JAG1 c.1615C > T mutation impairs bile duct regeneration but not differentiation in hepatic organoids derived from a patient with Alagille syndrome Authors: Zirui Wan, Haiyan Zhang, Lu Zhang, Xiaofang Cao, Lisha An, Bin Wang, Jingyun Guan, Jinlin Jia, Chen Liu, Xiaohua Jin, Yue Li, Guohua Liu, Xu Ma Published: 2026 ABSTRACT Abstract Objective: Alagille syndrome (ALGS) is an autosomal dominant disorder caused primarily by mutations in JAG1. Intrahepatic bile duct paucity is the most consistently reported feature of ALGS; however, the mechanisms associated with biliary defects remain unclear. Methods: Here, we reverted or introduced the JAG1 c.1615C > T mutation to patient-specific induced pluripotent stem cells (iPSC) or human embryonic stem cells (hESCs) using base editors to generate isogenic cell lines and examine cellular phenotypes associated with this mutation in hepatic organoids. Results: We observed that JAG1 mutant lines showed no significant differences in cholangiocyte differentiation or function compared with the corresponding wild-type lines. By contrast, organoids derived from JAG1 mutant cells showed reduced cell survival, attenuated responsiveness to VEGF stimulation, and abnormal cell polarity. Conclusions: These findings offer a new insight into the biliary pathologies in ALGS and provide theoretical support for future proof-of-concept studies of novel therapeutic approaches for ALGS. Intrahepatic ductopenia is one of the most prominent features of Alagille syndrome. It remains unclear whether JAG1-associated biliary defects are due to cholangiocyte differentiation defects, biliary repair defects, or both. Here, we combined stem cell-derived hepatic organoids with a base editing system to model biliary phenotype of Alagille syndrome in isogenic systems. Our findings offer an insight into the biliary pathologies in Alagille syndrome and provide theoretical support for future proof-of-concept studies of novel therapeutic approaches for Alagille syndrome. Introduction Alagille syndrome (ALGS), an autosomal dominant disorder with an incidence of approximately 1:30 000 live births, was first described by Daniel Alagille in 1969 as a multisystem disorder with variable phenotypic penetrance, including chronic cholestasis, cardiovascular abnormalities, butterfly vertebrae, posterior embryo toxic, renal anomalies, vascular involvement, and characteristic facies.1,2 JAG1, which is involved in a variety of cell-fate decisions that affect organ development and function, has been identified as the predominant disease-associated gene, with pathogenic variants occurring in nearly 95% of individuals with ALGS.3,4 In addition, approximately 2.5% of patients with ALGS carry pathogenic NOTCH2 variants.5 Although approximately 700 pathogenic JAG1 and NOTCH2 variants have been reported in ALGS, no consistent genotype-phenotype correlation has been observed, and the basis for the heterogeneity of clinical features remains unclear.6–8 To date, therapeutic interventions for ALGS have been mainly supportive and directed toward specific clinical manifestations rather than the underlying pathophysiology of disordered NOTCH signaling.9 Congenital intrahepatic ductopenia was characterized as the most prominent and nearly constant feature of ALGS and as the criterion of this disease for early diagnosis at one time.10 However, analyses of sequential liver biopsies have shown that intrahepatic bile ducts can appear normal during the fetal period but decline during postnatal liver growth in some reports.11–13 Previous studies have demonstrated that cholangiocyte differentiation defects and/or ductal plate malformation are involved in the biliary abnormalities of ALGS.14–16 Moreover, the absence of periphery bile ducts despite preserved central bile ducts in some patients with ALGS has suggested that a biliary repair/regeneration defect may also be involved.13,17 A number of studies have shown that JAG1 regulates biliary tree development during fetal and postnatal periods,18 but how JAG1 mutations are associated with biliary defects in ALGS remains incompletely understood. Rodent and zebrafish models have been generated to investigate the phenotype and mechanism of intrahepatic bile duct defective in ALGS; however, these models do not reproduce the full syndrome,16,19 and more importantly, pathogenic mechanisms are quite different among species, which animal models are based on homozygous mutation of Jag1 or Notch2 or Jag1/Notch2 compound heterozygous, whereas ALGS in humans is generally caused by heterozygous for either JAG1 or NOTCH2 mutations.20 Therefore, a system that more closely reflects the genetic context of ALGS is needed to study how these mutations are associated with human liver development. Over the past decade, a variety of hepatic organoids have been derived from liver stem cells and pluripotent stem cells.21 Recently, Guan et al.22 established a hepatic organoid culture system that mimics human bile duct development in vitro. This system differentiates human induced pluripotent stem cells (iPSCs) into primary hepatic organoids (HO1s) to model embryonic bile duct development. After HO1s’ dissociation and reseeding in Matrigel, secondary HOs (HO2s) with luminal structures can be generated. These HO2s express early biliary lineage markers but not mature hepatocyte markers, indicating that they resemble reactive ductular cells (RDCs). Given the role of these progenitor cells in injury-induced biliary responses, this HO2 system provides