Bringing doctors together

or

A free, private network for verified physicians

  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    NQO1 acts as a context-dependent regulator of oxidative stress and biliary fibrosis in primary sclerosing cholangitis.

    Oct 23, 2026DOI: 10.1097/hc9.0000000000001076Journal ArticleHepatology Communications

    Written by: Juhyun KimSeung Hee WooDa-Hyeun YoonHeedoo LeeChul-Ho LeeJung Hwan Hwang

    INTRODUCTION

    Primary sclerosing cholangitis (PSC) is a chronic, progressive cholestatic liver disorder defined by persistent biliary inflammation and fibrotic remodeling of the biliary tree.1,2 These pathological hallmarks inevitably culminate in biliary cirrhosis, portal hypertension, and eventual hepatic failure.3 Despite its classification as a rare disease, the clinical burden of PSC is profound; it predominantly affects young to middle-aged populations and frequently necessitates liver transplantation.4 Furthermore, patients with PSC face a significantly elevated risk of hepatobiliary malignancies, most notably cholangiocarcinoma, which remains a primary driver of mortality in this cohort.1,5The precise etiology and molecular pathogenesis of PSC remain elusive, despite decades of intensive clinical and translational research. Current paradigms suggest that PSC arises from a multifaceted interplay between genetic predisposition, immune dysregulation, environmental factors, and disrupted bile acid homeostasis.6 Notably, the strong clinical association with inflammatory bowel disease—observed in 60%–80% of cases—underscores the potential role of gut–liver axis-mediated immune responses in disease progression.1 However, the specific signaling pathways that bridge chronic inflammation to biliary injury and progressive fibrosis have yet to be fully elucidated.Growing evidence also indicates that oxidative stress plays a critical role in PSC pathogenesis.7,8 Cholestatic conditions promote the intrahepatic accumulation of hydrophobic bile acids and pro-inflammatory cytokines, which trigger the excessive production of reactive oxygen species (ROS) in both hepatocytes and cholangiocytes.1 This surge in ROS induces cellular damage through lipid peroxidation and mitochondrial impairment, while simultaneously activating fibro-inflammatory cascades such as the NF-κB and MAPK pathways.9 Moreover, oxidative stress facilitates cholangiocyte apoptosis, recruitment of inflammatory cells, and the activation of hepatic stellate cells, all of which are essential drivers of biliary fibrosis.10,11To counteract such oxidative insults, cells employ endogenous antioxidant defense systems to preserve the redox equilibrium. NAD(P)H:quinone oxidoreductase 1 (NQO1) is a critical cytoprotective enzyme that safeguards cells by catalyzing the two-electron reduction of quinones to hydroquinones, thereby neutralizing ROS and preventing membrane lipid peroxidation.12 The expression of NQO1 is primarily governed by nuclear factor erythroid 2-related factor 2 (Nrf2) during oxidative stress or by aryl hydrocarbon receptor (AhR) in response to xenobiotics.13,14 Although NQO1 has demonstrated protective effects in various inflammatory liver diseases by mitigating ROS-induced injury,15 its functional significance in the context of PSC remains largely uncharacterized.
    In the present study, we aimed to investigate the impact of NQO1 deficiency on the cardinal pathogenic features of PSC, including apoptosis, inflammation, and fibrosis. Furthermore, we explored the underlying molecular mechanisms to determine whether NQO1-associated signaling serves as a key regulatory node in PSC pathogenesis.

