Potential conflict of interest: Nothing to report.
Supported by the National Research Foundation of Korea (NRF) grant (2010-0017987) and a grant from the National R&D Program for Cancer Control (0920040) funded by the MEST and Ministry for Health, Welfare and Family Affairs of the Korean government, respectively (to I.C.).
V-set and Ig domain-containing 4 (VSIG4, CRIg, or Z39Ig), a newly identified B7-related cosignaling molecule, is a complement receptor and a coinhibitory ligand that negatively regulates T-cell immunity. Despite its exclusive expression on liver Kupffer cells (KCs) that play key roles in liver tolerance, the physiological role of VSIG4 in liver tolerance remains undefined. Mice lacking VSIG4 had poor survival rates and severe liver pathology in a concanavalin A (ConA)-induced hepatitis (CIH) model, which could be prevented by adoptive transfer of VSIG4+ KCs. The absence of VSIG4 rendered endogenous liver T- and natural killer T (NKT)-cells more responsive to antigen-specific stimulation and impaired tolerance induction in those cells against their cognate antigens. T-cell costimulation with VSIG4.Ig suppressed Th1-, Th2-, and Th17-type cytokine production and arrested the cell cycle at the G0/G1 phase but did not induce apoptosis in vitro. VSIG4-mediated tolerance induction and cell-cycle arrest were further supported by down-regulation of G1 phase-specific Cdk2, Cdk4, and Cdk6, and up-regulation of tolerance-inducing p27KIP-1 in VSIG4.Ig-stimulated T-cells. Administration of soluble VSIG4.Ig to wildtype mice prevented CIH development and prolonged the survival of mice with established CIH. Conclusion: Collectively, our results suggest that VSIG4+ KCs play a critical role in the induction and maintenance of liver T- and NKT-cell tolerance, and that modulation of the VSIG4 pathway using a VSIG4.Ig fusion protein may provide useful immunological therapies against immune-mediated liver injury including autoimmune hepatitis. (HEPATOLOGY 2012;56:1838–1848)
Despite the risk of immune activation by continuous exposure to potential antigens, the liver avoids overactivation of the innate and adaptive immune responses by inducing tolerance.1, 2 Many studies have investigated the molecular and cellular basis of liver tolerance. Initial studies focused on identifying tolerance-inducing soluble factors from liver nonparenchymal cells, including hepatic stellate cells (HSCs) and liver-resident antigen-presenting cells (APCs), such as liver sinusoid endothelial cells (LSECs), hepatic dendritic cells (DCs), and Kupffer cells (KCs).3 Among them, KCs are believed to induce liver tolerance by producing an immunosuppressive cytokine, interleukin (IL)-10, and immunosuppressive metabolites including nitric oxide, prostaglandin E2 (PGE2), and 15-deoxy-delta 12,14-PGJ2 (15d-PGJ2).3–6
Alternative mechanisms for liver tolerance have also been suggested. KCs prime CD4+ T-cells to be converted to regulatory T cells (Tregs) with a CD25low FoxP3neg phenotype that can inhibit the proliferation of naïve CD4+ T-cells.7, 8 The functional significance of B7-H1 (PD-L1 or CD274), a coinhibitory ligand, in liver tolerance was demonstrated by showing that B7-H1-expressing KCs directly suppress T-cell proliferation and cytokine production by way of the B7-H1:PD-1 pathway.9 These results suggest that coinhibitory ligands in the liver microenvironment are important for regulating local immune responses. Despite the increasing number of coinhibitory ligands that play negative roles in T-cell responses, few studies have focused on the cellular and molecular mechanisms of liver tolerance mediated by these coinhibitory ligands.