an opportunity to investigate whether ALGS biliary defects are related to differentiation defects, repair/regeneration capacity impairment, or both. Additionally, the CRISPR-Cas9-mediated base editors (BE) have emerged as powerful tools for precisely reverting or introducing disease-causing mutations in iPSCs with high efficiency.23 Therefore, combining iPSCs-derived hepatic organoids with genome editing technologies may help determine how disease-causing mutations are associated with biliary phenotypes in patients with ALGS. In this report, we generated iPSC and human embryonic stem cell (hESC) cell lines carrying the JAG1 c.1615C > T mutation by reprogramming a patient’s peripheral blood mononuclear cells (PBMCs) or by genome engineering to investigate cellular phenotypes associated with intrahepatic bile duct defects in ALGS. We also evaluated whether correction of this pathogenic variant could reverse the organoid phenotype in vitro. We found that JAG1 c.1615C > T was not associated with overt disruption of cholangiocyte differentiation but was associated with altered regenerative-associated properties in this system. Our results suggest potential cellular mechanisms related to biliary defects and provide proof-of-principle data that gene correction of the JAG1 mutation may attenuate disease-associated phenotypes in vitro. Methods Human ESCs and iPSCs culture IPSC carries a heterozygous JAG1 c.1615C > T mutation was derived from a 3-month male patient with ALGS.24 The H9mut line was generated by using the cytosine base editor to introduce JAG1 c.1615C > T mutation in the H9 lines as described in our previous study.25 Collection of clinical information and venous blood from the patient was approved by the Medical Ethics Committee of Children’s Hospital Affiliated to Shandong University, approval number: SDFE-IRB/P-2022015. All cell lines were maintained on Matrigel-coated plates in mTeSR plus medium (StemCell Technologies) and cultured at 37 °C in 5% CO2. All cells are passaged with ReLeSR™ (StemCell Technologies) at a ratio of 1:6 every week. Base editor construction and transfection The plasmids, NGG-ABEmax (addgene #112095), NG-ABE8e (addgene #138491), and pGL3-U6-sgRNA-PGK-puromycin (addgene #51133) were obtained from Addgene (https://www.addgene.org/). The target-specific sgRNA with the protospacer sequence 5′-gcactaggcaccgttctggc-3′ or 5′-actaggcaccgttctggcag-3′ corresponding to NGG-ABEmax and NG-ABE8e was cloned into the pGL3-U6-sgRNA-PGK-puromycin vector, respectively. To correct JAG1 c.1615C > T mutation in ALGS iPSCs line, cells were electroporated with plasmids as follows. 1 × 106 single cells were electroporated with the 4D Nucleofector X System (Lonza) by using the P3 Primary Cell Kit with the program CM138, the quantity of the BE and sgRNA in each reaction was 7.2 μg and 2.25 μg, respectively. Electroporated iPSCs were cultured and selected by using 1 μg/mL puromycin for 24 h. Antibiotic-selected cells were harvested for editing efficiency analyses and single-cell flow cytometry sorting. When cell colonies were large enough to be picked, clones were passaged and analyzed using Sanger sequencing for genotype detection. SgRNA-dependent off-target analysis and deep sequencing Base-edited and unedited iPSCs were harvested for sgRNA-dependent off-target analysis. The on-target and potential off-target sites were predicted by Cas-OFFinder (http://www.rgenome.net/cas-offinder/). One on-target and 21 off-target sites were selected ultimately and amplified using DNA polymerase. The purified PCR products were submitted to sequence on Illumina platforms with PE150 strategy in Novogene Bioinformatics Technology Co., Ltd (Beijing, China). Approximately 1 GB clean data was generated for each sample. The Genome Center platform (National Human Genetic Resources Center) was used to process the data. Editing efficiencies of NGG-ABEmax and NG-ABE8e were calculated using silico analysis EditR. Primers used for deep sequencing are listed in Supplementary Table S1. Whole-genome sequencing Whole-genome sequencing (WGS) of NGG-ABEmax-treated and untreated iPSCs were analyzed to search for base editing-induced SNVs and InDels variants. The genomic library was prepared from 0.2 μg of genomic DNA for each sample. The sequencing (depth of 30x) was performed on Illumina HiSeq X Ten (2 × 150 PE) at Novogene Bioinformatics Technology Co., Ltd. Valid sequencing data was mapped to the reference genome (GRCh37/hg19) by Burrows-Wheeler Aligner software to get the original mapping result in BAM format. SAMtools were used to do variant calling and identify SNPs and InDels. The common SNPs between the edited and unedited iPSCs were excluded depending on the database of single nucleotide polymorphisms (dbSNPs). Human hepatic organoids differentiation Hepatic organoids were generated according to a previously described differentiation protocol.22 In brief, iPSCs were dissociated into single cells and cultured in suspension to form endoderm spheres in DMEM/F12 medium containing growth factors (Activin A, BMP-4, CHIR99021, bFGF, and LY294002). Foregut- and hepatoblast-like spheroids were then formed in the presence of FGF10 combined with BMP-4. After 10 days of cultured in HCM medium containing HGF, oncostatin M, and dexamethasone, spheroids were developed into a more mature type named HO1s. These HO1s have the capacity to be dissociated and reformed secondary HOs (HO2s) to generate reactive ductal cells (RDC)-like cells. These progenitor-like cells possess