Human Umbilical Cord Mesenchymal Stem Cells Alleviate Concanavalin A-induced Fulminant Hepatitis by Repressing NF-κB Signaling and Glycolysis

Background: Fulminant hepatitis is a severe life-threatening clinical condition with rapid progressive loss of liver function. It is characterised by massive activation and inltration of immune cells into the liver and disturbance of inammatory cytokine production. Mesenchymal stem cells (MSCs) showed potent immunomodulatory properties, and transplantation of MSCs is suggested as a promising therapeutic approach for a host of inammatory conditions. Methods: In the current study, a well-established Concanavalin A (Con A)-induced fulminant hepatitis mouse model was used to investigate the effects of transplanting human umbilical cord MSC (hUC-MSC) on fulminant hepatitis. Results: We showed that hUC-MSCs effectively alleviate fulminant hepatitis in mouse models, primarily through inhibiting T cell immunity. RNA-sequencing of liver tissues and human T cells co-cultured with hUC-MSCs showed that NF-κB signaling and glycolysis are two main pathways involved in mediating the protective role of hUC-MSCs on both Con A-induced hepatitis in vivo and T cell activation in vitro. Conclusion: In summary, our data conrmed the potent therapeutic role of MSCs-derived from human umbilical cord on Con A-induced fulminant hepatitis, and uncovered new mechanisms that glycolysis metabolic shift mediates suppression of hUC-MSCs on T cell immunity.


Introduction
Fulminant hepatitis is an emergency clinical condition usually caused by hepatitis viral infection, or toxin (or drug) exposure that mediates rapid liver function destruction. It is an in ammatory lesion of the liver parenchymal mediated by the immune responses (1). No speci c therapies are currently available for fulminant hepatitis, and orthotopic liver transplantation is the only de nitive therapy for end-stage fulminant hepatitis patients. However, liver transplantation's practical application is limited by the desperate shortage of donor organs and high costs.
Mesenchymal stem cells (MSCs) are adult stem cells found in almost all tissues and organs (2,3). These cells could accumulate at the sites of damage or in ammation and exhibit multipotency potential into various types of differentiated cells (4)(5)(6). Moreover, MSCs exhibited potent immunomodulatory functions both in vitro and in vivo, and thus demonstrated the therapeutic potential for many in ammatory disorders, such as acute graft-versus-host disease (aGVHD) (7), rheumatic arthritis (8), in ammatory bowel diseases (9), and so on. MSCs may modulate the immune responses through either direct cell-to-cell contact or paracrine secretion of soluble factors (10)(11)(12)(13)(14). The therapeutic potential of MSCs in the treatment of liver failure has been con rmed in both preclinical and clinical trials (15)(16)(17).
Concanavalin A (Con A)-induced fulminant liver injury is a well-documented murine model closelyresembling the pathological process of human viral hepatitis and autoimmune liver disease (18-21). Previous studies showed that bone marrow-derived MSCs (17, 22-28), tonsil-derived MSCs (29), and adipose tissue-derived MSCs (30-33) effectively protect mice from Con A-induced liver injury. However, the source of MSCs in all these tissues is quite limited, and isolating MSCs from these tissues is invasive. Human UC-MSCs isolated from the umbilical cord (which is considered waste material) provide an attractive MSCs source alternative. UC-MSCs share a similar morphology, surface markers, tri-lineage differentiation potential, and potent immunomodulatory properties with MSCs from other origins. Moreover, there are su cient sources, suitable materials, easy isolation and culture procedures, and no ethical or moral restrictions for utilising hUC-MSCs (34). These advantages make hUC-MSCs attractive for extensive studies in the eld of stem cell-based therapy research recently.
In this study, the therapeutic effect of hUC-MSCs on Con A-induced fulminant liver injury was investigated, demonstrating that hUC-MSCs could e ciently alleviate Con A-induced liver injury. Mechanistically, hUC-MSCs suppress the hepatic T cell in ltration and activation by inhibiting NF-κB signaling and glycolysis. The results con rmed that hUC-MSCs transplantation is a potential therapeutic approach for treating fulminant hepatitis while uncovering new mechanisms for the immunomodulatory functions of hUC-MSCs.

