Human Placental Mesenchymal Stem Cells Relieve Primary Sclerosing Cholangitis via Upregulation of TGR5 in Mdr2−/− Mice and Human Intrahepatic Cholangiocyte Organoid Models

Primary sclerosing cholangitis (PSC) is a biliary disease accompanied by chronic inflammation of the liver and biliary stricture. Mesenchymal stem cells (MSCs) are used to treat liver diseases because of their immune regulation and regeneration-promoting functions. This study was performed to explore the therapeutic potential of human placental MSCs (hP-MSCs) in PSC through the Takeda G protein-coupled receptor 5 (TGR5) receptor pathway. Liver tissues were collected from patients with PSC and healthy donors (n = 4) for RNA sequencing and intrahepatic cholangiocyte organoid construction. hP-MSCs were injected via the tail vein into Mdr2−/−, bile duct ligation (BDL), and 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) mouse models or co-cultured with organoids to confirm their therapeutic effect on biliary cholangitis. Changes in bile acid metabolic profile were analyzed by liquid chromatography/tandem mass spectrometry (LC-MS/MS). Compared with healthy controls, liver tissues and intrahepatic cholangiocyte organoids from PSC patients were characterized by inflammation and cholestasis, and marked downregulation of bile acid receptor TGR5 expression. hP-MSC treatment apparently reduced the inflammation, cholestasis, and fibrosis in Mdr2−/−, BDL, and DDC model mice. By activating the phosphatidylinositol 3 kinase/extracellular signal-regulated protein kinase pathway, hP-MSC treatment promoted the proliferation of cholangiocytes, and affected the transcription of downstream nuclear factor κB through regulation of the binding of TGR5 and Pellino3, thereby affecting the cholangiocyte inflammatory phenotype.


Introduction
Primary sclerosing cholangitis (PSC) is a biliary disease often accompanied by chronic inflammation of the liver and progressive biliary stricture [1]. No genetic or environmental factors have yet been established for this disease, and there are as yet no known triggers for disease manifestation [2]. Takeda G protein-coupled receptor 5 (TGR5, also known as G proteincoupled bile acid receptor 1 [GPBAR1]) is a G protein-coupled receptor that is widely expressed in large and small bile duct epithelial cells in the liver and is activated by stimulation with bile acid. PSC patients show TGR5 sequence variation, which may be one of the susceptibility factors [3][4][5]. In addition to promoting bile acid secretion, TGR5 can protect the liver from bile acid overload by regulating the permeability of the bile duct epithelium [6]. TGR5 activation also promotes proliferation and inhibits apoptosis of cholangiocytes [7]. Bile-derived cholangiocyte organoids derived from patients with PSC can recapitulate the inflammatory profile of this disease [8]. The relationship between TGR5 and the inflammatory response of cholangiocytes has yet to be elucidated.
No effective pharmacotherapy has yet been developed for treatment of patients with PSC [9]. Liver transplantation remains the only life-saving option, but disease recurrence occurs in approximately 25% of graft recipients [1]. Mesenchymal stem cells (MSCs), characterized by self-renewal and multi-directional differentiation potential, have shown therapeutic potential for treating a number of diseases, including cardiovascular and autoimmune diseases [10,11]. Human placental-derived MSCs (hP-MSCs) are located in the fetal membrane of the full-term placenta, and can be collected both easily and noninvasively. Previous studies showed that hP-MSCs have a strong immunosuppressive ability and promote cell regeneration in patients with COVID-19 [12], acute liver failure [13], and liver cirrhosis [14]. Extracellular vesicles derived from human bone marrow MSCs can target the liver, reduce bile acid and alanine aminotransferase levels, decrease the content of T cells, and ameliorate liver fibrosis [15]. The effectiveness of hP-MSC treatment has not been evaluated in PSC animal models, and there have been insufficient studies of the mechanism by which hP-MSCs regulate cholangiocytes.
In this study, we collected liver tissue from PSC patients, constructed intrahepatic cholangiocyte organoids, and developed a number of cholestasis disease models to clarify the localization and differential expression of TGR5. The differential expression of interleukin (IL)-8 and its homologues C-X-C motif chemokine ligand 1/2 (CXCL1/2), and their effects on the expression of TGR5 were examined in human patients and mouse models. Intrahepatic cholangiocyte organoids and mouse models were used to explore the effectiveness of hP-MSCs for treatment of PSC and the associated changes in bile acid metabolism. Finally, we examined the potential mechanism underlying changes in TGR5 expression and downstream regulation of cholangiocytes in PSC after hP-MSC treatment. This study was performed to explore the pathomechanism of PSC and clarify the feasibility and mechanism of action of hP-MSCs in the treatment of PSC.

