Graphical Abstract
INTRODUCTION
Primary sclerosing cholangitis (PSC) is a rare cholestatic liver disease of unknown cause, marked by chronic fibrosis and inflammation that leads to the destruction and narrowing of the bile ducts [1]. This condition increases the risk for cholangiocarcinoma, cirrhosis, and colorectal cancer, making liver transplantation the only end-stage treatment option. Symptoms of PSC vary widely, from asymptomatic to severe, underscoring the urgent need for therapeutic advances [2].
Several drugs are in development for PSC, but challenges such as lack of efficacy or severe side effects have led to the discontinuation of some, like obeticholic acid (OCA) and cilofexor, despite their initial promise as farnesoid X receptor (FXR) agonists [3]. The development of OCA for PSC is uncertain after a phase 2 study, and cilofexor was discontinued during phase 3 [4,5]. Both drugs were associated with a high incidence of pruritus in treated patients [3,6], indicating challenges for future FXR agonist-based therapies.
Treatment strategies exploring PPAR-delta or PPAR-alpha/delta targeting for reducing bile acid synthesis have shown potential. Activation of these pathways increases fibroblast growth factor 19 (FGF19) and FGF21 expression, reducing CYP7A1 and bile acid synthesis [7]. Although Seladelpar, a PPAR-delta agonist, was discontinued during a phase 2 trial for non-alcoholic steatohepatitis due to liver damage concerns [8], it showed promising results in a phase 3 trial for primary biliary cholangitis (PBC) [9]. Ursodeoxycholic acid (UDCA), used for supportive care in PSC, has shown mixed results. Low doses of UDCA improve serum alkaline phosphatase (ALP) levels but do not improve liver histology or survival rates, whereas high doses of UDCA increase the risks of adverse reactions and colorectal neoplasms [10,11], necessitating further research on its efficacy. In 2017, a phase 2 trial of norUDCA showed promising results, including a dose-dependent reduction in ALP levels and comparable adverse events to the placebo group [12]. However, phase 3 trials of norUDCA for PSC have not been published yet.
Overall, while multiple potential therapeutic avenues are being explored, none have yet been definitively successful, and liver transplantation remains the only curative option for advanced disease. Ongoing clinical trials and research are crucial for finding effective treatments for PSC.
Based on a previous in vivo study using a 3,5-diethoxycarbonyl-1,4-dihydrocollidine-induced PSC mouse model, beta-lapachone effectively inhibited key features of PSC pathogenesis. It reduced aspartate aminotransferase (AST) and ALP levels and improved pathogenic characteristics, such as inflammation (IL-1β, TNF-α, and TGF-β1) and fibrosis (Col1a2). Furthermore, beta-lapachone significantly decreased liver tissue fibrosis [13]. Additionally, beta-lapachone was shown to stimulate the expression of genes related to NAD synthesis and Abcb4, which contributed to improved mitochondrial stability and overall amelioration of PSC pathogenesis [13].
HK-660S, a synthetic beta-lapachone compound, explicitly targets the NAD(P)H quinone dehydrogenase 1 (NQO1) enzyme, enhancing intracellular NAD+ levels [14]. This elevation activates the NAD-dependent sirtuin family (SIRT1-SIRT7), which is crucial for regulating energy metabolism. Notably, SIRT1 boosts AMP-activated protein kinase by influencing liver kinase B1, serving as an energy sensor.
The effect of HK-660S on sirtuins results in anti-inflammatory and anti-fibrotic actions. Specifically, activated SIRT1 curtails inflammation by reducing the activity of key transcription factors, such as nuclear factor-kB and activator protein 1 [15], and combats liver fibrosis by inhibiting hepatic stellate cells [16]. SIRT1 reduces fibrotic and inflammatory liver damage caused by cholestatic liver injury in mice fed cholic acid [17], with SIRT6 also reported to protect against such damage by deacetylating estrogen-related receptor γ, affecting bile secretion.
MATERIALS AND METHODS
Study design and participants
The present study was designed as a multicenter, randomized, double-blind, placebo-controlled, parallel-group, phase 2a, proof-of-concept study (Supplementary Fig. 1).
The study protocol was approved by the Korean Ministry of Food and Drug Safety and institutional review boards. All patients provided written informed consent before study participation. The study was conducted according to Good Clinical Practice guidelines of the International Council for Harmonization and the Declaration of Helsinki. The study is registered with cris.nih.go.kr (KCT0006590). All authors had access to the study data and reviewed and approved the final manuscript.