    METHODS

    Male WT or NKO mice, aged 8–10 weeks, were utilized in this study. To induce a PSC-like phenotype, the mice were administered a diet supplemented with 0.1% (w/w) 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC; Sigma-Aldrich, MO, USA) or a standard control diet for 1 or 3 weeks. All mice were maintained in a specific pathogen-free facility under controlled environmental conditions (20–22 °C, 12-h light/dark cycle) with ad libitum access to water and food. All animal experiments were conducted in accordance with the institutional guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee of the Korea Research Institute of Bioscience and Biotechnology (KRIBB-AEC-23065 and KRIBB-AEC-24194). The animal experiments were reported in accordance with the ARRIVE guidelines.
    To analyze the biochemical indicators of PSC, blood samples were harvested from mice and centrifuged at 10,000g for 5 minutes at 4 °C. The resulting plasma was isolated and maintained at −80 °C until further use. Systemic liver function was evaluated by measuring plasma levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), and total bilirubin (TBIL) using an automatic chemical analyzer (Hitachi, Tokyo, Japan). Furthermore, variations in the total bile acid (TBA) concentrations induced by DDC were quantified using a commercial TBA assay kit (Sigma-Aldrich) in strict accordance with the provider’s guidelines. Briefly, plasma was incubated with the supplied enzymatic reagents containing 3α-hydroxysteroid dehydrogenase and NAD for 20 minutes in the dark. The fluorescence generated by the NADH-dependent reaction was measured using a microplate reader at an excitation wavelength of 530 nm and an emission wavelength of 585 nm. TBA concentrations were calculated according to the manufacturer’s internal standard method.
    Publicly available human PSC and non-diseased liver scRNA-seq data from Andrews et al16 were re-analyzed using processed and annotated datasets obtained from CZ CELLxGENE Discover (collection ID: 0c8a364b-97b5-4cc8-a593-23c38c6f0ac5; GEO GSE243981). Analyses were performed in Python using Scanpy and AnnData.17 Hepatocytes, cholangiocytes, hepatic stellate cells, and macrophages/Kupffer cells were identified according to the original cell-type annotations, and the expression of NQO1, NFE2L2 (Nrf2), and AhR was evaluated. The data represent the percentage of cells with detectable transcript expression (expression >0). Individual donors were treated as biological replicates. PSC and non-diseased groups were compared using a two−sided Mann–Whitney U test, with p˂0.05 considered statistically significant.
    Total RNA was extracted from liver tissues employing TRIzol reagent (Thermo Fisher Scientific, MA, USA) as per the manufacturer’s instructions. The concentration and integrity of the extracted RNA were assessed via NanoDrop spectrophotometer (NanoDrop Technologies, DE, USA). Subsequently, 1 μg of total RNA was synthesized into cDNA using the UltraScript 2.0 cDNA Synthesis Kit (PCR BioSystems, London, UK). Quantitative PCR was conducted using the AccuPower 2× Greenstar qPCR Master Mix (Bioneer, Daejeon, Korea) on a StepOnePlus Real-Time PCR system (Applied Biosystems, CA, USA). Relative mRNA expression levels were determined using the 2−ΔΔC t method, with 18S rRNA as the internal control. Data are presented as fold changes relative to the control group. A comprehensive list of the primer sequences is provided in Table 1.
    List of primers used in this study
    Liver tissues were homogenized in radioimmunoprecipitation assay (RIPA) lysis buffer (Sigma-Aldrich) containing protease (Roche Applied Science, Mannheim, Germany) and phosphatase inhibitors (Sigma-Aldrich). The protein yields were quantified using a BCA protein assay kit (Thermo Fisher Scientific). Equal protein loads (20–40 μg) were resolved by SDS-PAGE on 8%–12% polyacrylamide gels and subsequently transferred to nitrocellulose membranes using a semi-dry transfer apparatus (Bio-Rad Laboratories, CA, USA). After blocking with 5% non-fat skim milk in Tris-buffered saline with Tween 20 (TBST) for 1 hour at room temperature, the membranes were probed overnight at 4 °C with primary antibodies against NQO1 (Abcam, MA, USA), HO-1 (Abcam), α-SMA (Abcam), fibronectin (Abcam), Prdx1 (Cell Signaling Technology, MA, USA), Prdx2 (Cell Signaling Technology), Prdx6 (MyBioSource, CA, USA), catalase (Cell Signaling Technology), MnSOD1 (BD Biosciences, CA, USA), phosphor-AKT (Ser473) and total AKT (Cell signaling Technology), and β-actin (Cell Signaling Technology). Following 3 washes with TBST, the membranes were incubated with HRP-conjugated secondary antibodies (1:2000) for 1 hour. Protein immunoreactivity was detected using enhanced chemiluminescence (ECL) reagents and captured using the Ez-Capture digital imaging system (ATTO, Tokyo, Japan).
    To measure porphyrin accumulation in the PSC mouse model, liver tissues were homogenized in an ice-cold 0.25 M sucrose solution supplemented with phosphatase and protease inhibitors. The resulting homogenates were centrifuged at 13,000 rpm for 10 minutes at 4 °C, and the supernatants were collected for further analysis. Protein concentrations were determined using a BCA protein assay kit. Porphyrin fluorescence intensity in the supernatants was initially detected in 96-well black microplates using an in vivo imaging system (Lumina System; PerkinElmer, MA, USA) with Cy5.5 excitation/emission filters. Subsequent quantification was performed using a microplate reader at an excitation wavelength of 405 nm and emission wavelength of 635 nm.
    Liver tissues were fixed in 10% neutral-buffered formalin, embedded in paraffin, and sectioned at a thickness of 5 μm. Sections were deparaffinized in xylene and rehydrated through a descending ethanol gradient. For morphological evaluation, tissue slides were stained with hematoxylin and eosin (H&E). Histological activity was assessed using a modified histological activity index (mHAI) scoring system based on the Ishak criteria. Four pathological features were scored: interface hepatitis (0–4), confluent necrosis (0–6), focal lytic necrosis/apoptosis with focal inflammatory activity (0–4), and portal inflammation (0–4). The individual scores were summed to generate an overall histological activity score ranging from 0 to 18. Histological scoring was conducted by a pathologist blinded to the experimental group allocation.18 To assess inflammatory cell infiltration, oxidative damage, and NQO1 expression, tissue sections were subjected to heat-induced antigen retrieval in citrate solution for 15 minutes, followed by blocking for 1 hour. The slides were incubated overnight at 4 °C with primary antibodies against F4/80 (Cell Signaling Technology), 4-hydroxynonenal (4-HNE; Abcam), or NQO1 (Abcam). After 3 washes in PBS, secondary antibody incubation was performed using the Vectastain Elite ABC kit (Vector Laboratories, CA, USA) as per the manufacturer’s instructions. Immunoreactivity was visualized using diaminobenzidine (DAB; Vector Laboratories), and the slides were mounted with an antifade medium (Sigma-Aldrich). To quantify biliary fibrosis, the Masson trichrome staining was conducted using a commercial staining kit (Abcam), following the provider’s protocol. To detect apoptotic cell death, a TdT-mediated dUTP nick-end labeling (TUNEL) assay was performed using the In Situ Cell Death Detection Kit (Sigma-Aldrich) in accordance with the manufacturer’s instructions.
