomized patients). The percentage change in serum GGT from baseline to Week 24 was analysed using an Analysis of Covariance (ANCOVA) with treatment and disease severity as fixed effects, and baseline GGT as a continuous covariate. The estimated difference between the setanaxib 400 mg BID group and placebo was calculated, along with 95% confidence intervals (CI). Reported p values compare setanaxib 400 mg BID and placebo groups. In the event of missing data at Week 24, the last observation carried forward (LOCF) method was applied, by imputing Week 24 visit data with the last non-missing post-baseline data available from Week 12 onwards, where Week 12 data were collected. A secondary analysis of GGT including all post-baseline visits over the 24-week treatment period (assessments at Weeks 2, 6, 12, 18, and 24) was performed using a repeated measures ANOVA with treatment, visit and disease severity as fixed effects and baseline GGT as a continuous covariate plus an interaction term for treatment and visit. With the inclusion of longitudinal data from the same subjects within each arm, variability is reduced and statistical power is increased when compared with an ANCOVA that compares data only at Week 24.
The secondary endpoint of percentage change in serum ALP was analysed using the same statistical methods as the primary endpoint.
No significant tests were performed for other secondary liver disease outcomes. For change in fatigue score from baseline to Week 24, treatment differences and associated 95% CI are presented with p values to compare setanaxib 400 mg BID and placebo treatment groups. Data were summarized descriptively. For pre-planned subgroup analyses, patients were grouped at baseline according to a pre-defined liver stiffness cut-off value of 9.6 kilopascals (kPa), as recommended by the European Association for the Study of the Liver (EASL) Clinical Practice Guidelines. 4 These cut-offs were chosen due to a previous study reporting that patients with liver stiffness >9.6 kPa have a five-fold increased risk of PBC-related complications, indicating a higher risk patient group with greater unmet need. 17 Post-hoc analyses were conducted to explore the effect of treatment in subgroups according to baseline liver stiffness (<9.6 kPa and ≥9.6 kPa). The safety population included all patients who received ≥1 dose of treatment and had ≥1 safety assessment.
Safety data were periodically reviewed by the Safety Monitoring Committee. All analyses were performed using SAS Version 9.4.  Table 2). Three further patients discontinued due to change in UDCA dose (n = 1), withdrawal of consent (n = 1) or administrative reasons (n = 1; Figure S1). The study ended after the last patient attended their Week 28 follow-up visit.

| Baseline characteristics
Baseline characteristics were similar between treatment groups, although patients randomized to setanaxib 400 mg BID had higher serum ALP (p = .046) and ALT (p = .042) levels than placebo (Table 1).
Those randomized to setanaxib (400 mg OD and BID) had been diagnosed approximately 5 years prior to the placebo group, though the difference in time since first PBC diagnosis was only significant for patients randomized to setanaxib 400 mg OD (p = .049). Patients randomized to setanaxib 400 mg OD had a higher median baseline UDCA dose (mg/kg/day) than placebo (p = .017), although there was no significant difference between patients randomized to setanaxib 400 mg BID versus placebo (p = .085). No significant differences were found between the setanaxib 400 mg OD and setanaxib 400 mg BID treatment groups compared with placebo for other baseline characteristics in Table 1. For all treatment groups, patients were predominantly female (n = 99; 89.2%) and mean age was 56.2 years, in line with reported demographic characteristics of patients with PBC. 5 Median baseline UDCA doses were 14.6, 14.9, and 13.4 mg/kg/day for setanaxib 400 mg OD, setanaxib 400 mg BID and placebo, respectively (Table 1; doses ranged from 3-68 mg/kg/day). Using vibrationcontrolled transient elastography, 91 patients (82.0%) had valid liver stiffness measurements at baseline and Week 24. Patients in the placebo group had higher median baseline liver stiffness than in the setanaxib 400 mg OD and BID groups (8.9 kPa versus 8.0 kPa, and 7.5 kPa, respectively); however, the differences were not statistically significant. Mean fatigue scores at baseline were generally comparable across treatment groups. Baseline characteristics for liver stiffness subgroups (<9.6/≥9.6 kPa) are available in Table S1.