PSC patients aged >17 years were diagnosed according to the following criteria: i) serum ALP levels >1.5 times the upper limit of normal (ULN) at screening; ii) patients previously diagnosed with PSC due to multifocal bile duct involvement, as determined by magnetic resonance cholangiopancreatography (MRCP) or endoscopic retrograde cholangiopancreatography, and who had baseline MRCP images within 6 months of screening. Critical exclusion criteria included chronic liver diseases other than PSC, obstacles to MRCP implementation, and any clinically significant cardiovascular diseases. Furthermore, patients with PSC unsuitable for participation in the study at the investigator’s discretion, those administered herbal medicine to improve fatty liver disease within 2 weeks of screening, alanine aminotransferase (ALT) or AST >10 times the ULN, and serum creatinine ≥2 mg/dL were excluded. Participants with chronic liver diseases other than PSC, including non-alcoholic fatty liver disease, viral chronic hepatitis, alcoholic liver disease, PBC, biliary obstruction, autoimmune hepatitis, hemoglobin deposition, Wilson’s disease, α-1 antitrypsin deficiency, thyroid diseases, autoimmune diseases, and medication history for steatohepatitis within five times the half-life of the respective drug before screening were also excluded. All medication possibly related to PSC except low-dose UDCA, azathioprine, and prednisone were prohibited.
Randomization and masking
The randomization codes for this clinical study were created using SAS® (version 9.4; SAS Institute Inc., Cary, NC, USA) by a statistician not directly related to this study, in order to assign the participants to the HK-660S or placebo group in a 2:1 ratio by stratified block randomization method using the NQO1 genotypes (T/T, C/T, C/C type) as a stratification factor. The sponsor packed the study drugs according to the randomization codes and distributed the study drugs to clinical sites prior to the start of the clinical study. Participants who met the inclusion/exclusion criteria were assigned to each treatment group in the order of enrollment according to the randomization codes using the interactive web response system.
To maintain the blinding, the placebo has the same formulation and appearance as the investigational product to make them visually indistinguishable. Furthermore, the study drugs were indistinguishable by taste, appearance, scent, and color.
Procedures
After participants voluntarily consented to participate in the clinical study, those who met the inclusion/exclusion criteria were evaluated through a 5-week screening period. HK-660S or placebo were provided as a 100 mg tablet, which was administered orally twice a day before morning and evening meals. The participants visited the study sites every 4 weeks during the treatment period. After treatment, safety follow-up was conducted for an additional 4 weeks. Study drugs were provided for 12 weeks in a double-blind manner.
Outcomes
The primary endpoints were 1) the percentage of patients who showed improvement in PSC severity, as assessed by the Anali score in MRCP at week 12, and 2) the change in serum ALP levels 12 weeks from baseline. The Anali scores were calculated as follows: Anali score=Without gadol inium: (1×Di latat ion of intrahepat ic duct)+ (2×Dysmorphy)+(1×Portal hypertension)+With gadolinium: (1×Dysmorphy)+(1×Parenchymal enhancement heterogeneity) [MRCP Anali score range: 0 to 7, intrahepatic bile duct dilatation: 0 (≤3 mm), 1 (<3 mm and <5 mm), 2 (≥5 mm), Dysmorphy, Portal hypertension: 0 (absent), 1 (present)].
The scores are graded and totaled from 0 to 7 and evaluated on a separate form (evaluation form). Improvement in the severity of PSC was defined as a decrease of 1 or lower in the MRCP Anali score. After viewing the MRCP videos for each participant, an independent blinded radiologist graded the Anali score [20]. To address the issues in MRCP evaluation, such as poor inter- and intra-observer agreement and variability in technique between exams [21], this clinical trial employed two independent radiologists at a central imaging lab who assessed MRCP videos independently. If there was a disagreement in the Anali score between two radiologists, an independent adjudicator made the final decision on the Anali score.
Secondary efficacy endpoints included: 1) the percentage of participants who showed improvement of bile duct strictures in MRCP, 2) liver elasticity values obtained from vibration controlled transient elastography (VCTE), 3) hepatic function and biochemistry tests such as ALT, AST, gamma-glutamyl transferase (γ-GT), and total bilirubin, 4) fibrosis as assessed by the enhanced liver fibrosis (ELF) score and serum C4 (7α-hydroxy-4-cholesten-3-one) levels.