    Total RNA was isolated from the livers of NQO1 WT and KO mice after 1 week of DDC administration. Briefly, mRNA was extracted from the total RNA pool using poly T oligo-conjugated magnetic beads, followed by thermal fragmentation in the presence of divalent cations. First-strand cDNA was synthesized using a random hexamer, whereas second-strand synthesis incorporated dUTP for strand-specific marking. The final cDNA libraries were constructed using end-repair, A-tailing, adapter ligation, and PCR enrichment sequences. Sequencing was executed on the Illumina X Ten platform, as previously described.19 Volcano plot, gene set enrichment analysis (GSEA), and heatmap were performed on the differentially expressed genes (DEGs) using R software (version 4.5.1). For heatmap visualization, genes were filtered according to the gene annotation information provided in the RNA-seq dataset. Genes annotated as pseudogenes based on gene biotype information were excluded. In addition, genes with symbols beginning with Gm, BC, or Rik were removed before heatmap generation.
    The hepatic NAD+ content was quantified using a commercial NAD/NADH assay kit (Abcam). Briefly, liver extracts were homogenized and centrifuged in an extraction buffer to obtain clear lysates. To ensure accuracy, proteins were removed by filtration through a 10-kDa spin column (Abcam) before the enzymatic assay. The concentrations of free NAD+ and NADH were determined in accordance with the manufacturer’s protocol.
    Primary hepatocytes were harvested from WT and NKO mice following established protocols with slight adjustments. Briefly, after euthanasia, mice underwent in situ liver perfusion. The central vein was cannulated using a 24-gauge catheter, and the liver was first cleared of blood and cell–cell junctions by perfusing 50 mL of prewarmed EDTA buffer at a constant flow rate of 27.5 mL/min. Subsequently, the liver was digested with HBSS (Invitrogen) containing 4% Type I collagenase (Worthington, NJ, USA) and 0.002% DNase I (Roche) at 37 °C until the hepatic tissue softened. The digested liver was carefully excised, and the capsule was disrupted in RPMI medium to liberate the hepatocytes. The resulting cell suspension was filtered through a 70-μm nylon cell strainer and centrifuged at 50g for 5 minutes at 4 °C. For the purification of viable hepatocytes, an isotonic Percoll gradient was prepared by mixing Percoll, 10× PBS, and HEPES at a ratio of 45:5:0.5 (v/v/v). After centrifugation, the resulting pellet containing viable hepatocytes was collected from the bottom of the tube, washed with RPMI medium, and processed for subsequent culturing.
    HepG2 (human hepatocellular carcinoma), LX-2 (human hepatic stellate), and primary hepatocytes were cultured in the Dulbecco Modified Eagle Medium (HyClone GE Healthcare Life Sciences, UT, USA) supplemented with 10% fetal bovine serum at 37 °C in a 5% CO2 environment. Upon reaching 70% confluence in 6-well plates, the cells were serum-starved overnight and subsequently stimulated with 100 ng/mL recombinant periostin (Sino Biological, Beijing, China) for the specified durations. HepG2 cells were exposed to taurocholic acid (TCA, 100 μM, Sigma-Aldrich), TGF-β1 (10 ng/mL; Sigma-Aldrich), or TNF-α (10 ng/mL; PeproTech, NJ, USA) for 24 hours. Cells were then rinsed with cold PBS and lysed with sodium dodecyl sulfate sample buffer for subsequent protein analysis. For the inhibition study, HepG2 cells were seeded in 6-well plates and cultured for 24 hours. The cells were then serum-starved overnight and pretreated with BAY 11-7082 (10 μM, Sigma-Aldrich), LY294002 (10 μM, Sigma-Aldrich), PF-573228 (5 μM, Selleck Chemicals), or N-acetyl-L-cysteine (NAC; 5 mM, Sigma-Aldrich) for 45 minutes before periostin or insulin (100 nM, Sigma-Aldrich) treatment.
    To achieve transient silencing of NQO1, HepG2 and LX-2 cells were seeded in 6-well plates and grown to 50% confluence. Transfection was performed using 20 nM NQO1-specific siRNA or non-targeting scrambled siRNA (Bioneer). Transfection was mediated by Lipofectamine RNAiMAX Reagent (Invitrogen) according to the manufacturer’s protocol. The specific siRNA sequences were 5′-AAACCAGCCUUUCAGAAUGGCUGGC-3′ (sense) and 5′-GCCAGCCAUUCUGAAAGGCUGGUUU-3′ (antisense). Cells were further cultured until they reached 70% confluence, after which the intracellular ROS levels were measured. To investigate the role of NQO1 in TGF-β signaling, LX-2 cells were transfected with scrambled siRNA or NQO1-targeting siRNA. Following transfection, the cells were stimulated with TGF-β1 for the indicated time periods.
    Intracellular ROS production was evaluated using a 2′,7′-dichlorofluorescein diacetate (DCF-DA) fluorescent probe. HepG2 and primary hepatocytes were treated with recombinant periostin for 6 hours and then incubated with DCF-DA to quantify the oxidative stress response.
    Quantitative data are presented as mean ± standard error of the mean (SEM). The ImageJ software (version 1.53t) was used to quantify the positive cell areas and cell counts. Statistical comparisons between 2 experimental groups were performed using the unpaired Student t test with GraphPad Prism software (version 8.0; Northampton, MA, USA). For comparisons between 2 independent experimental groups, an unpaired two-tailed Student t test was used. When multiple individual genes were simultaneously compared between 2 genotypes under the same experimental condition, multiple unpaired t tests with Holm–Šídák correction for multiple comparisons were applied. For experiments involving both genotype and diet as independent variables, data were analyzed by two-way analysis of variance (ANOVA), followed by the Šídák multiple-comparisons test for the relevant post hoc comparisons. A two-sided p-value <0.05 was considered statistically significant.