Recently, V-set and Ig domain-containing 4 (VSIG4, also referred to as CRIg or Z39Ig) was identified as a B7-related immunoglobulin superfamily member that is exclusively expressed on tissue-resident macrophages and particularly on liver KCs.10 VSIG4 is a complement receptor for C3b and iC3b, and its binding to the convertase subunit C3b interferes with C5 binding to C3b, thus blocking the alternative complement pathway and subsequent suppression of inflammatory responses.10 VSIG4 also acts as a coinhibitory ligand that negatively modulates adaptive immunity.11 Based on the immunosuppressive function of VSIG4 and its exclusive expression on tolerogenic KCs, we hypothesized that VSIG4+ KCs play a critical role in the induction and maintenance of liver tolerance.
In this study, we addressed these issues by examining the protective role of VSIG4 in concanavalin A (ConA)-induced hepatitis using VSIG4 wildtype (WT) and knockout (KO) mice and analyzing the effect of VSIG4+ KCs on the induction of liver T- and NKT-cell tolerance using ovalbumin (OVA)-induced and α-galactosylceramide (α-GalCer)-induced tolerance models. We demonstrated that the absence of VSIG4 greatly reduced survival rates and resulted in severe hepatitis upon ConA challenge, and impaired the induction of liver T- and NKT-cell tolerance. We also found that G1 phase-specific Cdk2, Cdk4, and Cdk6 were down-regulated and tolerance-inducing p27KIP-1 was up-regulated in T-cells costimulated with VSIG4.Ig. Thus, the present study provides evidence that VSIG4 contributes to KC-mediated liver T- and NKT-cell tolerance.
Balb/c and C57BL/6 mice (8-10 weeks old) and LY5.1 congenic mice were purchased from the Jackson Laboratory. VSIG4 KO mice were generously provided by Dr. Campagne (Genentech). Mice were maintained under specific pathogen-free conditions in our animal facilities and received humane care under a protocol approved by the Institutional Animal Care and Use Committee of Inje University.
ConA, bromodeoxyuridine (BrdU), OVA protein, complete Freund's adjuvant, DNase, and collagenase were purchased from Sigma-Aldrich. α-GalCer (KRN 7000) was purchased from Alexis Biochemicals. OVA323-339 peptide was synthesized by Peptron. FITC-, PE-, PE Cy5-, or APC-conjugated anti-CD45.1 (A20), anti-TCR-β (H57-597), anti-NK1.1 (NKR-PIC), anti-CD11b (M1/70), anti-CD11c (N418), anti-F4/80 (BM8), anti-CD146 (ME-9F1), anti-I-Ad (39-10-8), anti-CD80 (16-10A1), anti-CD86 (GL1), anti-B7-H1 (M1H5), anti-CD44 (IM7), anti-CD62L (MEL-14), anti-CD4 (L3T4), anti-IFN-γ (XMG1.2), anti-tumor necrosis factor alpha (TNF-α) (MP6-XT22), anti-IL-4 (11B11), anti-IL-17A (TC11-18H10.1), anti-FoxP3 (FJK-16S), anti-BrdU, and 7-AAD were obtained from eBioscience or BD Pharmingen. Anti-FITC, anti-CD11c, and anti-CD90.2 microbeads were purchased from Miltenyi Biotech. Antibodies against Rb, p130, E2F-1, E2F-4, cyclin D1, cyclin E, Cdk2, Cdk4, Cdk6, p53, p16INI4a, p21Waf1/Cip1, and p27Kip1 were purchased from Santa Cruz Biotechnology. Antibody against VSIG4 (14G8) that blocks the binding of C3b to VSIG4 was a kind gift from Dr. Campagne (Genentech).
To construct a plasmid to express VSIG4 fusion protein, the extracellular domain (aa 1-186) region was first cloned into pFlag-CMV (Sigma-Aldrich) encoding the Fc region of human IgG1 to generate the VSIG4-human Fc fusion protein (VSIG4.Ig). The recombinant plasmid was transfected into HEK293 cells using Lipofectamine 2000 (Invitrogen), and then VSIG4.Ig was purified from the culture supernatant using HiTrap Protein G HP Columns according to the manufacturer's recommendations (GE Healthcare).