Exploratory endpoints included: 1) lipid profiles such as total cholesterol, triglycerides, free fatty acid, LDL-C, and HDL-C; 2) HOMA-IR; 3) body mass index (BMI), waist circumference, and fatigue severity scale; 4) PSC-related biomarkers such as CK-18, CCL-2, CCL-5, hs-CRP, CRP, FGF21, and GDF15; 5) inflammatory cytokines including TNF-α, TGF-β, IL-6, and adiponectin, and 6) mechanism-related biomarkers (lactate, pyruvate, and lactate/pyruvate ratio). The medication adherence was assessed by patients bringing their clinical medication bottles to each regular visit, during which the pharmacist confirmed the amount of medication remaining. Safety was monitored by evaluating adverse events and serious adverse events (SAEs).
Statistical analysis
A study size of at least 12 participants in the treatment group was determined without formal statistical hypothesis testing according to the precedent [22]. A 2:1 randomization ratio was used to reduce the exposure rate to placebo for ethical reasons. Thus, considering a drop-out rate of approximately 20%, 25 participants were targeted for enrollment, with a minimum of 14 randomized to HK-660S and seven randomized to placebo.
The efficacy analysis was performed for both the full analysis set (FAS) and the per-protocol set (PPS). However, the interpretation of results for the final efficacy evaluation was based on the FAS. All statistical analyses used SAS® (version 9.4; SAS Institute Inc.). In principle, the sig-nificance test between treatment groups was a two-sided test under a significance level of 10%. Descriptive statistics were presented for continuous variables, and a t-test or Wilcoxon rank sum test was performed to test the difference between groups. For categorical variables, frequency and ratio were given, and a chi-square or Fisher’s exact test was performed to test the difference between groups. When missing values occurred in the efficacy endpoint measurements in the FAS population, the last observation carried forward method was applied. Efficacy endpoints measured only once after the baseline were analyzed using the observed values. For safety and other data, the observed data were analyzed without correction.
A subgroup analysis was performed based on NQO1 genotype stratification. Ad-hoc analyses were performed to identify meaningful clinical data. Specifically, studies were performed for the change in ALP levels based on the following criteria: 1) PSC duration, 2) Inflammatory bowel disease (IBD) diagnosis, 3) baseline ALP > 2 times the ULN, and 4) ALP reduction of 20% or more.
The primary efficacy analysis compared the change in ALP levels between the HK-660S and placebo groups from baseline to week 12, using an analysis of covariance (ANCOVA) model, with treatment as a fixed effect and baseline ALP, age, and BMI as covariates. ANCOVA and repeatedmeasures ANCOVA in a mixed model were used to provide estimates of ALP change from baseline and correspond to a 95% confidence interval (CI) at each visit. Changes from biochemical parameters obtained from the model are reported as least-square mean with standard error (SE) for statistical testing.
RESULTS
Patient population
A total of 23 patients were enrolled and randomized into the placebo (n=7) or treatment group (n=16) as the safety set. One patient from each of the placebo and treatment groups met the exclusion criteria (change in UDCA dosage during the screening period) and was eliminated from the FAS (n=21; placebo, 6; study group, 15). Out of the 21 patients in the FAS, four patients in the treatment group were disqualified owing to medication with prohibited drugs, whereas one patient in the placebo group was disqualified owing to a randomization error. Thus, the PPS (n=16) consisted of five patients in the placebo group and 11 in the study group (Fig. 1).
Demographics and baseline characteristics resembled those typically observed in patients with PSC (Table 1). Medication compliance was more than 80% for all patients. Except for two patients, all were concomitantly administered UDCA (with one patient from each group not receiving it). Most patients were administered 900 mg/day or less, whereas two patients in the HK-660S group exceeded 900 mg/day (1,200 and 1,600 mg/day). The range of administered UDCA doses was from 5.0 to 16.9 mg/kg/day.
Primary outcomes
The primary efficacy analysis in the FAS population showed a favorable improvement rate of PSC severity on the MRCP Anali score in the HK-660S group, with an im-provement rate of 13.3% (90% CI 2.42–36.34) compared to 0% (90% CI 0–39.30) in the placebo group without statistical significance (Fig. 2A). In the PPS population, the improvement rate of PSC severity in the HK-660S group was 9.1% (90% CI 0.47–36.44) compared with 0% (90% CI 0–45.07) for the placebo group (Fig. 2B). Both patients with a decrease in Anali scores showed a dilation of the bile duct (baseline-12 weeks-16 weeks; 6-5-5 and 4-3-2) (Fig. 2C).