Characteristics of liver tissue-derived intrahepatic cholangiocyte organoid from PSC patients and healthy donors
Human liver tissue formed intrahepatic cholangiocyte organoids (named as organoid PSC liver and organoid healthy liver ) (Fig.  1A). Organoid PSC liver and organoid healthy liver showed higher expression levels of the bile duct cell marker CK19 and stemness markers SOX9 when compared with liver tissue, but did not express the liver markers albumin (ALB) and hepatocyte nuclear factor 4 alpha (HNF4A) as determined by quantitative polymerase chain reaction (qPCR; Fig. S1A). Single-layer spherical structure was observed under light microscopy and scanning electron microscopy (SEM) ( Fig. 1B and C). Compared with organoid healthy liver , organoid PSC liver was more chaotic, with more cells tending to senescence, and bile duct cholangiocyte cilia were reduced as determined by transmission electron microscopy (Fig. 1D). After adding Rhodamine 123 for 2 h with or without Verapamil, organoid PSC liver showed faster uptake of Rhodamine 123 as determined by laser confocal microscopy, indicating that the barrier function was weakened in organoids derived from tissue of PSC patients (Fig.  1E). Immunofluorescence staining showed that intrahepatic cholangiocyte organoids expressed CK7, CK19, SRY-Box transcription factor 9 (SOX9), and epithelial cell adhesion molecule (EPCAM) (Fig. 1F). Compared with organoid PSC liver , organoid healthy liver showed higher levels of expression of the bile duct cell marker CK19 and stemness markers SOX9 but did not express marked difference in the epithelial markers (EPCAM). The same result was also repeated in quantitative analysis of SOX9, CK7, CK19, and EPCAM fluorescence intensity between organoid healthy liver and organoid PSC liver (Fig.  1G). Besides, the appearance time of organoid PSC liver was significantly longer than that for organoid healthy liver (5.80 ± 0.84 vs. 3.29 ± 1.11 days, respectively, n = 4), and the diameter of organoid PSC liver on day 8 was significantly smaller than that of organoid healthy liver (191.50 ± 63.43 vs. 256.00 ± 94.45 μm, respectively, n = 30) (Fig. 1H).
After digestion, isolation, and cultivation of Mdr2 −/− mouse liver tissue for 3 to 5 days, intrahepatic cholangiocyte organoids (organoid Mdr2KO liver ) formed with morphology similar to that of organoid PSC liver on light microscopy and SEM (Fig. 1I, J). The qPCR analysis showed that organoid Mdr2KO liver had apparently increased LGR5, SOX9, and CK19 expression, and apparently reduced ALB and HNF4A expression in comparison with Mdr2 −/− mouse liver tissue (Fig. S1B). Immunofluorescence analyses showed that organoid WT liver expressed stronger fluorescence in CK7 and CK19, compared with organoid Mdr2KO liver (Fig. 1K). The same result was also repeated in the quantitative analysis of CK7 and CK19 fluorescence intensity between organoid healthy liver and organoid PSC liver (Fig. S2A). The appearance time of organoid Mdr2KO liver was significantly longer than that for organoid WT liver (5.00 ± 0.76 vs. 3.25 ± 1.04 days, respectively, n = 8), and the diameter of organoid Mdr2KO liver on day 8 was significantly smaller than that of organoid WT liver (208.57 ± 55.36 vs. 253.57 ± 59.46 μm, respectively, n = 30) (Fig. S2B). Table S1 summarizes the clinical characteristics of 4 PSC patients. Hematoxylin and eosin (HE) staining and Sirius Red staining revealed robust liver damage and collagen fibers (Fig. S3). RNA sequencing analysis revealed the differences in gene expression between liver and organoid specimens from the PSC patient and healthy donor groups. Data from 24 cases were examined using StringTie software for the sequences of known genes, and FPKM was used to calculate the measurement index for the expression of known genes (Fig. S4A). Differentially expressed genes (DEGs) were defined by P < 0.05 and fold change > 1.