properties associated with regeneration. Whole-mount staining Spheroids were devoid of matrix and fixed by 4% paraformaldehyde for 1 h at room temperature. After washed with PBS, spheroids were incubated in a blocking solution containing 3% BSA and 0.5% Triton X-100 for 2 h at room temperature. Then organoids were incubated in primary antibody at the appropriate dilution in a blocking solution overnight at 4 °C. After three washes with PBS at room temperature for 10 min, organoids were incubated with the diluted secondary antibody in a blocking solution in the dark for 1 h. Organoids were then washed three times and counterstained with DAPI for 10 min, and then organoids were placed in a glass-bottom cell culture plate and imaged using a Leica LSM800 confocal microscope. Primary antibodies and secondary antibodies were shown in Supplementary Table S3. Rhodamine 123 transport assay Organoids were incubated with 25 μM rhodamine 123 for 1 h at 0 °C and washed with pre-cooled RPMI/B27 medium three times. To demonstrate that rhodamine 123 transfer indeed reflected the activity of the membrane channel MDR1, a transport protein expressed in the membrane of mature bile duct cells, organoids were incubated with 10 µM verapamil at 37 °C for 1 h and the rhodamine assay was repeated. After washing three times, images were taken using a confocal microscope. VEGF stimulate cell proliferation assay A VEGF stimulates cell proliferation assay was performed as described previously with slight modifications.26 In brief, twelve 50-μL droplets of Matrigel, each mixed with 40 000 single cells dissociated from HO2, formed a Matrigel dome to seed in 12 wells of a 24-well plate. VEGF at a concentration of 50 ng/mL was added to 6 of the wells with every medium change. After 5 days culture, Matrigel was digested and organoids were dissociated into single cells. The number of cells for each Matrigel dome was then counted using a hemocytometer. Measurement of HO2 To quantify the number of HO2s, three Matrigel domes were randomly selected per independent differentiation experiment, and a total of nine Matrigel domes were counted for each group. For organoid size quantifications, a total of 300 organoids were randomly selected and quantified for each group. Diameter measurements of HO2s were made using a CapStudio software. Statistical analyses All statistical analyses were performed using SPSS (Version 23). Data were presented as mean ± SD. For comparison between two mean values, an unpaired Student’s t-test was used to calculate statistical significance. For comparison between multiple values, one-way ANOVA was used with the post hoc Bonferroni test for experiments with more than two groups. P-values less than .05 were considered significant. Results Generation and characterization of iPSCs-derived hepatic organoids To mimic ALGS-related disease model in a dish, we differentiated iPSCs into human hepatic organoids (HOs) following a previously described method (Figure 1A).22 Cells or organoids at each stage exhibited specific morphological features (Figure 1B) and expressed signature markers, including RDC markers CK19 and SOX9, which indicate regenerative potential. (Figure 1C). Furthermore, iPSC-derived hepatic organoids exhibited gamma-glutamyl transferase (GGT), alkaline phosphatase activity (Figure 1D and E) and secretory potential, as assessed by Rhodamine 123 transport activity (Figure 1F). These results supported the use of this organoid model to assess cholangiocyte differentiation and progenitor-like organoid properties relevant to biliary repair/regeneration. Generation and characterization of hepatic organoids (HOs) derived from iPSCs. (A) Overview of the protocol used to differentiate iPSCs to HOs. EN, definitive endoderm; FG, foregut; HB, hepatoblast; HO1s, primary HOs; HO2s, secondary HOs. (B) Bright field microscopy images reveal the morphological changes of cell line at each stage of differentiation. Scale bars, 200 µm. (C) Immunofluorescence analyses demonstrating the signature markers of each stage. Scale bars, 50 µm. (D) ALP staining revealing ALP activity in HO1s. MEF, mouse embryonic feeders. Scale bars, 200 µm. (E) HO1s exhibit GGT activity. Positive control was provided as part of the kit components and mouse embryonic feeders (MEF) were used as a negative control. Mean ± SD, ns P > .05, ***P < .001 (unpaired Student’s t test, n = 3 independent differentiation experiments). (F) Fluorescence intensity measurements demonstrating the rhodamine123 detected in the lumen of HOs confirming MDR1 functionality. Scale bars, 50 µm. Revert JAG1 c.1615C > T mutation in patient-derived iPSCs using adenine base editors To investigate cellular phenotypes associated with the JAG1 mutation during bile duct development, we generated iPSC lines (referred to as “ALGS”) from PBMCs donated from a patient with ALGS carrying a heterozygous JAG1 c.1615C > T mutation. To confirm that the phenotypes were attributable to JAG1 c.1615C > T, we used an adenine base editing system to correct this mutation in ALGS iPSC lines to generate isogenic iPSC lines (referred to as “ALGScorr”). Gene targeting at JAG1 c.1615C > T mutation site in ALGS iPSCs was performed using the adenine BE system. sgRNAs corresponding to NGG-ABEmax and NG-ABE8e were designed and constructed (Figure 2A). Results showed that both NGG-ABEmax and NG-ABE8e could correct JAG1 c.1615C > T mutation effectively with an approximate 30% editing efficiency (Supplementary Figure S1A and B). However, the application