Human placenta mesenchymal stem cell-derived exosomes delay H2O2-induced aging in mouse cholangioids

Cholangiocyte senescence is an important pathological process in diseases such as primary sclerosing cholangitis (PSC) and primary biliary cirrhosis (PBC). Stem cell/induced pluripotent stem cell-derived exosomes have shown anti-senescence effects in various diseases. We applied novel organoid culture technology to establish and characterize cholangiocyte organoids (cholangioids) with oxidative stress-induced senescence and then investigated whether human placenta mesenchymal stem cell (hPMSC)-derived exosomes exerted a protective effect in senescent cholangioids. We identified the growth characteristics of cholangioids by light microscopy and confocal microscopy. Exosomes were introduced concurrently with H2O2 into the cholangioids. Using immunohistochemistry and immunofluorescence staining analyses, we assessed the expression patterns of the senescence markers p16INK4a, p21WAF1/Cip1, and senescence-associated β-galactosidase (SA-β-gal) and then characterized the mRNA and protein expression levels of chemokines and senescence-associated secretory phenotype (SASP) components. Well-established cholangioids expressed cholangiocyte-specific markers. Oxidative stress-induced senescence enhanced the expression of the senescence-associated proteins p16INK4a, p21WAF1/Cip1, and SA-β-gal in senescent cholangioids compared with the control group. Treatment with hPMSC-derived exosomes delayed the cholangioid aging progress and reduced the levels of SASP components (i.e., interleukin-6 and chemokine CC ligand 2). Senescent organoids are a potential novel model for better understanding senescence progression in cholangiocytes. hPMSC-derived exosomes exert protective effects against senescent cholangioids under oxidative stress-induced injury by delaying aging and reducing SASP components, which might have therapeutic potential for PSC or PBC.


Background
Senescent cells usually remain metabolically active but show growth arrest in G1 phase [1,2], such that they express the expressing cell cycle related inhibitors p21 Waf1/Cip1 and p16 INK4a . Cell senescence characterized by activation of the senescence-associated secretory phenotype (SASP), which activates and reinforces the inflammatory reaction of bystander cells and attracts immune cells that influence the tissue repair process [3][4][5]. This also facilitates the progression of chronic inflammatory diseases.
Cholangiocytes are targets of various biliary diseases [6,7]. Continuous stimulation leads to cholangiocyte damage, which in turn leads to apoptosis or entry into a senescent state when the damage is not repaired [1]. Cellular senescence is an important pathological process in diseases such as primary sclerosing cholangitis (PSC) and primary biliary cirrhosis (PBC) [7][8][9]. Currently, there is no effective therapy for PSC or PBC, aside from liver transplantation [6]. Nearly half of affected patients experience one or more episodes of acute cellular rejection, and~25% develop recurrent disease [10] after transplantation.
Notably, there is evidence that stem cell-derived exosomes exert a positive effect on diseases such as hepatotoxic injury [11], myocardial ischemia/reperfusion injury [12,13], and aging-induced cardiac dysfunction [14]. Watanabe et al. and Lee et al. found that stem cell/induced pluripotent stem cell-derived exosomes significantly reduced the activity of senescence-associated βgalactosidase (SA-β-gal) and the levels of cell cycle inhibitors p21 Waf1/Cip1 , in a model of bisphosphonaterelated osteonecrosis of the jaw [15] and in senescent skin fibroblasts [16], respectively. However, the protective effects of stem cell-derived exosomes against PSC and PBC remain unclear.

Materials and methods
Cell culture 293T-HA-Rspon1-Fc and Wnt-3a cells were used as described previously to generate conditional medium of Rspon1 and Wnt-3a proteins [27]. Stably transfected 293T cells producing mouse Rspon1-Fc fusion protein were obtained as a gift from Professor Enkui Duan and Xiaohua Lei at the Chinese Academy of Sciences. Wnt3a mouse subcutaneous connective tissue cells were purchased from Bei Na Lian Chuang Biotechnology Co. Ltd. (Beijing, China). Placentae were obtained from donors at The First Affiliated Hospital, Zhejiang University School of Medicine.