    RESULTS

    To investigate the expression profile of NQO1 under cholestatic conditions, mice were fed a DDC-supplemented diet for 1 week. Successful establishment of the PSC model was evidenced by a systemic increase in liver injury markers (AST, ALT, ALP, and T-bil), the histological emergence of representative ductular reactions, and a significant increase in the mHAI score in DDC-fed mice compared with CD-fed mice (Figures 1A, B). Under these pathological conditions, we observed a significant upregulation of NQO1 at both the mRNA and protein levels (Figures 1C, D). Immunohistochemical analysis further revealed increased NQO1 protein expression in hepatocytes of DDC-fed mice compared with CD-fed mice (Figure 1E). This induction was accompanied by increased levels of Nrf2 and AhR, which are the primary upstream regulators of NQO1, suggesting that NQO1 is a key stress-responsive element in DDC-induced injury (Figure 1F). To assess the translational relevance, we re-analyzed available human liver scRNA-seq data from PSC patients. NQO1 expression was significantly increased in hepatocytes, with nonsignificant upward trends in HSCs and Kupffer cells (Figure 1G). Nrf2 was significantly elevated across all examined cell types, whereas AhR was increased in cholangiocytes, HSCs, and Kupffer cells (Figures 1H, I). These findings indicate that PSC is associated with enhanced NQO1-related stress responses in a cell-type-dependent manner.
    Characterization of NQO1 induction in a DDC-fed PSC mouse model of PSC. (A) Pathogenic parameters such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), and total bilirubin (T-bil) were quantified in mice maintained on either CD (n=5) or a DDC diet (n=5) for 1 week. (B) Representative H&E-stained liver sections from the CD (n=5) and DDC (n=5) groups and quantitative assessment of histological injury using the modified histological activity index (mHAI). Additional H&E-stained fields used for mHAI assessment from individual mice are shown in Supplemental Figure S1, https://links.lww.com/HC9/C508. Scale bar: 200 μm. (C) RT-qPCR analysis of hepatic Nqo1 mRNA expression levels (CD, n=5; DDC, n=6). (D) Western blot analysis displaying the protein levels of NQO1 and HO-1, with β-actin as an internal control (left). Densitometric quantification of NQO1 and HO-1 protein levels normalized to β-actin is shown as bar graphs (right). (E) Representative immunohistochemical staining of liver sections with antibodies against NQO1 to determine the spatial distribution of NQO1. Scale bar: 200 μm. (F) Transcriptional levels of the upstream regulators of aryl hydrocarbon receptor (AhR) and nuclear factor erythroid 2-related factor 2 (Nrf2) measured via RT-qPCR (CD, n=5; DDC, n=6). (G–I) Re-analysis of publicly available human liver scRNA-seq data from healthy individuals and patients with PSC. The percentages of cells with detectable transcripts for (G) Nqo1, (H) Nrf2, and (I) AhR were compared between healthy and PSC livers across hepatocytes, cholangiocytes, HSCs, and Kupffer cells; p<0.05, p<0.01, and ***p*<0.001. Abbreviations: AhR, aryl hydrocarbon receptor; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CD, control diet; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; H&E, hematoxylin and eosin; HO-1, heme oxygenase 1; HSCs, hepatic stellate cells; mHAI, modified histological activity index; NQO1, NAD(P)H quinone dehydrogenase 1; Nrf2, nuclear factor erythroid 2-related factor 2; PSC, primary sclerosing cholangitis; RT-qPCR, reverse transcription-quantitative PCR; scRNA-seq, single-cell RNA sequencing; T-bil, total bilirubin.
    To investigate the contribution of NQO1 to the hepatic antioxidant response during cholestatic liver injury, WT and NKO mice were fed a control diet (CD) or DDC diet for 7 days. DDC feeding substantially altered the hepatic expression of several antioxidant genes, including catalase, MnSOD, Gpx, Prdx1, Prdx2, Prdx6, and HO-1. Notably, the transcript levels of these genes were significantly lower in DDC-fed NKO mice than in their WT counterparts (Figure 2A). At the protein level, Prdx1 expression remained unchanged following DDC administration, whereas Prdx2, Prdx6, MnSOD, and catalase generally showed decreasing trends in both genotypes (Figure 2B). Among these proteins, catalase was reduced in NKO mice relative to WT mice under DDC-fed conditions (Figure 2B). The antioxidant protein profile in isolated primary hepatocytes differed partially from that observed in whole-liver lysates (Figure 2C). Prdx1 and Prdx2 levels were comparable between hepatocytes isolated from WT-CD and WT-DDC, whereas Prdx6, MnSOD, and catalase were reduced following DDC treatment. In addition, MnSOD expression was higher in NKO hepatocytes than in WT hepatocytes in both conditions (Figure 2C). These differences suggest that non-parenchymal cells or hepatic microenvironmental factors may contribute to the antioxidant response observed in the whole liver. DDC markedly increased hepatic 4-HNE accumulation, whereas this increase was significantly reduced in NKO mice, indicating that NQO1 deficiency attenuated DDC-induced lipid peroxidation despite the altered antioxidant profile (Figure 2D). Although hepatic NAD+ and NADH levels tended to decrease following DDC treatment in both genotypes, the differences relative to the corresponding CD groups were not statistically significant. Furthermore, no significant genotype-dependent differences were observed under either dietary condition (Figure 2E). These results indicate that NQO1 deficiency attenuates DDC-induced hepatic lipid peroxidation despite reduced expression of several antioxidant genes and proteins, without significantly affecting hepatic NAD+ and NADH levels.