of NG-ABE8e would result in bystander editing within the active base editing window (Supplementary Figure S1C). To revert the mutant site accurately, we performed gene editing with NGG-ABEmax in subsequent experiments. ABE-mediated correction of a heterozygous mutation c.1615C > T in the JAG1 gene in patient-specific iPSCs. (A) Design of sgRNAs for correction of JAG1 c.1615C > T mutation. The patient with ALGS harbors a heterozygous mutation at nucleotide position c.1615 (shown in red font) causing a glutamine to stop codon mutation in exon 13 at amino acid residue 539. The blue or red frame was indicated the target strand, and blue or red dashed lines connecting the sgRNA-1/2 which would displaced the PAM-containing genomic DNA strand to form a single-stranded DNA R-loop. (B) Schematic representation of gene editing and flow cytometry sorting. (C) Sequencing chromatograms of JAG1 sequence fragments in ALGS and ALGScorr iPSC lines. Red arrows indicate heterozygous mutation and the corresponding corrected nucleotides. (D) Western blot analyzes the JAG1 expression in gene mutant and corrected lines. Mean ± SD, *P < .05 (unpaired Student’s t test, n = 3 independent differentiation experiments). After confirming the strategy for JAG1 c.1615C > T mutation correction, NGG-ABEmax and the corresponding sgRNA-1 plasmid were nucleofected into ALGS iPSCs. The workflow of antibiotic selection and single-cell clone preparation is presented in Figure 2B. A total of 24 colonies were isolated and evaluated by Sanger sequencing, and 18 (75%) of them confirmed that the targeted mutation was corrected (Supplementary Figure S1D). We then selected gene-corrected clones randomly for further characterization and experiments. We observed that ALGScorr lines had undergone on-target editing at JAG1 c.1615C > T mutation site, and the JAG1 protein level was significantly increased following gene correction (Figure 2C and D). Furthermore, ALGScorr iPSCs were characterized, and the results shown in Supplementary Figure S2 described that although ALGScorr iPSCs had undergone base editing, cells displayed typical iPSCs characteristics. Off-target detection in base editing-corrected iPSC lines To investigate whether sgRNA-dependent off-target occurred during gene correction, 21 potential off-target sites with up to 5-nt mismatch for the correctional sgRNA were selected preferentially using Cas-OFFinder (Figure 3A). After these potential off-target sites were amplified by PCR from six corrected clones (edited iPSC 1-6) and three unedited clones (Control-1, Control-2, and Control-3), the PCR products were equally mixed and subjected to deep sequencing. Data analysis showed that no obvious off-target editing or InDels were detected (Figure 3A). We then performed WGS to identify potential off-target sites in edited and unedited clones. After excluding false-positive sites based on the database of single nucleotide polymorphisms (dbSNPs), no off-target sites (SNPs or InDels) were detected (Figure 3B and C and Supplementary Table S1). These results revealed that genome editing using NGG-ABEmax did not give rise to obvious off-target events in the modified iPSCs. Off-target analysis of base edited iPSCs. (A) Deep sequencing to assess sgRNA-dependent off-target sites in edited and unedited cells. The tables in left column show the potential 21 off-target sites identified by Cas-OFFinder. And the percentage of A-to-G within a 21-nt window near the PAM-distal end of the protospacer at potential off-target sites was calculated (right column). Letters in lower case represent mismatches between the on- and off- target sites. Letters in bold represent the PAM sequence. Data are from six edited and three unedited clones, error bars show standard error of measurement. (B, C) WGS to detected SNPs and InDels among the whole genome in edited and control cells. Data are from 6 edited and 2 unedited clones, error bars show standard error of measurement. JAG1 c.1615C > T did not inhibit cholangiocytes differentiation Previous studies have shown that some JAG1 mutations can impair cholangiocyte differentiation during bile duct development.14,15 To assess whether JAG1 c.1615C > T affects cholangiocyte differentiation, we differentiated ALGS and its isogenic cell line, ALGScorr, into HOs using the established protocol in biological triplicate. We also differentiated H9 and its isogenic cell lines (referred to as “H9mut”), in which JAG1 c.1615C > T was introduced into H9 using a cytosine base editing system,25 to further confirm the phenotypes were indeed a consequence of JAG1 c.1615C > T mutation. After 20 days of differentiation, HO1s derived from both wild-type lines (H9 and ALGScorr) and JAG1 mutant lines (ALGS and H9mut) developed intact structures with no significant differences observed in stereomicroscopic images (Figure 4A). Immunofluorescence analyses showed that organoids produced from all cell lines reached a cholangiocyte endpoint, as indicated by expression of the mature biliary markers CK7 and CFTR (Figure 4B and C). Meanwhile, real-time quantitative analyses of SOX17, HNF4α, and CK7 expression in definitive endoderm (EN), hepatoblast (HB), and HO1s, respectively, showed no significant differences during the differentiation process in each group (Figure 4D-F). Moreover, mRNA expression of NOTCH2 (the receptor of JAG1) and its downstream target genes NRARP, HES1, and HEY1 was not significantly altered at the HB stage, further supporting that differentiation was not impaired