HLSCs were isolated from either unused deceased donor liver grafts that were unsuitable for liver transplant, or the non-tumor portion of livers from patients undergone surgical resection for benign liver tumors. Written informed consent were obtained from all participants that provided tissue samples. This study was approved by the Institutional Review Board at Samsung Medical Center (No. 2003-10-021, No. 2017-04-042). We analyzed the characteristics of HLSCs from 14 donors for this study and ultimately used HLSCs from 3 donors (HLSC7, HLSC15, HLSC26) for transplantation. Hepatocyte preparation was conducted using a minor modification of the two-step collagenase perfusion technique (8, 9). In brief, the liver was perfused via the portal vein with a warm oxygenized perfusion buffer (NaCl 8 g/l, KCl 0.4 g/l, NaH₂PO₄∙2H₂O 0.06 g/l, Na₂HPO₄∙12H₂O 0.06 g/l, glucose 2.7 g/l, NaHCO₃ 1.05 g/l, HEPES 2.38 g/l, and EDTA 0.19 g/l) and liver digestion was accomplished by recirculating a perfusion buffer supplemented with collagenase (0.3 g/l) and calcium chloride (0.735 g/l) at 37℃ to break down the ECM. The cell suspension was centrifuged three times at 50×g for 3 minutes at 4℃ to remove hepatocytes. Then the supernatant was collected from the liver cell suspensions and the cells were collected by centrifugation at 200×g for 5 minutes at 4℃. The pellet was resuspended in phosphate-buffered saline (PBS), layered on a Histopaque cushion, and centrifuged at 800×g for 20 minutes at 4℃. Cells at the interface were collected and washed. HLSCs were cultured on plastic dishes at a density of 5×10⁴ cells per cm² in Dulbecco’s modified eagle medium with 10% fetal bovine serum. In the experiment, we used the sixth passage HLSCs.

Sprague Dawley (SD) rats were obtained from OrientBio (Seongnam-si, Korea). All rats were bred under conditions in accordance with the ethical guidelines for the care of mice at the Laboratory Animal Research Center at Samsung Medical Center (Seoul, Korea).

The experimental scheme is outlined in Fig. 1. Cirrhosis was induced by intraperitoneal injection of Carbon tetrachloride (CCl₄, Aldrich chemical Co., Milwaukee, USA) dissolved in paraffin oil (1：1). Liver fibrosis was induced in rats using 2 protocols with varying durations of CCl₄ treatment (Supplementary Table S1). Rats received 0.65 ml/kg of CCl₄ twice weekly for 8 weeks and underwent HLSC transplantation (1×10⁷ cells) with simultaneous left hepatic lobectomy in the moderate toxicity protocol. CCl₄ treatment was continued for 6 subsequent weeks at 0.5 ml/kg twice weekly. In the intense toxicity protocol, rats received 0.75 ml/kg of CCl₄ twice weekly for 10 weeks and underwent HLSC transplantation (1×10⁷ cells) with simultaneous left hepatic lobectomy. CCl₄ treatment was continued for 4 subsequent weeks at 0.5 ml/kg twice weekly. Animals in both groups were sacrificed following completion of the 14-weeks course of CCl₄ treatment before and after HLSC transplantation. Liver specimens from the sacrificed rats (n=5 in each group) were analyzed (Fig. 1). Histological and immunohistochemistry analyses of the left hepatic lobe was done to provide baseline parameters of liver fibrosis at the time of HLSC transplantation. Rats in the control group underwent 14 weeks of CCl₄ treatment without HLSC transplantation.

HLSCs were infused at a dose of 1×10⁶ cells and saline was infused at the same volume into the portal vein. Three schemes for HLSC transplantation were applied. First was the standard transplant scheme, where 1×10⁷ cells were transplanted once following 10 weeks of CCl₄ treatment (HLSCx group). Second was a double-dose scheme, where 2×10⁷ cells were transplanted following 10 weeks of CCl₄ treatment (HLSC2x group). Third was a tandem transplant scheme, where two HLSC transplant sessions with 1×10⁷ cells were done in 2-week intervals, following 10 weeks of CCl₄ treatment (HLSCxx group). Rats continued to be treated with CCl₄. After cell infusion, rats were also given cyclosporine at 5 mg/kg body weight daily for two weeks to suppress immunity. At 4 weeks after HLSC transplantation, rats were euthanized and the liver and blood sample were obtained.

Rat liver tissue samples were immediately fixed in 10% neutral buffered formalin at 4℃ and routinely processed for paraffin embedding and cross-sectioned to obtain 4 μm-thick sections with circular layer and myenteric ganglia cut longitudinally. Before use, sections were deparaffini-zed, rehydrated and processed for routine haematoxylin/eosin and histochemical staining.