Transcriptome analysis revealed the bile acid metabolic profile and inflammation in PSC patients and the organoid PSC liver model
A total of 9,973 DEGs were identified between the liver tissues of PSC patients and healthy controls ( Fig. 2A). Hierarchical clustering was performed to determine the overall differences between PSC and healthy control liver tissue (Fig. S4B). With the functional annotation and classification of disease types in the DisGeNET disease database, DEGs were enriched in liver cirrhosis (Fig. S5A). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses showed that DEGs were enriched in bile acid metabolism and immune activation ( Fig. 2B and Fig. S5B). The results of RNA-Seq showed that bile secretion-related genes (Mrp2, Bsep, Oatp2, and Ntcp) were significantly downregulated (Fig. 2C), while immune activationrelated genes (tumor necrosis factor-α [TNF-α], IL-1β, IL-6, CXCL1, and CCL2) were significantly elevated in PSC liver tissue compared with healthy control liver tissue (Fig. 2D).
Taken together, the results of transcriptome analyses showed that the disease annotations of DEGs from PSC were enriched in liver fibrosis and inflammation. The DEGs were enriched in bile acid metabolism and immune activation-related pathways, similar to the results in PSC intrahepatic cholangiocyte organoids.

hP-MSC treatment alleviated liver fibrosis and inflammation in mouse models of sclerosing cholangitis
The results of phenotypic identification and multilineage differentiation of hP-MSCs are shown in Figs. S8 and S9. hP-MSC treatment of sclerosing cholangitis mouse models (bile duct ligation [BDL] and Mdr2 −/− ) is shown in Fig. 3A and Fig. S10A, respectively. The MSC homing capacity toward inflammation and injured sites is an essential process to a successful cell therapy. Mdr2 −/− mice and Mdr2 +/+ mice were injected intravenously with hP-MSCs labeled with the fluorescent dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotricarbo-cyanine-iodide (DiR) for longterm follow-up. Fluorescence of DiR-labeled hP-MSCs could be detected at 6 h and 1, 4, 7, 9, 15, and 21 days after injection, and decreased gradually thereafter ( Fig. 3B and C). In general, the fluorescence intensity was stronger in Mdr2 −/− mice (approximately 150% to 200%), compared with Mdr2 +/+ mice every detected time. DiR-labeled cells were mainly distributed in the lungs and liver, while no labeled cells were found in the heart, spleen, and intestine (Fig. S11). In addition, hP-MSCs were observed to differentiate into hepatocytes at week 4 in Mdr2 −/− mice (Fig. S12). After hP-MSC administration via the tail vein in the 3 mouse models, immune infiltration and necrosis of liver tissue were apparently reduced, while the fibrosis level was apparently decreased in the hP-MSC treatment group ( Fig. 3D and G and Fig. S10B). Histological activity index (HAI) score and liver/ body weight ratio were also decreased in the hP-MSC treatment group ( Fig. 3E and F and Fig. S10C). hP-MSCs significantly reduced mortality in BDL models with greater alleviation of disease progression (Fig. S10D). Serum biochemical indexes, such as ALT, AST, ALP, and TBIL, were significantly decreased ( Treatment with hP-MSCs apparently ameliorated hepatic fibrosis and inflammation in mouse models of sclerosing cholangitis. hP-MSC treatment of 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) mouse models is shown in Fig. 4A. After hP-MSC administration via the tail vein in the mouse models, immune infiltration and necrosis of liver tissue were significantly reduced, while the fibrosis level was significantly decreased in the hP-MSC treatment group (Fig. 4B). HAI score was also decreased in the hP-MSC treatment group (Fig. 4C). hP-MSCs significantly reduced mortality in DDC models with greater alleviation of disease progression (Fig. 4D). Serum biochemical indexes, such as ALT, AST, ALP, and TBIL, were significantly decreased (Fig. 4E).