Mice were injected intravenously with either a lethal dose (25-30 mg/kg) or a sublethal dose (15 mg/kg) of ConA (Sigma-Aldrich). Serum alanine aminotransferase (ALT) levels were measured using a transaminase kit (Asan Pharmaceutical) according to the manufacturer's instructions. For adoptive transfer of KCs, KCs (3 × 106) isolated from VSIG4 WT or KO mice were injected intravenously into VSIG4 KO mice by way of the tail vein.
Hematoxylin and Eosin Staining.
Mouse livers were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. 5-μm sections were stained with hematoxylin and eosin using a standard procedure and analyzed by light microscopy.
Isolation of Mononuclear Nonparenchymal Cells (MNCs), KCs, Liver T- and NKT-Cells.
Liver MNCs were isolated by the collagenase digestion method with some modification.12–13 Briefly, mouse liver was perfused in situ with Hank's buffered salt solution (HBSS) containing 0.025% collagenase, removed, and passed through 70-μm stainless steel mesh. Initial cell suspension that was resuspended in 40% Percoll was overlaid onto 70% Percoll and centrifuged at 750g for 20 minutes. MNCs were collected from the interface. For purification of KCs, liver MNC suspension was overlaid onto Percoll gradient (25%/50%), and centrifuged at 1,800g for 30 minutes. KC-enriched MNCs located in the interface were harvested and stained with FITC-conjugated anti-F4/80 (clone BM8, eBioscience). F4/80 positive KCs were purified using anti-FITC Microbeads (Miltenyi Biotech) according to the manufacturer's protocols. KC isolates were 95% pure and KCs were the only cell fraction expressing VSIG4 among liver APCs (Supporting Fig. 1). For purification of splenic DCs, splenocytes were incubated with anti-CD11c Microbeads (Miltenyi Biotech) and enriched by the MACS system according to the manufacturer's protocols. For purification of liver T- and NKT-cells, liver MNCs were stained with FITC-conjugated-NK1.1 mAb and PE Cy5-conjugated anti-TCR-β mAb, and then TCR-β+NK1.1+ NKT and TCR-β+NK1.1− T-cells were sorted using a BD FACSAria.
T-Cell Proliferation Assay.
T-cells (105) were plated in 96-well flat-bottom plates that were precoated with indicated concentrations of mouse anti-CD3e antibody (145-2C11) together with VSIG4.Ig or control Ig (10 μg/mL). [3H]-Thymidine (1 μCi/well) was added 16 hours prior to harvesting of the cultures. [3H]-Thymidine incorporation was measured with a Wallac MicroBeta TriLux Liquid Scintillation counter (PerkinElmer). In some experiments, purified DO11.10 T-cells (105) were incubated with KCs (1-10 × 103) in the presence of OVA323-339 (10 μg/mL) for 3 days before [3H]-thymidine incorporation.
In Vivo Tolerance Induction.
For liver NKT-cell tolerance induction, mice (Balb/c background) were injected intraperitoneally with α-GalCer (0.3 mg/mouse) or phosphate-buffered saline (PBS). To test NKT-cell tolerance induction, liver NKT-cells were isolated 16 days after α-GalCer injection, and then activated in vitro for 2 days in the presence of α-GalCer (1-100 ng/mL) in the context of WT splenic CD11c+ DCs (2 × 105). For liver T-cell tolerance induction, mice (Balb/c background) were given 0.5 mg of OVA intragastrically every other day for 8 days. Three days after the last feeding, mice were immunized with OVA emulsified with Complete Freund's adjuvant. To examine T-cell tolerance induction, liver or splenic T-cells were isolated 14 days after the last OVA immunization, and then cultured with WT splenic CD11c+ DCs in the presence of OVA (1-100 μg/mL) for 2-3 days. The culture supernatants were assayed for cytokines including interferon-gamma (IFN-γ), IL-2, or IL-4 by enzyme-linked immunosorbent assay (ELISA).