    Impact of NQO1 deficiency on antioxidant defense and oxidative stress. (A) Comparison of antioxidant-related gene transcripts in DDC-treated WT (n=6) and NKO (n=5) livers; p<0.05, p<0.01, and ***p*<0.001. (B) Representative immunoblots showing hepatic expression of antioxidant proteins in WT and NKO mice under CD or DDC-challenged conditions are presented on the left (n=2 per group). Densitometric quantification of the corresponding protein bands was performed using ImageJ software (right graph). The experiment was independently performed twice. (C) Schematic overview of the experimental procedure (left). WT and NKO mice were fed either a CD or a DDC for 7 days. Primary hepatocytes were subsequently isolated, plated, and harvested for protein extraction immediately after cell attachment. Representative immunoblots showing the expression of antioxidant proteins in the isolated primary hepatocytes (middle) and densitometric quantification of the corresponding protein bands (right). Quantification of antioxidant concentrations in primary hepatocytes harvested from WT and NKO mice after administration of DDC for 1 week. (D) Immunohistochemical detection and ImageJ-based quantification of 4-HNE, a marker of lipid peroxidation. Scale bar: 200 μm. (E) Hepatic NAD+ and NADH levels were measured in all experimental groups (n=5 per group). Abbreviations: 4-HNE, 4-hydroxynonenal; CD, control diet; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; NAD+, nicotinamide adenine dinucleotide (oxidized form); NADH, reduced nicotinamide adenine dinucleotide; NKO, NQO1 knockout; NQO1, NAD(P)H:quinone dehydrogenase 1; NS, not significant; WT, wild type.
    To determine the acute effects of NQO1 deficiency, WT and NKO mice were fed either a CD or a DDC-containing diet for 1 week. Plasma markers of liver injury were then evaluated. Among DDC-fed mice, AST and ALP did not differ significantly between the 2 genotypes (Figures 3A, C). In contrast, plasma ALT levels were significantly lower in NKO mice than in WT mice, while total bilirubin levels showed a decreasing trend in the NKO group (Figures 3B, D). Histological examination by H&E staining revealed attenuated liver injury in NKO mice, which was further supported by a significantly lower mHAI score in NKO-DDC mice compared with WT-DDC mice (Figure 3E). Although TUNEL-positive hepatocytes were rare relative to the widespread induction of NQO1, their frequency was significantly lower in NKO-DDC mice than in WT-DDC mice (Figure 3F). Notably, the accumulation of porphyrin and TBA remained unchanged between the groups, suggesting that NQO1 deficiency protects the liver through mechanisms independent of bile acid levels or porphyrin clearance (Figures 3G, H).
    Mitigation of DDC-induced hepatobiliary damage in NKO mice. To determine whether NQO1 deficiency protects against acute cholestatic liver injury, WT and NKO mice were fed either a CD or DDC for 7 days. (A–D) Analysis of plasma AST, ALT, ALP, and total bilirubin (T-bil) levels to assess liver function in the WT and NKO groups; p*<0.05. (E) Representative images of H&E-stained liver sections and mHAI scoring illustrating DDC-induced pathological changes. Additional H&E-stained fields used for mHAI assessment from individual mice are shown in Supplemental Figure S2, https://links.lww.com/HC9/C508. Scale bar, 200 μm; p<0.05. (F) Representative images of TUNEL-stained liver sections and quantification of TUNEL-positive cells using ImageJ software. The bar graphs show the percentages of total TUNEL-positive cells and TUNEL-positive hepatocytes, each calculated relative to the total number of cells within the analyzed microscopic field; ***p*<0.001. (G) IVIS-based visualization of porphyrin accumulation and subsequent RFU measurement using a VICTOR multimode plate reader. (H) Determination of TBA concentrations in liver tissues. All experimental groups consisted of 5 mice (n=5 per group). Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CD, control diet; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; H&E, hematoxylin and eosin; IVIS, in vivo imaging system; mHAI, modified histological activity index; NKO, NQO1 knockout; NQO1, NAD(P)H quinone dehydrogenase 1; RFU, relative fluorescence units; TBA, total bile acids; T-bil, total bilirubin; TUNEL, TdT-mediated dUTP nick-end labeling; WT, wild type.
    Next, we evaluated the impact of NQO1 deficiency on inflammation and fibrosis, which are the defining pathological features of PSC. We performed immunohistochemical staining of F4/80 to assess inflammatory cell infiltration. Quantification of F4/80-positive macrophages revealed a modest decrease in the livers of NKO compared with those of WT mice (Figure 4A). Despite this reduction, hepatic mRNA levels of the pro-inflammatory cytokines Il-1β, Il-6, and TNFα did not differ significantly between the 2 groups (Figure 4B). In contrast, NKO mice exhibited a substantial reduction in hepatic fibrosis. Masson staining demonstrated significantly lower collagen deposition in NKO mice than in WT mice (Figure 4C). This histological improvement was accompanied by marked downregulation of fibrogenic genes, such as Tgf-β1, Tgf-β2, Col1α1, and Fibronectin (Figure 4D). Corroborating these transcriptomic shifts, the protein levels of fibronectin and α-SMA were also notably lower in DDC-fed NKO livers (Figure 4E). Collectively, these results suggest that NQO1 specifically promotes DDC-induced hepatic fibrosis, independent of the initial inflammatory response.