Isolation and culture of human UC-MSCs
Human umbilical cord collection and processing were conducted after receiving informed consent from all participants included in the study. The hUC-MSCs were obtained and expanded as previously described with minor modi cations (35). In summary, to isolate hUC-MSCs, the umbilical cord was dissected into short sections of about 3cm in length, and the vessels were manually stripped. The Wharton's jelly was then separated from the umbilical cord and minced into small fragments with an approximate diameter of 2 mm. The minced tissue was digested by collagenase for 1 hour and cultured in low glucose DMEM supplemented with 10% heat-inactivated Fetal Bovine Serum (FBS), 2 mM L-glutamine and 1% Penicillin-Streptomycin at 37°C in a 5% CO 2 humidi ed incubator. The medium was changed every three days, and non-adherent cells were discarded by completely changing the medium. Fibroblast-like cells that were considered to be hUC-MSCs grew around the explants in about ten days.
Adherent hUC-MSCs were detached using 0.125% Trypsin-EDTA and passaged at a split ratio of 1:3 when they reached 80-90% con uence. All of the hUC-MSCs used in this study were characterised by ow cytometry and multi-directional differentiation. This study was approved by the Institutional Human Ethics Committee of The Third A liated Hospital of Sun Yat-sen University.

Differentiation of hUC-MSCs to osteoblasts and adipocytes
For differentiation to osteoblasts, hUC-MSCs were seeded onto a 6-well plate at a density of 4×10 3 /cm 2 .

Flow cytometry and cell sorting
Cells were washed and stained with labeled antibody (or isotype control) according to the manufacturer's instructions. Flow cytometric analysis was performed on either an LSR II (BD) or CytoFLEX ow cytometer (Beckman Coulter, Fullerton, CA, USA), and the data were analysed with FlowJo7.6 software (Treestar, Ashland, OR, USA). For ow cytometric cell sorting, cells were stained with the appropriate antibodies and isolated on a BD in ux cell sorter (BD Biosciences, USA).

Co-culture of hUC-MSCs with T cells and the proliferation assay
First, human peripheral blood mononuclear cells PBMC were isolated by density gradient centrifugation. The T lymphocytes were then stained with CD3-FITC and sorted on a BD in ux cell sorter (BD Biosciences, USA). For co-culture, hUC-MSCs were seeded in a 24-well plate and co-cultured at the ratio of hUC-MSCs: CD3 + T cell = 1:5. The stimulant was phytohaemagglutinin (PHA, Sigma-Aldrich) at a nal concentration of 2.5 μg/mL. Following co-culture for 48 hours, CD3 + T cells that were stimulated separately and the CD3 + cells co-cultured with hUC-MSCs were collected and analyzed using ow cytometry (CytoFLEX, Beckman Coulter).

Hematoxylin and eosin (H&E) staining
Mouse liver tissues at the equivalent locations were xed in 4% paraformaldehyde overnight and embedded in para n. Sections of 4µm were prepared for H&E staining.
Serum alanine aminotransferase (ALT) and Aspartate Transaminase (AST) measurements Mouse blood was sampled at various time points after Con A administration and then centrifuged at 3000 rpm at 4°C for 15 min to separate the serum. ALT and AST levels were measured using a Hitachi 7020 automatic biochemical analyzer (Hitachi, Tokyo, Japan).

Isolation of liver non-parenchymal cells
Mouse liver tissues were collected and perfused with Collagenase I to dissociate cells, followed by ltration through a 200 µM cell strainer. Hepatocytes were removed by centrifuging at 50 g for three min at 4°C. Supernatants were collected and liver non-parenchymal cells were isolated with mouse lymphocytes separation medium (tdbscience, China).

RNA sequencing
The sample preparation, sequencing, and data analysis were performed as previously described (36). GSEA was conducted using the Gene set enrichment analysis (GSEA) software (https://www.broadinstitute.org/gsea/) (37). Differentially expressed genes (DEGs) were identi ed using DESeq2 (38) on the following cutoff criteria: P < 0.05 and |Fc| > 1.5. GO and KEGG pathway analyses of the DEGs were performed using the online Database for Annotation Visualization and Integrated Discovery (DAVID) (https://david. ncifcrf.gov/), with P < 0.05 as the threshold for statistical signi cance. Sequencing data are available in the Gene Expression Omnibus (GEO) database under the accession number GSE166327.