Flow Cytometric Analysis.
For intracellular cytokine staining, liver MNCs were fixed and permeabilized using the Cytofix/Cytoperm Plus kit (BD PharMingen), stained with FITC-conjugated anti-IFN-γ, anti-TNF-α, and anti-IL-17A, and flow cytometrically analyzed using CellQuest software (Becton Dickinson) or FlowJo software (Tree Star). For cell cycle analysis, cells were labeled with Bride using the Bride Flow Kit (BD Bioscience) according to the manufacturer's instructions. Cells were stained with FITC-conjugated anti-Bride antibody for 20 minutes at room temperature. 7-Amino-actinomycin D (7-AAD, 5 μg/mL) was added to the cell suspension before flow cytometry analysis.
Cytokine levels of TNF-α, IFN-γ, IL-2, IL-4, IL-10, IL-17A, and transforming growth factor beta (TGF-β) were determined by sandwich ELISA using capture and detection antibody pairs according to the manufacturer's instructions (e-Biosciences).
The data from at least two independent experiments were expressed as the mean ± standard error of the mean (SEM) or standard error (SD). Statistical significance was analyzed with an unpaired two-sample t test, Gehan-Breslow-Wilcoxon test, or one-way analysis of variance (ANOVA), followed by the Newman-Keuls post-hoc test.
Mice Lacking VSIG4 Are More Susceptible Than WT Mice to ConA-Induced Liver Injury.
To demonstrate the physiological role of VSIG4 in vivo, VSIG4 WT or KO mice were injected intravenously with ConA. Within 24 hours of ConA injection, all of the VSIG4-deficient mice died of acute hepatitis, whereas half of the WT mice survived indefinitely (P = 0.0075; Fig. 1A). VSIG4 KO mice showed significantly elevated serum ALT levels that peaked at 9 hours after ConA challenge (P < 0.01; Fig. 1B), and exhibited significantly increased TNF-α and IFN-γ levels that peaked at 3 and 9 hours after ConA challenge, respectively, as compared to WT mice (P < 0.001) (Fig. 1C). VSIG4 KO mice also showed massive parenchymal necrosis in liver (Fig. 1D). To rule out the possibility that the difference in liver damage between VSIG4 WT and KO mice following ConA challenge was attributable to intrinsic cell death in VSIG4 KO mice, we performed a Transferase-Mediated dUTP Nick-End Labeling (TUNEL) assay. There was no difference in the number of TUNEL-positive cells in liver tissues between the groups of mice (Supporting Fig. 2).
VSIG4 Suppresses Liver T- and NKT-Cell Production of Proinflammatory Cytokines In Vivo.
Liver damage during conA-induced hepatitis (CIH) depends mainly on the activities of liver T- and NKT-cells and their production of a variety of cytokines, including TNF-α and IFN-γ.14, 15 To examine the effect of VSIG4 on cytokine production by liver T- and NKT-cells, we performed intracellular cytokine staining after ConA challenge. The frequency of NKT-cells producing intracellular proinflammatory cytokines such as IFN-γ, TNF-α, and IL-17A was significantly higher in VSIG4 KO mice than in WT mice (Fig. 2A; Supporting Table 1). A similar pattern of intracellular cytokine production was observed in liver T-cells from VSIG4 WT and KO mice. Furthermore, in vivo administration of soluble VSIG4.Ig to B6 WT mice 2 hours before ConA challenge greatly decreased the frequency of proinflammatory cytokine-producing NKT-cells compared to those given control Ig (Fig. 2B; Supporting Table 1). A similar pattern of cytokine production was observed in liver T-cells from mice given control Ig and VSIG4.Ig. We found that VSIG4.Ig bound naïve liver T- and NKT-cells in a binding assay, which was not blocked by 14G8 mAb, an antibody that blocks C3b binding to VSIG4 (Supporting Fig. 3),16 suggesting that VSIG4 may directly suppress liver T- and NKT-cells by way of unidentified receptor(s).
