Comparative effects of hepatocyte growth factor and tacrolimus on acute liver allograft early tolerance

Allostimulated CD8+ T cells (aCD8+ T cells), as the main mediators of acute liver rejection (ARJ), are hyposensitive to apoptosis due to the inactivation of death receptor FAS-mediated pathways and fail to allow tolerance induction, eventually leading to acute graft rejection. Although tacrolimus (FK506), the most commonly used immunosuppressant (IS) in the clinic, allows tolerance induction, its use is limited because its target immune cells are unknown and it is associated with increased incidences of malignancy, infection, and nephrotoxicity, which substantially impact long-term liver transplantation (LTx) outcomes. The dark agouti (DA)-to-Lewis rat LTx model is a well-known ARJ model and was hence chosen for the present study. We show that both hepatocyte growth factor (HGF) (cHGF, containing the main form of promoting HGF production) and recombinant HGF (h-rHGF) exert immunoregulatory effects mainly on allogeneic aCD8+ T cell suppression through FAS-mediated apoptotic pathways by inhibiting cMet to FAS antagonism and Fas trimerization, leading to acute tolerance induction. We also showed that such inhibition can be abrogated by treatment with neutralizing antibodies against cMet (HGF-only receptor). In contrast, we did not observe these effects in rats treated with FK506. However, we observed that the effect of anti-rejection by FK506 was mainly on allostimulated CD4+ T cell (aCD4+ T cell) suppression and regulatory T cell (Treg) promotion, in contrast to the mechanism of HGF. In addition, the protective mechanism of HGF in FK506-mediated nephrotoxicity was addressed. Therefore, HGF as a tolerance inducer, whether used in combination with FK506 or as monotherapy, may have good clinical value. Additional roles of these T-cell subpopulations in other biological systems and studies in these fields will also be meaningful.


Introduction
Liver transplantation (LTx) is a life-saving procedure for endstage liver disease (ESLD), and donor-specific tolerance induction and suitable use of immunosuppressants (ISs) are essential for achieving success. Tacrolimus (FK506) is currently the most commonly used IS in the clinic (1). It was discovered in 1984 and was first used in the clinic as an IS by Starzl et al. in 1989 (2). Its classic mechanism through IL-2 inhibition is well known (3). Interference through other inflammatory cytokines such as interferon-gamma (IFN-g) and tumor necrosis factor (TNF)-ɑ has also been reported (4). Although FK506 has been the primary immunosuppressive agent used in multiple organ transplantation since it was initially shown to be more effective in preventing graft loss than other ISs in a clinical trial (1,5), life-long FK506 treatment increases the incidences of malignancy, infection, and nephrotoxicity, thus limiting its efficacy. As such, identification of more reliable ISs to improve graft outcomes in transplant recipients is urgently needed. In recent years, hepatocyte growth factor (HGF) has attracted attention because of its studies in the field of LTx, in particular, its showing a possibility to go across the CD8 + T cell alloimmune barrier (6,7).
HGF is a multifunctional, pleiotropic protein with mitogenic, motogenic, and morphogenic effects in a wide variety of cells in developing and regenerative livers (8). Based on its recent immunomodulatory effect on xenogeneic rodents and pigs during organ transplantation (9)(10)(11), we investigated whether HGF has a similar immunomodulatory role in the DA-to-Lewis allogeneic liver transplant rodents, a well-defined allogeneic acute liver rejection model with a well-done operative procedure of orthotopic liver transplantation (OLT) (namely, allo-OLT here). Taking into account the results of our previous study on xenostimulated CD8 + T cell apoptosis induced by HGF-secreting stem cells in a hamster-to-Lewis rodents during hyper-rejection after LTx (1), we herein used the allo-OLT model to compare HGF with FK506, aiming to address whether HGF is more beneficial than FK506 in acute tolerance induction in the allo-OLT rodents and the mechanism interpretation. We chose two types of HGF agents: one is an HGF-like drug (namely, cHGF in the present study) for in vivo studies that mainly promotes HGF production in blood after intravenous injection (12), and another is human-sourced recombinant HGF (h-rHGF) obtained from commercial sources for in vitro coculture studies with magnetic activated cell sorting (MACS)-purified allogeneic aCD4 + and aCD8 + T cells with HGFand FK506, respectively, by those we compared the identify differences.
The obtained data indicate that HGF is much more efficacious than FK506 in acute tolerance induction for prevention from liver allograft rejection. Notably, humans (livers) with acute liver rejection (ARJ) show similar expression patterns of CD8 and HGF as allo-OLT rats (livers). To our knowledge, this is the first demonstration of the different proneness of HGF and FK506 in the immunomodulation of T cell subpopulations (CD8, CD4, and Tregs) during acute tolerance induction after LTx.

Materials and methods Animals
Male DA (RT1 n RT2 a ) rats which were 8 to 10 weeks old and weighing 150-250 g were used as liver donors, and male Lewis (RT1 l RT2 b ) rats weighing 250-350 g were used as recipients. The animals were purchased from Shanghai Laboratory Animal Research Center, China (http://www.slaccas.com), and housed under specific pathogen-free conditions at the Army Medical University animal facility on a 12-h light/12-h dark cycle for 2 weeks before experimentation. All animal experiments were conducted in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines. The institutional ethics committee of Army Medical University approved the experimental protocol (registration number: #SYXK YU 2012-0012).

Human specimen collection and ethical approval Human blood samples
Blood samples were routinely collected from 75 patients who received FK506 therapy 1 and 3 weeks after LTx at the Department of Hepatobiliary Surgery between January 2012 and December 2017 for analysis of leukocytes, neutrophils, and, in particular, the immune lymphocyte population, the results of which are shown in Table 1.

Human liver samples
We were fortunate to have six archival paraffin-embedded human liver specimens (~1 cm 3 ) from patients with ARJ; the samples were histologically confirmed by two independent histopathologists. Six hemangioma liver tissue samples, in which the degree of infiltration was less than 5% (denoted as HA), were used as controls (distal,~1 cm 3 ). Each sample was split into 12 parts and processed for Immunohistochemistry (IHC) and immunofluorescence (IF) analyses at three parts per person. All samples were obtained from the Army Medical University Tissue Bank following a protocol approved by the University's Institutional Review Board (registration number: Ky2022114).

Establishment of a DA-to-Lewis allogeneic allo-OLT rat model
We used the three-cuff technique for performing LTx. It was performed by using DA rats as donors and Lewis rats as recipients to establish the allo-OLT rat liver acute rejection model based on Kamada's two-cuff technique. Briefly, these rats were anesthetized by isoflurane inhalation (in O 2 ), and hepatectomized donor livers were flushed with 15 mL of University of Wisconsin (UW) solution (40 units/mL) (13,14) and stored at 4°C until transplantation. The recipient weight and total surgical time (65.0 ± 13.9 min) were similar among all groups. The suprahepatic vena cava was anastomosed end to end. The portal vein (PV), infrahepatic vena cava (IVC), and bile ducts were anastomosed using the cuff technique. The allo-OLT was performed by skilled microsurgeons, and the mean recipient cold ischemic (anhepatic) time was 35.0 ± 1.3 (SD) min, the warm ischemic time was 0.25 ± 0.05 (SD) min, and the graft ischemic time was approximately 31.0 min. Liver allograft survival was measured as the mean survival time (MST) of recipients in days following LTx. After completion of the operation, the animals' health status was monitored throughout the subsequent experiments. The technical failure rate of LTx surgery was approximately 10%. After completion of the operation, all transplant recipients were kept in single cages and given ad libitum access to a standard pellet diet and water. Each transplant recipient was assigned a numerical code, and the individuals responsible for caring for the animals and analyzing tissues were blinded to the treatment groups. The whole surgical procedure is shown in Supplementary Figures S1A-C, and acute allograft rejection was assessed based on Banff's criteria.
Banff's criteria for assessing acute liver allograft rejection Two experienced board-certified pathologists (YHC and GJD) independently and semiquantitatively assessed the acute rejection status for both patients and rats by determining the rejection activity index (RAI) score (RAIS) and overall or total RAISs (tRAISs) based on Banff's criteria. To this end, three (humans) or eight (rats) liver sections containing portal veins [PVs (a), boxes], bile ducts [BDs (a), boxes], and vascular endothelia [VEs (c), arrows, mainly marking the central vein part in this study] were selected from each individual, and six subjects were subjected to H&E staining to determine infiltrative cells. Each section contains at least five triads for staining analysis and then graded on a scale of 0 (absence) to 3 (severe) for RAISs. Furthermore, three individual RAISs were added together to derive the overall or total final RAIS (tRAIS) on a scale of 0 (absent) to 9 (severe) (15,16) as follows: 0 to 2: no rejection; 3: borderline, consistent with allo-OLT in rats or ARJ in humans respectively; 4 to 5: mild; 6 to 7: moderate; and 8 to 9: severe. The H&E-stained tissue sections were graded in a "blinded" fashion by a transplant pathologist (AJD) using the Banff criteria for acute liver rejection (15). The significance of differences between means was ascertained using an unpaired Student's t-test, two-way analysis of variance, or a log-rank test using Prism version 5.00 software; P < 0.05 was considered significant.