Statistical analysis
All results were expressed as mean ± SD. Statistical comparison was made using the two-tailed Student's t-test between two groups or one-way ANOVA for multi-group comparison. Survival was analyzed by the Kaplan-Meier log-rank test. P < 0.05 was considered signi cant. Analyses and graphs were performed using GraphPad Prism version 7 (San Diego, CA).

hUC-MSCs effectively protected mice from Con A-induced fulminant hepatitis
Like MSCs of other origins and previous reports (2, 3), hUC-MSCs were strongly positive for CD29, CD73, CD105, and CD90, whereas they were negative for hematopoietic stem cell markers such CD34 and CD45 (Supplementary Figure 1A). The cells showed typical spindle-shaped broblastic morphology (Supplementary Figure 1B) and were successfully induced to differentiate into osteoblasts and adipocytes (Supplementary Figure 1C).
To evaluate the therapeutic effects of hUC-MSCs on Con A-induced fulminant hepatitis, fulminant liver failure was induced using Con A (20 mg/kg) in mice, and either PBS or hUC-MSCs was administered intravenously 0.5 hours later ( Figure 1A). Compared with the PBS group, hUC-MSCs demonstrated signi cant protection on Con A-induced death ( Figure 1B). A sub-lethal dose of Con A (15 mg/kg) was then administrated to study the mechanism of hUC-MSCs mediated protection on Con A-induced liver injury. The hUC-MSCs group showed signi cantly decreased serum ALT and AST activity (two liver injury markers) at 6, 12, 24, and 48 hours after Con A administration, indicating and immediate, effective improvement of Con A-induced liver injury by hUC-MSCs ( Figure 1C and 1D). Observation of gross liver specimens and histological analysis at 24 hours showed that hUC-MSCs effectively ameliorated Con Ainduced hepatocyte necrosis and disseminated hemorrhage ( Figure 1E and 1F). These results con rm that hUC-MSCs effectively protected the mouse liver from Con A-induced liver injury in vivo.

hUC-MSCs suppressed the activation of CD3 + hepatic T cells in Con A-induced fulminant hepatitis
The Con A-induced liver injury model is characterised by lymphocyte-induced liver damage, especially CD4 + T lymphocytes activation and in ltration in the liver (39). Therefore, this model is considered a suitable mouse model for studying the pathological mechanisms of liver damage due to, for example, human viral hepatitis and autoimmune liver disease (18). However, activation of other immune cells (including B cells, macrophages, NK cells, and NKT cells) has also contributed to Con A-induced hepatitis (19,(40)(41)(42). To determine which immune cells hUC-MSCs modulate in the liver, liver non-parenchymal cells of control, Con A + PBS group and Con A + hUC-MSCs group were isolated 24 hours after Con A administration. A ow cytometry analysis was undertaken to determine which subpopulations were changed by the hUC-MSCs (Figure 2A). The results showed that the number of CD3 + T cells and F4/80+ macrophages was increased after Con A administration, while CD19 + B lymphocytes and NK1.1 + NK/NKT cells were decreased ( Figure 2B). However, only the CD3 + T cell (especially CD3 + CD4 + T cells) ratio recovered after the treatment with hUC-MSCs ( Figure 2B and 2C). These results infer that hUC-MSCs alleviate Con A-induced hepatitis primarily through suppressing hepatic CD3+ T cells.
Transcriptional analysis of hUC-MSCs mechanisms to alleviate Con A-induced fulminant hepatitis in vivo The RNA-sequencing for liver tissues of the control group, Con A + PBS group, and Con A + hUC-MSCs group 24 hours after Con A administration were determined, to further characterise the mechanisms via which hUC-MSCs alleviate Con A-induced hepatitis (Supplementary Table 1). Gene set enrichment analysis (GSEA) showed that almost all (35/36) pathways signi cantly changed in the Con A + PBS group versus the control group were recovered in the Con A + hUC-MSCs group ( Figure 3A). We found that crucial liver functions, such as bile acid metabolism, xenobiotic metabolism, and fatty acid metabolism, were repressed in the Con A + PBS group and reconstituted in the Con A + hUC-MSCs group ( Figure 3B-3D). Also, in ammatory response pathways including IL-2-STAT5, TNFα signaling via NF-κB, IFNα, and IFNγ response were among the most signi cantly enriched gene pathways in the Con A + PBS group. They were repressed in the Con A + hUC-MSCs group ( Figure 3B, 3E). Consistent with the phenomenon of massive hepatocyte necrosis and activation of immune cells, apoptosis pathway and cell cycle-related pathways (including E2F targets, G2M checkpoints, and Myc targets) were activated in the Con A + PBS group ( Supplementary Figure 2A and 2B). Both pathways returned to the resting state in the Con A + hUC-MSCs group (Supplementary Figure 2A and 2B), consistent with the observation of decreased hepatocyte necrosis ( Figure 1F) and suppressed T cell in ltration in the liver ( Figure 2B) after hUC-MSCs treatment.
Notably, we also observed a signi cant change in other metabolism pathways, including activation of glycolysis, mTORC1, and hypoxia response after Con A administration and corresponding recovery following hUC-MSCs infusion. The oxidative phosphorylation pathways showed the opposite tendency ( Figure 3F, Supplementary Figure 2C). In total, 562/674 upregulated and 546/726 downregulated genes in Con A + hUC-MSCs group compared to Con A + PBS group showed opposite expression tendencies in the Con A + PBS group versus the control group (Supplementary Figure 2D and 2E), demonstrating signi cant molecular changes following hUC-MSC treatment. GO and KEGG pathway analysis of differentially expressed genes (DEGs) showed similar results to the GSEA analysis, con rming metabolism and in ammatory pathways are the primary pathways mediating the therapeutic effects of hUC-MSCs on Con A-induced hepatitis (Supplementary Figure 3 and 4).