| Changes in liver biochemistry
Patients receiving setanaxib 400 mg BID showed a mean (SD) [16.0%]); the difference between setanaxib 400 mg BID and placebo was −9.0% (95% CI -17.9% to 0.0%; Figure 2B). At Week 24, the difference between setanaxib 400 mg BID and placebo (p = .049; significance level 0.047) did not reach statistical significance after adjustment for multiple comparisons. However, over the full 24-week treatment period (repeated measures analysis), the difference between setanaxib 400 mg BID and placebo was statistically significant (p = .002). For both serum GGT and ALP, the difference in mean percentage change from baseline between the setanaxib 400 mg BID and placebo treatment groups was most pronounced at Week 6. Median absolute change from baseline data for GGT and ALP is reported in Table S2. Week 24 in fibrinogen, FIB-4, APRI and ELF scores, serum AST and ALT and conjugated and total bilirubin are shown in Table S3. No p values were calculated for these secondary outcomes.

| Changes in liver stiffness and markers of fibrosis
In total, 91 patients had available liver stiffness readings at baseline and Week 24. Of these patients, those treated with setanaxib 400 mg OD or BID experienced +0.1 kPa and −0.4 kPa absolute changes in median liver stiffness, respectively, compared to +0.4 kPa for placebo. Median absolute liver stiffness at baseline and Week 24 for each treatment group is presented in Figure S2A.

| Changes in fatigue scores
Baseline fatigue scores for the three treatment groups were comparable (Table 1)

| DISCUSS ION
This phase 2 study was the first trial to investigate the effect of the first-in-class selective NOX 1/4 inhibitor, setanaxib (400 mg OD or BID), on markers of cholestasis and liver fibrosis in patients with PBC.
Our results, particularly those from secondary endpoints, provide preliminary evidence for the potential anti-cholestatic, anti-inflammatory and, anti-fibrotic effects of setanaxib (400 mg OD or BID, used alongside the standard of care UDCA), as well as the reduction of fatigue.
Although the primary endpoint was not met, patients treated with setanaxib 400 mg BID showed a larger reduction in serum GGT after 24 weeks of treatment than patients receiving placebo. The mean results for serum GGT levels in patients treated with setanaxib F I G U R E 3 Mean percentage change in serum GGT and ALP between baseline and Week 24, stratified by baseline liver stiffness.
Intention-to-treat population. n numbers given underneath figure represent patients included in the analysis. Liver stiffness assessments were performed at baseline and Week 24 using vibrationcontrolled transient elastography at study sites with available equipment. In total, 16 patients were not included in the subgroup analysis either due to missing liver stiffness data at baseline (n = 7) or missing serum GGT and ALP data at Week 24 (n = 9 In this study, patients with elevated liver stiffness (≥9.6 kPa) treated with setanaxib 400 mg BID experienced a median absolute reduction of −3.0 kPa liver stiffness at Week 24, possibly suggesting the potential for reduced fibrosis when treated with higher doses of setanaxib in this patient group. This patient population is at higher risk of worse long-term disease outcomes than patients with lower baseline liver stiffness, thus representing patients with greater unmet need. 17,27 Although a recent open-label study of bezafibrate plus UDCA showed that 48% of patients achieved fibrosis regression after five years' treatment, to date no large-scale, randomized controlled trials have been conducted to validate this outcome. 28,29 Elafibranor and seladelpar have shown potential for reducing mark- therapies were at the early stages of investigation and not currently approved for treatment.
It should be noted that liver stiffness can also be affected by pathological processes such as inflammation and oedema, and setanaxib is also known to have anti-inflammatory properties. 24 Collagen fragment serum levels were assessed to further investigate the potential anti-fibrotic effect of setanaxib 400 mg OD or BID. 31 It was previously reported that cirrhotic livers contain approximately four-fold as much type III collagen as healthy livers (8 mg/g versus 2 mg/g wet weight in a healthy liver). 32 Our results showed reduced markers of types III and V collagen formation (PRO-C3 and PRO-C5) after 24 weeks across all three treatment groups, but this was more pronounced in the setanaxib groups versus placebo (no p value calculated). When analysed by liver stiffness subgroups, the greatest median reduction in PRO-C3 was observed in patients with liver stiffness ≥9.6 kPa at baseline receiving setanaxib 400 mg BID, which was accompanied by an increase in serum levels of C3M (no p value calculated). Given that these patients also had advanced fibrogenesis at baseline, as indicated by median baseline PRO-C3 levels of >20 ng/mL, this finding is in line with previous results from a phase 2 clinical trial in primary sclerosing cholangitis. 33 Together, these results may support the theory that in patients with clinically significant fibrogenesis at baseline, the improvements in liver stiffness observed after setanaxib 400 mg BID treatment may be partly due to fibrosis reduction, either with or without additional reduction of inflammation. However, due to small patient numbers in this subgroup (n = 14), these results should be interpreted with caution and further research is warranted to investigate the potential antifibrotic effect of setanaxib.
Chronic fatigue is one of the most common symptoms associated with PBC and is highly prioritized by patients as a treatment target, due to its debilitating effects on QoL and limited treatment options. 19 Our results suggest that setanaxib 400 mg BID can produce clinically significant improvements (defined as a mean reduc- The impact of our findings is limited by the study being insufficiently powered for the analysis of secondary efficacy outcomes.
Throughout treatment, patients received different stable concomitant doses of UDCA and had differing lengths of exposure; this could have affected study outcomes. The use of vibration-controlled transient elastography to assess liver stiffness provided a less invasive alternative compared to liver biopsy, but it is not without limitations. 37,38 Although one large multicentre study concluded that Fibroscan® provides an accurate assessment of liver fibrosis, 37 it is also accepted that liver stiffness measurements via Fibroscan® can be affected by inflammation, venous pressure, and cholestasis. 38 Thus, these early findings should be interpreted with caution.
Further studies will be beneficial to fully assess the benefits of setanaxib treatment (alongside UDCA) on liver fibrosis and fatigue, as well as overall control of the disease, in patients with PBC.
This study enrolled 111 patients with PBC across nine countries, making it one of the largest and longest phase 2 trials con- Setanaxib was shown to be well-tolerated at both doses (400 mg OD and BID). These results support the further evaluation of setanaxib in a phase 2b/3 trial in patients with PBC and elevated liver stiffness.