in the JAG1 mutant lines (Figure 4G-J). JAG1 c.1615C > T influence on HO1s differentiation and function in an isogenic system. (A) Bright-field images taken on day 20 show the morphology of JAG1 mutant and wild-type cells derived HO1s. Scale bars, 200 µm. (B, C) Immunofluorescence analyses demonstrating the expression of mature cholangiocyte markers, CK7 and CFTR, in JAG1 c.1615C > T mutant cell lines and their isogenic lines. Scale bars, 50 µm. (D-F) qPCR analyses of EN, HB and HO1 signature marker expression in JAG1 mutant and wild-type lines. Mean ± SD, ns P > .05 (unpaired Student’s t test, n = 3 independent differentiation experiments). (G-J) NOTCH2, NRARP, HES1, and HEY1 expression of JAG1 mutant and wild-type line-derived cells at the hepatoblast stage. Mean ± SD, ns P > .05 (unpaired Student’s t test, n = 3 independent differentiation experiments). (K) Fluorescence intensity within the lumen of organoids carrying JAG1 c.1615C > T mutation or wild-type alleles with or without verapamil treatment. Mean ± SD, ns P > .05, **P < .01 (One-way ANOVA with post hoc analysis, n = 3 independent differentiation experiments). (L) GGT activity assay in gene mutant and wild-type organoids. Positive control was provided as part of the kit components, and mouse embryonic feeders (MEF) were used as a negative control. Mean ± SD, ns P > .05 (One-way ANOVA with post hoc analysis, n = 3 independent differentiation experiments). Next, we characterized the functionality of the generated organoids. Rhodamine 123 transport experiments showed that JAG1 mutant lines-derived (ALGS and H9mut) HO1s could transport rhodamine 123 into their intraluminal space while luminal dye accumulation was prevented by the MDR1 inhibitor verapamil, in keeping with wild-type lines (H9 and ALGScorr) (Supplementary Figure S3). Quantification of luminal fluorescence also showed no significant differences among groups (Figure 4K). Similarly, GGT activities exhibited no significant differences among HO1s derived from H9, H9mut, ALGS, and ALGScorr lines (Figure 4L). These results suggested that HO1s could be differentiated from JAG1-deficient lines and maintain mature cholangiocyte-like functions comparable to wild-type lines. JAG1 c.1615C > T impairs cell survival in HO2s Following bile duct injury, activation and proliferation of RDCs are considered important for biliary repair and regenerative responses.18,27,28 We generated HO2s from ALGS, ALGScorr, H9, and H9mut line-derived HO1s to assess whether JAG1 c.1615C > T affects the regenerative-associated properties of the progenitor-like organoid. After 10 days of expansion post dissociation, organoids with cyst-like structures derived from H9 and ALGS lines reassembled (Figure 5A) and expressed RDC markers CK19 and SOX9 but not the hepatic lineage marker ALB (Figure 5B). Although the diameter of newly generated organoids did not differ significantly among these lines, the number of HO2s derived from ALGS lines was significantly lower than that derived from H9 lines (48% decrease, P < .001). After reversion of the JAG1 c.1615C > T mutation, the number of intact HO2s derived from ALGScorr lines recovered (284% increase, P < .001; Figure 5C-E). Cleaved caspase-3 staining revealed apoptosis in the organoid lumen of HO2s derived from ALGS lines, whereas H9 and ALGScorr-derived HO2s showed little evidence of apoptosis (Figure 5F and G). Notably, H9mut lines showed a more severe phenotype, in which there were no intact structures formed from H9mut line-derived organoids (Figure 5A). Together, these results suggest that HO2s generation from ALGS and H9mut line is impaired to varying degrees, that caspase-3-associated apoptosis may contribute to this process, and that gene correction of JAG1 c.1615C > T can partially rescue HO2 formation in vitro. JAG1 c.1615C > T mutation impairs the RDCs-like cells survival in HO2s. (A) Bright-field images of HO2s formed after JAG1 mutant or wild-type organoids were dissociated. The graph shows the intact organoids or vesicles. Scale bars, 200 µm. (B) Immunofluorescence analyses demonstrating the expression of RDC markers, CK19 and SOX9, but not the hepatic lineage marker ALB, in JAG1 c.1615C > T mutant lines and its isogenic lines derived organoids. Scale bars, 50 µm. (C-E) Representative bright-field images of HO2s (C) used to count the number (D) and measure the diameter (E) of intact organoids derived from JAG1 c.1615C > T mutant and wild-type lines. Mean ± SD, ns P > .05, ***P < .001 (One-way ANOVA with post hoc analysis, n = 3 independent differentiation experiments). (F) Immunofluorescence of cleaved caspase-3 expression in differentiated JAG1 mutant and wild-type organoids. Scale bars, 50 µm. (G) Quantification of apoptotic cells in wild-type and JAG1 mutant lines derived organoids, mean ± SD, ns P > .05, ***P < .001 (One-way ANOVA with post hoc analysis, n = 3 independent differentiation experiments). JAG1 c.1615C > T mutation is associated with reduced HO2s responsiveness to VEGF Previous studies have shown that vascular endothelial growth factor (VEGF) plays an important role in biliary repair and regenerative responses.29–31 We evaluated the responsiveness of HO2s derived from JAG1 mutant and wild-type cell lines to VEGF. VEGF administration significantly increased the number of RDC-containing cells derived from the H9 lines (27% increase in expansion, P < .001), whereas ALGS lines showed only a modest and statistically nonsignificant increase after VEGF treatment (4% increase in expansion, P > .05). After correction of JAG1 c.1615C > T, ALGScorr lines-derived RDCs-like cells regained partial responsiveness to VEGF (13% increase in expansion, P < .001; Figure 6A). EdU analyses of cell proliferation in the presence or absence of VEGF were consistent with these findings (Figure 6B and C). These results indicate that ALGS lines-derived HO2s are associated with attenuated responsiveness to VEGF, whereas H9 and ALGScorr lines-derived RDCs-like cells respond positively to VEGF stimulation. Decreased proliferation in response to VEGF stimulation in JAG1 mutant HO2s. (A) Fold change over starting number of cells after 5 days culture in the presence or absence of VEGF. Mean ± SD, ns P > .05, **P < .01, ***P < .001 (unpaired Student’s t test, n = 3 independent differentiation experiments). (B, C) EdU assays were used to determine the proliferation ability of HO2s with or without VEGF treatment. Immunofluorescence (B) and quantification (C) of cells taking up EdU indicated that JAG1 c.1615C > T mutant organoids show impaired cell proliferation ability in response to VEGF stimulation. Scale bars, 50 µm. Mean ± SD, ns P > .05, *P < .05 (unpaired Student’s t test, n = 3 independent differentiation experiments). (D and E) VEGF and VEGFR2 expression in JAG1 mutant and wild-type lines derived HO2s. Mean ± SD, ns P > .05, *P < .05, (unpaired Student’s t test, n = 3 independent differentiation experiments). To explore a potential mechanism associated with this phenotype, we quantified VEGF and VEGFR2 expression in JAG1 mutant and wild-type organoids. VEGFR2 expression was significantly decreased in JAG1 mutant organoids compared with wild-type organoids, and after gene correction, VEGFR2 expression was restored (Figure 6D). In contrast, VEGF expression did not differ significantly among groups (Figure 6E). Given the critical role of VEGF/VEGFR2 signaling in stimulating bile duct proliferation in autocrine and paracrine manners,32 our findings suggest that reduced VEGFR2 expression may contribute to the attenuated response of RDC-like cells to VEGF stimulation and to impaired regenerative-associated properties in ALGS organoids. JAG1 c.1615C > T mutation is associated abnormal cell polarity and intercellular adhesion in HO2s Transcriptomic analyses of patients with ALGS and mouse models have shown that cell polarity disruption is associated with morphogenesis defects during bile duct development, repair, and regeneration.20 To determine whether cell polarity was disrupted in HO2s derived from ALGS iPSCs, we performed whole-mount immunostaining of organoids with F-actin and β-catenin, markers of the apical and basolateral domains, respectively. Immunofluorescence confocal microscopy showed apical localization of F-actin and basolateral localization of β-catenin in H9-derived HO2s, whereas ALGS-derived HO2s showed expanded F-actin staining at both the apical and basal sides of the HO2s lumen. After gene correction, ALGScorr-derived HO2s showed an improved apical-basolateral staining pattern, with only a few focal areas of aberrant F-actin expression (Figure 7A). Quantification of F-actin and β-catenin fluorescence intensity at the apical and basolateral domains of HO2s also showed abnormal polarity in JAG1 mutant organoids (Figure 7B and C). As intercellular junctions are essential for the establishment of epithelial cell polarity, we further investigated the distribution of zona occludens 1 (ZO-1), a marker of apical junctions, in HO2s. We observed that ZO-1 was aberrantly localized with some junctions missing the signal, while other cell junctions had extra ZO-1 puncta in JAG1 mutant HO2s compared with wild-type HO2s. After gene correction, abnormal ZO-1 localization was partially restored (Figure 7D-F). Subsequently, we quantified the mRNA expression of adhesion molecules in HO2s (Figure S4) and observed significantly decreased NCAM expression in JAG1 mutant HO2s (Figure 7G). These results suggest that HO2s derived from the patient with ALGS carrying JAG1 c.1615C > T are associated with abnormal cell polarity and altered cell-cell adhesion, which may contribute to intrahepatic bile duct defects. Abnormal cell polarity and cell-cell junctions in JAG1 mutant HO2s. (A) Representative images demonstrating localization of F-actin as an apical marker and β-catenin as a basolateral marker in wild-type organoids, whereas JAG1 mutant organoids show expanded distribution of these proteins. Scale bars, 50 µm. (B, C) Fluorescence intensity measurements of F-actin and β-catenin in wild-type and JAG1 mutant organoids. Mean ± SD, ns P > .05, **P < .01, ***P < .001 (unpaired Student’s t test, n = 3 independent differentiation experiments). (D) Representative images demonstrating localization of ZO-1 in JAG1 mutant and wild-type HO2s. The white dashed frame indicates the magnified region. Scale bars, 50 µm. (E, F) Percentages of junctions missing ZO-1 and with extra ZO-1 in JAG1 mutant and wild-type HO2s. Mean ± SD, *P < .05, **P < .01 (unpaired Student’s t test, n = 3 independent differentiation experiments). (G) Expression of NCAM in JAG1 mutant and wild-type HO2s. Mean ± SD, *P < .05, **P < .01 (unpaired Student’s t test, n = 3 independent differentiation experiments). Discussion It was once thought that biliary defects typically present in the fetal period in patients with ALGS, while in some recent cases, the development of intrahepatic bile ducts appears to be normal during the neonatal period but damaged in postnatal liver growth.13,17 Based on