Adoptive Transfer of VSIG4-Expressing KCs Alleviates Immune-Mediated Liver Injury.
To demonstrate the role of VSIG4+ KCs in the regulation of liver inflammation in vivo, we adoptively transferred VSIG4 WT or KO KCs into mice lacking VSIG4 7 days before ConA injection. Our preliminary study using CFSE-labeled KCs showed that ∼45% of adoptively transferred KCs were localized in liver until 1 week after adoptive transfer (Supporting Fig. 4). The purified KCs expressed comparable levels of CD80, CD86, MHC class II, and B7-H1, and produced similar levels of IL-10 and TGF-β between KCs from VSIG4 WT and KO mice (Supporting Fig. 5A,B). Serum ALT levels were significantly reduced in VSIG4 KO mice that received VSIG4+ WT KCs compared to those that received VSIG4 KO KCs (P < 0.001; Fig. 3A). Similar results were obtained for serum IFN-γ. Consistently, VSIG4 KO mice were largely free of hepatic parenchymal necrosis after receiving VSIG4+ WT KCs compared to those given VSIG4 KO KCs (Fig. 3B).
VSIG4+ KCs Directly Inhibit Liver T- and NKT-Cell Cytokine Production In Vitro.
To examine whether VSIG4+ KCs directly regulate liver T- and NKT-cell cytokine production in vitro, we cocultured WT liver NKT-cells with KCs isolated from VSIG4 WT or KO mice in the presence of α-GalCer (KRN 7000) for 2 days. Liver NKT-cells produced more IFN-γ and IL-4 when cocultured with VSIG4 KO KCs than with WT KCs (at KC:NKT ratio of 1:1; IFN-γ, P < 0.01; IL-4, P < 0.001; Fig. 4A). NKT-cells did not produce detectable levels of IFN-γ or IL-4 in a coculture without α-GalCer stimulation (data not shown). Intracellular TNF-α was increased more in liver NKT-cells cocultured with VSIG4 KO KCs than in counterpart NKT-cells (Fig. 4B). To determine whether VSIG4+ KCs are directly involved in the suppression of T-cell cytokine production in an antigen-specific manner, we cocultured T-cells from TCR transgenic DO11.10 mice with VSIG4 WT or KO KCs in the presence of OVA323-339 for 2 days. DO11.10 T-cells produced more TNF-α and IFN-γ, and to a lesser extent, IL-4, when cocultured with VSIG4 KO KCs rather than with WT KCs (Fig. 4C,D).
Endogenous VSIG4 Is Involved in the Induction and Maintenance of Liver T- and NKT-Cell Tolerance.
We investigated the potential role of VSIG4 in the induction of liver NKT-cell tolerance in vivo by using an α-GalCer-induced NKT-cell tolerance model in which NKT-cells acquire an anergic phenotype following in vivo stimulation with α-GalCer.17 Liver NKT-cells isolated from α-GalCer-tolerized WT mice did not produce IFN-γ and IL-4 in response to in vitro restimulation with a low dose of α-GalCer (10 ng/mL), whereas liver NKT-cells from α-GalCer-tolerized VSIG4 KO mice produced higher levels of IFN-γ and IL-4 (P < 0.001; Fig. 5A). However, the cytokine levels of NKT-cells from α-GalCer-tolerized VSIG4 KO mice in response to in vitro α-GalCer restimulation were still lower than those from WT liver NKT-cells tolerized with vehicle alone (Fig. 5A, inset). Next, to examine the role of endogenous VSIG4 in the induction of liver T-cell tolerance, we used orally tolerized mice with multiple low doses of soluble OVA protein (0.5 mg/mouse). Liver T-cells from orally tolerized WT mice did not produce detectable levels of IFN-γ and IL-2 in response to in vitro restimulation with OVA protein, whereas liver T-cells from orally tolerized VSIG4 KO mice produced significant levels of IFN-γ and IL-2 even at a high concentration of OVA protein (IFN-γ, P < 0.001; IL-2, P < 0.001; Fig. 5B).