Bioinformatics analysis of hUC-MSCs mechanisms to inhibit the activation and proliferation of T cells in vitro
Since Con A-induced fulminant hepatitis is mainly T cell-mediated hepatic injury, treatment with hUC-MSCs could ameliorate Con A-induced liver injury principally by reducing CD3 + T cell in ltration ( Figure  2B) Table 2). Like liver tissues, most (28/31) of the signi cantly enriched pathways (Figure 4A), indicating a strong suppression of T cell immunity by hUC-MSCs. GSEA analysis showed that similar to Con A-induced liver injury models, in ammatory pathways, such as IL-2-STAT5, TNFα signaling via NF-κB ( Figure 4B and C), cell cycle-related pathways (Supplementary Figure 5A), and metabolism pathways (including glucose metabolic process, hypoxia, and mTORC1 signaling) ( Figure 4D, Supplementary Figure 5B) were enriched in activated T cells and repressed in activated T cells co-cultured with hUC-MSCs. Interestingly, unlike in liver tissues, oxidative phosphorylation pathways were activated in activated T cells (Supplementary Figure 5C), consistent with previous reports that both aerobic glycolysis and oxidative phosphorylation were increased in activated T cells (43). Most of the differentially regulated genes after T cell activation recovered in activated T cells co-cultured with hUC-MSCs ( Figure 4E and 4F). The KEGG pathway analysis of these differentially regulated genes con rmed signi cant metabolic changes among all three groups (Supplementary Figure  5D).

NF-κB signaling and glycolysis are shared pathways between the in vivo and in vitro models
The overlapped genes of the four categories were considered and analysed: 1) downregulated in Con A + hUC-MSCs versus Con A + PBS group, 2) upregulated in Con A + PBS versus control group, 3) downregulated in activated T cells co-cultured with hUC-MSCs versus activated T cells and 4) upregulated in activated T cells versus naïve T cells. Eighty-nine genes met the criteria, and the KEGG pathway analysis showed that the NF-κB signaling pathway, HIF-1α signaling, and glycolysis pathways were among the most signi cantly enriched pathways ( Figure 5A). Key genes involved in NF-kB signaling and glycolysis were upregulated in Con A-treated mouse liver tissues and activated human T cells, while all were repressed following treatment with hUC-MSCs. As HIF-1α signaling could directly induce glycolysis under hypoxic conditions (44,45), these results con rm that NF-κB signaling and glycolysis metabolic processes are two important pathways mediating the therapeutic function of hUC-MSCs both in Con A-induced fulminant hepatitis in vivo and T cell activation in vitro ( Figure 5B and 5C). RT-qPCR analysis of liver tissues con rmed that Con A administration indeed induced key enzymes involved in glycolysis and MSC treatment partially blunted this effect ( Figure 5D).

