IL-25 prevents and cures fulminant hepatitis in mice through a myeloid-derived suppressor cell-dependent mechanism

Authors


  • Potential conflict of interest: Nothing to report.

Address reprint requests to: Giovanni Monteleone, M.D., Ph.D., Dipartimento di Medicina dei Sistemi, Università Tor Vergata, Via Montpellier 1, 00133 Rome, Italy. E-mail: Gi.Monteleone@Med.uniroma2.it; fax: +39-06-72596391.

Abstract

Fulminant hepatitis (FH) is a disease characterized by massive destruction of hepatocytes with severe impairment of liver function. The pathogenesis of FH is not fully understood, but hyperactivity of T cells and macrophages with excessive production of cytokines are important hallmarks of the condition. In this study, we investigated the role of interleukin (IL)−25 in FH. IL-25 expression was evaluated in patients with FH and in livers of mice with FH induced by D-galactosamine (D-Gal) and lipopolysaccharide (LPS). Mice were treated with IL-25 before D-Gal/LPS-induced FH and before or after concanavalin A (ConA)-induced FH. Mononuclear cells were isolated from livers of mice treated with or without IL-25 and analyzed for GR1+CD11b+ cells. CFSE-labeled T cells were cocultured with GR1+CD11b+ cells and their proliferation was evaluated by flow cytometry. Mice were also treated with a depleting anti-GR1 antibody before IL-25 and D-Gal/LPS administration. IL-25 was constitutively expressed in mouse and human liver and down-regulated during FH. IL-25 prevented D-Gal/LPS-induced FH and this effect was associated with increased infiltration of the liver with cells coexpressing GR1 and CD11b. In vitro studies showed that GR1+CD11b+ cells isolated from mice given IL-25 inhibited T-cell proliferation. Consistently, in vivo depletion of GR1+ cells abrogated the protective effect of IL-25 in experimental D-Gal/LPS-induced FH. IL-25 was both preventive and therapeutic in ConA-induced FH. Conclusions: IL-25 expression is markedly reduced during human and experimental FH. IL-25 promotes liver accumulation of GR1+CD11b+cells with immunoregulatory properties. (Hepatology 2013;58:1436–1450)

Abbreviations
Ab

antibody

AFP

alpha-fetoprotein

ALB

albumin

ALF

acute liver failure

ALT

alanine aminotransferase

AMD

actinomycin D

APC

allophycocyanin

AST

aspartate aminotransferase

CFSE

carboxyfluorescein diacetate succinimidyl ester

ConA

concanavalin A

DCs

dendritic cells

D-Gal

D-galctosamine

ELISA

enzyme-linked immunosorbent assay

FBS

fetal bovine serum

FCM

flow cytometry

FH

fulminant hepatitis

FITC

fluorescein isothiocyanate

Gr

granulocytic

H&E

hematoxylin and eosin

HMNCs

hepatic mononuclear cells

IF

immunofluorescence

IFN-γ

interferon-gamma

IgG

immunoglobulin G

IL

interleukin

iNOS

inducible nitric oxide synthase

IP

intraperitoneal

IV

intravenous

LPS

lipopolysaccharide

MDSCs

myeloid-derived suppressor cells

Mo

monocytic

NK

natural killer

PBS

phosphate-buffered saline

PCR

polymerase chain reaction

PE

phycoerythrin

PMA

phorbol-12-myristate-13-acetate

TGF-β

transforming growth factor beta

Th

T helper

TNF-α

tumor necrosis factor alpha

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

Fulminant hepatitis (FH) (also termed fulminant liver failure or acute liver failure [ALF]), in patients without previous liver disease, is caused by massive destruction of hepatocytes with resultant severe impairment of liver function, followed by hepatic encephalopathy, and, in many cases, progressive multiorgan failure.[1] Viruses, drugs, and toxins are the major causes of FH.[1] Although many pharmacological approaches have been proposed to recover liver function, transplantation is the only definitive treatment for FH.[2] However, transplantation-related problems, such as lack of donors, surgery-associated complications, risk of rejection, and side effects of immunosuppressive drugs suggest the necessity of novel effective treatments.[1, 2]