Experimental design
Both in vivo and in vitro experiments were performed. For the in vivo animal studies, the rats were divided into two transplantation groups and studied on postoperative day (POD) 7 or 10-12. The transplantation groups were as follows: (1) the Lewis rat-to-Lewis rat syngeneic transplant group (Syn, n = 20, n = 10/ time point)-rats that undergo syngeneic LTx are commonly used as long-term grafts acceptance controls (17). The rats in this group survived for more than 1 year, and they were sacrificed on day 128 in this study, n = 20, 10/time point; (2) The DA rat-to-Lewis rat transplantation group (allo-OLT as acute liver rejection group, n = 24, 12/time point)-168 rats from the allo-OLT group exhibited rejection within 9-12 days, which is consistent with previous observations (18,19). The remaining allo-OLT rats were randomly assigned to five group treatments: (1) (21), at a dosage of 0.2 mg/kg/day, which is 2 days before grafting and once every day for 2 weeks after grafting. The rats received anti-cMet-Ab (2x, 200 ng/animal/20 mL, intrasplenic injection) to block cMet function or control antibody IgG1 (the same concentration as the anti-cMet-AB) through the same way as the cHGF but without IV injection approach. Tissues from liver allografts, recipient spleens, and kidneys were collected from rats in these groups (at least n = 6) following euthanasia and used for further analysis of allostimulated T lymphocyte subpopulations (CD3, CD4, and CD8) and regulatory T cells (Tregs and FoxP3/CD25/CD4), the levels of hepatic protein, pro-and anti-cytokine profiles, HGF and cytotoxic kidney injury molecule 1 (KIM-1, also known as TIM-1), etc., via IHC and IF staining. For the in vitro studies, fresh naive splenic CD4 + and CD8 + T cells and recipients of these cells (aCD4 + T cells and aCD8 + T cells) were purified by magnetic bead cell sorting (MACS) and analyzed using flow cytometry (FCM). These cells were then stimulated with 0.5%-2% phytohemagglutinin (PHA) and treated with a relatively higher dose of h-rHGF (400 ng/mL) (being published) and FK506 (1-8 ng/mL) (22) for 24 h. The number of expanded clones were counted by investigators blinded to the groups. The counting criteria were based on the clonal size of small (a white arrow), middle (a yellow arrow), and large (a pink arrow) as shown in Supplementary Figure S3D (top panels).

Intrasplenic injection
Before LTx, a surgery was performed for intrasplenic injection based on previously reported studies (23,24) with slight modifications. Briefly, the animals were placed on the operating table under anesthesia [isoflurane inhalation (in O 2 )], and the fur was shaved along the torso and then sterilized with povidoneiodine until dry. The region just below the ribs was palpated, and an approximately 1.5-cm single incision was made to access the peritoneal cavity and gently explore the spleen. A dry/sterile square of gauze was placed underneath the organ to hold it in place and get ready for injection. The 27-G injection needle was slowly inserted into the spleen (approximately 3 mm, close to the spleen hilum) over a period of 40 s when splenic distension/paling was visible. Notably, needle removal is extremely important, and careful manipulation can completely avoid bleeding when the needle is removed: (1) a piece of sterilized wet gauze was placed over the injection site and used to hold the gauze in place while very slowly withdrawing the needle; (2) after needle removal, the injected area was gently pressed with a sterilized cotton swab for 3 min until the area visibly returned to normal. Finally, the suture and disinfection steps were completed.

IF and IHC staining
Liver graft, splenic, and kidney tissues were either cryoprotected in 30% sucrose overnight and frozen in optimal cutting temperature medium for IF staining (25) or embedded in paraffin (26) for IHC staining.

IF staining
Cryoprotected tissues were cut into serial 5-mm sections on a Leica cryostat (CM1950, Germany), blocked with 5% bovine serum albumin (BSA) for 20 min, and then incubated overnight with primary rat monoclonal antibodies against CD4 and CD8a, a mouse monoclonal antibody against Ki-67, and rabbit polyclonal antibodies against CD25, FoxP3, cMet, and FAS diluted 1:100 in phosphate-buffered saline (PBS) (pH 7.4). Normal mouse and rabbit IgG were used at the same dilution as the primary antibody controls. The slides were then washed three times for 5 min each with PBS after primary antibody incubation. Then, the sections were incubated in the dark with Alexa Fluor 488conjugated anti-rabbit IgG (green) or Cy3-conjugated (red) antimouse IgG secondary antibodies diluted 1:2,000 in 5% normal goat serum (NGS) for 1 h and mounted in fluorescence mounting medium. The mounted slides were photographed under a fluorescence microscope (SMZ25/SMZ18; Nikon, Tokyo, Japan). The number of cells from a quarter area per field-with a sum of four areas per field, six fields from one section, eight sections per animal for six animals per group-was quantified by investigators blinded to the status of each animal and compared with the average value obtained for vehicle-treated animals.

H&E staining
Paraffin-embedded tissues were stained with H&E according to routine protocols (27). Briefly, after deparaffinization and rehydration, 5-mm longitudinal sections were stained with hematoxylin solution for 5 min and then dipped 5 times in 1% acid ethanol [1% hydrochloric acid (HCl) in 70% ethanol] and then rinsed in distilled water, stained with eosin solution for 3 min, dehydrated in graded alcohol solutions, cleared in xylene, and then examined using a light microscope. The staining intensity of infiltrating cells around the portal vein (PVs), bile duct (BDs), and vascular region (VEs) was determined based on Banff's criteria and RAI scores in randomly selected at least five triad fields per section, three (human) or eight sections (animal) at the same tissue level from six subjects in each group.
IHC staining for cytokine profiles and functional hepatic proteins Sections of 5 mm from paraffin-embedded liver graft and kidney tissues were deparaffinized and subjected to antigen retrieval using citrate sodium solution (pH 6.0) in a 100-to 800-W microwave three times for 5 min and cooled at room temperature (RT) for 30 min. Endogenous peroxidase activity was blocked with 1% hydrogen peroxide (H 2 O 2 ) for 10 min. The slides were incubated with 3% BSA and 0.5% Triton in Tris-buffered saline (TBS) for 1 h. The tissues were then incubated overnight at 4°C with appropriately diluted primary antibodies (1:400) against HGF; CD8a; the proinflammatory cytokines IL-2, IL-12, TNF-ɑ, and IFN-g and the anti-inflammatory cytokines IL-4 and IL-10; and the functional hepatic proteins aspartate transaminase (AST), alkaline phosphatase (ALP), albumin (ALB), and KIM-1/TIM-1. The sections were then washed in TBS, incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at RT, developed with 3,3′-diaminobenzidine (DAB) for 5 min, counterstained with hematoxylin, and mounted with neutral balsam. The control sections were incubated without substrate or with medium to which 1 mM levamisole was added. A third control was incubated with DAB alone (28). The positive staining area for each marker in each section was calculated by determining the percentage of the total area exhibiting positive staining (brown color) at a magnification of ×100-400. The density of IHC staining was determined randomly. For the correlation analysis of CD8 and HGF, at least 25 positive areas/fields were randomly selected, six fields/sections, three (human) or eight (animal) sections each; others like cytokines and hepatic proteins, sum of one-sixth area/ field (shown as boxes in figures), analysis of all fields, six fields/ section, eight sections/animal, total of six animals/group. The stained images were observed under a light microscope (BX41 OLYMPUS, Japan) and analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA), and integrated optical density values are presented.