Discussion
MSCs are a heterogeneous population of broblast-like progenitor cells derived from mesoderm and ectoderm in early development. Their multidirectional differentiation potential, ease of isolation and gene modi cation, low immunogenicity, chemotaxis to damaged tissues, potent immunoregulatory functions, and many other advantages infer that MSCs have tremendous unrealised potential in cell therapy and regenerative medicine (5,(46)(47)(48)(49). MSCs can be derived from bone marrow, umbilical cord, amniotic uid, placenta, muscle, adipose tissues, dermis, liver, and almost all tissues and organs (2,3). Since the isolation of MSCs, most preclinical and clinical studies utilised bone-marrow MSCs for cell therapy. Sources of bone marrow-derived MSCs are quite limited, and the isolation procedure is invasive. In the present study, human umbilical cord-derived MSCs were used to treat fulminant hepatitis. The hUC-MSCs are rapidly isolated in large numbers from Wharton's jelly of the umbilical cord, which is considered medical waste, providing a robust and e cient source for MSC cell therapy. The current study con rmed that hUC-MSCs have signi cant therapeutic effects on Con A-induced fulminant hepatitis in vivo and potently inhibit T cell immunity in vitro and are as functional as MSCs from other origins (27,29,30).
Concanavalin A (Con A) is a plant lectin extracted from Brazilian rubber Canavalia. Con A-induced liver injury is a commonly used murine liver injury model considered suitable for studying the pathological mechanisms of human viral hepatitis and autoimmune liver disease. The changes in hemodynamics, morphology, and biochemical metabolism are similar to human viral hepatitis and autoimmune liver disease, etc. In Con A-induced liver injury, lymphocyte (especially CD4 + T cell) activation and in ltration play a central role in inducing liver damage (18, 19,40,42,[50][51][52]. Con A can activate many macrophages in the liver sinusoid to produce cytokines, directly damaging liver cells (53), or activate T lymphocytes to produce cytokines and in ltrate the liver leading to immune liver damage (54). Several studies previously reported that MSCs from different origins could alleviate Con A-induced hepatitis by acting on different types of immune cells, including switching Con A-stimulated macrophages to the M2 phenotype by bone marrow-derived MSCs (27), suppression of myeloid-lineage cells and CD4 + T cell by adipose-derived MSCs (30), suppression of T cells by both bone marrow-derived MSCs and tonsil-derived MSCs (26, 29), and suppression of NKT cells by bone marrow-derived MSCs (24,28). Interestingly, our results demonstrated that hUC-MSCs alleviate Con A-induced fulminant hepatitis primarily by suppressing hepatic CD3 + T cells with no signi cant change of macrophages, CD19 + B lymphocytes and NK1.1 + NK/NKT cells. These results indicated that MSCs from different origins might exert therapeutic outcomes through several disparate mechanisms.
To further elucidate the regulatory mechanisms by which hUC-MSCs protect Con A-induced fulminant hepatitis, RNA-sequencing analysis of liver tissues was undertaken. As expected, noticeable improvement in liver functions was observed following hUC-MSCs treatment, such as fatty acid metabolism, bile acid metabolism and xenobiotic metabolism. Furthermore, in ammatory response and cell cycle-related pathways were signi cantly inhibited by hUC-MSCs administration. Interestingly, we also observed the induction of glycolysis by Con A and its suppression after hUC-MSCs administration. Notably, the changes of in ammatory response pathway, cell cycle-related pathway, and glycolysis were recapitulated in T cells activated in vitro and co-cultured with hUC-MSCs.
Analysis of merged DEGs in mouse liver tissues and human T cells showed that NF-κB signaling and glycolysis were two critical pathways mediating the therapeutic function of hUC-MSCs both in Con A-induced fulminant hepatitis in vivo and T cell activation in vitro. NF-κB is a common in ammatory signaling pathway, previously reported to be activated in Con A-induced fulminant hepatitis and suppressed by MSCs (25). Glycolysis plays an essential role in T cell immunity (55)(56)(57)(58)(59). However, to the best of our knowledge, no previous studies have detailed the involvement of glycolysis in Con A-induced liver injury. Additionally, few studies reported MSCs modulating glycolysis in target cells/tissues (60). Our data showed that glycolysis is upregulated in Con A-treated mouse liver tissues and activated human T cells. Simultaneously, glycoysis is repressed following treatment with hUC-MSCs in both systems, indicating that glycolysis plays an important role in mediating the therapeutic outcomes of hUC-MSCs. Further investigations are warranted to see whether targeting glycolysis with other therapeutics will alleviate Con A-induced fulminant hepatitis. Equally of interest are future studies to determine whether hUC-MSCs could mitigate other T cell-mediated tissue injuries by inhibiting glycolysis.

Conclusion
In this study, we con rmed the potent therapeutic role of MSCs-derived from human umbilical cord on Availability of data and materials All data will be made available upon reasonable request to the corresponding author.  shown as mean ± SD (n = 9-12/group) with the indicated signi cance (*p < 0.05, **p < 0.01, ***p < 0.001). for CD3+CD4+ and CD3+CD8+ T cells ratio in CD45+ liver non-parenchymal cells after Con A administration in indicated groups. Data in (B) and (C) is shown as mean ± SD (n = 5-9/group) with the indicated signi cance (*p < 0.05, **p < 0.01).