this phenomenon, researchers speculate that the ductal plate malformations resulting from cholangiocyte differentiation defects in the fetal period and/or inability of patients with ALGS to mount a ductular reaction to repair biliary damage during chronic cholestasis after birth may contribute to the progression of bile duct paucity, although the exact mechanisms remain unclear. Although animal models, such as mice and zebrafish, have been generated to mimic ALGS by targeting Jag1 and/or Notch2,16,33–37 an important genetic distinction should be noted: ALGS in humans is generally caused by heterozygosity for JAG1, whereas in mice the phenotype often arises in the homozygous state. Moreover, the mouse liver microenvironment differs from that of humans, raising the possibility that conclusions based on rodent models may not fully translate across species.38 Therefore, patient-specific iPSCs-derived organoids provide a practical tool to mimic ALGS in vitro and to study the relationship between mutations and cellular pathomechanisms. In the present study, we used stem cell-derived hepatic organoids combined with base editing system to investigate how JAG1 c.1615C > T mutation of the patient with ALGS is associated with biliary phenotypes. When JAG1 mutant cell lines (ALGS and H9mut) were differentiated into hepatic organoids, organoid morphology, expression of representative markers at each differentiation stage, and cholangiocyte-like cell function were indistinguishable from those of the wild-type cell lines (ALGScorr and H9). These results suggest that the JAG1 c.1615C > T mutation from this patient with ALGS is not associated with impaired differentiation or function of cholangiocyte-like cells in this model. The JAG1/NOTCH2/HES1 and/or JAG1/NOTCH2/HEY1 signaling pathways play important roles in intrahepatic bile duct development. However, unexpectedly, expression of NOTCH2, NRARP, HES1, and HEY1 was not affected in JAG1 mutant lines. Similar phenotypes have also been observed in mouse models, patient biopsies, and patient-specific organoids.13,17,20,39 We hypothesize that additional factors may contribute to activation of the JAG1/NOTCH2 signaling pathway during intrahepatic biliary development. Collectively, these results suggest that, at least in a subset of patients with ALGS, disease-causing mutations may not substantially affect cholangiocyte differentiation. We then investigated whether JAG1 c.1615C > T is associated with altered regenerative-associated properties of progenitor-like organoids. Cells in the HO2 stage predominantly expressed the early biliary lineage markers CK19 and SOX9, but not the hepatic lineage marker ALB, indicating that they resemble reactive ductular cell (RDC)-like cells.22 Although the definition of RDCs remains inconsistent across previous studies, CK19+/SOX9+ cells are commonly used as a specific marker to evaluate ductular reactions during biliary injury.28,40 Accordingly, the model employed in this study may be used, with caution, to investigate how JAG1 mutations are associated with progenitor-like organoid properties relevant to biliary repair and regeneration. After HO1s were dissociated into single cells and reseeded in Matrigel, both JAG1 mutant and wild-type lines reassembled into intact organoids with luminal structures and expanded for several days in growth medium. Previous studies have reported that a number of organoids derived from Jag1 mutant mice collapsed during the culture process, which may indicate inherent structural instability.20 However, no such phenomenon was observed in the present study. One possible explanation is that organoids investigated in prior reports were generated from primary hepatic tissue, whereas those established in our study were derived from iPSCs, thus, the observed discrepancies may be attributable to distinct cellular origins. Although organoid morphology did not differ significantly among groups, the organoid count was significantly decreased in JAG1 mutant lines. More severely, H9mut line-derived HO2s were almost completely unable to form intact organoids with luminal structures. Additional findings suggest that caspase-3-associated reduction in RDC-like cell viability may contribute to this process. These findings are consistent with previous observations.22,41 A number of previous studies have demonstrated that, following biliary injury, RDCs can be activated and proliferate to restore bile duct mass and rebuild the biliary network in humans and rodents.17,18,42,43 However, in some patients with ALGS, this reaction is not preserved, and RDCs are almost absent in liver biopsy.17 Given the role of the JAG1/NOTCH2 signaling in ductular reactions and biliary repair during postnatal life, impaired survival of RDCs-like cells may contribute to ductopenia in patients with ALGS. Furthermore, we analyzed HO2s responsiveness to VEGF in JAG1 mutant lines and isogenic wild-type lines. Our results showed that the number of RDC-like cells derived from ALGS lines did not increase significantly after VEGF treatment, whereas after correction of JAG1 c.1615C > T, cell proliferation in response to VEGF stimulation was partially restored. Further analysis showed that VEGFR2, but not VEGF, expression was decreased in HO2s carrying JAG1 c.1615C > T. During biliary damage, VEGF/VEGFR2 signaling was the key inducer of ductular reaction, which unlocked the proliferative potential of RDCs via autocrine and paracrine mechanisms to repair bile ducts.32,43,44 We therefore propose