The pathogenesis of FH is not fully understood, but circumstantial evidence suggests that an exaggerated, poorly controlled immune response plays a major role in the pathological process.[3] FH is characterized by infiltration of immune cells into the liver and the production of inflammatory cytokines and reactive oxygen species, which promote apoptosis and necrosis of hepatocytes.[3, 4] Most animal models of acute liver injury, including the administration of hepatotoxins (e.g., D-galactosamine; D-Gal) and injection of immune cell-activating substances (e.g., concanavalin A; ConA), do not exactly reproduce the complexity of hepatocyte-damaging mechanisms in patients with FH, but have delineated some of the major pathways of liver injury.[5, 6] T cells, natural killer (NK) cells, NKT cells, and macrophages all play a crucial role in experimental FH, and molecules or compounds, inhibiting the function of these cells, attenuate liver injury.[7]

IL-25 (also known as IL-17E), a member of the IL-17 cytokine family, is highly expressed by polarized T-helper (Th)2 cells and plays a key role in the expansion of Th2 cell responses in various organs.[8] On the other hand, IL-25 can target and deliver negative signals to macrophages and dendritic cells (DCs) with the downstream effect of suppressing the production of proinflammatory cytokines.[9-13] Studies in human and mouse systems have shown that IL-25 inhibits the development and/or amplification of Th1 and Th17 cell responses and exerts therapeutic effects in murine models of autoimmunity.[9, 14] Because an imbalance between dominant Th1 and Th17 responses and reduced Th2 responses has been documented in FH,[3, 15, 16] we hypothesized that defective IL-25 production could play a role in the condition. Therefore, this present study investigated the role of IL-25 in FH. Here, we show that IL-25 is produced by human and murine hepatocytes, and that induction of ALF is associated with a marked down-regulation of IL-25 expression. In vivo in mice, administration of IL-25 protects and reverses acute liver damage through a mechanism mediated by GR1-CD11b-positive myeloid-derived suppressor cells (MDSCs).

Materials and Methods

Mice

Male BALB/c mice (8-10 weeks old) were obtained from Harlan Laboratories (Udine, Italy) and maintained in standard animal cages under specific pathogen-free conditions in the animal facility at the University of “Tor Vergata” (Rome, Italy). The study was approved by the local ethics committee.

Mouse FH

All reagents were purchased from Sigma-Aldrich (Milan, Italy), unless specified. Mice were injected intraperitoneally (IP) with IL-25 (10 µg/mouse; R&D Systems, Minneapolis, MN) 1 hour before IP administration of D-Gal (20 mg/mouse) and lipopolysaccharide (LPS; 0.5 µg/mouse), dissolved in 200 µL of phosphate-buffered saline (PBS; Lonza, Treviglio, Italy). Blood samples were collected 6 hours after D-Gal/LPS administration by retro-orbital bleeding, and mice were sacrificed 2 hours later. Livers were harvested for RNA and protein extraction, isolation of hepatic mononuclear cells (HMNCs), and histopathological analysis.

For ConA-induced FH studies, mice were given IL-25 IP 1 hour before (preventive model) or 6 hours after (therapeutic model) intravenous (IV) injection of ConA (0.4 mg/mouse). Blood samples were collected at different time points (6-48 hours) after ConA administration. Mice were sacrificed and livers were harvested for isolation of mononuclear cells, RNA and protein extraction, and histopathology at the indicated time points. In both mouse models of FH, controls included mice injected with PBS (vehicle).

Cell Isolation From Liver Tissue

HMNCs were isolated from livers of mice using the procedure described by Dong et al., with minor modifications.[17] Briefly, mouse livers were minced using the gentleMACS Dissociator (Miltenyi Biotec, Bologna, Italy), according to the manufacturer's instruction, and the resulting suspension was centrifuged at 50×g for 5 minutes. Supernatants containing HMNCs were collected, washed in PBS, and resuspended in 45% Percoll solution. The cell suspension was gently overlaid onto 70% Percoll and centrifuged for 20 minutes at 750×g. HMNCs were collected from the interphase and washed twice in PBS.

Expression of IL-25 in parenchymal and nonparenchymal liver fractions was analyzed by western blotting. Livers were explanted from normal mice, minced with a scalpel, and incubated in RPMI supplemented with 10% fetal bovine serum (FBS) and 0.04% collagenase D (Worthington Biochemical Corporation, Lakewood, NJ), stirring at 37°C for 30 minutes. The resulting suspension was centrifuged at 30×g for 3 minutes. Pellets containing parenchymal fraction were washed once in PBS and then used for protein and RNA extraction (see Supporting Materials). Supernatants, containing nonparenchymal fraction, were collected, washed in PBS, and resuspended in 45% Percoll solution. The cell suspension was gently overlaid onto 70% Percoll and centrifuged for 20 minutes at 750×g. HMNCs were collected from the interphase, washed twice in PBS, and used for protein and RNA extraction (see Supporting Materials).