Evaluation of recipient T cell survival
To assess the proliferation of recipient rat aCD4 + and aCD8 + T cells in allogeneic liver grafts and recipient's spleens, we performed double IF staining for Ki-67 within aCD4 + or aCD8 + T cells. To assess cell death in the tissues, terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling (TUNEL) staining (to indicate apoptosis) was performed with the One Step TUNEL Apoptosis Assay Kit according to the manufacturer's instructions.
Briefly, the sections were incubated with 10 g/mL DNase I for 10 min at RT to induce DNA nicking as positive and negative controls were not treated with TUNEL enzyme. Then, the sections were washed three times for 3 min each time with PBS, blocked with 5% BSA for 20 min, and incubated with 50 mL of TUNEL labeling mixture for 60 min at 37°C in a humidified chamber. The slides were washed three times in TBS for 3 min after incubation. Redstained cells with 4′,6-diamidino-2-phenylindole (DAPI)-labeled nuclei (blue) were considered TUNEL (green)-positive cells. The number of these cells in 200 DAPI+ cells from a quarter region per field (shown as boxes in figures) was counted. Then, with a sum of four fields per section and eight sections/animal, a total of six animals per group were analyzed under an inverted fluorescence microscope (Nikon Eclipse Ti series, Tokyo, Japan) (original magnification, ×400) in a blinded manner.
MACS bead isolation of recipient splenic aCD4 + and aCD8 + T cells treated with h-rHGF and FK506 in cultures CD4 + and CD8 + T cells were obtained from fresh recipient rat spleens on POD 7 or 10-12 by MACS beads (29). A cell suspension was loaded in a column, which was placed in the magnetic field of a MACS separator to negatively purify the unlabeled cells. According to FCM, the purity of the CD4 + cells was greater than 96%, and that of the CD8 + cells was greater than 93%. With a total of 1 × 10 4 , the cells were first stimulated with 0.5%-1% PHA and treated with different concentrations of h-rHGF (50-400 ng/mL) and FK506 (0.25-40 ng/mL) in 96-well plates containing 200 mL of serum-free Neurobasal-A medium (Gibco, 10888) for 2-6 h. The cells treated with PHA only were used as controls. In this kind of experiment, we selected the concentrations of h-rHGF as 400 ng/mL and FK506 as 1-8 ng/m, respectively, for further coculture experiments, in which the same number of cells were put in 24-well plates with 300 mL of the serum-free medium for an additional 18-24 h. The response was assessed by determining the number of expanded colonies. The counting criteria were based on the more frequent sphere size appearing during cultures-as small, middle, and large, indicated as white, yellow, and pink arrows, respectively, as shown in Supplementary Figure S3D.
FCM analysis of CD8 + T cell activity in blood cHGF-, FK506-treated, and untreated CD8 + T cells (5 × 10 5 -1 × 10 6 ) from recipient rat blood were isolated and incubated for 30 min on ice with fluorescein isothiocyanate (FITC)-conjugated rat anti-rat CD8a antibody (1:100). After three washes, the stained cells were incubated with rabbit polyclonal PE-conjugated rat IgG (1:100) for an additional 30 min. After three additional washes, the cells were incubated with 7-AAD (1:100) for 30 min. The percentage of T cells among total cells was then assessed by FCM. Cells incubated with isotype-matched antibodies were used as controls.

Western blot analysis
Protein lysates were obtained by incubating liver allograft tissues in RIPA buffer with 1 mM sodium orthovanadate and protease inhibitor cocktail for 60 min on ice and cleared by microcentrifugation at 10,000 × g for 20 min at 4°C. The supernatant proteins (50 to 100 µg) of each sample were separated on 6%-and 12.5%-gradient SDS-PAGE gel and transferred to a PVDF membrane. Then, the membranes were blocked in TBS/0.1% Tween 20/5% BSA and immunoblotted with an antibody against HGF (1:200). After three washes in TBS/0.1% Tween 20, the membranes were incubated for 1 h at RT with HRP-conjugated anti-rabbit or anti-mouse secondary antibodies (1:5,000), developed with enhanced chemiluminescence reagent (Solarbio, China) system and imaged with X-ray films (LAS MINI 4000, Japan). All experiments were independently repeated three times. The protein expression levels were normalized to the expression level of the housekeeping protein glyceraldehyde-3-phosphate dehydrogenase.

Quantitative polymerase chain reaction analysis
The mRNA expression of CD8 in liver allografts was measured at POD 7 using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Shiga, Japan, no. RR047A) and TB Green Premix Ex Taq II (TaKaRa, Shiga, Japan, no. RR820A) in accordance with the manufacturer's instructions (Applied Biosystems). The primer sequences were as follows: b-actin: forward, 5′-CCGCGAGTACAACCTTCTTGC, and reverse, ATACCCACCATCACACCCTG-3′; CD8: forward, 5′-GCTTTTTGCCTTCGAGCTATCG, and reverse, GACTTCGTAGC GTACCTCTGG-3′. Reverse transcription was performed at 37°C for 15 min and 85°C for 5 s. The products were amplified at 95°C for 30 s followed by 40 cycles of 95°C for 5 s and 60°C for 30 s on a Touch Real-Time PCR Detection System (Bio-Rad, CA, USA, CFX96). The data were analyzed using Bio-Rad CFX96 Manager. The expression levels were normalized to the expression level of the housekeeping gene bactin. We used the comparative Ct (DDCt) method to calculate the relative mRNA expression.

Analysis of serum cytokine profiles and functional hepatic proteins by ELISA
Blood samples were obtained from the aorta abdominals under short-term inhalation anesthesia on POD 7 and 10-12 after LTx using standard enzyme-linked immunosorbent assay (ELISA) (30) following the manufacturer's instructions to measure the levels of the hepatic proteins AST, total bilirubin, ALP, and ALB; proinflammatory cytokines IL-2, TNF-a, and IFN-g; and anti-inflammatory cytokine IL-10.

Coimmunoprecipitation and immunoblotting
The binding between Fas and c-Met on naive splenocytepurified CD8 + T cells was evaluated by co-IP and IB (31). A total of 1 × 10 6 of these cell lysates were immunoprecipitated with an antibody against either Fas (IP) or cMet (IP) and then subjected to IB with either an anti-cMet-Ab to detect cMet (145 kDa) or an anti-Fas antibody to detect Fas (35 kDa). Briefly, equal amounts of cell lysates (500 µg to 1 mg) were incubated with protein A/G magnetic beads in IP lysis buffer at RT for 2 h and washed three times with 0.1% Tween 20 in TBS for 3 min. Then, protein A/G magnetic beads were incubated with either a mouse IgG1 isotype control antibody (2 µg/mL), a rabbit IgG isotype control antibody (2 µg/mL), or a mouse anti-cMet monoclonal or rabbit polyclonal anti-Fas antibody (2 µg/mL) for 1 h at RT. The protein A/G magnetic beads were centrifuged at 10,000 × g and washed three times with lysis buffer. The pellet was resuspended in Western blot loading buffer and heated at 95°C for 5 min. After centrifugation, the supernatant was resolved on 6% or 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel for western blot analysis.
Localized surface plasmon resonance spectroscopy of metallic nanoparticles for the analysis of chemical and biological sensing Localized surface plasmon resonance (LSPR) is a powerful technique for the analysis of chemical and biological sensing (32) and protein-protein interactions (33). In this study, we used this method to deduce whether FAS antagonized by cMet on active aCD8 + T cells in allo-OLT recipient rats can be abrogated by HGF (in the form of cHGF) and whether the cHGF treatment induces FAS-mediated apoptosis of recipient aCD8 + T cells recruiting Fasassociated protein with death domain (FADD) and caspase-8. Briefly, 50 mg/mL Fas antibody was immobilized on a COOH sensor chip that was activated by injection of 1-ethyl-3-(3dimethyl propyl)-carbodiimide and N-hydroxysuccinimide at a flow rate of 20 ml/min in running buffer. Free activated carboxyl groups were inactivated by the addition of 100 ml blocking buffer. The immobilized protein was washed with running buffer until a stable baseline was achieved. Total protein (50 mg/mL) isolated from cell lysates was injected into the flow at a rate of 20 ml/min, Fas protein was captured by the sensor, and 50 mg/mL IgG1 was used as a negative control. Antibodies against cMet, FADD, and caspase-8 were added to the sensor chip to assess the interactions of these target proteins with Fas for 5 min, and dissociation was evaluated for an additional 7 min. The kinetic parameters of the binding reactions were calculated and analyzed using Trace Drawer software (Ridgeview Instruments, Uppsala, Sweden).

Whole-slide imaging and image acquisition
Whole-slide images (WSIs) of liver tissues were obtained with a multicolor automatic digital slide scanner (34), and the images were analyzed with algorithms (35) using ZEN 2.3 (blue edition) software (Carl Zeiss Microscopy GmbH 2011). This software allows the user to navigate images and perform geometric and quantitative measurements. After scanning the liver sections with an Axio Scanner, five regions of interest (ROIs) per slide (images were taken at full resolution with a ROI of 1,364 × 1,364 pixels and pixel at 4.55 µm × 4.55 µm) were selected randomly from across the entire digital slide, with the centrilobular vein being in the center of each ROI, for a total of 250 ROIs. The area of each ROI was quantified. A series of analyses were performed for these 250 ROIs by different investigators. The images were subsequently analyzed using inhouse macros in batch mode in ImageJ® software (version 1.53e, National Institutes of Health, USA).

Iodine I-125 FasL labeling
Labeling with 125I-FasL was performed according to the manufacturer's instructions (Iodobeads, Knowledge Management, and Development Bioscience, Tianjin, China). FasL protein (5-10 mg) was placed in a siliconized Eppendorf tube containing 100 µl of 100 mM sodium phosphate reaction buffer (pH 6.8) and 0.5 mCi of sodium 125 I-FasL (Knowledge Management and Development Bioscience, Tianjin, China) and incubated for 8-15 min. The specific activity of 125 I-FasL was approximately 5 × 10 4 corn/ng. Briefly, the 125 I-FasL reaction mixture was then added to duplicate wells at the desired concentration in a total of 1 ml of binding buffer with or without h-rHGF and loaded onto a 5-ml G-25 desalting column in cold binding buffer containing Neurobasal-A medium (catalog no. 10888022, Thermo Fisher, USA), 25 mM N-2hydroxyethylpiperazine-N′-2-ethanesulfonic acid (catalog no. 15630080, Thermo Fisher, USA), and 1% BSA (catalog no. A8010, Solarbio, China). Splenic aCD8 + T cells were isolated from recipient rats using MACS, treated with either cHGF or FK506 and stimulated with 0.5-2% PHA as described above, washed two times with cold buffer, and solubilized with 1 mL of 1% Triton X-100 to extract receptor-associated 125 I-FasL. Radioactivity was measured with a gamma scintillation counter (PerkinElmer, TRI-CARB 4810TR, MA, USA). Nonspecific binding was determined in the presence of 5-0 mg of pure unlabeled FasL (unless otherwise indicated, all the binding data were corrected for nonspecific binding) (36).