Isolation of cholangiocytes from mouse liver
Cholangiocytes were isolated from adult C57BL/6 mouse liver. After exposing the liver, blood was washed out using 1× phosphate-buffered saline (PBS) via inferior vena cava injection with a portal vein cutoff. Tissues were cut into small pieces and washed with precooled Dulbecco's modified Eagle's medium (DMEM) (high glucose, with GlutaMAX and pyruvate, 1% fetal bovine serum, and 1% penicillin/streptomycin), then transferred to prewarmed digestion medium (DMEM with 0.1 mg/ ml DNase I, 0.125 mg/ml collagenase D, and 0.125 mg/ ml Dispase II) with shaking for 1.5 h at 37°C. During digestion, the appearance of the bile ducts was checked by light microscopy at intervals of 20-30 min. Digestion was stopped using precooled DMEM and cell pellets were collected by centrifugation at 200-300×g. Precooled DMEM was added to centrifuge tubes, which were then kept on ice for 30 min to collect cell pellets. The second digestion was performed with TrypLE solution and DNase I 50 μg/ml for~20 min at 37°C. Finally, pellets were washed with Advanced DMEM/F-12 (with 1% penicillin/streptomycin, 1% GlutaMAX, and 10 mM HEPES) and filtered through a 70-μm filter to obtain single cholangiocytes, as described previously [27].

Cholangioid passage
At 7-14 days after seeding, precooled Advanced DMEM/F-12 (with 1% penicillin/streptomycin, 1% Glu-taMAX, and 10 mM HEPES) was added and then pipetted up and down to disrupt the Matrigel. Organoids were collected, digested into single cells by incubation with TrypLE solution at 37°C, and reseeded into new Matrigel. Organoids were usually passaged with a split ratio of 1:3-4 at intervals of 5-7 days after the first passage. Organoids at passages 1-4 were used in the study. Detailed reagent information is shown in Table S1.

In vitro modeling of cholangioid senescence and administration of exosomes
After stable growth of cholangioids (2-3 days), expansion medium containing H 2 O 2 (50 nM) was added to organoids in the senescent (Sen) group, while organoids cultured in expansion medium containing both H 2 O 2 (50 nM) and exosomes (0.1 μg/ml) represented the exosome-treated (Exo) group. Organoids in the control group (Ctrl) were grown in expansion medium alone. All medium was replaced at intervals of 24 or 48 h for up to 120 h.

SA-β-gal activity assessment
For complete collection of organoids, the expansion medium was aspirated and replaced with 500-1000 μl precooled Cell Recovery Solution (Dow Corning, Corning, NY, USA) per well. The plates were shaken gently at 4°C to disrupt the Matrigel. The supernatant was carefully removed and precooled 1× PBS was used to wash the cells three times. SA-β-gal activity was detected using the senescence detection kit (BioVision, Mountain View, CA, USA). The percentage of SA-β-gal-positive organoids was determined for each condition using a 20× objective and bright field illumination of 30 randomly selected areas. In this study, 30% was set as the cutoff value: if one organoid contained more than 30% green-stained cells, it was considered SA-β-gal stainingpositive.

Immunofluorescence and immunohistochemical analyses
Organoids were collected carefully as described in the SA-β-gal activity subsection above, then fixed in 4% (wt/ vol) paraformaldehyde at 4°C and subjected to immunofluorescence or immunohistochemistry analysis.

Immunofluorescence
Fixed organoids were washed in cold PBS-Tween (0.1% vol/vol) solution and blocked in cold OWB (1 ml Triton X-100 and 2 g bovine serum albumin in 1 L 1× PBS). Organoids were transferred to a 24-well plate. Primary antibodies were added (anti-CK7, anti-Ki67, anti-CK19, and anti-p21 WAF1/Cip1 ) and incubated with the organoids overnight at 4°C. After the organoids had been washed in 1× PBS, secondary antibodies were added and organoids were incubated overnight at 4°C. Organoids were removed to fructose-glycerol clearing solution (60% vol/ vol glycerol and 2.5 M fructose) and incubated at room temperature for 20 min, then transferred to slides with coverslips and subjected to imaging as described previously [28].