For flow cytometry (FCM) analysis of IL-25 expression, cell suspensions, obtained after collagenase digestion of mouse liver, were centrifuged 200×g for 5 minutes, washed once in PBS, and left untreated or incubated with monensin (2 µM; eBioscience, San Diego, CA) in the presence or absence of phorbol-12-myristate-13-acetate (PMA; 40 ng/mL) and ionomycin (1 mg/mL) for 5 hours and then analyzed for CD45, CD3, and IL-25.

FCM

All antibodies (Abs) were purchased from Becton Dickinson (Milan, Italy), unless specified. Freshly isolated HMNCs were initially incubated with Fc-block (1:100 final dilution; Becton Dickinson) for 15 minutes at room temperature, then washed and stained with the following monoclonal antimouse Abs: allophycocyanin (APC)-Cy7 anti-Ly6G (used 1:50 final dilution); phycoerythrin (PE) anti-Ly6C (used 1:200 final dilution); fluorescein isothiocyanate (FITC) anti-CD11b (used 1:50 final dilution); PercP anti-GR1 (used 1:200 final dilution); PE anti-CD3 (used 1:50 final dilution); APC-Cy7 CD45 (used 1:50 final dilution), and FITC-anti IL-25 (R&D Systems). In all experiments, appropriate isotype control immunoglobulin Gs (IgGs; Becton Dickinson or R&D Systems) were used. For intracellular staining, cells were fixed and permeabilized using IC Fixation buffer and the permeabilization buffer (both from eBioscience), according to the manufacturer's instruction. Fluorescence-activated cell sorting analysis was performed using FACScalibur and/or FACSverse (both from Becton Dickinson).

Cell Sorting and T-Cell Proliferation Assay

Mice were injected with IL-25 and D-Gal/LPS, as described above, and HMNCs were isolated 8 hours later. Cells were blocked using Fc-block and stained using the following monoclonal antimouse Abs: APC-Cy7 anti-Ly6G; PE anti-Ly6C; V450 anti-GR1 (all from Becton Dickinson); and PerCP anti-CD11b (BioLegend, San Diego, CA, USA). CD11b+GR1+ cells as well as CD11b+Ly6GhighLy-6C+ and CD11b+Ly6G-Ly-6Chigh subsets were sorted using FACSAria II (Becton Dickinson). Purity of sorted cells was >90%, as evaluated by FCM. T cells were purified from spleen of BALB/c mice by magnetic cell separation. Briefly, splenocytes were passed through a 70-μm nylon mesh cell strainer, and T cells were isolated using a CD90.2+ cell isolation kit (Miltenyi Biotec), according to the manufacturer's instruction. The resulting cell preparations contained more than 95% CD3+ T cells, as assessed by FCM.

T-cell proliferation was assessed using carboxyfluorescein diacetate succinimidyl ester (CFSE; Molecular Probes, Eugene, OR). Purified CD11b+GR1+ cells as well as CD11b+Ly6GhighLy-6C+ or CD11b+Ly6G-Ly-6C+ subsets were cultured at different ratios with syngenic purified CFSE-positive T cells stimulated with anti-CD3/CD28-activating Abs (MiltenyiBiotec). Coculture was performed in a 96-well plate in RPMI 1640 medium supplemented with 10% FBS and 1% penicillin-streptomycin (all from Lonza, Milan, Italy). T-cell proliferation was determined after 72 hours of culture by FCM. RNA was extracted from GR1/CD11b+ and GR1/CD11b cells isolated from livers of mice injected with IL-25 and D-Gal/LPS, as described above, and used for inducible nitric oxide synthase (iNOS) and arginase II RNA expression by real-time polymerase chain reaction (PCR; see Supporting Materials).

In Vivo Depletion of GR1+ Cells

Mice were given IP a depleting antimouse GR1 Ab (250 µg/mouse) or control IgG (250 µg/mouse; both from R&D Systems) 36 hours before IL-25 administration. Blood samples were collected 6 hours after D-Gal/LPS hepatitis induction by retro-orbital bleeding, and sera were stored at −80°C. Mice were euthanized 2 hours later, and livers were explanted for RNA and protein extraction, isolation of mononuclear cells, evaluation of GR-1 cell depletion, and histopathological analysis.

Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling Assay

Please see the Supporting Materials for details on terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay.

Real-Time PCR

Please see the Supporting Materials for details on real-time PCR.

Protein Extraction, Western Blotting, and Enzyme-Linked Immunosorbent Assay

Please see the Supporting Materials for details on protein extraction, western blotting and enzyme-linked immunosorbent assay (ELISA).

