The preparation of PD-MSCs was approved by the CHA Gangnam Medical Center Institutional Review Board (Seoul, Korea, No.: IRB 07-18) and isolated as previously described [13]. WB-F344 (a diploid epithelial cell line from normal adult rat liver with phenotypic properties of “oval” cells) and T-HSC/Cl-6 (rat HSCs) were cultured in α-MEM (HyClone, Logan, UT, USA) with 10% FBS (Gibco, Carlsbad, CA, USA) and 1% penicillin/streptomycin (P/S, Gibco). Primary hepatocytes from 7-week-old male Sprague–Dawley rats (Orient Bio Inc., Seongnam, Korea) were isolated via a two-step pronase/collagenase perfusion and cultured in William’s E medium (Sigma-Aldrich, St. Louis, MO, USA) with 10% FBS (Gibco), 1% P/S (Gibco), and 4 mM L-glutamine (Gibco). All cells were cultured at 37°C in 5% CO₂. To analyze EMT induction and its reversal, each cell was treated with TGFB1 (2 ng/mL, PeproTech, Rocky Hill, NJ, USA) for 48 and 72 hours. Cells were then cocultured with PD-MSCs or PD-MSCs^PRL-1 (5×10³ cells/cm²) in Transwell inserts (8-μm pore, Corning, NY, USA) using α-MEM (HyClone) with 10% FBS (Gibco) and 1% P/S (Gibco) for 24 hours.

To induce overexpression in PD-MSCs, a human PRL-1 plasmid was purchased from Origene Inc. (Rockville, MD, USA) and transfected into naïve PD-MSCs using the 4D Nucleofector™ system (Lonza, Basel, Switzerland), as previously described [14]. For gene knockdown, small interfering (si) RNA targeting PRL-1 and BMP7, along with scrambled RNA, were obtained from Integrated DNA Technologies (San Jose, CA, USA). Cells at 80% confluence were transfected with 50 nM siRNA using Lipofectamine RNAiMAX following the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA).

A cytokine antibody array analysis was performed using the Proteome Profiler Human XL Cytokine Array Kit (R&D Systems, Minneapolis, MN, USA). In brief, the supernatants from cultured PD-MSCs and PD-MSCs^PRL-1 were collected. The array procedure strictly adhered to the manufacturer’s instructions. Each membrane was equipped with 105 cytokine antibodies, each of which was spotted in duplicate.

Animal protocols were approved by the CHA University Animal Care Committee (IACUC-190048). Male Sprague–Dawley rats (7 week-old, 200 g; Orient Bio Inc.) were maintained under specific pathogen-free conditions with a 12-hour light/dark cycle at 22–25°C. Af ter 1 week of acclimatization, liver cirrhosis was induced by BDL for 10 days [15]. Rats were randomly divided into groups: BDL-induced non-transplanted (BDL NTx; n=20) and sham controls (Con; n=6). PKH67 (Sigma-Aldrich)-labeled PD-MSCs (n=20) and PD-MSCs^PRL-1 (n=20) were transplanted intravenously (2×10⁶ cells/animal). Rats were sacrificed at 1, 2, 3, and 5 weeks for analysis.

The levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, and albumin were assessed in individual serum samples from each group (n=5). Each serum sample was subjected to analysis upon test request (Southeast Medi-Chem Institute, Busan, Korea).

For histopathological analysis, liver specimens (n=5/group) were fixed in 10% neutral buffered formalin, embedded in paraffin, and stained with hematoxylin & eosin (H&E), Masson’s trichrome, and Sirius red. For immunohistochemistry, liver sections were deparaffinized, hydrated, and treated with 3% hydrogen peroxide in methanol. After antigen retrieval, specimens were incubated in Protein Block solution (DAKO, Santa Clara, CA, USA), followed by antibodies against PCNA (1:50, Santa Cruz Biotechnology Inc., Dallas, Texas, USA), PRL-1 (1:200, Abcam, Cambridge, UK), and BMP7 (1:100, Abcam) at 4°C overnight. 3,3-Diaminobenzidine (EnVision Systems, Santa Clara, CA, USA) was used to generate chromatic signals, and the samples were counterstained with Mayer’s hematoxylin (DAKO). Whole liver tissue images were analyzed with a digital slide scanner (3DHISTECH Ltd., Budapest, Hungary). For immunofluorescence (IF), primary antibodies against E-cadherin (1:200, Cell Signaling Technology), BMP7 (1:200, Abcam), p-SMAD2 (1:200, Cell Signaling Technology), Ki -67 (DAKO), and HNF4A (Novus Biologicals, Littleton, CO, USA) were used in diluent solution (DAKO). 4’,6-Diamidino-2-phenylindole (DAPI) (Invitrogen) was used as a counterstain and slides were visualized by confocal microscopy (LSM 700). Images were analyzed with ZEN blue software (Zeiss).