Evaluation of calcineurin inhibitor toxicity in the kidneys
Sections of 5 mm from recipient kidney tissues were fixed and stained with H&E to visualize renal lesion by CNIT scores (37) under light microscopy, whose diagnostic features of tubular isometric vacuoles (tv, i), tubular atrophy with a striped interstitial fibrosis (ii) and subendothelial, medial, and peripheral arteriolar hyalinosis (ah, iii) were identified. The CNIT score was determined by semiquantitatively grading the degree of these three features (scale 0 to 3), with a total score of 9 in tissues where epithelial necrosis, loss of the brush border, cast formation, and tubular dilation were present. A five-point scale was used to assess tubular necrosis in the kidneys (0: normal; 1: damage to 1%-25% of tubule cells; 2: damage to 26%-50% of tubule cells; 3: damage to 51%-75% of tubule cells; and 4: damage >50%) (37) ( Table 2). Moreover, the kidney sections were subjected to IHC staining for KIM-1; a kidney injury marker (38) is present in tubule cells (39). A rat monoclonal antibody was used for staining, and KIM-1 expression in each section was determined by selecting the expression of positively stained tubule cells in six random-field pictures taken from one section, eight sections/animal, for a total of six animals in each group, at a magnification of ×200. All information about the antibodies and kits used in this study is provided in Table 3.

Statistical analysis
All summary data for the different groups are expressed as the mean ± SD of at least three independent experiments, and representative images are also presented. Unpaired t-test was used for statistical analyses, and differences were considered significant when the p-value was <0.05. For multiple comparisons, individual comparisons of group means were made by one-way analysis of variance (ANOVA) followed by Dunnett's test using the SPSS 13.0 software package (SPSS Inc., Chicago, IL, USA). When quantifying the cellular composition of cultures, at least three trials were performed in duplicate for each group of cells in each experiment. Linear regression analysis was used to determine the correlation between two parameters. Group means and standard deviations were calculated for each parameter. All statistical analyses were performed with Prism 5.00 (GraphPad Software, La Jolla, CA, USA).

Parameters of inflammatory cellular infiltration after LTx in patients who received FK506 therapy
The immunomodulation effect of FK506 during tolerance induction after LTx has been studied in animals (40). To determine whether the effects observed in animals reflect those in humans, we selected 75 patients who received FK506 therapy post-LTx and analyzed the proportions of not only key inflammatory cells (leukocytes neutrophils) 1 and 3 weeks after FK506 administration but also, for the first time, the proportions of lymphocyte population ( Table 1). As expected, the proportions of leukocytes and neutrophils were dramatically increased at 1 week after FK506 treatment compared with those at baseline (1 week before surgery) (Figure 1, panel 1), and these cells returned to normal after 3 weeks (Figure 1, panel 2). However, when analyzing lymphocytes, we found that the proportion of these cells was significantly decreased 3 weeks post-FK506 compared with the baseline (Figure 1, panel 2, *p < 0.05). These findings suggest that the early systematic inflammation (1 week post-LTx) may not be due to FK506 in this stage but may be due to lymphocytes, although which lymphocyte subpopulation remains unclear. These preliminary clinical outcomes provide new information about the correlation between FK506 and lymphocyte population that would be clinically meaningful.  Differences in the effects of cHGF and FK506 in regulating the proportions of infiltrating cells and total CD3+ T cells on POD 7 and 10-12 in the allo-OLT rat model HGF has also been reported to have suppressive effects on CD8 + T lymphocytes (41,42). To compare the effects of the two reagents in allogeneic LTx recipients, we used the established allo-OLT liver rejection rat model based on Kamada's two-cuff technique  PVs and BDs (Figure 2Ai, left two panels, boxes, ii, ns p > 0.05). When observing the inflammation around VEs (Figure 2Ai, right bottom two panels, arrows), we found that cHGF significantly decreased the number of infiltrating cells compared with FK506 ( Figure 2Aii, *p < 0.05). However, upon analyzing the scores of tRAIS, cHGF showed significant (*p < 0.05) effects on the tRAIS in all areas compared with FK506 ( Figure 2B), suggesting that cHGF induced a greater pathological improvement than FK506, especially around the injured blood vessels. For lymphocytes, the proportion of total T cells was also significantly decreased in both liver allografts ( Figure 2C) and recipient spleens ( Figure 2D) from the rats in the cHGF monotherapy group compared with those from the rats in the FK506 group, as indicated by IF staining with an antibody against CD3 (green). Approximately 6.25% ± 1.77 of CD3 + T cells in liver allografts ( Figure 2Ci; the boxes indicate a quarter of field) and 3.27% ± 2.35 of these cells in recipient spleens (Figure 2Di, boxes) were found in the cHGF group, while 14.13% ± 5.04% of cells in grafts and 10.31% ± 2.35 of CD3 + T cells in the recipient's spleens were found in the FK506 group; compared with FK506 (blue bars), cHGF (red bars) decreased significantly (Figures 2Ci, Di, *p < 0.05) and induced an almost 2.26 times greater decrease of cHGF over FK506 in the proportion of CD3 + T cells in liver allografts (Figure 2Cii, striped bar) and a 3.14-fold greater decrease in recipient spleens (Figure 2Dii, striped bar), although both FK506 and cHGF induced a greater decrease in the proportion of these cells than allo-OLT alone (not shown), suggesting that the HGF-promoting drug cHGF has a more robust effect on T lymphocytes than FK506. These observations indicate that, at early time points (prior to 3 weeks, POD 10-12), cHGF exerts a superior effect over FK506 on anti-inflammatory and T lymphocyte suppressive effect.
cHGF mainly affects the proportion of aCD8 + T cells while FK506 mainly affects the proportion of aCD4 + T cells Given the more potent effect of cHGF than FK506 in decreasing the proportion of total CD3 + T cells, the FK506-mediated CD4 + T cells/IL-2 tolerance (43), and our previous evidence that HGF prolongs liver xenograft survival by promoting xenostimulated CD8 + T cell apoptosis (6), we speculated that allostimulated aCD4 + and aCD8 + T cell subpopulations may respond differently to cHGF and FK506. To assess this hypothesis, we immunostained liver allografts ( Figure 3) and recipient immune organs (spleens) (Supplementary Figure S2). We first found that both aCD4 + (red) and aCD8 + (red) T cells were more strongly activated in the allo-OLT group (Figure 3Ai; the boxes indicate a quarter of one field) on POD 10-12, but the proportion of aCD8 + T cells was shown to be significantly higher than that of aCD4 + T cells (Figure 3Aii, a red striped bar); this confirmed that aCD8 + T cells are the main effectors in acute liver rejection (44). We next confirmed that the proportion of aCD4 + T cells was significantly decreased in the FK506 group (Figure 3Bi, a box) compared with the cHGF group (Figure 3Bii, a box, iii, #p < 0.05), while aCD8 + T cells were more sensitive to cHGF than FK506 (Biii, *p < 0.05), suggesting that aCD4 + T cells and aCD8 + T cells in liver allografts are sensitive to FK506 and cHGF, respectively. Interestingly, compared with allo-OLT (Figures 3C, D, black bars) which had significantly increased CD4 and CD8 T cell subpopulations compared with the Syn group ( Figures 3C, D, gray bars), cHGF (red bars) had a significant effect on the proportions of both aCD4 + ( Figure 3C, red bar, *p < 0.05) and CD8 + T cells ( Figure 3D, **p < 0.001), while FK506 (blue bars) affected only aCD4 + T cells ( Figure 3C, #p < 0.05) and did not affect the proportion of aCD8 + T cells ( Figure 3D, ns p > 0.05), suggesting that cHGF is able to decrease both CD4 and CD8 T cell subpopulations, while FK506 does so only with the CD4 T cell subpopulation during early tolerance induction. Similar biological phenomena were observed in recipient spleens ( Supplementary  Figures S2A-D). Moreover, double staining for Ki-67 and TUNEL showed a greater decrease in Ki-67 + /CD4 + T cells (Figure 3Ei, arrows, merged images) and an increase in the proportion of TUNEL + /CD4 + T cells (Figure 3Fi, arrows, merged images) in the FK506-treated rats compared with the cHGF-treated rats (Figures 3Eiii, Fiii, #p < 0.05). In contrast, a decreased proportion of Ki-67 + cells in the aCD8 + T cells (Figure 3Eii, arrows, merged images) and increased TUNEL expression (Figure 3Fii, arrows, merged images) were observed in the cHGF- Mononuclear cell infiltration at 1 and 3 weeks after LTx compared with 1 week before surgery in patients who received FK506 therapy (panels 1 and 2). Clinical routine blood analysis in numbers of leukocytes, neutrophils, and lymphocytes at 1 week (panel 1) and 3 weeks after surgery compared with 1 week before surgery (panel 2) in 75 patients who underwent LTx. *p < 0.05 vs. 1 week before surgery.
To determine whether h-rHGF treatment has a similar effect in vitro as cHGF has in vivo, we analyzed MACS-purified splenic CD4 + T cells and CD8 + T cells from allo-OLT model rats (on POD 7), and the FCM analysis showed that the purity of CD4 + and CD8 + T cells were more than 96% and 93%, respectively (Supplementary Figure S3A). These cells were stimulated with 0.5%-1% PHA (denoted as paCD4 + T cells and paCD8 + T cells) and cocultured with FK506 on paCD4 + T cells (Supplementary Figure S3B Figure S3D-ii, right panel, v, ns p > 0.05) compared with the PHA control treatment, suggesting that FK506 and h-rHGF have different abilities to affect CD4 + and CD8 + T cell subpopulations: FK506 preferentially suppressed paCD4 + T cells, while h-rHGF preferentially suppressed paCD8 + T cells. These observations indicate that aCD4 + T cells are sensitive to both FK506 and cHGF but are more sensitive to FK506, while aCD8 + T cells are only sensitive to cHGF.