Histopathological Analysis and Immunofluorecence

Please see the Supporting Materials for details on histopathological analysis and immunofluorecence.

AML12 Cell Culture and Apoptosis Assay

Please see the Supporting Materials for details on AML12 cell culture and apoptosis assay.

Organ Culture

Information on organ culture can be found in the Supporting Materials.

In Vivo Transfer of MDSCs in D-Gal/LPS-Treated Mice

For information on in vivo transfer of MDSCs in D-Gal/LPS-treated mice, please see the Supporting Materials.

Statistical Analysis

Differences between groups were compared using the Student t test or Mann Whitney's U test.

Results

IL-25 Is Highly Expressed in Livers of BALB/c Mice and Its Production Decreases During Acute Liver Damage

Initially, we measured IL-25 in proteins extracted from various organs of healthy BALB/c mice by ELISA. IL-25 was detectable in extracts from liver, kidney, intestine, spleen, and lung, but the highest concentrations of the cytokine were noted in liver and kidney (Fig. 1A). Western blotting analysis of total liver extracts showed that content of IL-25 was more pronounced in the parenchymal fraction in comparison to the nonparenchymal fraction (Fig. 1B). To exclude the possibility that the high IL-25 noted in the hepatocyte fraction was the result of contaminating leukocytes, albumin (ALB) and CD3 RNA transcripts were evaluated in both hepatocytes and mononuclear cell fractions by real-time PCR. ALB was detected only in hepatocyte-enriched preparations, whereas CD3 RNA expression was markedly higher in mononuclear cells (Supporting Fig. 1A,B). Further analysis of IL-25 expression in hepatocyte-enriched fractions by FCM revealed that the cytokine was mostly produced by CD45-negative cells (Fig. 1C), thus confirming that hepatocytes were the major source of IL-25 in this cell preparation. Moreover, comparison of IL-25 expression in hepatocyte-enriched and mononuclear cell preparations confirmed that IL-25 is mostly produced by hepatocytes and that few CD3-positive cells expressed IL-25 (Fig. 1 C-D). To further prove that IL-25 is constitutively produced by murine hepatocytes, we measured IL-25 in supernatants of AML12 cells, a normal murine hepatocyte line, cultured in the presence or absence of transforming growth factor beta (TGF-β)1, a cytokine that positively regulates IL-25 production in other systems.[18] AML12 spontaneously secreted IL-25 and responded to TGF-β1 with enhanced IL-25 production (Fig. 1E). To evaluate whether induction of acute liver damage changes expression of IL-25, mice were injected IP with D-Gal/LPS, because this experimental model of acute liver damage shows biochemical and immunological changes in the liver similar to those observed in human FH.[19] Mice given D-Gal/LPS exhibited a time-dependent reduction of IL-25 levels in the liver, compared to PBS-treated (control) mice (Fig. 1F), whereas D-Gal/LPS-induced liver damage was associated with no significant change in IL-6 production (not shown). Consistently, RNA transcripts for Fizz, a molecule positively regulated by IL-25,[12] was reduced in livers of D-Gal/LPS-treated mice (Supporting Fig. 2A). In contrast, RNA expression of hepatocyte-derived alpha-fetoprotein (AFP) remained unchanged (Supporting Fig. 2B), suggesting that the decline in IL-25 production in D-Gal/LPS-injected mice was not simply the result of necrosis of hepatocytes. To confirm that the reduced expression of IL-25 was caused directly by the combined activity of D-Gal/LPS, we performed ex vivo organ cultures of liver specimens of BALB/c mice in the presence or absence of D-Gal plus LPS or PBS. Livers cultured with D-Gal plus LPS exhibited a significant decrease in IL-25 production (Supporting Fig. 2C). Together, these observations indicate that induction of D-Gal/LPS-mediated liver damage is accompanied by decreased IL-25 production.

Figure 1.