The slides were incubated with a 0.1% Direct Red 80 (Sigma-Aldrich, Missouri, USA) dissolved in aqueous saturated picric acid (Sigma-Aldrich, Missouri, USA) for 1 hour, washed in acidified water (0.5% acetic acid (glacial). Dehydrated and mounted with VECTOR Mounting (VECTOR, California, USA). Collagen and non-collagen components were red- and orange-stained, respectively.

Digital images were captured with an Aperio Scanscope System AT (Leica, Wetzlar, Germany) and the area of fibrosis was quantified in 20×-magnification fields five area per specimen. An RGB (Red, Green, Blue) threshold was used to identify the areas of fibrosis using Imagescope software (Leica, Wetzlar, Germany). Fibrosis was expressed as percentage (%) of total area.

For IHC, formalin-fixed rat liver tissue sections were deparaffinized and rehydrated, then heated at 100℃ for 30 min in a HIER Citrate Buffer (0.01 M), (Zytomed, Berlin, Germany) for antigen retrieval. The sliced tissues were later dealt with Hydrogen Peroxide Block (Thermo, Massachusetts, USA) to block the endogenous peroxidase activity. Next, the slides were blocked with Ultra V Block (Thermo, Massachusetts, USA). After rinsing in wash buffer, the sections were incubated with anti-alpha 1 antitrypsin antibody (Abcam, Cat. No. ab9373, Cambridge, UK), rabbit polyclonal to human serum albumin (Abcam, Cat. No. ab2406, Cambridge, UK), and monoclonal mouse anti-human cytokeratin 18 (Dako, Cat. No. M7010, Santa Clara, USA) for 1 hour at room temperature. The slides were then rinsed in wash buffer and incubated for 15 min with secondary antibody horseradish peroxidase-conjuga-ted (HRP Polymer, Thermo, Massachusetts, USA). Immune complexes were detected through the standard substrate detection of HRP. Last, the slides were stained with hematoxylin and dehydrated in graded alcohols and xylene. Digital images were captured with an Aperio Scanscope System AT (Leica, Wetzlar, Germany).

Total RNAs were extracted from fresh rat liver tissue with Trizol reagent (Invitrogen Life Technologies, Carlsberg, USA), following the manufacturer’s protocol (Invitrogen). To remove Genomic DNA, the RNA samples were treated with DNase I (Promega, Wisconstin, USA). Total RNA was extracted from HLSCs with RNeasy mini kit (Qiagen, Hilden, Germany). The A₂₆₀/A₂₈₀ ratio of the optical density of the RNA samples (measured with NanoDrop 1000; peQLab, Erlangen, Germany) was between 1.95 and 2.05 indicating adequate RNA quality. The reverse transcrip-tion was performed with the total RNAs (500 ng) and the Superscript III Reverse Transcriptase (Invitrogen Life Technologies, Carlsberg, USA).

Real-time qPCR was performed with the SensiFAST SYBR HI-ROX 2X Mix Kit (BioLIne, Toronto, Canada) by the 7900HT Fast Real-Time PCR system (Applied Biosystem, California, USA). The sequences of the specific primers used were as follows: Collagen 1A2 (Col1A2), 5’-TAAGGCTGTCATTCTGCAGGGC-3’ and 5’-GTCCAG AGGTGCAATGTCAAGG-3’; MMP2, 5’-CTGATAACCTG GATGCAGTCGTG-3’ and 5’-CAGCCAGTCCGATTTGA TGCTTC-3’; MMP1, 5’-AGCTGAAGGTAGAGCCAGGG TAC-3’ and 5’-AGAACGACAGCAGCTGCACTCAC-3’; TIMP1, 5’-TCCCTTGCAAACTGGAGAGTGAC-3’ and 5’-GGTATTGCCAGGTGCACAAATC-3’; TIMP2, 5’-ATCA CACGCTGCCCTATGATCC-3’ and 5’-CATCACTTCTC TTGATGCAGGC-3’; Transforming growth factor beta 1 (TGF-b1), 5’-TAGGAAGGACCTGGGTTGGAAG-3’ and 5’-CACCCGGGTTGTGTTGGTTGTAG-3’; hepatocyte growth factor (HGF), 5’-AGCTAGTCTATGGACCTGAAGGC-3’ and 5’-GTGTAGCCCCAGCCGTAAATACT-3’. Each sample contained 0.8 μl cDNA, 2X SensiFAST SYBR HI-ROX Mix, and 10 μM each primer in final volume of 20 μl. Amplification conditions were as follows: one cycle of denaturation at 95℃ for 2 minutes, 40 cycles denaturation at 95℃ for 5 seconds, annealing at 60℃ for 30 seconds, extension at 72℃ for 15 seconds, and melting curve at 55℃ with the temperature gradually increased for 0.5℃, up to 95℃. β-actin mRNA was used as an internal control for real-time qPCR. The results were analyzed with the 2^−ΔΔCt method. To prove the correct lengths of the qPCR products, the samples were analyzed with agarose gel electrophoresis (Bioneer, Daejeon, Korea).