To examine the in vivo tolerant state of liver NKT-cells, we stimulated liver MNCs containing NKT-cells and APCs with α-GalCer. VSIG4 KO liver MNCs produced more IFN-γ than WT counterparts (P < 0.001; Fig. 5C). The observation was not due to a difference between VSIG4 WT and KO mice in the frequencies of responding cells in liver MNCs, including NKT-cells, KCs, DCs, and Treg cells (Supporting Fig. 6A-C). Next, we purified Thy1.2+ liver T-cells using anti-CD90.2 microbeads and stimulated the cells with various concentrations of anti-CD3. The liver T-cells from VSIG4 KO mice produced more IFN-γ and IL-2 than WT counterparts (at 1 μg/mL anti-CD3; IFN-γ, P < 0.001; IL-2, P < 0.01; Fig. 5D). Despite enhanced responsiveness of liver T- and NKT-cells from VSIG4 KO mice against cognate antigens, there was no significant difference between VSIG4 WT and KO mice in the frequencies of liver T- and NKT-cells with activated phenotypes, including CD44hi and CD62Llo (Supporting Fig. 6D).
VSIG4 Inhibits T-Cell Proliferation by Suppressing Cell Division and Arresting Cell Cycle at G0/G1 Phase, But Not by Inducing Apoptosis.
To examine the ability of VSIG4-expressing KCs to regulate T-cell proliferation, we cocultured DO11.10 T-cells with KCs from VSIG4 WT and KO mice in the presence of OVA peptide. A thymidine incorporation assay showed that VSIG4 WT KCs significantly inhibit DO11.10 T-cell proliferation compared to KO KCs (P < 0.01; Fig. 6A).
VSIG4.Ig costimulation (20 μg/mL) in the presence of anti-CD3 greatly suppressed T-cell division compared to control Ig treatment (Fig. 6B). Cell cycle analysis revealed that the inclusion of VSIG4.Ig significantly diminished the frequency of cells in S phase, whereas the percentage of cells in the G0/G1 phase increased (Fig. 6C). Interestingly, the apoptotic cells, as indicated by the subdiploid (sub-G0) population, did not increase in T-cells treated with VSIG4.Ig versus control Ig. Similar cell cycle arrest at G0/G1 phase was observed in DO11.10 T-cells cocultured with VSIG4+ KCs in the presence of OVA peptide compared to counterpart DO11.10 T-cells (Supporting Table 2). In addition, expression levels of cell cycle-regulating kinases such as Cdk2, Cdk4, and Cdk6 were down-regulated, and total Rb and p27KIP-1 expression levels were increased over time in T-cells costimulated with VSIG4.Ig (Fig. 6D).
Contact-Dependent T-Cell Inhibition by VSIG4+ KCs Is Prevented and Rescued by CD28 Costimulation.
To determine whether VSIG4+ KC-mediated T-cell inhibition requires direct contact between KCs and T-cells, we performed a transwell assay in the presence of anti-CD3. In the absence of contact with WT KCs, T-cells produced more than twice as much IFN-γ as those directly contacted with WT KCs (P < 0.001; Fig. 7A). In contrast, there was no difference in IFN-γ production between T-cells in the presence or absence of contact with VSIG4 KO KCs. Interestingly, there was no significant difference in the production of immunosuppressive cytokines including IL-6, IL-10, and TGF-β between T-cell cocultures with VSIG4 WT and KO KCs (Supporting Fig. 7). We further tested whether VSIG4-mediated T-cell suppression could be prevented by T-cell activation signaling, such as CD28 costimulation. Addition of agonistic anti-CD28 to T-cells cocultured with WT KCs significantly increased IL-2 production that was otherwise suppressed in control Ig-treated counterpart T-cells (P < 0.001; Fig. 7B). We also examined whether T-cell hyporesponsiveness induced by VSIG4+ KCs could be rescued by CD28 costimulation. OVA-primed DO11.10 T-cells that were cocultured with WT KCs for 7 days produced significantly less IL-2 than those cocultured with VSIG4 KO KCs upon restimulation with anti-CD3 (P < 0.05; Fig. 7C), whose IL-2 reduction was rescued by treatment with anti-CD28 versus control Ig (P < 0.01).