Treatment with hP-MSCs ameliorated alterations in bile acid metabolism in Mdr2 −/− mice by activation of TGR5 rather than FXR in cholangiocytes
RNA-Seq analysis showed that there were no differences in bile acid receptor TGR5 and farnesoid X receptor (FXR) mRNA expression between PSC and healthy liver tissues (Fig. 5A). However, the levels of TGR5 gene expression in organoid PSC liver were significantly lower than those in organoid healthy liver (Fig. 5B). Compared with healthy donor tissue, the fluorescence staining intensity and immunohistochemical color depth of TGR5 in cholangiocytes of PSC patients were decreased ( Fig. 5C and D and Fig. S14A), and the fluorescence of TGR5 in the organoid PSC liver was also reduced ( Fig. 5E and F), while the level of FXR protein was not significantly different in liver tissues or organoids between PSC and healthy control. As TGR5 is also highly expressed in hepatocytes and macrophages of PSC liver (Fig. S14B), TGR5 expression may be specifically decreased in cholangiocytes of PSC patients.
The liver tissues of Mdr2 +/+ mice (control group), Mdr2 −/− mice (model group), and hP-MSC-treated Mdr2 −/− mice (treatment group) were analyzed by liquid chromatography/mass spectrometry. On 2-dimensional (2D) principal component analysis (PCA), the 3 groups could be well distinguished and clustered, and the characteristics of the hP-MSC treatment group were intermediate between those of the control group and model group (Fig. 5G). The heat map of bile acid hierarchical clustering among the 3 groups showed that bile acid deposition was higher in the model group than the control group, and was ameliorated in the treatment group (Fig. 5H); in particular, α-muricholic acid and β-muricholic acid were significantly decreased in the treatment group (Fig. 5I). In KEGG analysis, DEGs were enriched in primary bile acid synthesis and bile secretion (Fig. S15). qPCR of mouse liver tissue showed that hP-MSCs significantly improved bile acid metabolism (TGR5 mRNA level was significantly increased), reduced bile acid synthesis (levels of CYP7A1 and CYP27B1 mRNA were significantly decreased), and increased bile acid secretion (MRP2 mRNA level was significantly increased). However, there were no differences in FXR mRNA level expression ( Fig. 5J and Fig. S16). Compared with Mdr2 +/+ mice, senescence-related P16 and P21 protein expression were specifically decreased in Mdr2 −/− mice, and could be rescued by hP-MSC treatment in the Mdr2 −/− model ( Fig. S17A and B). TGR5 protein expression may be specifically decreased in cholangiocytes of Mdr2 −/− mice, and a significant increase was also observed in the Mdr2 −/− model after hP-MSC treatment (Fig.  S17C), while the level of FXR protein was not significantly different in liver tissues or organoids among 3 groups (Fig. S17D).
(data not shown). Interestingly, a significant increase in TGR5 mRNA level was also observed in the DDC model after hP-MSC treatment, but not in the more severe cholestasis BDL model. This may have been because a large number of cholangiocytes stimulated the expression of TGR5, leading to a compensatory increase in TGR5 (Fig. S18). Therefore, hP-MSCs improved the bile acid metabolism of Mdr2 −/− mice, which may play a role by activating the bile acid receptor TGR5 rather than FXR on cholangiocytes.

hP-MSC treatment improved the expression of TGR5 in organoid Mdr2KO liver models by downregulating CXCL1/2
The IL-8 level was significantly increased in the serum of PSC patients (Fig. 6A), similar to its homologues CXCL1/2 in the serum of Mdr2 −/− mice. hP-MSC treatment significantly reduced CXCL1/2 levels in the serum of Mdr2 −/− mice (Fig. 6B). To examine the relationships between CXCL1/2 and TGR5, we added CXCL1, CXCL2, TNF-α, and IFN-γ separately to organoid Mdr2KO liver culture medium for 24 h, followed by co-culture with hP-MSCs for a further 24 h. qPCR analysis showed that addition of CXCL1 and CXCL2 led to reductions in the level of TGR5 mRNA in the 2 groups of organoids. There was no difference in TGR5 mRNA level in both the IFN-γ and TNF-α groups (Fig. 6C). Western blotting and immunofluorescence analyses were performed to determine the changes in TGR5 protein level of organoids in the co-culture experiment (Fig. 6D to J), and the results were consistent with the mRNA analyses. These observations showed that the mouse homologues of IL-8,    To identify TGR5-dependent downstream phenotypic signals, the expression of the inflammation-related genes (TNF-α, IL-1β, IL-6, and TGF-β1) and proliferation-related genes (Ki-67 and proliferating cell nuclear antigen [PCNA]) were examined. A heat map of the above genes is shown in Fig. 7A. Compared with the organoid Mdr2KO liver -CXCL1 group, TNF-α, IL-1β, and IL-6 levels were significantly decreased in the hP-MSC-organoid Mdr2KO liver -CXCL1 group, while Ki-67 and PCNA expression were significantly increased (Fig. 7B). There was no significant difference in TGF-β1, p16, and p21 expression between the 2 groups (Fig. S19A). Compared with the organoid Mdr2KO liver -CXCL2 group, IL-6 levels were significantly decreased in the hP-MSC-organoid Mdr2KO liver -CXCL2 group, while Ki-67 and PCNA expression were significantly increased. There was no significant difference in TNF-α, IL-1β, TGF-β1, p16, and p21 expression between the 2 groups (Fig. S19B). Typical organoids of the control, organoid Mdr2KO liver -CXCL1, and hP-MSC-organoid Mdr2KO liver -CXCL1 groups are shown in Fig. 7C and D. Immunofluorescence staining for Ki-67 showed that CXCL1 stimulation significantly reduced the proliferation of organoid Mdr2KO liver , and hP-MSC co-culture significantly ameliorated this effect ( Fig. 7E and F).
Next, siRNA was used to knock down TGR5 expression in organoid Mdr2KO liver . Western blotting and qPCR analyses were performed to determine the knockdown efficiency of the 3 siRNA sequences used here, and the results showed that SiTGR5_1 and SiTGR5_2 significantly reduced TGR5 expression at the protein and mRNA levels ( Fig. 8E and F). The effects on the downstream inflammatory and proliferative phenotype were examined, and the results showed that SiTGR5_2 was the most consistent with expectations (Fig. 8G). In addition, the organoid Mdr2KO liver with SiTGR5_2 knockdown was cultured for 48 h, co-cultured with hP-MSCs for 24 h, and then stimulated with CXCL1. Western blotting showed that the expression of TGR5 first decreased, then increased, and later decreased; the ERK pathway showed the same trend (Fig. 8H).