Immunohistochemistry
Organoids were washed with 1× PBS and resuspended in 70% ethanol, then dehydrated and stained in 0.5% eosin (dissolved in 96% ethanol) for 30 min. The organoids were dehydrated in 100% ethanol and embedded in paraffin; the following processes were the same as the normal immunohistochemistry steps. Detailed antibody information is shown in Suppl. Table 1.

Preparation of hPMSC-derived exosomes
Exosomes were collected from hPMSCs at passages 3-5. After hPMSCs reached 80-90% confluence, they were washed once with PBS and covered with serum-free DMEM (low glucose), then cultured for another 48 h. Subsequently, the supernatant was collected and centrifuged for 30 min at 10,000×g to remove cell debris, filtered through a 0.22-μm filter, and concentrated using a 100-kDa molecular weight cutoff ultrafiltration tube (Merck Millipore, Hessen, Germany). Ultracentrifugation was performed at 120,000×g at 4°C for 70 min using preparative ultracentrifuges (Beckman Coulter, Brea, CA, USA). Small extracellular vesicles (exosomes) were obtained and washed once in 1× PBS, then stored at − 80°C until use. Exosome concentrations were determined using a Micro BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA).

Organoid viability assay
Before organoid treatment with H 2 O 2 and exosomes, or when organoids had reached the desired culture duration for analysis, expansion medium was replaced with working fluids, which contained 2 μM calcein and 8 μM propidium iodide, in accordance with the instructions of the Live & Dead Viability/Cytotoxicity Assay Kit for Animal Cells (Keygen Biotech, Nanjing, China). Organoids were observed and imaged using a confocal laser scanning microscopy (Zeiss LSM710; Carl Zeiss AG, Jena, Germany).

Software packages
Images were cropped using Adobe Photoshop CS6 (version 13.0; Adobe Systems Software Ireland Ltd., Dublin, Ireland) and formatted using Adobe Illustrator 2020 (version 24.0.1; Adobe Systems Software Ireland Ltd.).
Flow cytometry data were analyzed by FlowJo V10 (BD Biosciences, Franklin Lakes, NJ, USA). Immunohistochemical staining results were analyzed by Image-Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA). The number of organoids was counted by Image J (https:// imagej.net/Download).

Statistical analysis
Statistical analyses of the p16 INK4a , p21 WAF1/Cip1 , and Ki67 immunohistochemical staining results, and the percentage of SA-β-gal-positive organoids, were performed with two-tailed Student's t test; other date was performed with one-way or two-way analysis of variance (ANOVA), and the Sidak multiple comparisons test was used to confirm the results of the ANOVA. Data analysis was conducted using Graph-Pad Prism 8.0.1 (GraphPad Software Inc., La Jolla, CA, USA). P < 0.05 was considered to indicate statistical significance.

Additional methods
Western blotting procedures, transmission electron microscopy (TEM) assessment of exosomes, RNA isolation and real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) procedures, and hPMSC culture are described in the supplementary material.

Features of long-term cultured cholangioids
After 10-14 days of culture, the cholangiocytes in organoids remained in close contact and developed into more than one cell layers (Fig. 2b). Some organoids shrank in a manner like deflation of a balloon (Fig. 2a, left bottom). Other organoids had cells scattered inside and outside their lumen (Fig. 2a, top and right bottom), which were confirmed to be dead cells by viability/cytotoxicity analysis (Fig. 2d, red). Furthermore, long-term cultured organoids expressed SA-β-gal (Fig. 2c). After digestion into single cells and subsequent passaging, some single cells developed into organoids (black arrow), while others did not (white arrow) (Fig. 2e). Therefore, it is suggestable to passage organoids before the appearance of cell aggregates inside the lumen to increase the organoid formation rate from single cells. In our study, the plating efficiency of the different passages varied from 36.5 to 41.8% (data not shown), consistent with previous findings [27,29].