DATA AVA I L A B I L I T Y S TAT E M E N T
Study data, including individual participant data, will not be made available to others after publication. The study protocol is available at https://clini caltr ials.gov/ct2/show/study/ NCT03 226067. No additional related documents will be made available.

E TH I C A L A PPROVA L S TATEM ENT
This study was performed in accordance with the provisions of the Declaration of Helsinki, and all revisions thereof, and in accordance with USA Food and Drug Administration (FDA) regulations.
Ethics approval was granted by the relevant central and regional ethics committees. The trial was conducted in agreement with the International Conference on Harmonization (ICH) Guidelines on Good Clinical Practice (GCP).

PATI ENT CO N S ENT S TATEM ENT
All patients provided written informed consent to participate in the study prior to screening.

PE R M I SS I O N TO R E PRO D U CE M ATE R I A L FRO M OTH E R S O U RCE S
This is an open-access article which permits use, distribution, and reproduction in any medium, provided the original work is properly cited. and other agencies maintain current lists, which can be referred to via their websites. Furthermore, these lists are anticipated to evolve as new drugs come to market and more is learned about the pharmacology of GKT137831 and other medications. Therefore, they are to be regarded as a minimum set of excluded and pre-cautioned concomitant medications.

O RCI D
A PPEN D I X 3

Study outcomes
Primary efficacy endpoint: • Percentage change from baseline to Week 24 in serum GGT.
Secondary efficacy endpoints: • Absolute and percentage change in serum GGT from baseline to each assessment.
• Absolute change in ELF score from baseline to Weeks 12 and 24.
• Absolute and percentage change in serum ALP from baseline to each assessment.
• Absolute and percentage change in serum levels of hsCRP and fibrinogen, from baseline to each assessment.
• Absolute and percentage change in serum ALT, AST, and conjugated and total bilirubin, from baseline to each assessment.
• Absolute and percentage change in fibrosis (FIB)-4 and AST to platelet ratio index (APRI) scores, from baseline to each assessment.
• Tertiary efficacy endpoints: • Absolute and percentage change in total bile acid levels from baseline to Weeks 12 and 24.
• Proportion of subjects achieving a 15, 20, 30 and 40% reduction in serum ALP from baseline to each assessment.
• Proportion of subjects who meet the definition of PBC responder criteria applying the Paris I, Toronto I, Toronto II, Toronto III,   Toronto IV, Mayo II, and Barcelona disease prognostic risk criteria at Weeks 12 and 24.
Exploratory endpoints: • Absolute and percentage change in serum C4 and FGF19 from baseline to Weeks 12 and 24.
• Assessment of metabolomics signatures.
• Assessment of additional biomarkers of interest.