In the FAS population, mean ALP levels decreased in the HK-660S treatment group by 78.6±60.35 U/L (vs. 33±24.00 U/L in placebo) after 12 weeks. However, owing to the low sample size, statistical significance could not be derived (P=0.73; Wilcoxon rank sum test) (Fig. 3A). In the PPS population, mean ALP levels decreased in the HK-660S treatment group by 148.6±71.43 U/L (vs. 46.8±24.05 U/L in placebo) after 12 weeks (Fig. 3B).
The HK-660S group showed a decrease of 15.2% in mean ALP levels at 12 weeks compared with 6.6% in the placebo group (Fig. 3C). An ad-hoc analysis was performed to identify a meaningful response rate to drug treatment. When the patient population was stratified by baseline ALP levels, a statistically significant 50% responder rate (percentage of patients achieving a 20% or greater reduction in ALP levels at 12 weeks from baseline) was observed in patients with a baseline ALP >2 times the ULN and treated with HK-660S (P=0.05; Fisher’s exact test, significant at a two-tailed α=0.1). This contrasted with the placebo group, where none of the patients showed a reduction in ALP greater than 20%.
Secondary outcomes
The improvement rate of biliary stricture based on MRCP in the FAS population was 26.7% (90% CI 9.7–51.1%) of patients in the HK-660S group showed an improvement, whereas none (90% CI 0–39.3%) of the patients in the placebo group showed improvement (P=0.28). In the PPS population, 27.3% (90% CI 7.9–56.4%) of patients in the HK-660S group showed an improvement, whereas none (90% CI 0–45.1%) of the patients in the placebo group showed improvement (P=0.51).
Liver elasticity also showed improvement. Three patients in the HK-660S group showed improvement from severe (F3) to normal (F0–1), severe to moderate (F2), and moderate to normal. In contrast, one patient in the placebo group worsened from moderate to severe. Regarding the ELF score, four patients (26.7%) in the HK-660S group improved from severe to moderate fibrosis at 12 weeks, compared to none in the placebo.
Although serum AST, ALT, and total bilirubin levels did not show noticeable differences between the treatment and placebo groups, γ-GTP showed decreasing trends in the treatment group (Supplementary Fig. 2).
Exploratory outcomes
Exploratory biomarkers were obtained at baseline and 12 weeks. These included the inflammatory biomarkers TNF-α, TGF-β, IL-6, CCL-2, and CCL-5 (Supplementary Fig. 3A–E). Additionally, PSC-related biomarkers, including CK-18, GDF15, and FGF21 were also measured (Supplementary Fig. 3F–H). Except for TNF-α, FGF21, and CK-18, other biomarkers showed decreasing trends in the HK-660S treatment group. Most notably, CCL-2 showed a statistically significant decrease (P=0.04; Wilcoxon rank sum test) in the treatment group.
Safety
During the study period, the rate of at least one treatment-emergent adverse event (TEAE) was 42.9% (3/7, nine events) in the placebo group and 62.5% (10/16, 36 events) in the HK-660S group (Table 2). Most TEAEs were mild to moderate in severity. Adverse drug reactions (ADRs) occurred in 18.8% (3/16, seven events) and were only observed in the HK-660S group. Three patients experienced seven ADRs, including abdominal pain (two instances), abdominal distension, diarrhea, nausea, fatigue, and dizziness. All ADRs were mild in severity except for one instance of moderate diarrhea, which resolved after treatment discontinuation. An SAE was reported for one patient (bacteremia) in the HK-660S group but was assessed as not related to drug treatment. No deaths were reported during the study.
DISCUSSION
PSC remains a challenging condition with limited effective therapeutic options, highlighting the urgent need for new treatments. Specifically, the U.S. National Institutes of Health lists clinical trials such as a study evaluating the safety and efficacy of DUR-928 in subjects with PSC (NCT03394781) and a study on the antimicrobial and immunomodulating effects of oral vancomycin in PSC (NCT01802073) [23]. Additionally, therapies targeting the gutliver axis, such as antibiotics and integrin blockers like vedolizumab, are being explored due to the association of PSC with inflammatory bowel disease and the gut microbiome (NCT03035058) [24]. These trials reflect a broad spectrum of therapeutic targets, from immunomodulation and fibrosis inhibition to modification of bile acid metabolism, highlighting the diverse pathophysiological insights that are currently guiding PSC research.