IL-25 is constitutively expressed in mouse liver. (A) IL-25 was measured in total protein extracts from liver, kidney, small intestine, large intestine, spleen, and lung of mice. Data indicate the mean ± standard deviation of all mice (N = 4). (B) Representative western blottings showing IL-25 expression and β-actin in proteins extracted from the parenchymal (PF) and nonparenchymal fractions (NPF) of livers harvested from normal mice. Figure is representative of three independent experiments in which similar results were obtained. (C) Hepatocytes were purified from livers of normal mice and either left untreated or cultured with monensin in the presence or absence of PMA and inonomycin for 5 hours. Cells were then analyzed for expression of CD45 and IL-25 by FCM. Representative dot plot showing percentages of CD45- and IL-25-positive cells; numbers in quadrants indicate the percentage of positive cells. Isotype control staining is also shown. One of two experiments analyzing cells isolated from at least four mice is shown. (D) Mononuclear cells were purified from livers of normal mice and either left untreated or cultured with monensin in the presence or absence of PMA and inonomycin for 5 hours. Representative dot plot showing percentages of CD3- and IL-25-positive cells; numbers in quadrants indicate the percentage of positive cells. Isotype control staining is also shown. One of two experiments analyzing cells isolated from at least four mice is shown. (E) AML12 cells were cultured in the presence or absence of TGF-β1 (2-10 ng/ml). After 24 hours, cell-free supernatants were collected and analyzed for content of IL-25 by ELISA. (F) Mice, injected IP with D-Gal and LPS, were sacrificed at the indicated time points and livers were explanted and used for extracting total proteins. IL-25 content was then measured by ELISA. Horizontal bars indicate median value (10 mice per each time point were analyzed); *P < 0.05.

IL-25 Protects Mice Against Gal/LPS-Induced FH

Next, we examined whether IL-25 could prevent D-Gal/LPS-driven acute liver damage. Mice were pretreated IP with IL-25 or vehicle 1 hour before D-Gal/LPS administration; blood samples were collected 6 hours later and mice were sacrificed at hour 8. The dose of IL-25 we selected for this study was the same as that we previously used to suppress experimental colitis in mice.[9, 10, 12] As expected, serum levels of both alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were increased in D-Gal/LPS-injected mice and IL-25 significantly reduced D-Gal/LPS-induced transaminases (Fig. 2A,B). Histopathology of liver sections showed severe organ damage in mice treated with D-Gal/LPS, characterized by a confluent hemorrhagic pattern, mononuclear cell infiltrate, and large areas of necrosis (Fig. 2C, left panel). Pretreatment with IL-25 reduced D-Gal/LPS-induced liver damage. In particular, mice receiving IL-25 showed less vessel congestion and reduced infiltration of the liver with inflammatory cells and minimal necrosis (Fig. 2C, left panels). TUNEL assay confirmed massive necrosis of the liver in D-Gal/LPS-injected mice and the preventative effect exerted by IL-25 (Fig. 2C, right panels). In line with these data, western blotting showed activation of caspase-3 in total proteins extracted from mice treated with D-Gal/LPS, but not in proteins extracted from control or IL-25-treated D-Gal/LPS-injected mice (Fig. 2D). Because tumor necrosis factor alpha (TNF-α) is involved in the pathogenesis of D-Gal/LPS-induced liver damage,[20] we next assessed whether IL-25 reduced in vivo TNF-α expression. Pretreatment of mice with IL-25 significantly reduced D-Gal/LPS-induced TNF-α synthesis (Fig. 2E). Moreover, pretreatment of mice with IL-25 significantly reduced D-Gal/LPS-induced IL-23p19 RNA expression (Supporting Fig. 3A), a cytokine known to be negatively regulated by IL-25.[9] Induction of FH by D-Gal/LPS was associated with enhanced IL-17A, but not IL-22 expression (Supporting Fig. 3B,C). IL-25 did not reduce IL-17A induction (Supporting Fig. 3B). Although it has been previously shown that IL-25 reduces Th17 cell responses by suppression of IL-23,[21] the reason why IL-17A was unchanged in IL-25-treated mice, despite down-regulation of IL-23, remains unknown. A possibility is that reduction of IL-23 in IL-25-treated mice occurred in a time frame (i.e., 6-8 hours) that was not sufficient to cause down-regulation of IL-17A. It is also conceivable that, in this model, IL-17A is produced by cell types (e.g., CD8+ cells, NK cells, and neutrophils) that are not directly targeted by IL-23. Altogether, these data support the protective effect of IL-25 in D-Gal/LPS-driven FH.

Figure 2.

IL-25 protects mice against D-Gal/LPS-induced FH. Mice were given IP IL-25 (10 µg) or vehicle 1 hour before treatment with D-Gal and LPS. Control mice were given PBS alone. (A and B) Serum levels of ALT (A) and AST (B) were measured 6 hours after D-Gal/LPS treatment. Data indicate the mean ± standard error of the mean of all mice (N = 8 per group); *P < 0.05; **P < 0.01; ***P < 0.001. (C) Representative photomicrographs of H&E- (left; n = 8 per group) and TUNEL-stained (right; n = 4 per group) frozen liver sections of mice treated as described above and euthanized 8 hours after D-Gal/LPS treatment (original magnification: ×100). Insets show higher magnification (×200) images. (D) Representative western blottings showing active caspase-3 and β-actin in proteins extracted from the livers of mice treated as described above. (E) Total proteins were extracted from mice treated as indicated above and used to measure TNF-α by ELISA. Data are the mean ± standard deviation of the content of TNF-α in livers of eight mice per group. *P < 0.05.