Administration of Soluble VSIG4.Ig Rescues Mice from Severe CIH.
Next, we examined VSIG4 expression in inflamed liver tissues throughout the course of CIH. VSIG4 transcripts rapidly disappeared in liver tissues within 12 hours of ConA injection (r2 = 0.703, P < 0.001), whereas serum ALT levels peaked 12 hours after ConA injection and gradually decreased thereafter (r2 = 0.538, P < 0.01; Fig. 8A). The decrease of VSIG4 transcript expression was not due to a direct effect of ConA because there was no significant difference in VSIG4 expression levels between spleen cells 6 hours after treatment with PBS and ConA in vitro (Supporting Fig. 8A). This inverse correlation between VSIG4 expression and the degree of immune-mediated liver damage was also seen in liver biopsies from autoimmune hepatitis patients, where weaker VSIG4 expression was observed on KCs compared to control liver tissues obtained from patients with cholangiocarcinoma that lacked liver pathology (Supporting Fig. 8B). In vivo injection of VSIG4.Ig (400 μg/mouse) significantly protected mice from acute hepatitis, leading to prolonged survival in 60% of mice treated with a lethal dose of ConA, whereas 100% of the mice given control Ig died within 24 hours of ConA injection (P = 0.0108; Fig. 8B). Liver histological studies revealed that VSIG4.Ig pretreatment greatly reduced hemorrhagic necrosis and inflammatory infiltration in the livers of mice treated with ConA, thus maintaining the integrity of the normal liver microarchitecture (Fig. 8C). To examine the therapeutic effect of VSIG4.Ig in mice with established CIH, we injected control Ig or VSIG4.Ig (400 μg/mouse) into mice 3 hours after injection with a lethal dose of ConA. All control mice given control Ig died of ConA-induced liver damage within 14 hours of ConA injection, whereas infusion of VSIG4.Ig prolonged the survival to 28 hours after ConA injection (P = 0.0336; Fig. 8D).
Here we demonstrate for the first time that endogenous VSIG4 that is exclusively expressed on KCs is involved in the induction of KC-mediated liver tolerance. ConA-challenged mice lacking VSIG4 showed reduced survival and severe liver pathologies that was prevented when VSIG4+ KCs were adoptively transferred. Furthermore, VSIG4-deficient mice failed to induce liver T- and NKT-cell tolerance toward their cognate antigens, and in vivo administration of soluble VSIG4.Ig prevented ConA-induced liver damage. Thus, VSIG4+ KCs likely exert their protective role in immune-mediated liver injury through inducing liver T- and NKT-cell tolerance.
Despite several studies focusing on KC-mediated liver T-cell tolerance, few studies have addressed the physiological role of KCs in the induction of liver NKT-cell tolerance. We provided several lines of evidence supporting the notion that VSIG4 mediates liver NKT-cell tolerance induction. First, liver NKT-cells from α-GalCer-tolerized VSIG4 KO mice, but not those from α-GalCer-tolerized WT mice, failed to suppress IFN-γ and IL-4 production in response to in vitro restimulation with α-GalCer. Second, IFN-γ production was greatly inhibited in WT liver MNCs stimulated with α-GalCer, whereas IFN-γ production was significantly elevated in similar cells from VSIG4 KO mice. Third, infusion of soluble VSIG4.Ig caused intrahepatic endogenous NKT-cells to produce lower levels of proinflammatory cytokines than did infusion of control Ig. Although the latter result could be interpreted as an agonistic effect of VSIG4.Ig, we cannot exclude the possibility that VSIG4.Ig also interferes with the physical interaction between VSIG4 expressed on KCs and its putative receptor on NKT-cells.