    Effects of NQO1 deficiency on hepatic inflammatory and fibrotic responses following acute DDC exposure. To assess whether NQO1 deficiency alters hepatic inflammation and early fibrotic responses, WT and NKO mice were fed either a CD (n=5) or DDC (n=6) diet for 7 days. (A) Representative images of F4/80-immunostated liver sections showing hepatic macrophage infiltration (upper), accompanied by quantification of the F4/80-positive area as a percentage of the total tissue area (lower). Scale bar, 200 μm; p<0.01. (B) RT-qPCR analysis of hepatic mRNA expression levels of representative pro-inflammatory cytokines and inflammatory markers. (C) Representative images of Masson trichrome-stained liver sections showing collagen deposition, accompanied by quantification of the trichrome-positive area as a percentage of the total tissue area. Scale bar, 100 μm; *p<0.01. (D) RT-qPCR quantification of fibrosis-related transcripts: TGF-β1, α-SMA, Col1α1, and fibronectin; *p<0.05 or ***p*<0.001. (E) Representative immunoblots for fibrotic markers (fibronectin and α-SMA), with GAPDH as a loading control. Band intensities were quantified using ImageJ software and are presented as bar graphs. This experiment was independently performed twice. Abbreviations: α-SMA, alpha-smooth muscle actin; CD, control diet; Col1α1, collagen type I alpha 1 chain; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; mRNA, messenger RNA; NKO, NQO1 knockout; NQO1, NAD(P)H:quinone dehydrogenase 1; RT-qPCR, reverse transcription-quantitative PCR; TGF-β1, transforming growth factor beta 1; WT, wild type.
    To further elucidate the molecular mechanism by which NQO1 deficiency modulates PSC pathogenesis, we conducted comparative RNA sequencing (RNA-seq) of WT and NKO liver tissues. Transcriptomic profiling identified 101 DEGs, comprising 43 upregulated and 58 downregulated genes, in NKO mice (Figure 5A). A heatmap of the top 20 most significantly altered genes highlighted the induction of Abca3, Fam213b, Nudt7, Cyp2a22, Mup14, and Mup11, whereas Nqo1, Cd300ld, Postn, Lgals1, and Inmt were repressed (Figure 5B). Among the DEGs, several were validated by reverse transcription-quantitative PCR (RT-qPCR) and exhibited expression patterns consistent with the RNA-seq data (Figure 5C). GSEA provided deeper insights into functional shifts driven by NQO1 ablation. We observed a positive enrichment of gene sets involved in “Bile acid metabolism,” “Oxidative phosphorylation,” and “Xenobiotic metabolism” in NKO mice. Conversely, pathways critical to disease progression, such as “EMT,” “G2M checkpoint,” and “TNFα signaling via NF-kB,” were downregulated (Figures 5D, E). These signatures underscore the pivotal role of NQO1 in orchestrating pro-fibrogenic signaling during chronic liver injury.
    Transcriptomic profiling of WT and NKO livers following DDC administration. To characterize the effects of NQO1 deficiency on hepatic gene expression under cholestatic conditions, RNA sequencing was performed using liver tissues from DDC-fed WT (n=3) and NKO (n=3) mice. (A) Volcano plot representing the distribution of differentially expressed genes (DEGs) between WT and NKO livers. (B) Clustered heatmap of DEGs exhibiting a greater than twofold difference in expression between the 2 groups. (C) RT-qPCR validation of several genes, including Bud13, Mup11, Mup14, Nudt7, Nqo1, and Postn, identified as differentially expressed in the RNA-seq dataset; p<0.05 or *p*<0.001. (D) GSEA plots for biological pathways significantly enriched in the NKO-DDC group. (E) GSEA plots for pathways downregulated in NKO mice compared with WT mice following DDC challenge. Abbreviations: DEGs, differentially expressed genes; GSEA, gene set enrichment analysis; NKO, NQO1 knockout; NQO1, NAD(P)H:quinone dehydrogenase 1; RNA-seq, RNA sequencing; RT-qPCR, reverse transcription-quantitative PCR; WT, wild type.
    Our DEG analysis identified a significant modulation of Postn (periostin), a known driver of tissue fibrosis.20 To explore the functional link between periostin and NQO1, we treated HepG2, LX-2, and primary hepatocytes with recombinant periostin. Western blot analysis revealed a time-dependent induction of NQO1 protein levels across all cell types (Figure 6A). Periostin is known to activate multiple signaling pathways, including FAK1, NF-κB, and PKB/AKT.21 To determine the mechanism by which periostin induces NQO1 protein expression, we performed pharmacological inhibition experiments. Interestingly, the periostin-induced NQO1 protein levels were attenuated by BAY 11-7082 and NAC, suggesting the involvement of redox-sensitive signaling in this response (Figure 6B). To confirm inhibitor activity, AKT phosphorylation, a downstream target of PI3K and FAK1 signaling, was examined. Although periostin did not induce AKT phosphorylation, both LY294002 and PF-573228 suppressed basal AKT phosphorylation, confirming their activity under our experimental conditions (Figure 6C). NQO1 protein expression remained unaffected by other pathological stimuli, such as TCA, TGF-β1, or TNFα, suggesting specificity in the periostin–NQO1 axis (Figure 6D). Functional assays further demonstrated that periostin challenge markedly elevated ROS production in both HepG2 cells and primary hepatocytes. Critically, this oxidative surge was effectively neutralized by either siRNA-mediated knockdown or genetic knockout of Nqo1 (Figure 6E). To further elucidate the impact of NQO1 deficiency on the antioxidant defense system, we examined the expression of various antioxidant enzymes that were previously found to be altered in DDC-treated livers. In HepG2 cells, siRNA-mediated Nqo1 knockdown robustly increased MnSOD expression regardless of periostin treatment, although other antioxidant proteins showed limited correlation with in vivo data (Figure 6F). In contrast, primary hepatocytes from NKO mice exhibited a protein profile consistent with DDC-injured livers, characterized by elevations in Prdx1, Prdx6, and MnSOD levels (Figure 6G). These results suggest that while NQO1 deficiency triggers a compensatory antioxidant response, a comprehensive and stable upregulation of these defenses, as observed in vivo, may depend on the sustained loss of NQO1 rather than a transient reduction.