Superior effects of FK506 over cHGF in modulating Tregs in the livers of allo-OLT model rats
The Treg subpopulation is one of the key contributors to peripheral tolerance induction (45). Although cHGF and FK506 have both been reported to contribute to tolerance induction after LTx through these cells, we are the first group to compare their effects on Tregs after allogeneic LTx. Double IF staining for CD4/CD25, which are Treg markers (46), showed that the proportion of CD25 + (green) within CD4 + T cells (red) was significantly higher in both liver allografts ( Figure 4A; the boxes indicate half of the field) and recipient spleens ( Figure 4B, boxes) of the rats treated with FK506 (blue bars) compared with those of the allo-OLT model rats treated with cHGF (red bars) on POD 10-12 ( Figures 4C, D, #p < 0.05), although both drugs promote the Treg population (not shown), suggesting that FK506 has a stronger effect than cHGF in regulating Tregs. Stained sections for FoxP3 (red), another Treg marker (47), in the same recipient tissues (Figures 4Ei, Fi, boxes) revealed a trend similar to the proportion of CD25 + /CD4 + T cells. The number of FoxP3 + cells (red) was dramatically higher in the FK506 group (blue-striped bars) than in the cHGF group (red-striped bars) (Figures 4Eii, Fii, #p < 0.05), further confirming our findings. These observations indicate that, unlike cHGF, FK506 may induce allogeneic tolerance after LTx by affecting mainly the populations of both CD4 + T cells and Tregs.

Superior effects of cHGF over FK506 in prolonging allograft survival and inducing functional recovery in the allo-OLT rat model
Given the preferential effects of cHGF and FK506 on CD8 + T cells and Tregs, respectively, the key subpopulations in acute cHGF mainly suppressed the aCD8 + T cell subpopulation, while FK506 mainly affected the aCD4 + T cell subpopulation in liver allografts. (Ai, ii) IF staining for CD4 + (red) and CD8 + (red, stripe) T cells (i, boxes) and quantitative analysis (ii). (Bi-iii) IF staining for CD4 + (red) and CD8 + (red) T cells from the FK506 (i, boxes) and cHGF (ii, boxes) groups and comparative analysis of the proportions of CD4 + and CD8 + T cells between the FK506 (blue bars) and cHGF (red bars) groups (iii, n = 6). (C, D) Analysis of aCD4 + T cells (C) and CD8 + T cells (D) in the Syn (gray bars) and allo-OLT recipient rats (black bars) treated with FK506 (blue bars) and cHGF (red bars); n = 6. (Ei-iv) Immunostaining (i, ii) and analysis for Ki-67 (green) within CD4 + (red, i, iii) and CD8 + (red, ii, iv) T cells (merged images, arrows) and their comparative analysis between the FK506 (blue bars) and cHGF (red bars) groups. (Fi-iv) TUNEL (green) staining of CD4 + (red, i, iii) and CD8 + (red, ii, iv) T cells (merged images, arrows) and comparative analysis of recipient liver allografts between the FK506 (blue-striped bars) and cHGF (red-striped bars) groups; the boxes represent a quarter of the whole stained field. Original magnification = ×400; scale bar = 200 µm. The data are presented as the mean ± SD of at least three independent experiments, n = 6. *#p < 0.05 vs. CD4 and FK506; **p < 0.001 vs. the allo-OLT group; ns denotes no significance.
Looking over the prolonged graft survival in cHGF group than in the FK506 group compared the rats at early time points (before 2 weeks, before~6-14 days, Figure 5Ci), typical period of lethality days after LTx, and also at later time points (after 6 weeks,~45-73 days (Figure 5Cii; the typical period of acute rejection time after LTx), a significantly longer MST was found in the cHGF-treated group (red) than in the FK506-treated group (blue) at both periods (Figure 5Ciii, *p < 0.05), suggesting that cHGF has a stronger effect than FK506 on inducing early tolerance and maintaining the acute tolerance. In our study, the rats in the Syn group survived for more than 1 year, considered to be equivalent to indefinite survival after LTx ( Figures 5A, B, gray, n = 20, 10/time point). IL-2 is commonly accepted to promote tolerance mechanisms by FK506 after organ transplantation (48); thus, we compared the effect on the allo-OLT rats after the two drug treatments. Using HIC staining (brown), we identified decreased IL-2 expression more obviously in the cHGF (red-striped bar) than the FK506 (a blue-striped bar) groups ( Figure 5Di; the boxes indicate one-sixth area per field, ii, *p < 0.05), although both cHGF and FK506 decreased the IL-2 expression compared with that in the allo-OLT group (Figures 5Di, ii, black), which dramatically increased compared with that in the Syn group (Figure 5Di, ii, gray). Similar decreases in the expression of other proinflammatory cytokines, e.g., IL-12, TNF-a, and IFN-g, and the anti-inflammatory cytokine IL-4 were observed in those subgroups (Supplementary Figures S3E-i, ii, 5E -i-iii; the boxes indicate one-sixth area/field). Moreover, the hepatic proteins AST and ALB were observed by the same  Figure S3F, 5Fi, ii), such were consistent with histological changes (Figures 2A, B), suggesting the advantageous effects of cHGF over FK506 in improving functional recovery after LTx. Differently, in terms of IHC staining in these subgroups, hepatic ALP was stronger with FK506 (blue latticed bar) than with cHGF (red latticed bar) (Figure 5Fiii

cHGF induces aCD8 + T cell apoptosis through the HGF-only receptor cMet
The main target of cHGF in other biological systems is HGF (12), and given our previous findings that HGF can induce the apoptosis of recipient xenogeneic CD8 + T cells (6), we speculate that cHGF may also exert its effects in the allo-OLT model through HGF. To test this hypothesis, we first assessed the expression level of HGF in liver allografts after cHGF administration. Western blotting revealed higher HGF levels in the allografts of the cHGFtreated recipients (Figures 6Aiii, iv) than in those of both the OLT model and FK506-treated recipients (Figures 6Ai, ii) on POD 10-12 (n = 3), suggesting that cHGF administration enhanced local endogenous HGF production. Then, we surmised that decreased HGF levels in allo-OLT rats may cause strong CD8 + T cell activity. IHC staining (Figure 6Bi, boxes) in the same liver tissue fields of liver allografts confirmed that the expressions of HGF (on POD 10-12) (Figure 6Bii) were significantly lower and CD8 (on both POD7 and 10-12; the boxes indicate one-sixth area/field) was significantly higher (#P < 0.05) in the liver grafts from the allo-OLT recipient rats than in those from the syngeneic recipient rats (gray) (Figures 6Biiii). An analysis of the immunostaining data from the allo-OLT recipient rats revealed that approximately 25%-40% of the CD8 + T cells were within DAPI + cells (Supplementary Figure S4A, red bar) compared with the naive livers (Supplementary Figure S4B, less than 4% CD8 + T cells, red), and within the active aCD8 + T cells in liver grafts of allo-OLT model rats, less than 4% of these cells were targeted by TUNEL (Supplementary Figure S4A, green bar).
Notably, we observed that the decrease in HGF expression and increase in CD8 expression on days 10-12 ( Figure 6Biii) were significantly negatively correlated (R 2 = 0.6691, Figure 6Biv, P < 0.0001). The results of the Western blot analysis of HGF and CD8 expression in liver allografts from the same animals revealed similar results (Figure 6Bv, red; numbers in the red boxes), suggesting that the lack of HGF signals in allo-OLT rats suppresses aCD8 + T cell activity. In other systems, HGF signals through the tyrosine receptor kinase cMet, the best-characterized receptor for HGF (50). To determine whether cHGF/HGF signaling is also mediated by cMet, we treated allo-OLT model rats with a neutralizing antibody that blocks cMet function (anti-cMet-Ab) (2 × 200 ng/animal, intrasplenic, 2 days before and during grafting, mixed with cHGF treatment). HGF expression, which was assessed by Western blot analysis, was not increased ( cHGF). We detect cMet (green) to be expressed in MACS-purified aCD8 + T cells (red) using double IF staining (Figure 6Fiv, arrow), and by using Western blotting (Figure 6Fv), it was confirmed that the biological effects of cHGF in allo-OLT is mediated through cMet. These observations strongly suggest that cHGF/HGF/cMet signaling acts as a critical player in aCD8 + T cell-mediated acute tolerance induction by suppressing cMet-expressing aCD8 + T cell activity, which leads to graft survival and functional recovery in allo-OLT model animals.