To examine whether the IL-25-mediated antiapoptotic effect in hepatitic mice was the result of a direct effect of IL-25 on hepatocytes, AML12 cells were cultured with TNF-α in the presence of actinomycin D (AMD), a potent inhibitor of RNA transcription, which is known to make hepatocytes sensitive to TNF-α-driven apoptosis.[22] Neither AMD nor TNF-α alone induced significantly hepatocyte apoptosis, whereas the combined stimulation enhanced apoptosis (Supporting Fig. 4). Treatment of cells with IL-25 did not inhibit the AMD- plus TNF-α-driven apoptosis (Supporting Fig. 4).

IL-25 Promotes Accumulation of GR1+CD11b+ Cells in Mice With D-Gal/LPS-Induced FH

We next evaluated whether the protective effect of IL-25 against D-Gal/LPS-driven acute liver damage was associated with changes in the composition of immune cells in the liver. Immunofluorescence (IF) analysis revealed that induction of liver damage by D-Gal/LPS was associated with marked infiltration of the liver with T lymphocytes, macrophages, and neutrophils (Fig. 3A-D). IL-25 alone did not modify the number of lymphocytes, macrophages, and neutrophils, as compared with mice given PBS alone (Fig. 3A-D), but significantly increased the number of both GR1- and CD11b-positive cells in mice treated with D-Gal/LPS (Fig. 3A-D).

Figure 3.

GR1- and CD11b-positive cells accumulate in liver of mice treated with D-Gal/LPS and IL-25. Mice were given IL-25 (10 µg) or vehicle IP 1 hour before treatment with D-Gal/LPS or PBS. Mice were sacrificed 8 hours after D-Gal/LPS treatment, livers were explanted, and IF was performed on liver sections using CD3, F4/80, GR1, and CD11b Abs. Left panels show representative photomicrographs (original magnification: ×100) of sections stained with anti-CD3 (A), -F4/80 (B), -GR1 (C) and -CD11b (D) (right panels). Histograms show the mean ± standard error of the mean of the number of positive cells per high power field (hpf) (N = 6 mice per group); *P < 0.05, **P < 0.01.

IL-25 Enhances the Fraction of GR1/CD11b-Positive Cells With Immunosuppressive Activity in Livers of Mice Treated With D-Gal/LPS

GR1+CD11b+ cells are a heterogeneous subset of myeloid cells.[23] Part of this subset comprises a population of cells with a potent immune-suppressive ability, termed MDSCs.[23] To evaluate whether protection against acute liver damage noted in IL-25-treated mice was accompanied by an increase in the fraction of GR1/CD11b double-positive cells, we assessed the expression of GR1 and CD11b in HMNCs isolated from mice pretreated with IL-25 or vehicle and then injected with D-Gal/LPS or PBS by FCM. D-Gal/LPS, but not IL-25, increased the percentage of GR1/CD11b-positive cells, as compared to PBS-treated mice (Fig. 4A). However, IL-25 significantly enhanced the percentage of GR1/CD11b-positive cells induced by D-Gal/LPS (Fig. 4A). Consistently, real-time PCR analysis showed that expression of transcripts associated with MDSC immune-suppressive activity, such as arginase II and iNOS, was more pronounced in livers of D-Gal/LPS-injected mice pretreated with IL-25, compared to animals pretreated with PBS (Supporting Fig. 5A,B). Moreover, in livers of D-Gal/LPS-injected mice pretreated with IL-25, arginase II and iNOS RNA content was higher in GR1/CD11b-positive cells than in GR1/CD11b-negative cells (Supporting Fig. 5C,D).

Figure 4.