Persistent treatment of H 2 O 2 induced senescence in cholangioids
After establishment of the organoid culture system, we tested whether exposure to oxidative stress could induce cell senescence. Before treatment with H 2 O 2 , organoid viability was verified through viability/cytotoxicity assay in which live cells were stained green (Fig. S1a). Then, 50 nM H 2 O 2 with expansion medium was added to organoids and cultured for 120 h, as shown in Fig. 3a. The diameter of H 2 O 2 -induced organoids increased rapidly during the first 48 h, consistent with the presentation of normal organoids. After 48 h, the appearance of some organoids was like that of long-term cultured organoids (Fig. 2a): shrunken, scattered dead cells, and aggregated cells inside and outside the lumen (Figs. 3b and S1b). H 2 O 2 -treated organoids partially expressed SA-β-gal (Fig. 3d), as well as the cell cycle arrest markers p21 WAF1/Cip1 (Fig. 3c) and p16 INK4a (Fig. 5b, top row,  Senescence).
SASP components and some chemokines, another feature of cell senescence, were detected by qRT-PCR, and included IL-6, IL-8 (Fig. S4b), CCL2, CXCL2, CX3CL1, CXCL9, and CXCL16 (Fig. 3e). Detection of SASP components at different time points indicated that the mRNA expression levels of these components were significantly higher at 120 h than at 24, 48, or 96 h. Compared with the control group, the mRNA expression of CCL2 increased 16.65-fold with 120 h, and 1.110-3.626fold with 24-96 h, exposure to H 2 O 2 treatment (P < 0.05). These results suggested that cell senescence was successfully induced by H 2 O 2 treatment and that 120-h culture was adequate for establishment of senescent cholangioids. These cholangioids could be used for further research, including as a 3D model for PSC in vitro.
Exosomes were derived from hPMSCs and co-cultured with organoids Stem cell/induced pluripotent stem cell-derived exosomes have been shown to play protective roles in senescence-related diseases [15,16]. Therefore, we hypothesized that mesenchymal stem cell (MSC)-derived exosomes may play a similar role in PSC.
The MSCs used in our research were derived from human placenta. They were identified through imaging (Fig. S3a), assessment of multilineage differentiation potential, and detection of MSC-specific markers. hPMSC pluripotency was verified by differentiation into osteoblasts (Fig. S3b) and adipocytes (Fig. S3c). The MSCspecific markers CD73, CD90, and CD105 were verified by flow cytometry analysis (Fig. S3d). Small extracellular vesicles were then obtained using the serial ultracentrifugation protocol shown in Fig. 4a. Western blotting revealed that MSC-derived vesicles expressed the exosome-specific markers CD9 and CD81 (Fig. 4b).
Morphology analysis by transmission electron microscopy (TEM) (Fig. 4c) indicated that the small extracellular vesicles were exosomes.
We co-cultured 0.1 μg/ml exosomes with 50 nM H 2 O 2 -stimulated cholangioids to verify the protective effect reported previously [11]. Details are shown in Fig.  4d. Exosomes were confirmed to be absorbed into cholangioids after culturing for 72 and 120 h, based on immunofluorescence staining (Fig. 4e).

hPMSC-derived exosomes delay aging in senescent cholangioids
After oxidative stress-induced injury and exosome treatment, we recorded changes in organoid growth from 0 to 120 h via light microscopy. We found that the number of senescent cholangioids increased gradually over time. As shown in Figs. 5a and S2a, senescent cholangioids at 120 h noted by white lines suggested that the number of senescent organoids was relatively lower in the Exo group than in the Sen and Ctrl groups. Further analysis showed significant differences in the number of senescent organoids between the Sen and Exo groups at 120 h (p = 0.0262) (Fig. S2b), suggesting that exosomes may delay organoid aging during exposure to oxidative stress.
To investigate this possibility, we examined the expression patterns of SA-β-gal and cell cycle arrest proteins in the Sen, Exo, and Ctrl groups. We found that the expression level of the cell cycle arrest protein p16 INK4a (Fig. 5b, top row) was reduced significantly by exosome treatment relative to the Sen and Ctrl groups (Fig. 5c). Exosome-treated organoids showed lower expression of p21 WAF1/Cip1 protein than the Sen group (Fig. 5b, middle  row, and 5d). This was confirmed by immunofluorescent staining (Figs. 5g and S2c). However, semi-quantitative comparison of p21 WAF1/Cip1 expression between the Ctrl and Sen groups did not show a significant difference. Furthermore, organoids in the Exo group exhibited greater proliferative activity compared with the Ctrl group (Fig. 5b, bottom row, and 5e). These results demonstrated the ability of exosome treatment to relieve senescence and promote proliferation. Similarly, the percentage of SA-β-gal-positive organoids decreased significantly ( Fig. 5f) (P < 0.05), as typical SA-β-gal-positive cholangioids shown in Fig. 5h. Interestingly, when the organoids from Sen, Exo, and Ctrl groups were digested and cultured for 1 day and 5 days, the number of secondgeneration organoids in Sen group was significantly lower than that in Exo and Ctrl group (Fig. S2d-2e).
Taken together, these results supported our hypothesis that hPMSC-derived exosomes exert a protective effect against senescence in cholangioids.