Inhibition of Fas antagonism by cMet through cHGF/HGF in recipients with acute liver rejection triggers aCD8 + T cell FASmediated apoptosis
Given that the higher proportion of aCD8 + T cells is negatively correlated with the lower levels of HGF in allo-OLT recipient rats with acute liver rejection (Figures 6Biii-v) and our previous findings of HGF-mediated liver xenograft aCD8 + T cell apoptosis in xenogeneic liver transplantation rat model (6), we speculate that aCD8 + T cells exhibit elevated apoptosis in allogeneic recipients due to the crosstalk of HGF (from cHGF) to the death receptor FAS, resulting in FAS-mediated apoptosis and establishing tolerance, which can rapidly prevent graft rejection. Using a WSI automatic image scanner, we found Fas (green) antagonism by cMet (pink) in mouse liver allograft aCD8 + T cells (red), as shown by the merged brick color (Figure 7Ai, bottom panel, box; Supplementary Figure 6A, indicated by an arrow) which was stronger in the alloliver grafts on POD 7 than that from naive ones (Figure 7Ai HGF induced aCD8 + T cell apoptosis in a mechanism mediated by its only receptor cMet. (Ai-iv) Western blot analysis for HGF in liver allografts from recipient rats of the allo-OLT (i) that received FK506 (ii) and cHGF (iii iv) on POD 10-12; n = 6. (Bi-v) IHC staining (brown) for HGF and CD8 in the same area of tissue specimens from the rats subjected to allo-OLT on POD 10-12 (i, boxes) (n = 6/group) and a comparison to the Syn groups on both POD 7 (ii) and 10-12 (iii); the correlations between CD8 and HGF in allo-OLT recipient rats on POD 10-12 (iv), n = 25 points. Western blot analysis was used for the expression of HGF and CD8 in naïve livers (black) and livers from the rats subjected to allo-OLT (red in boxes) on POD 7 and 10-12 (Bv), densitometric quantification showing the decreasing trend in expression. The target protein expression was normalized to the expression of b-actin; n1 and n2 represent two individual samples; n = 3/time point. Original magnification ×200; scale bar = 100 µm. *#p < 0.05 vs. the Syn group. (Ci, ii) Western blot detection for HGF expression in liver allografts from the recipient rats treated with cHGF plus anti-cMet-Ab (i, the same treatment with cHGF but without the IV injection approach) on POD 7 (n = 6) and survival curve with the MST analysis (ii). n = 16. #p < 0.05 vs. the cHGF group (Student's t-test). (Di-v) IF staining for aCD8 + T cells (red) in the allo-OLT and cHGF groups at the protein (i, boxes, ii, n = 6) and mRNA (RT-qPCR, n = 3) (iii) levels on POD 7 and the cHGF plus anti-cMet-Ab group (iv, a box, v). (Ei, ii) Analysis of the IF staining of Ki-67 (green) within CD8 + T cells (red) in liver allografts from the allo-OLT and cHGF groups on POD 7 and 10-12 shown in Supplementary Figures S4D-i, ii (i, **p < 0.001 vs. allo-OLT); analysis of TUNEL staining (green) within CD8 + T cells (red) in liver allografts from the allo-OLT and cHGF groups on POD 7 and 10-12 shown in Supplementary Figures S4E-i, ii (ii, **p < 0.001 vs. allo-OLT). Analysis of cHGF plus anti-cMet-Ab (solid blue bars) for CD8 + T cell proliferation (Ki-67) and apoptosis (TUNEL) in Supplementary Figures S4D-iii, iv, Eiii, iv (i, ii, #P < 0.05 and #p < 0.05, ##p < 0.001 vs. the cHGF group), Student's t-test, n = 6. Original magnification, ×400; scale bar = 200 µm. (Fi-v) Analysis of the serum levels of AST, TBiL, and ALB (i); the cytokines IL-2, TNF-a, and IFN-g (ii) in the naive (light gray bars), Syn (gray bars), allo-OLT (black bars), cHGF (red bars), and cHGF plus anti-cMet-Ab (green bars) subgroups by ELISAs; histological assessment by RAI score [RAISs (0-3) and tRAISs (0-9)] around PVs (a, boxes), BDs (b, boxes) and VEs (c, arrows) from Supplementary Figures S4F-i, ii on POD 7 and POD 10-12 according to Banff's criteria (iii, #p < 0.05, ##p < 0.001 vs. the cHGF group, Student's t-test), n = 10/time point, IF staining (n = 6) of cMet (green) in MACS-purified Lewis rat naive splenic CD8 T cells (red) (F-iv, the arrow denotes merged brown staining) and confirmation of cMet expression by Western blot analysis (F-v, n = 3, n1-n3 represent each set of individual cells); the boxes represent one-sixth of whole stained field. Original magnification, ×400; scale bar = 200 µm. The data are presented as the means ± SDs of at least three independent experiments. The asterisks indicate a statistically significant difference.
of two molecules using Co-IP, we, for the first time, find the association of cMet and FAS with MACS-purified naive splenic mouse CD8 + T cells ( Figure 7B) (51). These findings may suggest that FAS antagonism by cMet on aCD8 + T cells correlates with aCD8 + T cells/higher activity/tolerance failure/rejection; we deduced that HGF/cHGF activated cMet through its ligand HGF from cHGF, parting the association, FAS-free and aggregation, and then undergo FAS-mediated apoptosis. Using 125 I labeling FasL ( 125 I labeling FasL) (Mat/Met), we deduced FAS assembly in sensitive response to Fas ligand (FASL) after HGF administration of 400 ng/mL h-rHGF significantly increased aCD8 + T cell binding to 125 I-FasL at 50 and 75 ng/mL during 72-h cultures (Figure 7Ci, red line, *p < 0.05), and this effect was blocked by anti-cMet-Ab (green line, Figure 7Cii, #p < 0.05), suggesting that cMet acts as a potent FasL antagonist and that its ligand HGF is a promoter of Fas trimerization (Supplementary Figures S4E, S5C, arrows indicate FAS trimerization). Additionally, by performing LSPR spectroscopy of metallic nanoparticles (Mat/Met), we showed in cultures that h-rHGF (red) significantly decreased cMet expression on MACSpurified splenic aCD8 + T cells (Figure 7Di, top panel) and promoted death-inducing signaling complex (DISC) formation, hallmarks of FAS-mediated apoptosis (52) through increased recruitment of FADD (Figure 7Dii, top panel) and caspase-8 (Figure 7Diii, top panel) to Fas within 72 h, compared with the cells from allo-OLT rats (Figures 7Di-iii, top panel, black). We used IgG1 (Figures 7Di-iii, gray lines) as controls. The effects of h-rHGF were inhibited by anti-cMet-Ab (Figures 7Di-iii, bottom panels,  green), suggesting that the effects of HGF (in the form of cHGF) in recipient rats triggered FAS-mediated aCD8 + T cell apoptosis through cMet. Interestingly, we found that the data obtained in  Figure S6D); the boxes represent a quarter area per stained field. The data are presented as the means ± SDs of at least three independent experiments. Original magnification, ×400; scale bar = 200 µm. *#p < 0.05 vs. the PHA-treated control or cHGF or HA group; **p < 0.001 vs. naive or allo-OLT or HA liver tissues (Student's t-test, two-way ANOVA).
animals are largely similar to those obtained in humans. We were fortunate to obtain 12 human recipient liver sections (~1 cm 3 ) from each individual with acute liver rejection (denoted as ARJ  Figure S6B-ii) and assessed by RAI scores of RAISs and tRAISs (Supplementary Figures S6B-iii, iv). We used distal noninflamed hemangioma liver tissues as controls (denoted as HA here, n = 6, Supplementary Figures S6B-i, iii, iv; n1-n3 denote individual samples). The WSI automatic image scanner detected similar expression patterns of Fas (red) and cMet (pink) in human allostimulated CD8 + T cells (h-aCD8 + T cells, green) ( Figure 7Ei; the boxes indicate a quarter of one field) to those in the rats (Figure 7Ai). Co-IP assays revealed that cMet and FAS interacted in healthy h-aCD8 + T cells isolated from peripheral blood (Figure 7Eii, n1-n2 denote individuals). Opposite expression patterns of CD8 and HGF were observed in ARJ livers (Supplementary Figure S6D) and HA tissues (Supplementary Figure S6C; the boxes indicate one-sixth area per field) at the same areas, identified by IHC staining (Figure 7Fi). A similar linear correlation to the allo-OLT rat model (Figure 6Biv) was observed between CD8 and HGF expression in the ARJ human liver allograft samples (Figure 7Fii, R 2 = 0.09115, p = 0.0005), suggesting that the resistance of h-aCD8 + T cells (green to apoptosis; TUNEL, red) in ARJ (Supplementary Figure S6E, merged, denoted as an arrow in the images and a chart; n1-n4 denote individuals) is also due to the lack of cMet activation by its ligand HGF. These data indicate that activation of cMet by its ligand HGF may allow aCD8 + T cells to cross immune barriers for acute tolerance induction in recipients with acute liver rejection.
cHGF protects against FK506-mediated nephrotoxicity in allo-OLT model rats through HGF via an IL-10related mechanism FK506 treatment after LTx commonly causes nephrotoxicity in the clinic (approximately 15%-20% of patients) (53), and here we show that cHGF can protect against FK506-mediated nephrotoxicity in allo-OLT model rats through HGF production. Nephrotoxicity was evaluated based on CNIT scores (54) via H&E staining (Mat/Met). FK506 indeed caused a significant injury to renal tissues on POD 10-12, which appeared as tubular isometric vacuolization (tv, i), striped interstitial fibrosis (ii), and peripheral or medial (ah, iii) as identified ( Figure 8A, blue arrows; Table 2, blue checks markers, CNIT scores = 2(i), 2(ii), and 1(iii) respectively). The scores were higher than those in the Syn and allo-OLT subgroups ( Figure 8A; Table 2, black check markers, CNIT scores = 0). The animals that received cHGF did not show any injury ( Figure 8B; Table 2 Figure 8C). These results further confirmed that cHGF alleviated injury. Interestingly, given that anti-cMet-Ab inhibited cHGF effect in blood (Supplementary Figure S4C-ii), also in liver allografts (Figures 6Aiii, iv) than FK506 (Figure 6Aii), we investigated whether HGF expression also increases in renal tissues after cHGF treatment compared with FK506 treatment. Using the same IHC staining in subgroups, we confirmed the hypothesis and found a higher expression of HGF in renal tissues of the cHGF-treated group than in the FK506-treated group ( Figure 8D, boxes, **p < 0.001), although HGF expression was also increased in the FK506-treated group which was higher than the Syn (gray bar) and allo-OLT (black bar) subgroups ( Figure 8D, #p < 0.05). Notably, although the staining in liver allografts detected dramatically increased the IL-10 expression in the cHGF-treated group over the allo-OLT group (Figure 8Ei; the boxes indicate onesixth area per field, *p < 0.05), this did not occur in the FK506treated group (blue bar, Figure 8Ei, ns p > 0.05). Using the same staining further detects the expression of IL-10 in renal tissues; we found a significantly higher level of IL-10 expression in the cHGFtreated rats than in the FK-506-treated rats (Figure 8Eii). The ELISA assay confirmed that IL-10 expression increased in the blood of the recipient rats treated with cHGF (red-striped bars) compared with the allo-OLT recipient rats (black) on both POD 7 and 10-12 ( Figure 8F, *p < 0.05), and these effects were abolished by anti-cMet-Ab ( Figure 8F, green-striped bars, #p < 0.05). These findings suggest that HGF, produced in response to cHGF, may systematically and locally promote IL-10 production to help protect against renal injury via cMet (HGF dependent). These observations indicate that the superior effect of cHGF to FK506 in protecting against nephrotoxicity is due to the influence of HGF signaling on IL-10 activation not only in liver allografts but also in the kidneys through the blood.
Taken together, these studies suggest that the superior ability of cHGF to induce acute tolerance in the allo-OLT rat model relative to that of FK506 is due to its ability to stimulate HGF/cMet signaling, resulting in the activation of cMet-expressing aCD8 + T cell apoptosis, enhancement of functional recovery, and promotion of IL-10 production to protect the kidneys.