    Periostin-induced NQO1 expression and its role in regulating cellular ROS responses. To investigate the signaling mechanisms responsible for periostin-mediated NQO1 induction and the contribution of NQO1 to oxidative stress, NQO1 expression and antioxidant responses were examined in vitro. (A) Representative immunoblots showing the time-dependent expression of NQO1 in HepG2, LX-2, and primary hepatocytes following recombinant periostin (100 ng/mL) treatment. The corresponding NQO1 bands are presented as graphs. This experiment was independently performed twice. (B) To identify the signaling pathways involved in periostin-induced NQO1 expression, cells were pretreated with BAY 11-7082 (10 μM), LY294002 (10 μM), PF-573228 (5 μM), or NAC (5 mM) before stimulation with periostin. NQO1 protein levels were determined by immunoblotting. This experiment was independently performed twice. (C) Immunoblot analysis of phospho-AKT and total AKT following pretreatment with BAY 11-7082, LY294002, or PF-573228 and subsequent periostin stimulation. Insulin was used as a positive control for AKT phosphorylation. This experiment was independently performed twice. (D) Immunoblot analysis of NQO1 expression following treatment with taurocholic acid (TCA; 100 μM), TGFβ1 (10 ng/mL), or TNF-α (10 ng/mL) for 24 hours. (E) Fluorescent DCF-DA imaging in periostin-stimulated HepG2 and primary hepatocytes. Scale bar: 100 μm. (F, G) Evaluation of antioxidant enzyme shifts in response to periostin following siRNA-mediated NQO1 knockdown and knockout. This experiment was independently performed twice. Abbreviations: DCF-DA, 2′,7′-dichlorofluorescin diacetate; FAK1, focal adhesion kinase 1; HepG2, human hepatocellular carcinoma cell line; NAC, N-acetyl-L-cysteine; NQO1, NAD(P)H:quinone dehydrogenase 1; ROS, reactive oxygen species; siRNA, small interfering RNA; TCA, taurocholic acid; TGF-β1, transforming growth factor beta 1; TNF-α, tumor necrosis factor alpha.
    To further delineate the functional contribution of NQO1 during prolonged cholestatic injury, we evaluated the effects of NQO1 deficiency after 3 weeks of DDC feeding. DDC feeding significantly elevated the plasma biochemical markers of liver injury; however, no significant differences were observed between the WT and NKO groups (Figure 7A). Histopathological assessment, including H&E and Masson trichrome staining, revealed comparable degrees of ductular reaction and collagen deposition in both genotypes, suggesting that NQO1 deficiency did not exacerbate chronic-phase fibrosis (Figure 7B). Furthermore, RT-qPCR analysis showed that NQO1-mediated suppression of pro-fibrotic genes, which was evident in acute models, was abolished after long-term DDC administration (Figure 7C). To explore a potential mechanism that might contribute to the loss of the early protective phenotype associated with NQO1 deficiency, LX-2 cells transfected with scrambled siRNA or siNQO1 were treated with TGF-β1 in a time-dependent manner (Figure 7D). Interestingly, TGF-β1–stimulated Smad2 phosphorylation was stronger and more sustained in NQO1-knockdown cells than in control cells (Figure 7D). These findings raise the possibility that altered TGF-β/Smad2 signaling may contribute to the loss of the early protective phenotype associated with NQO1 deficiency during prolonged cholestatic injury.
    Long-term effects of NQO1 deficiency in a DDC-fed PSC mouse model. To evaluate the effects of sustained NQO1 deficiency during chronic cholestatic liver injury, WT and NKO mice were maintained on either CD (n=5) or DDC (n=6) diet for 3 weeks. (A) Plasma levels of AST, ALT, ALP, and T-bil were quantified to assess hepatobiliary injury. (B) Histopathological evaluation of the liver via H&E and Masson trichrome staining in both groups. The Masson trichrome-positive area was quantified as a percentage of the total liver section area, and the results for DDC-fed WT and NKO groups are presented as a bar graph. Scale bar: 200 μm (H&E) or 100 μm (Masson trichrome). (C) RT-qPCR analysis of hepatic mRNA expression levels of the indicated genes. (D) LX-2 cells were transfected with scrambled siRNA or NQO1-targeting siRNA and subsequently stimulated with TGF-β1 for the indicated time periods. Representative immunoblots show the levels of phosphorylated Smad2 and total Smad2, with β-actin used as a loading control. Densitometric quantification of the corresponding protein bands is presented in the accompanying graphs. This experiment was independently performed twice. Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CD, control diet; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; H&E, hematoxylin and eosin; mRNA, messenger RNA; NKO, NQO1 knockout; NQO1, NAD(P)H:quinone dehydrogenase 1; RT-qPCR, reverse transcription-quantitative PCR; siRNA, small interfering RNA; T-bil, total bilirubin; TGF-β1, transforming growth factor beta 1; WT, wild type.