hPMSC-derived exosomes reduced expression of SASP components
Our previous studies showed significantly increased expression of SASP components in senescent cholangioids. Thus, we also assessed whether exosomes could affect SASP components in this study. We evaluated the mRNA expression levels of SASP components and chemokines in the Exo group, including IL-6, CCL2, CXCL9, CXCL16, CXCL10, and CX3CL1 (Fig. 6a, b), as well as IL-8 (Fig. S4b). All tested cytokines showed an  , 200 μm). c Typical organoids in the Sen and Ctrl groups expressing cell cycle arrest protein p21 WAF1/Cip1 (red) after exposure to oxidative stress for 120 h; CK19 protein was stained cyan and cell nuclei were stained blue (scale bar, 50 μm). d SA-β-gal-positive organoids (green) in the Sen and Ctrl groups following oxidative stress induction for 120 h, observed by light microscopy (scale bar, 100 μm). e mRNA expression levels of SASP components and chemokines (IL-6, CCL2, CXCL2, CXCL16, and CX3CL1) increased significantly at 120 h, compared with 24, 48, and 96 h. Data are presented as mean ± standard deviation (SD) (***P < 0.001, **P < 0.01, n = 4; ordinary one-way ANOVA) initial peak at 48 h, followed by a reduction at 96 h and a higher peak at 120 h (Fig. S4a) (all P < 0.001), in accordance with the trends in the Sen group (Fig. 3e). The reduction might have been caused by exchanging expansion medium at 48 h. Compared with the Ctrl group, significant reductions in mRNA expression levels were observed in the Exo group for IL-6 (0.1518-and 5.022-fold of Ctrl, respectively) and CCL2 (9.622-and 16.65-fold of Ctrl, respectively) (*P < 0.05, ***P < 0.001) (Fig. 6a).
We also detected the protein concentrations of SASP components by enzyme-linked immunosorbent assay and mouse proinflammatory chemokine detection kit. In the Exo group, we found that the TNF-α protein concentration also increased during the first 96 h, but was lower at 120 h, compared with the Sen and Ctrl groups (Fig. 6c). In addition, IL-6, CCL2, CXCL1, and CXCL10 protein concentrations showed decreasing trends at 120 h ( Fig. 6d-g), like the changes in their gene expression levels. Notably, IL-6, CXCL1, and CXCL10 were reduced significantly compared with the other groups (all P < 0.001). However, the protein expression levels of CX3CL1, CXCL2, CXCL15/IL-8, and CXCL16 in the Exo group were not reduced significantly compared with the other groups (Fig. S4c-f).
Overall, our results suggested a role for SASP in the pathogenesis of senescence, and demonstrate the ability of hPMSC-derived exosomes to reduce the expression of SASP components.