Discussion
LTx is a life-saving procedure for ESLD of various etiologies, and both central and peripheral tolerance have been studied in an effort to achieve acceptance of allograft tissues in liver transplant recipients (56). Early on, high rates of acute graft rejection were overcome by the introduction of potent IS regimens, which are still largely in use today. However, long-term improvements in LTx procedures and patient outcomes have been hampered by the cumulative toxicity of maintenance IS therapy (57). FK506, for instance, was confirmed to be an IS in early studies and is currently the most commonly used drug in the clinic for acute tolerance induction after LTx (58); however, it can cause infection, nephrotoxicity, and incomplete preservation of graft function during tolerance induction (59). Therefore, novel immunosuppressive strategies for active liver tolerance induction and long-term good prognosis are urgently needed. In the present study, we extensively investigated the effects of FK506 therapy on acute tolerance induction, functional recovery, and renal injury protection by comparing them with the effects of HGF from a commonly clinically used HGF-promoting drug (cHGF) in a rat allo-OLT model to assess the clinical relevance of HGF as a novel tolerance inducer.
HGF-induced tolerance in organ transplantation has been reported (6). In the present study, we showed that the advantages of cHGF monotherapy over FK506 in beneficial effects are sustained in the allo-OLT rat model through HGF signaling. We have previously reported that 10-12 days of HGF/CD8 + T cellmediated stem cell treatment facilitates the induction of xenogeneic tolerance for liver xenograft survival in a full MHC mismatch hamster-to-Lewis LTx rat model (MST, only 3-5 days) (6), suggesting the possibility of tolerance effect of HGF in allogeneic LTx. As expected, here using a well-established allo-OLT rat model based on Kamada's two-cuff technique, we showed that the MST in the cHGF/HGF-treated recipient rats was much longer (MST: 43.33 ± 23.91 days) than the rats treated with FK506 (MST: 20.22 ± 23.91 days), although FK506's MST was significantly longer than that of the untreated allo-OLT controls (MST: 9.45 ± 23.91 days). The 40-day liver allograft survival rate was approximately 21.81% in the rats treated with cHGF/HGF and 0% in the rats treated with FK506 alone, consistent with the results reported by Yamada et al. (60).
However, the mechanism underlying these differences is currently unclear. We show in the present study that the difference is due to T cell subpopulations: the action of HGF from cHGF is against the recipient aCD8 + T cells (6), which are the main mediators of graft acute rejection (61), and the effect of FK506 is on aCD4 + T cells and Tregs in recipients with LTx (6). The observations of the present studies showed that cHGF exerts a stronger effect on liver allograft survival in rats than FK506 mainly through the HGF-mediated suppression of proliferation (Supplementary Figures S4D, S5B, 6Ei) and promotion of apoptosis (Supplementary Figures S4E, S5C, 6Eii) in recipient aCD8 + T cells by binding the HGF-only receptor cMet on these cells ( Figure 6F), disturbing the spatial segregation of FAS by cMet (HGF-only receptor) ( Figures 7B, Eii) (62,63), which caused FAS assembly (Supplementary Figure S4E), arrows that indicate sensitivity to FAS-L (Figure 7Ci), and inducing these cells to cHGF protects against FK506-mediated nephrotoxicity in the allo-OLT rat model via an IL-10-related mechanism. (A) H&E staining to evaluate the CNIT score in renal tissues (Table 2) from the Syn, allo-OLT, and FK506 (scores of 2, 2, and 1) subgroups based on the histologic features shown in Table 2, specifically proximal tubules with cytoplasmic isometric vacuoles or tv [(i) a blue arrow], tubular atrophy with a striped pattern of interstitial fibrosis [(ii) a blue arrow], and subendothelial, medial, and peripheral arteriolar hyalinosis or ah [(iii) a blue arrow]; n = 6/subgroup. (B) Identical analysis in the cHGF group based on CNIT scores shown in Table 2  undergo FAS-mediated apoptosis ( Figure 7D). These results indicate that HGF is able to cross the aCD8 + T cell immune barrier for acute tolerance induction and leads to liver allograft survival, consistent with our previous studies in a xenogeneic rat model (6), while our observations show that FK506 prefer to suppress recipient aCD4 + T cells and cMet is not be expressed on the cells (64), indicating that FK506 may not be able to cross the aCD8 + T cell barrier. Furthermore, we investigated whether the mechanism(s) of proliferation and apoptosis of aCD8 + T cells is directly induced by HGF. A dose-dependent response of the recipient aCD8 + T cells to h-rHGF in cocultures suggests that the double-edged sword effect (pro-/antiapoptosis) of HGF (65-68) is also reflected in our allo-OLT recipient rats, and lower levels of HGF (<50 ng/mL)-induced CD8 + T cell DNA synthesis were not effective in dissociating cMet from Fas (Figures 7Ai, Ei), indicating that cMet and FAS preexist as a complex on CD8 + T cells from both rats ( Figure 7B) and humans ( Figure 7Eii); higher concentrations of HGF (>200 ng/mL) may disturb the antagonist effect of cMet (62) on FAS, causing FAS to dissociate from cMet and aggregation (Supplementary Figures S4E,  S5C, arrows), enhancing sensitivity to FasL binding (Figure 7Ci), recruiting FADD and caspase-8 (Figures 7Dii, iii), and eventually inducing Fas-mediated apoptosis (62), indicating that a lack of cMet activation by its ligand HGF leads to aggressive aCD8 + T cell activity in recipients with acute liver rejection ( Figures 6B, 7F).
Moreover, our clinical data showed for the first time that FK506 significantly suppressed lymphocyte proliferation at an early time point (3 weeks after FK506 administration), but which lymphocyte subpopulations are involved in these recipients remains unclear. We show herein the effect by FK506 preferentially on the CD4 + T cell subpopulation both in vivo and in vitro. However, the mechanism of aCD4 + T cell-mediated anti-rejection in allo-recipients is still unclear. FK506 may act on Tregs, either CD4 + Tregs (69, 70) or CD8 + CD45RC -Tregs (71), rather than directly acting on aCD8 + T cells. The finding in the present study that FK506 has a more potent effect in regulating CD4 + /CD25 + /FoxP3 + Tregs than cHGF/HGF ( Figure 4) may explain the mechanisms, although HGF has also been suggested to play a role in tolerance in Treg regulation after LTx (6,7,63,72).
To our knowledge, this study provides the first evidence that HGF (in the form of cHGF) and FK506 differ in their preferential targets in various T cell subpopulations in allogeneic recipients after LTx. The finding of cHGF/HGF-mediated environmental factorrelated functional recovery in allo-OLT recipient rats is consistent with the data that recipients release a variety of cytokines and haptic proteins and show an improved pathological load. Under multiple conditions, the biological effects of FK506 on tolerance induction for graft survival after LTx occurs through inhibiting the IL-2 mechanism (73,74). In this study, we showed that HGF resulted in an even more obvious inhibition of cytokine. In addition, IL-10 production could be another interesting issue to interpret the difference. We showed that cHGF/HGF treatment enhanced IL-10 production in liver allografts, consistent with previous studies using h-rHGF treatments in organ transplantation animal model (10,11), which did not occur with FK506 treatment. Additionally, we confirmed a protective role of IL-10 against FK506-mediated renal toxicity (75), which is related to HGF production after cHGF monotherapy, inconsistent with the results reported by Ido et al. (76), indicating as HGF/IL-10 a defense mechanism for FK506-mediated renal toxicity in recipients with LTx. Clinically, the investigated phenomena in allo-OLT animals are consistent with those in humans with ARJ, indicating clinically related differences that activate cMet signaling with its ligand HGF agonists to target recipient aCD8 + T cells. Meanwhile, protection of FK-506-mediated renal injury or regulating Tregs with FK506 or both to target recipient aCD4 + T cells could have a value for acute tolerance induction after LTx.
In conclusion, FK506 is currently a commonly used IS in the clinic after LTx, but its effects are not ideal-that is, this drug does not induce appropriate liver allograft tolerance while avoiding general immunosuppression and nonimmunologic adverse effects. Our data demonstrated the advantages of HGF over FK506 in acute tolerance induction after LTx. Induction of recipient aCD8 + T cell apoptosis via activation of cMet on these cells for acute tolerance induction by its ligand HGF also protected against FK506-mediated renal injury. Thus, HGF may become a novel tolerance inducer for strategies of acute tolerance induction after LTx. Further research will likely focus on approaches for stimulating HGF signaling to induce this phenomenon. As HGF has a short half-life in vivo (77), new methods for delivering HGF need to be considered. Additional studies aimed at increasing the effective duration of HGF and improving the bioavailability of HGF using antibodies or agonists will help to further elucidate the role of HGF as a potential tolerogenic agent. Finally, we propose that HGF may play a role in LTx either in combination with FK506 application or as a monotherapy in further clinical studies.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Ethics statement
The studies involving human participants were reviewed and approved by Human Tissue Bank (humans; registration number: Ky2022114) of Army Medical University. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by SYXK YU 2012-0012. Written informed consent was obtained from the individual (s) for the publication of any potentially identifiable images or data included in this article. thick liver allograft sections from allo-OLT model rats on POD 7 and quantification analysis of the percentages of aCD8 + T cell among DAPI+ cells (red bar) and TUNEL expression among aCD8 + T cells (green bar); n = 6. (B) IF staining for CD8 + T cells (red) in naive liver tissues (less than 4% of CD8 + T cells, a bar chart); n = 3. (C-i, ii) FCM detection of aCD8 + cells in blood from the recipient rats that received either allo-OLT (i, top two panels) or cHGF (i, bottoms two panels) and analysis for the cHGF (i, red bar, n = 6/ group), allo-OLT (i, black bar, n = 6/group), and cHGF plus anti-cMet-Ab (ii, green bar, n = 6/group on POD 7. (D-i-iv) Double IF staining for CD8 (red) and ki-67 (green) in liver allografts from the allo-OLT and cHGF groups on POD 7 (i) and 10-12 (ii), and the group of recipient rats received anti-cMet-Ab on the two time points (iii, iv), which were analyzed and the results are shown in (i), n = 6/group. (E-i-iv) TUNEL staining (green) of CD8 + cells (red) point) and anti-cMet-Ab (ii, 400 ng/mL, intrasplenic, boxes, n = 6/group/time point) 2 days before and during surgery and analyzed on POD7 and 10-12 (iii, iv), whose analysis is shown in , n = 6/group. Original magnification, ×400; scale bar = 200 µm; the boxes represent a quarter area of the whole stained field. (F-i, ii) H&E staining images of infiltrating cells around the PVs [(a) boxes], BDs [(b) boxes], and VEs [(c) arrows] in the cHGF and plus anti-cMet-Ab groups on POD 7 (i) and 10-12 (ii). The analyzed RAIS based on Banff's criteria is shown in . Original magnification ×200; scale bar = 100 µm. The data are presented as the mean ± SD of at least three independent experiments. **p < 0.001 vs. the allo-OLT group #p < 0.05 vs. the cHGF group, with Student's t-test.