    DISCUSSION

    The current study elucidates the complex and stage-specific role of NQO1 in PSC pathogenesis. Traditionally, NQO1 is hailed as a cornerstone of the phase II antioxidant response, primarily functioning under the transcriptional control of the Nrf2 pathway.12 Its canonical biochemical role involves the two-electron reduction of quinones to stable hydroquinones, a process that prevents the formation of highly reactive semiquinone intermediates and subsequent superoxide radical generation.22 Furthermore, NQO1 is known to maintain the cellular NAD+/NADH ratio and protect key proteins, such as p53 and PGC-1, from proteasomal degradation.23 Given this extensive protective repertoire, the upregulation of NQO1 is typically interpreted as a compensatory mechanism to mitigate oxidative damage. However, our findings suggest a more complex role for NQO1 in the context of biliary injury. During the acute phase of DDC-induced injury, NKO mice exhibited significantly lower plasma ALT levels and fewer TUNEL-positive hepatocytes, whereas AST, ALP, and total bilirubin levels were not significantly different between genotypes. These findings suggest that NQO1 deficiency may confer partial protection against acute cholestatic liver injury, although the underlying mechanisms and their relevance during prolonged injury remain to be established.
    The resistance of NKO mice to DDC-induced damage appears to stem from robust compensatory realignment of the hepatic antioxidant network. In the absence of NQO1, we observed baseline and stress-induced elevations in Prdx1 and MnSOD levels, specifically within hepatocytes. This localized enhancement of the peroxiredoxin and superoxide dismutase systems likely provides a more targeted defense against lipid peroxidation, as evidenced by the marked reduction in 4-HNE-positive cells in NKO livers. Interestingly, while NQO1 is often linked to mitochondrial bioenergetics via NAD+ metabolism,24 our data showed comparable nucleotide levels across genotypes, indicating that protection in NKO mice is primarily driven by superior ROS scavenging rather than systemic metabolic shifts. This highlights a high degree of functional redundancy and plasticity in hepatic redox homeostasis, where the loss of one major antioxidant enzyme triggers a more effective alternative defense suite.25
    A key mechanistic discovery of this study was the characterization of the periostin–NQO1 axis as a specific driver of oxidative stress. Periostin, a matricellular protein known for its role in subepithelial fibrosis and tissue remodeling in chronic inflammatory diseases,26 is a potent inducer of hepatic NQO1. Although pro-fibrotic or pro-inflammatory stimuli, such as TGF-β and TNF-α, had negligible effects on NQO1 protein expression, periostin treatment directly triggered an NQO1-dependent oxidative surge. The finding that NQO1 silencing or genetic deficiency may contribute to oxidative stress under specific pathological conditions, although the relevant redox substrates and mechanisms remain unclear. Importantly, NQO1 effects are substrate-dependent: menadione is generally detoxified by NQO1-mediated two-electron reduction, whereas β-lapachone can undergo NQO1-dependent futile redox cycling.27–29 In cholestatic liver injury, altered bile acid levels, oxidative stress, and mitochondrial dysfunction may further modify NQO1-dependent redox effects.30,31 Thus, our findings do not establish NQO1 as a universal pro-oxidant factor but suggest that its effects depend on the prevailing substrate and cellular redox environment. This may also explain the apparent difference from our previous study, in which β-lapachone ameliorated PSC pathogenesis through NAD-related metabolic changes and Abcb4 induction, with protection lost in Mdr2/Abcb4-deficient mice.19 Therefore, pharmacological β-lapachone treatment should not be assumed to reproduce the effects of genetic NQO1 deficiency or endogenous NQO1 activity during DDC-induced cholestatic injury.
    During acute DDC-induced injury, NKO mice displayed a substantial reduction in collagen deposition and downregulation of core fibrogenic genes, including Col1α1, fibronectin, and TGF-β1. Our RNA-seq data further corroborated this finding, showing negative enrichment of gene sets related to EMT and TNFα signaling via NF-κB in NKO livers. These results suggest an association between NQO1 and fibrogenic signaling during the early phase of DDC-induced injury. Notably, the early protective phenotype associated with NQO1 deficiency was no longer evident after 3 weeks of DDC administration, when WT and NKO mice exhibited comparable biochemical injury markers, collagen deposition, and fibrogenic gene expression. These findings indicate that the consequences of NQO1 deficiency differ according to the duration and pathological context of cholestatic injury. One possible mechanism contributing to this temporal difference may involve altered TGF-β/Smad signaling. Sustained TGF-β signaling can suppress the Nrf2/NQO1 antioxidant program in HSCs, as TGF-β1 reduces Nrf2 and NQO1 expression through a Smad-dependent mechanism.32 Conversely, activation of the Nrf2 antioxidant pathway, accompanied by the induction of NQO1, attenuates TGF-β/Smad signaling and fibrogenic responses in HSCs and experimental liver fibrosis.33 Direct NQO1 overexpression has also been shown to suppress TGF-β/Smad activation and extracellular matrix accumulation in an experimental fibrotic disease model.34 Consistent with these previous studies, our in vitro results demonstrated that NQO1 loss leads to sustained TGF-β/Smad2 signaling. This reciprocal cross-talk may partly explain why the protective effect of NQO1 deficiency observed in acute DDC exposure was attenuated or reversed during chronic exposure.
    Certain limitations of the present study warrant consideration. Primarily, although our data demonstrate NQO1-dependent oxidative stress, the identity of the endogenous substrates triggering ROS generation under cholestatic conditions remains unknown. Identifying whether bile acid metabolites, oxidized catecholamines, or ubiquinone derivatives undergo futile redox cycling via NQO1 will require comprehensive metabolomic profiling.27 In addition, the mechanistic transition from early NQO1-mediated damage to a chronic fibrotic stage driven by multiple factors requires further investigation. Finally, in vivo experiments were restricted to male mice, rationalized by the higher incidence of PSC in men and the hepatoprotective nature of estrogen.35–37 However, this choice limits the generalizability of our results. Because sex influences Nrf2 pathways, antioxidant defenses, and susceptibility to oxidative stress, the function of the periostin–NQO1 axis in females may vary. Subsequent studies incorporating both sexes will be imperative to elucidate potential sex dimorphism and ensure broader clinical relevance.
    In sum, our study reveals a stage-dependent and context-specific role for NQO1 in cholestatic liver injury. During acute DDC exposure, NQO1 deficiency was associated with reduced ALT levels, lipid peroxidation, and fibrogenic responses. During prolonged DDC-induced injury, however, these early protective effects were not evident, and WT and NKO mice exhibited comparable fibrotic responses. Collectively, our findings indicate that the biological consequences of NQO1 depend on disease stage, substrate availability, and cellular redox context, factors that should be considered when evaluating NQO1 as a potential therapeutic target in cholestatic liver disease.
    HCHepatology Communications

    Hepatology Communications

    Official Account
    https://journals.lww.com/hepcomm
    hepatologycommunications