Discussion
In this study, we established organoids from mouse liver duct cells and characterized them morphologically and biochemically as 3D spheroidal structures with high Comparison of the immunohistochemical staining results for proliferation marker Ki67, as well as the senescence-associated markers p16 INK4a and p21 WAF1/Cip1 among the Ctrl, Sen, and Exo groups at 120 h (× 40 objective, scale bar, 200 μm; rectangle indicates × 80 objective). c-e Semiquantitative analysis of p16 INK4a , p21 WAF1/Cip1 , and Ki67 immunohistochemical staining results at 120 h. These data were acquired with a × 20 objective, and Image-Pro Plus software was used to analyze 10 randomly selected areas (two-tailed t test, mean ± standard error of the mean, n = 10, ***P < 0.001, **P < 0.01, *P < 0.05). f Percentages of SA-β-gal-positive organoids in 30 randomly selected areas, counted under a × 20 objective with bright field illumination (two-tailed t test, mean ± SD, n = 30, ***P < 0.001, **P < 0.01). g Representative CK19-positive cholangioids (cyan) in 3D form. The area of p21 WAF1/Cip1 -positive staining was greater in the Sen group than in the other two groups, while the Exo group showed the smallest area (scale bar, 100 μm). h SA-β-gal-positive cholangioids. Organoids were collected and analyzed at 120 h (green, scale bar, 50 μm) expression of cholangiocyte markers. We also identified the characteristics of long-term cultured organoids, which exhibited an aged state. Persistent oxidative stress can induce senescence and, ultimately, SASP in cholangioids. Treatment with hPMSC-derived exosomes can delay the aging process, possibly through the regulation of the cell cycle-associated inhibitor proteins p21 WAF1/-Cip1 and p16 INK4a , thereby influencing cell cycle progression. hPMSC-derived exosomes positively affected the expression levels of SASP components and chemokines in senescent cholangioids. These results are important for establishing an in vitro PSC model, and for related pathogenesis research, as well as for the development of novel therapeutic approaches.
The widely used organoid culture system was first described by Broutier et al. and Huch et al. [27,30]. In our study, we simplified the collection of liver duct cells by gravitational settling and multistep centrifugation of digestion supernatant, rather than manual collection or flow cytometry sorting. Through the establishment of cholangioids, the growth process and morphological differences were easier to observe under light microscopy. Important, organoid culture systems enable studies of human development and disease, which are not easily modeled in animals [20,31,32].
Senescent cells exhibit growth arrest in the G1 phase and fail to enter the next stage of the cell cycle, which limits proliferation [3,21]. However, in our study, organoids with oxidative stress-induced senescence were able to grow. Staining for SA-β-gal and 3D confocal microscopy for detection of p21 WAF1/Cip1 protein did not reveal these markers in all organoid cells, which suggested that senescence was not uniform [22]. Therefore, we established a standard for easier identification and quantification of SAβ-gal-positive organoids. There is also evidence that senescent cholangiocytes can induce senescence in "bystander cholangiocytes" via paracrine signaling [33]. Thus, we suggest that cholangiocyte senescence was induced by H 2 O 2 treatment, which led to senescence in bystander cells that eventually spread to the entire organoid. However, hPMSC-derived exosomes may reduce the spread of senescence. The digestion of senescent organoids and the number of "secondary organoids" can also show the effect of hPMSC-derived exosomes on senescence. Fig. 6 Exosomes reduced the expression of SASP components and chemokines in senescent cholangioids. a mRNA expression levels of SASP components (IL-6, CCL2, CXCL2, CXCL9, CXCL16, CX3CL1, and CXCL1), which decreased significantly at 120 h in the Exo group compared with the Sen group. Data are presented as mean ± SD (n = 4). b Fold changes of CXCL10 mRNA expression in organoids at different time points. Data are presented as mean ± SD (n = 4). c Tumor necrosis factor-α protein concentrations in culture supernatant at 24, 48, 96, and 120 h. Data are presented as mean ± SD (n = 3). d-g IL-6, CCL2, CXCL1, and CXCL10 protein concentrations in culture supernatant at 120 h, which were significantly decreased in the Exo group compared with the Ctrl and Sen groups. Data are presented as mean ± SD (n = 3). One-way ANOVA, ***P < 0.001, **P < 0.01, *P < 0.05 Immunohistochemical staining showed significant morphological differences among the three