IL-25 induces the accumulation of GR1/CD11b-positive cells with immunosuppressive activity in liver of D-Gal/LPS-treated mice. (A) Histogram shows the mean ± standard deviation (SD) of the percentage of GR1/CD11b-positive cells in HMNCs isolated from 6 mice per group; *P = 0.03; **P = 0.004. Right panel: representative dot plots showing HMNCs, isolated from livers of mice given IL-25 or vehicle 1 hour before treatment with D-Gal/LPS or PBS and analyzed for the expression of GR1 and CD11b by FCM. Numbers in quadrants indicate percentages of GR1/CD11b-positive cells. (B) CFSE-positive T cells were cocultured in the presence or absence of anti-CD3/28-activating Ab with or without GR1/CD11b-positive cells isolated from livers of IL-25 treated D-Gal/LPS-administered mice at the indicated ratio. Histogram shows the mean ± SD of three experiments in which the percentage of CFSE-positive CD3 T cells was evaluated by FCM after 72 hours of culture; *P = 0.007; **P = 0.002; ***P = 0.0005. (C) Representative dot plots showing percentages of Ly6G- and Ly6C-expressing CD11b-positive HMNCs isolated from mice treated as indicated in (A). Numbers in quadrants indicate the percentage of Ly6GhighLy6Cint- (R1) and Ly6G-Ly6Chigh-expressing CD11b-positive cells (R2). (D) CFSE-positive T cells were cocultured in the presence or absence of anti-CD3/28-activating Ab with or without Ly6GhighLy6Cint-expressing CD11b-positive cells (R1), isolated from livers of mice treated as indicated in (A), at the indicated ratio. Histogram shows the mean ± SD of three experiments in which the percentage of CFSE-positive CD3 T cells was evaluated by FCM after 72 hours of culture; *P = 0.0005; **P < 0.0005. (E) CFSE-positive T cells were cocultured in the presence or absence of anti-CD3/28-activating Ab with or without Ly6G-Ly6Chigh-expressing CD11b-positive cells (R2), isolated from the liver of mice treated as indicated in (A), at the indicated ratio. Histogram shows the mean ± SD of three experiments in which the percentage of CFSE-positive CD3 T cells was evaluated by FCM after 72 hours of culture; *P = 0.01; **P = 0.001.

To confirm that GR1/CD11b-positive cells infiltrating livers of mice given IL-25 and, subsequently, D-Gal/LPS have immunosuppressive properties, we cocultured these cells with syngenic T cells at different effector and target ratios in the presence of activating anti-CD3/28 Abs. GR1/CD11b-positive cells significantly inhibited T-cell proliferation (Fig. 4B). GR1 recognizes both Ly6G and Ly6C epitopes, and two major subsets of MDSCs have been identified based on the expression of CD11b, Ly6G, and Ly6C antigens[23]: granulocytic (Gr) MDSCs, which express both Ly6G and Ly6C along with CD11b (CD11b+Ly6GhighLy6Cint), and monocytic (Mo) MDSC-expressing Ly6C and CD11b, but not Ly6G (CD11b+Ly6GLy6Chigh).[23] By FCM, we showed that both CD11b+Ly6GhighLy6Cint and CD11b+Ly6GLy6Chigh were present in livers of mice treated with IL-25 and D-Gal/LPS, even though the majority of MDSCs were Gr-MDSCs (Fig. 4C). These two subsets were then sorted and cocultured with activated T cells to determine their suppressive potential. Both Gr-MDSCs and Mo-MDSCs suppressed T-cell proliferation, but the inhibitory effect of Gr-MDSC was more pronounced, in comparison to Mo-MDSCs (Fig. 4D,E). To confirm that MDSCs inhibit D-Gal/LPS-induced FH, we isolated MDSC from spleen of IL-25-treated mice and injected IV 30 minutes before injecting mice with D-Gal/LPS. Mice transferred with MDSCs were largely protected from D-Gal/LPS FH, as revealed by reduced levels of serum transaminases (Supporting Fig. 5E) and histopathological analysis of liver sections (Supporting Fig. 5F).

Depletion of GR1-Positive Cells Abrogates the Hepatoprotective Effect of IL-25

To determine whether IL-25-induced protection was mediated by MDSCs, mice were given a depleting anti-GR1 Ab 36 hours before IL-25 treatment. GR1/CD11b-positive cells increased after treatment with IL-25 and D-Gal/LPS, but were virtually absent in the HMNC populations isolated from mice pretreated with anti-GR1 and IL-25 and then injected with D-Gal/LPS (Fig. 5A). Efficacy of the depleting Ab was also confirmed by IF analysis of liver sections (Fig. 5B). Analysis of serum transaminases (Fig. 5C,D) and hematoxylin and eosin (H&E) staining of liver sections (Fig. 5E) showed that depletion of GR1/CD11b-positive cells was accompanied by the lack of IL-25-mediated protective effect against D-Gal/LPS-driven acute liver damage.

Figure 5.