Liver tissues and cells were lysed with Pierce immunoprecipitation (IP) buffer (Thermo Fisher, Waltham, MA, USA) for Co-IP and RIPA buffer (Sigma-Aldrich) for immunoblotting containing protease and phosphatase inhibitors. For Co-IP, lysates were incubated with 2 μg of anti-PRL-1 (Abcam) overnight with gentle rotation. Protein G agarose (Thermo Fisher) was added to the antibody-labeled lysates, followed by the mixture bieng incubated for 2 hours at 4°C with gentle rotation. The bound proteins were eluted and immunoblotting was performed to detect BMP7 using anti-BMP7 (Abcam). For immunoblotting, total protein or nuclear-cytoplasmic fractioned (Thermo Fisher) samples (40 μg) from tissues, cells, nuclei, and cytoplasm were separated by SDS-PAGE. Primary antibodies against ALB, BCL2, COL1A1 (Novus Biologicals), ACTA2 (Sigma-Aldrich), BAX, VIM, SNAI1, p-SMAD1/5, GP130 (Santa Cruz Biotechnology), PARP, total/p-SMAD2, TJP, CDH1, total/p-STAT3, interleukin (IL)-6, CDK4, CCND1, CASPASE-3 (Cell Signaling Technology, Danvers, MA, USA), PRL-1, BMP7, and total SMAD1/5 (Abcam) were used. GAPDH (AbFrontier, Seoul, Korea) and LMNB1 (Cell Signaling Technology) were loading controls for cytoplasmic and nuclear proteins. HRP-conjugated secondary antibodies were applied for 1 hour incubation and chemiluminescence detection.

Total RNA was extracted from liver tissues from each group or cells with TRIzol LS reagent (Invitrogen), and Quantitative real-time polymerase chain reaction was performed using SYBR Green PCR Master Mix (Applied Biosystems, Waltham, MA, USA) on a CFX Connect™ Real-Time System (Bio-Rad, Hercules, CA, USA) as previously described [16]. The primers are listed in the Supplementary Table 1.

Exogenous levels were measured using Enzyme-linked immunosorbent assay (ELISA) kits using rat IL-6 and rat IL-10 (R&D Systems), rat TGFB1, rat/human BMP7 (Abcam) in rat serum, and human PRL-1 (MyBioSource, Inc., San Diego, CA, USA) in cell culture supernatant, following the manufacturer’s instructions. The absorbance was measured using an Epoch microplate reader (BioTek, Winooski, VT, USA) at 450 nm. The experiments were performed in triplicate.

WB-F344 were cultured in 6-well plates at a density of 2×10⁵ cells/well. The following day, a sterile 1,000 μl tip was used to create a straight-line scratch across the monolayer and subsequently they were treated with TGFB1 (PeproTech) for 48 hours, followed by treatment with PRL-1 (Novus Biologicals) and Pentamidine (Sigma-Aldrich), which is a PRL-1 inhibitor, respectively. The extent of wound closure was quantified by measuring the gap width at identical locations using ImageJ software (NIH).

All data were analyzed using GraphPad Prism version 9.0 (GraphPad Software, San Diego, CA USA), and represented as means±standard deviations. Statistical analysis was performed using two-tailed unpaired Student’s t-test and one-way ANOVA, and the difference was considered significant when the P-value was less than 0.05. All data are representative of at least three independent experiments.

In the BDL-induced cirrhosis model [15], rats were divided into four groups: BDL-induced nontransplantation (BDL NTx), naïve PD-MSCs, PD-MSCs^PRL-1 transplantation, and Con (Fig. 1A). Prior studies underscored the superior antifibrotic efficacy of PD-MSCs^PRL-1 transplantation, mediated by mitochondrial activation [12]. Histopathological evaluation revealed an elevated biliary ductal cell count in the BDL group, fostering tissue restitution while exacerbating fibrogenesis, inversely associated with survival rates (Fig. 1B). Collagen deposition was markedly augmented in the BDL NTx group compared to controls, whereas transplantation groups, particularly PD-MSCs^PRL-1, demonstrated substantial fibrosis attenuation, corroborated by Sirius red and Masson’s trichrome staining (Fig. 1C). Fibrotic markers COL1A1 and ACTA2 were significantly elevated in the BDL group but reduced in both transplantation groups, with PD-MSCs^PRL-1 showing greater reductions (Fig. 1D). To assess the recuperation of hepatic functions facilitated by transplantation, measurements of ALT, AST, and total bilirubin levels were conducted in the serum of rats. Liver function tests revealed that PD-MSCs^PRL-1 significantly decreased ALT, AST, and total bilirubin levels compared to naïve PD-MSCs (Fig. 1E). Additionally, PD-MSCs^PRL-1 increased IL-10 levels, indicating an enhanced anti-inflammatory response (Fig. 1F). These data demonstrate that PD-MSCs^PRL-1 transplantation effectively mitigated hepatic fibrosis and enhanced hepatic function in a rat model subjected to BDL, as compared to PD-MSCs transplantation.