groups, which might had been caused by multiple factors. In the Ctrl group, organoids were larger due to their more rapid growth and then became deformed under the physical pressure associated with paraffin embedding. In the Sen group, in addition to physical pressure, deformed organoids may also have been caused by senescence-related shrinkage. However, in the Exo group, because of the anti-senescence effect of exosome treatment, cell growth was slowed and organoids remained in a proliferative state, without shrinkage. Therefore, most organoids exhibited a round shape in the Exo group. Furthermore, the expression of p21 WAF1/-Cip1 was not significantly upregulated after oxidative stress-induced injury in the Sen group compared with the Ctrl group, suggesting that p16 INK4a is more suitable than p21 WAF1/Cip1 as a marker of aging.
MSCs/stem cells have been widely used to cure diseases. Recent studies have demonstrated the beneficial effects of MSC/stem cell-derived exosomes on senescence [14,16,34,35]. To the best of our knowledge, few studies have investigated the effect of MSC-derived exosomes on cholangiocyte senescence in PSC, such that the detailed mechanism remains unknown. Some studies have demonstrated a protective effect of stem cellderived exosomes on liver injury mice [11,36]. Our study provides a novel 3D cell model complementing in vivo experiments showing the protective effect of hPMSC-derived exosomes. An important limitation of this study was that it did not include in vivo analysis of the protective effect of exosomes, which would be included in future experiments.
Treatment with CCL2 and CXC3L1, components of SASP, has been shown to induce cellular senescence in biliary epithelial cells, suggesting a self-amplifying secretory network in which SASP reinforces cellular senescence [26,37,38]. In our study, hPMSC-derived exosomes reduced the expression of SASP components and chemokines, such as CCL2, CX3CL1, IL-6, TNF-α, CXCL1, and CXCL10, which are involved in the recruitment of immune cells, terminal differentiation of B cells, and the direct damage caused by cytotoxic T lymphocytes [3]. It is plausible that exosomes exert a protective effect by reducing certain SASP components and may weaken the self-amplifying effect in senescence, thereby influencing immune cell systems and regulating the senescent microenvironment [39]. The underlying mechanisms of these putative effects remain unclear.
In conclusion, our in vitro model of oxidative stressinduced senescence in cholangiocytes can be used to study PSC and PBC. It may also be useful for investigating the mechanisms underlying the protective effect of MSC-derived exosomes in senescent cholangioids. The insights provided by this study may lead to a novel model of the mechanisms of action of senescent cholangiocytes, which could in turn lead to treatments for diseases such as PSC and PBC.  Table S1. Regents of organoids analysis.
Additional file 3: Table S2. Primer sequences used in this study.
Additional file 4: Figure S1. Characteristics of organoids in H 2 O 2induced senescence. (a) Viability / cytotoxicity assay of organoids before oxidative stress induction. Almost all the organoids were alive (green), with a small number of single cells dead (red) (scale bar, 100 μm). (b) Typical appearance of organoids after 120 h H 2 O 2 induction, observed by light microscopy. Cells scattered and aggregated inside the lumen of organoids (scale bar, 200 μm).
Additional file 5: Figure S2. The protective effect of hPMSCs-derived exosomes on senescent cholangioids. (a) Supplement fields to Fig. 5a, organoids in Sen and Exo group at 120 h. Typical senescent organoids were noted with white lines circle (scale bar, 200 μm). (b) Analysis of randomly selective fields of senescent organoids in Sen and Exo group at 120 h (two tailed t-test, mean ± SD, n = 4, *P < 0.05). (c) Detailed immunofluorescent staining pictures of typical organoids in Ctrl, Sen, and Exo group after oxidative stress induction for 120 h. Cell-cycle-arrest protein p21 WAF1/Cip1 were stained red, CK19 protein were stained cyan, and the cell nuclei were stained blue (scale bar, 100 μm). (d) Typical photos of "secondary organoids" passage from Ctrl, Sen and Exo group, which were cultured for 1 and 5 days (scale bar, 100 μm). (e) Numbers of "secondary organoids". Organoids in Ctrl, Sen and Exo group were passaged at 120 h. The experiments were repeated for three times. After 1 and 5 days of culture, three fields of each experiment were randomly selected under 4 × object lens and counted using Image J software (ordinary one-way ANOVA, mean ± SD, n = 9, ***P < 0.001, **P < 0.01).