This study shows HK-660S as a potential treatment option, evidenced by significant and consistent ALP reductions. Notably, improvements began by week 4 and persisted, demonstrating its effectiveness in reducing PSC severity. Enhanced imaging techniques, such as MRCP and VCTE, suggested potential improvements in biliary stricture and fibrosis with HK-660S treatment, indicating therapeutic benefits in PSC liver pathology. Furthermore, regarding the ELF score, four patients (26.7%) in the HK-660S group improved from severe to moderate fibrosis at 12 weeks, compared to none in the placebo. However, it is important to note that the observed differences were not statistically significant.
HK-660S and other candidates showed promising results in liver injury biomarkers for PSC. In a clinical trial, norUDCA treatment significantly reduced ALT, AST, γ-GT, and ALP levels; however, SAEs were reported for seven patients in the low dose group [12]. OCA, a potent FXR agonist, decreased ALP and bilirubin levels in a clinical trial for primary biliary cholangitis [25,26]. In the phase 2 clinical study for PSC, OCA reduced ALP levels significantly within 24 weeks, with notable TEAEs related to pruritus [3]. NGM282, an engineered analog of FGF19, improved ALT, AST, and fibrosis biomarkers, including Pro-C3, and was generally well tolerated [27]. HK-660S prominently reduced ALP and γ-GT levels while improving fibrosis markers (ELF and VCTE), with most TEAEs related to mild and moderate gastrointestinal symptoms.
In contrast, four types of surrogate endpoints are used in therapeutic trials for PSC: serum biomarkers (such as ALP or ELF score), MRCP, liver elastography, and prognostic scores. Among these, MRCP is a powerful and non-invasive diagnostic method for evaluating bile duct strictures in PSC. Notably, the MRCP Anali score has been shown to correlate strongly with survival outcomes and is a valuable non-invasive prognostic tool for predicting disease progression in patients with PSC [28-30]. ALP is also widely used as a surrogate endpoint for both PSC and PBC. However, accumulating evidence indicates significant variability in serum ALP levels, both within and among patients, regardless of treatment. This variability may limit the utility of ALP as the sole surrogate endpoint for demonstrating clinical efficacy in trials [28]. Therefore, the results of this study, including the structural improvements in PSC severity indicated by the Anali score and the assessment of bile duct strictures by MRCP, provide strong evidence supporting the potential of HK-660S as a treatment option for PSC.
This study had several limitations. First, in the exploratory proof-of-concept study of HK-660S, the sample size was not formally calculated due to its initial clinical application to patients with PSC. In this phase 2a trial, the efficacy analysis was performed for both the FAS and the PPS, but the interpretation of results for the final efficacy evaluation was based on the FAS according to the study protocol. Although the number of participants in the PPS did not reach the predetermined size of 12 per treatment group, the number in the FAS was adequate for the final efficacy evaluation. Moreover, we consider the trend towards decreased ALP levels and improvement in biliary stricture observed on MRCP imaging to be promising indicators of the potential efficacy of HK-660S in treating PSC within the relatively short study period of 12 weeks. The improvement observed in MRCP images was particularly noteworthy. The use of exploratory endpoints is encouraged in PSC studies due to challenges in identifying appropriate study endpoints that adequately measure clinical benefit in PSC. Therefore, this proof-of-concept study for PSC was designed to explore any potential role of HK-660S.
Second, this pilot study would have benefitted from a longer treatment duration, such as 24 weeks, to evaluate cholangiographic changes using its proposed primary endpoint (Anali score) as well as the measurement of additional markers of fibrosis, such as pro-C3. Future confirmatory trials are planned to extend the treatment duration to at least 24 weeks and incorporate statistical sample size calculation, additional fibrosis markers (CCL-24, pro-C3), and prognostic factors (PRESTO, UK-PSC, Mayo-PSC).
Despite the limitations of this study, including the insufficiency to yield statistically significant results owing to the small sample size, subgroup analyses have provided intriguing insights into the differential responses to HK-660S based on patient characteristics. Patients with pre-existing IBD showed improvements in ALP levels, suggesting a potential synergistic effect between HK-660S and the management of IBD. Additionally, patients with baseline ALP levels exceeding twice the ULN showed a more potent reduction in ALP levels, suggesting that HK-660S may be particularly effective in patients with higher baseline ALP values.
In conclusion, in this phase 2a study of patients with PSC, HK-660S 100 mg twice daily over 12 weeks demonstrated promising trends in reducing serum ALP levels and improving biliary stricture on MRCP, although these results were not statistically significant. Moreover, HK-660S was well tolerated and potentially improved liver fibrosis. Its favorable safety profile and potential clinical benefits warrant further investigation in larger trials.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