hP-MSC treatment ameliorated the pathology of cholangiocytes in the human organoid model via TGR5/PI3K/ERK and TGR5/Pellino3/NF-κB signaling
Organoid PSC liver was used to confirm whether IL-8 downregulated TGR5 and whether hP-MSCs had a rescue effect. Similar to co-culture analysis of organoid Mdr2KO liver , RT-qPCR was performed to determine mRNA levels in organoid PSC liver . Expression levels of the inflammatory marker TNF-α, IL-1β, and IL-6 and the senescence-associated markers p16 INK4a and p21 WAF1/Cip1 were significantly reduced, and those of the proliferation markers Ki-67 were significantly increased in the hP-MSC-organoid PSC liver -IL-8 group compared with the organoid PSC liver -IL-8 group (Fig.  9A). Immunofluorescence analyses showed that TGR5 and Ki-67 expression were significantly decreased after IL-8 stimulation, while they were upregulated by hP-MSC treatment (Fig. 9B to E). Those of the proliferation markers Ki-67 were significantly increased in the hP-MSC-organoid PSC liver -IL-8 group compared with the organoid PSC liver -IL-8 group (Fig. 9F). Inhibition of senescence was demonstrated by senescence-associated (SA)-β-gal staining (Fig. 9G). The percentage of SA-β-gal-positive organoid PSC liver was also significantly decreased after co-culture with hP-MSCs (Fig. 9H). Western blotting analysis confirmed that hP-MSC treatment also ameliorated the inflammatory phenotype and inhibited the proliferation of organoid PSC liver through TGR5/ PI3K/ERK and TGR5/Pellino3/NF-κB pathways (Fig. 9I to K).