that reduced VEGFR2 expression may contribute to attenuated VEGF responsiveness and alter the regenerative-associated properties in organoids. Although the role of VEGF/VEGFR2 signaling in bile duct repair has been examined in liver-injured mouse models, our data in the present study should be considered associative and warrant further validation by direct pathway activity assays. It has been confirmed that cell polarity affects bile duct morphogenesis and maintenance at postnatal and adult stages in patients with ALGS and in mouse hepatic organoids.20 The present findings are consistent with this concept. We observed that organoids derived from patient-specific iPSCs showed abnormal cell polarity, whereas after correction of JAG1 c.1615C > T, the cell polarity was improved. Moreover, immunofluorescence staining of ZO-1 showed mislocalization, further supporting cell polarity defects. Previous studies have shown that aberrant cell polarity is associated with defective biliary morphogenesis and limited formation of a normal biliary network.45 Consistent associations between polarity defects and impaired morphogenesis have also been observed in neural tube and zebrafish lateral lines.46,47 Given the association between cell polarity and organ development, abnormal cell polarity may contribute to defective bile duct development in patients with ALGS. In addition, quantification of cell adhesion molecules revealed that NCAM expression was significantly decreased in JAG1 mutant HO2s. NCAM is a well‑characterized cell adhesion molecule implicated in the migration of RDCs from the periportal niche to the areas requiring repair during bile duct injury.48 Downregulation of NCAM may therefore contribute to impaired RDC migration toward peripheral hepatic regions during biliary repair and regeneration. These findings suggest a potential, but not directly proven, mechanism that may be related to the absence of bile ducts in peripheral rather than central regions of the liver in some patients with ALGS. The findings described above are consistent with the clinical phenotypes observed in the patient included in this study. This patient was initially hospitalized with hyperbilirubinemia (Supplementary Figure S5 and Table S5) at 1 month after birth. Whole-exome sequencing subsequently identified a heterozygous pathogenic JAG1 c.1615C > T variant. Initial abdominal ultrasound revealed normal liver dimensions and a normal physiological course of the intrahepatic bile ducts. However, the patient’s clinical manifestations showed progressive deterioration. At 15 months of age, the patient presented with diffuse hepatic parenchymal lesions, significant hepatomegaly, massive splenomegaly, and gallbladder sclerosis. Meanwhile, liver function tests from birth to 15 months showed progressively worsening abnormalities. These results suggest that this patient had relatively preserved hepatic function and a well-developed biliary network in early postnatal life; however, progressive hepatic dysfunction emerged over time, which may be associated with altered postnatal biliary maintenance or regenerative-associated responses. In this study, although H9mut and ALGS lines carried the same mutation, the HO2s phenotype derived from H9mut line was more severe, with scarcely any well-organized structures formed. It has been reported that individuals with the same mutations, including people who are in the same family, can display discordance in the overall phenotype and/or phenotype severity.5–7,49–52 We hypothesize that the observed discordance between genotype and phenotypic outcomes may be attributable to distinct genetic backgrounds across ALGS and H9mut lines. There are still limitations in our study. First, this study included iPSCs from only one patient with ALGS and one disease-causing mutation; genetic background effects and iPSC line-to-line variability may influence the results, and the findings should therefore be interpreted with caution. Second, our results are based on a cystic-like hepatic organoid model. Because intrahepatic bile ducts have a complex branching morphology in vivo, a human branching cholangiocyte organoid culture system39 should be used to further validate these findings. Third, direct pathway activity assays were not performed in the present study. Our interpretations regarding VEGFR2-mediated responsiveness, NCAM-associated peripheral bile duct defects, and polarity-related mechanisms should be regarded as associative and hypothesis-generating rather than definitive evidence of direct causality. The underlying molecular mechanism requires further elucidation through transcriptomic profiling. Conclusion In the present study, we established patient-specific iPSCs and hESCs-derived hepatic organoids combined with a base editing system to investigate cellular phenotypes associated with bile duct defects in ALGS. JAG1 c.1615C > T did not affect the cholangiocyte differentiation in our model but was associated with reduced survival of RDC-like cells, attenuated VEGF responsiveness, and abnormal cell polarity, all of which may contribute to reduced progenitor-like organoid regenerative-associated properties. This study provides a framework for investigating ALGS-associated biliary phenotypes and suggests that base editing correction of disease-causing mutations may ameliorate disease-associated organoid defects, providing theoretical support for future proof-of-concept studies of novel therapeutic approaches for ALGS. Supplementary Material