SUPPLEMENTARY FIGURE 5
cHGF signaling inhibits recipient rat splenic aCD8 + cell activity. (A-i, ii) IF staining for CD8 (red) in serial 5-mm-thick sections in groups of allo-OLT (black bar), cHGF (red bar) (i, boxes), and anti-cMet-Ab (ii, green bar) on POD7, and their quantitative analysis. (B-i-iv) Analysis of cell proliferation for Ki-67 (green) and CD8 (red) data from double IF-stained sections on POD 7 (i, boxed) and 10-12 (ii, boxes) in the allo-OLT and cHGF groups and quantification (green-striped bars); the staining results in the anti-cMet-Abtreated groups are shown in iii and iv (green bars); n = 6/subgroup/time point. (C-i-iv) Analysis of cell apoptosis by IF staining for TUNEL (green) and CD8 (red in the allo-OLT and cHGF groups on POD 7 (i, boxes) and 10-12 (ii, boxes), green-striped bars; staining and analysis of the anti-cMet-Ab group (iii, boxes, green bars); n = 6/subgroup/time point. Original magnification, ×400; scale bar = 200 µm. The data are presented as the mean ± SD of at least three independent experiments. **##P < 0.01 vs. the allo-OLT or cHGF groups.

SUPPLEMENTARY FIGURE 6
Assessment of aCD8 + T cell activity and its correlation with HGF in liver allografts from patients with ARJ. (A) A cartoon image of merged brick color of red, green, pink, and blue signals represented by an arrow. (B-i-iv) Assessment of acute liver rejection based on Banff's criteria by H&E staining for infiltration cells around the PVs [(a) boxes], BDs [(b) boxes], and VEs [(c) arrows] in liver samples from patients with ARJ (ii, n = 6) and identified analysis with RAI scores (iii, iv, dark pink bars, n = 6) using HA liver samples as controls (i, iii, iv, light pink bars, n = 6). Original magnification ×200; scale bar = 100 µm. RAI scores measured as RAISs (0-3) (iii) and tRAISs (0-9) (iv) based on Banff's criteria. (C, D) IHC staining for CD8 and HGF in liver specimens from patients with ARJ in the same field of view; n = 6/group. (D) and HA controls (C), n1-n3 represent individuals, and the boxes represent one-sixth area of the whole stained field. (E) Double IF staining and analysis for CD8 + T cells (green) in total cells (DAPI+) (green bar) and TUNEL+ cells (red, merged as brown, arrow) in CD8 + T cells (red bar) of patients with ARJ, n = 6; the box represents a quarter of whole stained field. Original magnification, ×400; scale bar = 200 µm. The data are presented as the means ± SDs of at least three independent experiments. **p < 0.001 vs. HA liver tissues (Student's ttest, two-way ANOVA).