Discussion
The results of this study showed that TGR5 was differentially expressed in PSC and normal human cholangiocytes, and that IL-8 reduced the level of TGR5. These differences were preserved in organoid PSC liver and organoid Mdr2KO liver . hP-MSC treatment alleviated sclerosing cholangitis in BDL, DDC, and Mdr2 −/− mouse models. The effects of hP-MSC treatment in Mdr2 −/− mice were mainly mediated by secretion of IGF-1 and downregulation of CXCL1/2 in vivo, targeting TGR5 in cholangiocytes. The results in organoids showed that co-culture with hP-MSCs improved the proliferation potential of cholangiocytes via the TGR5/PI3K/ERK pathway, inhibited inflammation by binding of TGR5 and Pellino3, activated NF-κB expression, and decreased the levels of the inflammatory factors TNF-α, IL-1β, and IL-6. This research strategy will contribute to exploration of the pathological mechanism of PSC and the clinical application of MSCs in PSC patients.  During disease progression, PSC shows abnormal bile secretion, chronic inflammation, liver fibrosis, and other manifestations [16]. The bile acid regulatory receptors TGR5 and FXR have attracted a great deal of attention in research on cholestasis [17]. TGR5 is mainly localized in the primary cilia, apical plasma membrane, and nuclear membrane of bile duct cells [18]. Bile acid-induced activation of TGR5 regulates cell proliferation, triggers cytoprotective mechanisms, and promotes chloride secretion [19]. In the liver, TGR5 is expressed in nonparenchymal cells, including bile duct cells, sinusoidal endothelial cells, natural killer (NK) cells, and Kupffer cells [20][21][22]. Many studies have shown that the decline of TGR5 in PSC is specific, and is not seen in primary biliary cirrhosis and nonalcoholic fatty liver [23][24][25]. RNA-Seq analysis of liver tissue from PSC patients showed that TGR5 was not significantly increased, but was decreased apparently in cholangiocytes and increased apparently in Kupffer cells as determined by tyramide signal amplification (TSA) multiple immunofluorescence staining. RNA-Seq analysis also showed that TGR5 was apparently decreased in intrahepatic cholangiocyte organoid PSC liver . These results showed that TGR5 expression is specifically reduced in cholangiocytes of PSC patients.
As a 3-dimensional (3D) cell culture system, organoids can be used to study human development and disease progression [26]. Biliary immunity response and cholangiocyte senescence contribute to the pathogenesis of cholangitis, which are the most distinguishing features of PSC. Compared with liver organoids, the characteristics of intrahepatic cholangiocyte organoids can better reflect the pathological changes of PSC. Soroka et al. [8] reported that cholangiocyte organoids derived from PSC patients showed marked biliary senescence with upregulation of the senescence-associated secretory phenotype (SASP) markers p21 and SERPINE2 (but not other SASP-associated molecules, including IL-6 and IL-8), and proinflammatory mediators, including CCL20, HLA-DMA, and CD74. We constructed intrahepatic cholangiocyte organoids from human and mouse liver tissues. qPCR and RNA-Seq analyses showed that organoid PSC liver reflected the senescence and inflammatory phenotype of cholangiocytes, the 2 most important characteristics of PSC. Moreover, the differences between organoid healthy liver and organoid PSC liver mimicked those of the original liver tissues, indicating that they are good in vitro models.
There is still a lack of effective treatment for PSC. MSCs have entered phase II or III clinical trials for use in a number of conditions, including graft versus host disease [27], Crohn's disease [28], and systemic lupus erythematosus [29]. Therapeutic strategies using MSCs may be applicable in patients with liver diseases. The therapeutic effects of MSCs are generally thought to be mainly due to their immunosuppressive effects and/or the release of nutritional factors, which are involved in cell regeneration and immune homeostasis. In the present study, based on evaluation of pathological sections, liver enzymes, total bile acids, and fibrosis, hP-MSC treatment delayed the progression of liver fibrosis and decreased inflammation in 3 mouse models of PSC. In addition, increased fluorescence was detected in the damaged liver in this study, indicating inflammatory chemotaxis of hP-MSCs. Surprisingly, hP-MSCs could differentiate into human hepatocyte-like cells in mice, although the number of these cells was very small.
PSC patients have disturbances in bile acid metabolism and inflammation, along with specifically reduced TGR5 expression in cholangiocytes. At present, little is known about the specific mechanisms underlying the effects of hP-MSC treatment in PSC. Therefore, we focused on the role of hP-MSCs in regulating bile acid metabolism and immune regulation. PSC patients have high serum IL-8 levels, which is a possible prognostic factor [30,31]. CXCL1/2, the mouse homologues of IL-8, were also increased apparently in the serum of Mdr2 −/− mice. Stimulation of organoid Mdr2KO liver with CXCL1/2 resulted in decreases in TGR5 mRNA and protein levels, respectively. Therefore, CXCL1/2 were considered to be TGR5 inhibitors. The decrease of TGR5 in organoid Mdr2KO liver by CXCL1/2 and the subsequent bile duct inflammatory response and weakening of proliferation capacity were ameliorated by co-culture with hP-MSCs. Next, we explored the specific signaling pathways downstream of TGR5 that may regulate the proliferation and inflammatory phenotype of cholangiocytes. Increased TGR5 expression activates the phosphorylation of PI3K, resulting in Ras activation. Activated Ras then activates Raf by binding its N-terminal domain. Raf, mitogen-activated protein kinase (MEK), and ERK are activated successively, and finally activated ERK enters the nucleus to increase cell proliferation. Liang et al. [32] reported that treatment with the TGR5 activator INT777 improved the neurological function of middle cerebral artery occlusion (MCAO) model mice. Colocalization of TGR5 and Pellino3, which can bind directly, increased after MCAO, apparently reducing Caspase-8 and NLRP3 levels. Pellino3 is a member of the mammalian Pellino family of E3 ubiquitin ligases [33]. To our knowledge, there have been no previous studies of Pellino3 in the liver. Our observations confirmed that TGR5 and Pellino3 bound directly to each other in the liver, and TGR5 activated Pellino3, which inhibited activation of NF-κB and decreased the levels of TNF-α, IL-1 β, IL-6, and other inflammatory factors. We also used IL-8 to damage organoid PSC liver , which was reversed by co-culture with hP-MSCs. IL-8 also reduced the proliferation and inflammatory phenotype of cholangiocytes in organoids. The effects of hP-MSCs were also mediated by the TGR5/PI3K/ERK and TGR5/Pellino3/NF-κB pathways.
MSC-derived IGF-1 has been shown to improve cell regeneration in a variety of disease models [34,35]. We confirmed that hP-MSCs secreted IGF-1 to stimulate the expression of TGR5 in organoid Mdr2KO liver , and could reverse the inflammatory and proliferative phenotypes. Some deficiencies cannot be ignored for this study. (a) Some findings should be regarded as tentative and more studies with larger PSC liver samples will be required to confirm the results. For example, there may be marked differences in FXR between PSC and healthy control, as the sample size increases. (b) Despite all this, although an organoid system has various advantages, organoids still have some critical limitation for functional studies including the Matrigel-based system and single-cell-based organoids. A multi-organoid system platform for automated and continual in situ monitoring of organoid behaviors may be a feasible option.