We further analyzed secreted factors by cytokine array in the supernatant from PD-MSCs^PRL-1 to demonstrate pathways associated with liver regeneration. Upregulated factors included FGF-2, GDF-15, HGF, and IL-6, while downregulated factors included DKK-1 and BDNF (Fig. 2A). To investigate hepatic regeneration, we focused on the IL-6/STAT3 pathway, known for its key role in liver regeneration [17]. IL-6 levels were significantly altered at 1 and 2 weeks post-PD-MSCs^PRL-1 transplantation (Fig. 2B), with activation of IL-6, GP130, and p-STAT3 observed up to 3 weeks. A strong correlation (R=0.5852) was found between IL-6 and p-STAT3 levels (Fig. 2C). Serum and endogenous albumin levels were elevated in the PD-MSCs^PRL-1 group compared to naïve. We further examined apoptosis and cell proliferation markers in liver tissues. BCL2 expression was reduced in the BDL-induced NTx group but elevated in both transplantation groups, with PD-MSCs^PRL-1 showing higher levels at 3 weeks. Proapoptotic markers c-PARP and c-CASPASE-3 were reduced in both transplantation groups, with a greater reduction in the PD-MSCs^PRL-1 group compared to PD-MSCs. Additionally, PD-MSCs^PRL-1 significantly suppressed BAX expression, indicating a reduction in apoptosis (Fig. 2D). In our earlier findings, PD-MSCs transplantation in a rat model of CCl₄-induced liver injury resulted in elevated BAX levels compared to the BDL-induced NTx group, indicating the initiation of an autophagic mechanism [7]. Cell cycle markers CDK4 and CCND1 were consistently elevated in the PD-MSCs^PRL-1 group across all weeks (Fig. 2E), and PCNA levels, assessed by immunohistochemistry, were also increased (Fig. 2F). Moreover, Ki-67 and HNF4A co-staining suppor ted a hepatocyte proliferation in WB-F344 cells, followed by pentamidine treatment , a pharmacological inhibitor of PRL-1 (Supplementary Fig. 1). These results suggest that PD-MSCs^PRL-1 transplantation enhances liver regeneration by inhibiting apoptosis and promoting hepatocyte proliferation in the BDL rat model.

Mesenchymal phenotypes, such as fibrosis or cirrhosis in liver tissues, often escalate under injurious conditions via EMT [18]. To determine whether MSC transplantation affects EMT progression by modulating TGFB1/SMAD2 signaling in the liver, we first measured TGFB1 levels in rat serum. The PD-MSCs^PRL-1 group showed a two-fold reduction in TGFB1 levels compared to the PD-MSCs group, with a strong positive correlation (R=0.6430, Fig. 3A). Liver tissue, consisting of hepatocytes and cholangiocytes, showed mesenchymal phenotypes near the central vein in BDL-induced rats (Fig. 3B). After transplantation, there was a significant reduction in EMT markers, including Tgfb1, vimentin (Vim), and Snai1, compared to the NTx group, with the PD-MSCs^PRL-1 group showing the most pronounced decrease (Fig. 3C). Protein expression of VIM and SNAI1, and especially phosphorylated SMAD2 (p-SMAD2) both total and nuclear, was also substantially reduced following PD-MSCs^PRL-1 transplantation (Fig. 3D). While p-SMAD2 localized in the cytoplasm of hepatocytes in normal liver tissue, it translocated to the nucleus in the BDL NTx group which was reversed by PD-MSCs and PD-MSCs^PRL-1 transplantation (Fig. 3E). These findings indicate that PD-MSCs^PRL-1 transplantation effectively suppresses EMT progression by modulating the TGFB1/SMAD2 signaling pathway in BDL-injured rat livers (Fig. 3F).