Our results indicate that VSIG4 likely inhibits liver T-cell proliferation through down-regulation of cyclin-dependent kinases involved in G1 phase progression such as Cdk2, Cdk4, and Cdk6, resulting in G0/G1 phase arrest. Interestingly, our results also showed that T-cells stimulated with VSIG4.Ig or VSIG4+ KCs did not enhance apoptosis, implying that programmed cell death may not be a major mechanism for VSIG4-mediated T-cell suppression. Furthermore, the finding that VSIG4 costimulation also up-regulates expression of p27KIP-1, which was recently shown to play a role as a tolerance inducer as well as a cell cycle inhibitor18, 19, indicates that p27KIP-1 up-regulation may be a key factor for VSIG4-mediated liver T-cell tolerance induction.
Additionally, our data from the transwell assays indicate that although VSIG4+ KCs suppressed T-cell production of effector cytokines by way of a contact-dependent pathway, immunosuppressive cytokines may also be involved in T-cell suppression because KCs secrete endogenous soluble IL-10 and TGF-β. CD28 costimulation can not only enhance T-cell activation,20, 21 but can also rescue T-cell tolerance induced by various coinhibitory pathways, including B7-H1:PD-1, B7-DC:PD-1, and HVEM:CD160.22–24 In this study, CD28 costimulation not only prevented T-cell suppression by VSIG4+ KCs, but also rescued IL-2 production in T-cells that were rendered hyporesponsive by VSIG4+ KCs. This suggests that CD28 costimulation may reprogram the inhibitory pathway established by VSIG4.
Our result showed VSIG4 expression on cells lining liver sinusoids, presumably KCs, was reduced in autoimmune hepatitis. In line with our finding, a previous report showed that VSIG4 is significantly down-regulated in CD68+ KCs in inflamed liver tissues from patients with chronic hepatitis B.25 Interestingly, these findings sharply contrast with a previous report showing that B7-H1 expression is greatly increased on LSECs and KCs during autoimmune liver diseases,26, 27 an observation that suggests differential roles for VSIG4 and B7-H1 in the pathogenesis of autoimmune hepatitis.
Recently, a report showed that VSIG4 expression on macrophages is down-regulated by proinflammatory cytokines such as IFN-γ and up-regulated by antiinflammatory cytokines including IL-1028, which explains an inverse correlation between the VSIG4 expression levels on KCs and the degree of immune-mediated liver injury in our CIH model. Further studies are needed to fully understand the factors that regulate VSIG4 expression in vivo, especially during the autoimmune hepatitis process.
In light of the dual functions of VSIG4 in host immune defense, we propose a model in which extracellular VSIG4 possesses at least two different binding sites: one for complement C3b and the other for putative receptor(s) on T-cells. This model is based on the observation that an anti-VSIG4 antibody (14G8) that blocks C3b binding did not reverse VSIG4.Ig-mediated T-cell suppression (Supporting Fig. 9A). The proposal was further supported by the findings in the CIH model that there was no survival benefit between mice given VSIG4.Ig together with mouse IgG or 14G8 (Supporting Fig. 9B), and 14G8 did not block the binding of VSIG4.Ig to T- and NKT-cells (Supporting Fig. 3). Collectively, our results suggest that VSIG4+ KCs play a critical role in the induction and maintenance of liver T- and NKT-cell tolerance, and that modulation of the VSIG4 pathway using a VSIG4.Ig fusion protein may provide useful immunological therapies against immune-mediated liver injury including autoimmune hepatitis.
We thank Dr. Menno van Lookeren Campagne (Genentech) for helpful discussions and generous provision of VSIG4 KO mice and 14G8 mAb.