Blood and liver sample collection from PSC patients and healthy controls
The study was performed in accordance with the Declaration of Helsinki. Four liver samples from 4 PSC patients who had undergone liver transplantation and 4 discarded tissues during donor liver repair as healthy controls were collected from Shulan (Hangzhou) Hospital. The sampling site was 5 cm away from the liver hilum and rich in intrahepatic cholangiocytes within 2 h after liver removal. Blood samples from 8 PSC patients and 35 healthy controls were obtained through The First Affiliated Hospital, Zhejiang University School of Medicine. The protocols using human tissues and blood samples were approved by the Research Ethics Committee of The First Affiliated Hospital, Zhejiang University School of Medicine (Approval No. 2021-158).

hP-MSCs and biliary cholangitis mouse models
hP-MSCs were obtained from the Cell Bank of State Key Laboratory for the Diagnosis and Treatment of Infectious Diseases, Zhejiang University. Culture, phenotypic identification, and multilineage differentiation were performed as described previously [36].

Intrahepatic cholangiocyte organoids derived from human and mouse liver
The media used for organoid culture are listed in Table S2, including wash medium (WM), basal medium (BM), human liver expansion medium (h-EM), human liver isolation medium (h-IM), human liver digestion solution medium (h-DM), mouse liver expansion medium (EM), mouse liver isolation medium (IM), and mouse liver digestion solution medium (DM). The isolation and culture of intrahepatic cholangiocyte organoids from livers of adult Mdr2 −/− and Mdr2 +/+ mice were performed as described previously [37,38]. Human liver tissues were washed twice with precooled WM and cut into pieces. The fragments were added to h-DM and digested at 37 °C in a shaking incubator for 1 to 1.5 h. After digestion, the supernatant was collected, filtered through a 70-μm filter (Falcon Plastics, Oxnard, CA, USA), and the filtrate was neutralized with WM. After centrifugation at 400 × g for 4 min, the particles were washed twice in BM. The final particles were resuspended in an appropriate amount of Matrigel (Corning, Bedford, MA, USA). Aliquots of 30 to 50 μl of the mixtures were plated into each well of 24-well cell culture plates (Corning) and incubated at 37 °C for 30 min to allow the Matrigel to polymerize. After polymerization, 500 μl of h-IM medium was added. Passage and medium exchange were performed as described for mouse organoids.

RNA-Seq analysis
Total RNA of human liver tissues and human liver tissue-derived organoids were extracted using an RNAmini kit (Qiagen, Hilden, Germany). Organoid (passages 3 and 7) were cultured for 96 h and then were used to harvest RNA for RNA-sequencing (RNAseq). Enrichment of mRNA, fragmentation, reverse transcription, library construction, sequencing using the Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, USA), and data analysis were performed by Genergy Biotechnology Co. Ltd. (Shanghai, China). The raw data were processed with Skewer and data quality was checked with FastQC v0.11.2 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). The read length was 2 × 150 bp. Clean reads were aligned to the human genome hg38 assembly using STAR and StringTie. Transcript expression was calculated as FPKM (fragments per kilobase of exon model per million mapped reads) using Perl. Differentially expressed transcripts (DETs) between different time points (12 h vs. 0 h, 36 h vs. 0 h, 72 h vs. 0 h) were determined using the MA-plot-based method with the Random Sampling (MARS) model in the DEGseq package. Generally, in the MARS model, M = log2C1 − log2C2, and A = (log2C1 + log2C2)/2 (where C1 and C2 denote the counts of reads mapped to a specific gene obtained from 2 samples). The thresholds for determining DETs were P < 0.05 and absolute fold change ≥ 1. Then, DETs were chosen for function and signaling pathway enrichment analysis using the KEGG and GO databases. The significantly enriched pathways were determined when P < 0.05 and at least 2 affiliated genes were included.

Bile acid detection by liquid chromatography/mass spectrometry
Liver tissues from 12-week-old Mdr2 +/+ , Mdr2 −/− , and hP-MSCtreated Mdr2 −/− mice were used to detect bile acids. Bile acid contents were determined quantitatively by MetWare (http://www. metware.cn) using the AB Sciex QTRAP 6500 LC-MS/MS platform (Applied Biosystems, Foster City, CA, USA). Unsupervised PCA was performed using the prcomp function in R (www.r-project.org). The data were unit variance scaled before unsupervised PCA. Metabolites showing significant regulation among the 3 groups were determined by Variable important in projection (VIP) and absolute log2FC (fold change). VIP values were extracted from the results of orthogonal partial least squares discriminant analysis (OPLS-DA), which also included score plots and permutation plots generated using the MetaboAnalystR R package. The data were log transformed (log2) and mean centered before OPLS-DA. To avoid overfitting, a permutation test (200 permutations) was performed.