Aspartate aminotransferase (AST, Fuji dri-chem slide GOT/AST, Fujifilm, Japan), alanine aminotransferase (ALT, Fuji dri-chem slide GPT/ALT, Fujifilm, Japan) were determined using an automatic biochemical analyzer (Dri-chem 7000i, Fujifilm, Japan).

Gross appearance of the liver observed at the time of left hepatic lobectomy with or without HLSC transplantation (at 8 weeks following moderate toxicity protocol and at 10 weeks following intense toxicity protocol, “baseline”) showed cholestatic discoloration with surface nodularity and blunt edges. Sirius red staining of the liver specimens revealed deposition of collagen fibers in the extracellular (Fig. 2A). The liver specimens of rats were then analyzed at post-14 weeks of CCl₄ treatment. In the HLSC transplantation group of both CCl₄ treatment protocols, the cholestatic discoloration and surface nodularity appeared to have been alleviated and Sirius red staining showed deposition of collagen fibers to a lesser degree. Presence of hALB secreting cells were detected by immunohistochemistry in all post-14 weeks liver specimens of HLSC transplantation group. In the control group, both gross appearance of the liver and Sirius red staining showed evidence of persistent liver fibrosis (Fig. 2A).

Quantitative measurements of fibrosis area stained by Sirius red were compared between baseline and post-14 weeks liver specimens. Fibrosis area significantly decreased from baseline to post-14 weeks in the control group as well as the HLSC transplantation group in the moderate toxicity protocol (Fig. 3A). In the intense toxicity protocol, fibrosis area increased in the control group during the final 4 weeks of CCl₄ treatment. However, fibrosis area significantly decreased from baseline to post-14 weeks in the HLSC transplantation group in the intense toxicity protocol (Fig. 3A).

RNA expression in liver tissues at baseline and post-14 weeks were analyzed (Fig. 2B). TIMP1 and TIMP2 expression were significantly decreased, while hepatocyte growth factor (HGF) expression was increased following HLSC transplantation in liver tissues of rats in the intense toxicity protocol.

Gene expression levels of factors contributing to ECM remodeling were analyzed among the HLSCs from multiple donors (Supplementary Fig. S1). Compared to MSCs, HLSCs expressed elevated levels of MMP1, TIMP2, and HGF, while MMP2, TIMP1, and TGF-b expression were decreased. Three HLSC strains (HLSC7, HLSC15, HLSC26) were selected for transplantation based upon elevated expression of MMP1 and decreased expression of TIMP1, MMP2, and TGF-b. An overall net effect towards ECM degradation by HLSCs may be expected with this pattern of mRNA expression. The comparative effects of transplantation of three HLSC strains were analyzed in rats undergone liver fibrosis induction using the intense toxicity protocol.

The liver specimens of rats (n=5 in each group) were analyzed at post-14 weeks of CCl₄ treatment (Fig. 4A). In the three HLSC transplantation groups, the cholestatic discoloration and surface nodularity appeared to have been alleviated and Sirius red staining showed deposition of collagen fibers to lesser degrees. Presence of hALB secreting cells were detected by immunohistochemistry in post-14 weeks liver specimens of all three HLSC strains. Fibrosis area significantly decreased from baseline to post-14 weeks following HLSC transplantation in HLSC7 group (p<0.01). Fibrosis area significantly increased in the control group from baseline to post-14 weeks (Fig. 3B). AST and ALT levels significantly decreased after 4 weeks following HLSC transplantation of all three HLSC strains. APRI, which is a surrogate marker for fibrosis, significantly decreased after 4 weeks following HLSC transplantation of all three HLSC strains. These changes were not observed in the control group (Table 1). RNA expression in liver tissues at baseline and post-14 weeks were analyzed (Fig. 4B). TIMP1 and TIMP2 expression were significantly decreased, while HGF expression was increased following HLSC7 and HLSC26 transplantation.

The liver specimens of rats (n=5 in each group) were then analyzed at post-14 weeks of CCl₄ treatment (Fig. 5A). Fibrosis area decreased from baseline to post-14 weeks following HLSC transplantation with all three schemes (HLSCx, HLSC2x, HLSCxx) compared to a significant increase in fibrosis area in control group (p<0.05). However, only HLSCx group showed significant decrease in fibrosis area following HLSC transplantation (p<0.05, Fig. 3C). AST, ALT levels and APRI significantly decreased after 4 weeks following HLSC transplantation in HLSCx scheme. No significant decrease in APRI were observed in the control group (Table 1). TIMP1 and TIMP2 expression were significantly decreased, while HGF expression was increased following single transplant of HLSC (HLSCx). TGF-b expression were also significantly decreased in this group (Fig. 5B).