EMT induced by TGFB1 can inhibit BMP signaling, while BMP7 promotes MET [19]. To evaluate the impact of PD-MSCs^PRL-1 on BMP7 levels in the BDL rat model, we measured serum BMP7 levels and its corresponding mRNA expression (Fig. 4A). A strong positive correlation (R=0.8121) was found between serum and mRNA Bmp7 levels. Additionally, phosphorylated SMAD1/5 (p-SMAD1/5), a downstream mediator of BMP7, was activated in both transplantation groups (Fig. 4B). Immunohistochemical analysis revealed that BMP7 and PRL-1 were predominantly localized in hepatocytes rather than cholangiocytes in rat liver tissues, with the PD-MSCs^PRL-1 group showing higher expression (R=0.7261) (Fig. 4C). In addition, the mRNA expression of epithelial markers (e.g., desmoplakin; Dsp, cadherin 1; Cdh1, and tight junction protein; Tjp1) was dramatically increased by PD-MSCs^PRL-1 transplantation compared to PD-MSCs transplantation (Fig. 4D) as well as protein levels (TJP and CDH1). IF confirmed elevated BMP7 and CDH1 levels in hepatocytes following PD-MSCs^PRL-1 transplantation (Fig. 4E). These findings suggest that PD-MSCs^PRL-1 activates BMP7 signaling in hepatocytes of cirrhotic livers, reducing EMT via the SMAD1/5 pathway (Fig. 4F).

To induce the progression of EMT in primary rat epithelial hepatocytes, a treatment with TGFB1 (2 ng/mL) for 48 hours was executed. To induce EMT in primary rat epithelial hepatocytes, TGFB1 (2 ng/mL) was applied for 48 hours. Subsequently, cells were cocultured with PD-MSCs or PD-MSCs^PRL-1 for 24 hours. Prior to experimentation, WB-F344 cells were pretreated with TGFB1 for 48 or 72 hours, resulting in a mesenchymal phenotype, confirmed by reduced Cdh1 and increased Vim and Snai1 expression (Supplementary Figs. 2, 3). TGFB1 treatment increased mesenchymal markers, but coculturing with PD-MSCs or PD-MSCs^PRL-1 significantly reduced TGFB1 expression (Fig. 5A). There was a strong negative correlation between TGFB1 and BMP7 expression (R=–0.9461). PRL-1 overexpression in PD-MSCs remarkably increased BMP7 and CDH1 levels, while PRL-1 knockdown led to the loss of the epithelial phenotype and decreased BMP7/SMAD1/5 signaling in primary hepatocytes (Fig. 5B). In the control group, both BMP7 and TGFB1 were secreted at approximately 500–700 pg/mL, but TGFB1-treated hepatocytes showed BMP7 loss and TGFB1 gain (Fig. 5C). Coculture with PD-MSCs or PD-MSCs^PRL-1 decreased mesenchymal markers (TGFB1, SNAI1, COL1A1) and increased epithelial markers (BMP7, DSP, CDH1) at both mRNA and protein levels (Fig. 5B, 5D). PRL-1 knockdown in PD-MSCs led to worsened EMT, indicated by increased p-SMAD2 activity in hepatocytes (Fig. 5E). These findings suggest that PD-MSCs^PRL-1 coculture suppresses the mesenchymal phenotype by downregulating p-SMAD2 and promotes the epithelial phenotype by upregulating BMP7/SMAD1/5 signaling (Fig. 5F).

Given the strong positive correlation between PRL-1 and BMP7 in liver tissues, our analysis aimed to identify the upstream regulator between these two factors. Initially, we confirmed elevated expressions of both PRL-1 and BMP7 in human PD-MSCs^PRL-1. Knockdown of PRL-1 in PD-MSCs resulted in diminished expression levels of both BMP7 and PRL-1, whereas silencing BMP7 had no discernible effect on PRL-1 expression, indicating the regulatory role of PRL-1 in modulating BMP7 levels. Recombinant PRL-1 treatment was administered to naïve PD-MSCs, resulting in increased PRL-1 and BMP7 mRNA and protein levels (Fig. 6A, Supplementary Fig. 4A–4C). Further validation by costaining and Co-IP assay confirmed the interaction between PRL-1 and BMP7 (Fig. 6B), not SMAD1/5. We conducted BMP7 knockdown in primary hepatocytes and modulated p-SMAD1/5 and p-SMAD2 expression (Fig. 6C). To substantiate the role of PRL-1, we observed that treatment with recombinant PRL-1 markedly suppressed cell migration in WB-F344, whereas the application of pentamidine exhibited an opposing effect by a wound healing assay (Fig. 6D). Importantly, PRL-1 treatment alone upregulated p-SMAD1/5 and downregulated p-SMAD2. Conversely, pentamidine increased p-SMAD2 expression (Fig. 6E). These findings suggest that PRL-1 directly regulates BMP7, and that BMP7, under the influence of PRL-1, modulates p-SMAD1/5 and p-SMAD2 activity in rat liver epithelial cells (Fig. 6F).