Transfection of organoids with small interfering RNA
Organoids were transfected with 3 short siRNAs (50 nM) targeting TGR5. The organoids were cultured as described previously [39]. At 80% to 90% confluence, they were digested and collected, and single cells were resuspended in 450 μl of IM. The mixture was kept on ice until transfection. Transfection was performed by adding 50 μl of DMEM to each of 2 1.5-ml microcentrifuge tubes for each condition. Then, 1 μl of small interfering RNA (siRNA) targeting TGR5 (Gene Pharma, Shanghai, China) was added to one of the tubes and 4 μl of siRNA-Mate mixture was added to the other tube. Both tubes were incubated at room temperature for 5 min. The contents of the 2 tubes were mixed together and incubated at room temperature for a further 15 min. Then, 50 μl of DNA-siRNA-Mate mixture was added to 450 μl of single-cell suspension and this new mixture was transferred to one well of a 24-well plate, centrifuged in a prewarmed centrifuge at 37 °C at 600 × g for 1 h, and incubated at 37 °C for 2 to 4 h. After centrifugation at 400 × g for 4 min, the particles were washed twice in BM. The final particles were resuspended in an appropriate amount of Matrigel (Corning, Bedford, MA, USA). The sequences of 3 short siRNAs (50 nM) targeting TGR5 are shown in Table S3. siRNAs that showed significant knockdown effects were used in subsequent analyses.

Statistical analysis
The data of all experiments are presented as the mean ± standard error of mean. The log-rank test was applied for survival analysis. Significance was determined by one-way analysis of variance (ANOVA) or 2-tailed Student's t test. All statistical analyses were performed using GraphPad Prism 8 software (GraphPad Software, San Diego, CA). In all analyses, P < 0.05 was taken to indicate statistical significance.
Details of biliary cholangitis models of mice established by DDC and BDL, hP-MSC administration route and dosage, measurement of IL-8 and CXCL1/2, serum biochemical indexes, inflammatory cytokines and chemokines, Sirius red staining, Rhodamine 123 staining, co-culture of hP-MSCs and organoid, RNA extraction, real-time fluorescent qPCR analysis, immunofluorescence and immunohistochemical staining, TSA staining, flow cytometry, and Western blotting analysis were performed as described in the Supplementary Information. Stimulants used in the co-culture of hP-MSCs and organoids are listed in Table  S4. The primer sequences used for tissue and organoid detection of human and mouse are listed in Table S5. The antibodies used in this experiment are all listed in Table S6. Laboratory (Nos. JNL-2022026C and JNL-2023003C). Author contributions: H.C. and Q.Y. contributed to study design and planning; Q.Y., W.C., Y.Y., F.G., J.Z., J. Yang, L.Z., and Q.P. contributed to the cellular and animal experiment; H.C., Q.Y., J.W., and J. Yu participated in analysis, interpretation of data, and manuscript writing; L.L. and H.C. contributed to study supervision. All authors reviewed and approved the final version of the manuscript. Competing interests: The authors declare that they have no competing interests.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request. Table S1. Patient characteristics. Table S2. Human and mouse organoid medium formulations.  Table S4. Stimulants used in the co-culture of hP-MSCs and organoids. Table S5. Primer sequences used for real-time PCR studies on tissues and organoids of human and mouse. Table S6. Antibody list. Fig. S1. Gene levels of liver tissues and organoids of human and mouse. Fig. S2. Construction of liver tissue-derived cholangiocyte organoid from Mdr2 −/− mice and Mdr2 +/+ mice. Fig. S3. Hepatic fibrosis and pathological damage are the pathological consequence of PSC liver injury. Fig. S4. RNA-Seq analysis of the livers and organoids from PSC patients and healthy donors. Fig. S5. RNA-Seq analysis of liver tissue from PSC patients and healthy controls. Fig. S6. RNA-Seq analysis of organoids. Fig. S7. Transcriptome analysis revealed the cholangiocyte senescence-related, profibrotic cytokine-related, and angiogenesis in organoid PSC liver . Fig. S8. Surface markers of hP-MSCs determined by flow cytometry. Fig. S9. Detection of hP-MSC multi-differential potential in vitro. Fig. S10. hP-MSC treatment ameliorated the pathological process of sclerosing cholangitis BDL mouse models.