In vivo depletion of GR1-positive cells attenuates IL-25-mediated hepatoprotection. Mice were given anti-GR1-depleting Ab (250 µg) or control IgG (250 µg), and 35 hours later, were treated with IL-25 (10 µg) or vehicle, followed by D-Gal/LPS or PBS. (A) Representative dot plots show percentages of GR1- and CD11b-positive cells in HMNCs isolated from livers of mice treated as described above and analyzed by FCM. Numbers in quadrants show the percentage of GR1/CD11b-positive cells. One of six experiments analyzing cells isolated from six mice per group is shown. (B) Representative photomicrographs (original magnification: ×100) of frozen liver sections of mice treated as described in (A) and stained with an anti-GR1 Ab (upper panels) and CD11b (bottom panels). One of six experiments analyzing six mice per group is shown. (C and D) Serum levels of ALT and AST were measured 6 hours after D-Gal/LPS treatment. Data indicate mean ± standard error of the mean of six mice per group; *P < 0.05; **P < 0.01. (E) Representative photomicrographs (original magnification: ×100) of frozen liver sections of mice treated as described in (A) and stained with H&E. Images are representative of six mice per group.

IL-25 Induces Production of CCL17 in the Liver

Several chemokines have been involved in the migration of MDSCs into tissues.[24] Mice treated with IL-25 alone showed a slight increase in CCL17 RNA compared to control mice (Fig. 6A). Induction of liver damage by D-Gal/LPS was accompanied by a significant up-regulation of CCL17 RNA and this was further increased by pretreatment with IL-25. Analysis of CCL17 protein by ELISA showed that both IL-25 and D-Gal/LPS treatments increased CCL17 protein expression, and that mice treated with IL-25 and D-Gal/LPS produced more CCL17 than mice receiving either IL-25 or D-Gal/LPS (Fig. 6B). In contrast, expression of CCL5 (Fig. 6C), CCL19 (Fig. 6D), CCL20 (Fig. 6E), and CCL22 (Fig. 6F) was increased in livers of mice treated with D-Gal/LPS, but not affected by IL-25. GR1/CD11b+ cells isolated from hepatitic mice expressed a high level of CCR4, the CCL17 receptor (Supporting Fig. 6).

Figure 6.

IL-25 induces CCL17 in livers of D-Gal/LPS-treated mice. Mice were given IP IL-25 or vehicle 1 hour before treatment with D-Gal/LPS or PBS. Mice were sacrificed 8 hours after D-Gal/LPS treatment, and livers were explanted and used for RNA and protein extraction. (A) Histogram showing the mean ± standard error of the mean (SEM; n = 6 mice per group) of CCL17 messenger RNA (mRNA) expression, relative to β-actin, as analyzed by real-time PCR; *P < 0.05. (B) Histogram showing the mean ± standard deviation (n = 6 mice per group) of the content of CCL17 protein, as measured by ELISA; *P < 0.05. (C-F) Histogram showing the mean ± SEM (n = 6 mice per group) of CCL5 (C), CCL19 (D), CCL20 (E), and CCL22 (F) mRNA expression, relative to β-actin, as analyzed by real-time PCR; *p < 0.05.

Preventive and Therapeutic Effect of IL-25 Against ConA-Induced Acute Liver Damage

We next explored whether IL-25 was also anti-inflammatory in mice with ConA-induced acute hepatitis. Induction of liver damage by ConA was followed by an increase in serum levels of both ALT and AST, which was evident as early as 6 hours after ConA administration and augmented further at 12 and 24 hours (Fig. 7A). As described elsewhere,[15, 16, 25] ConA-driven hepatitis was accompanied by up-regulation of interferon-gamma (IFN-γ) and IL-17A RNA transcripts and protein (Supporting Fig. 7). Pretreatment of mice with IL-25 significantly reduced serum levels of both transaminases (Fig. 7A), attenuated histological damage (Fig. 7B), and decreased RNA and protein expression of IFN-γ and IL-17A (Supporting Fig. 7). Induction of liver damage by ConA was associated with infiltration of the liver by GR1+ and CD11b+ cells, and the presence of both these cell types was further increased in hepatitic mice treated with IL-25 (Fig. 7C). FCM analysis revealed that the percentage of GR1/CD11b+ double positive cells was higher in IL-25/ConA-treated mice, compared to mice treated with ConA alone (Fig. 7D). Despite ConA inducing massive damage of the liver leading to the formation of large areas of necrosis, mice can survive more than 3 days. This allowed us to test the therapeutic effect of IL-25 by injecting mice 6 hours after ConA administration, a time that was sufficient to cause liver damage (as shown in Fig. 7A,B). Mice treated with ConA and receiving IL-25 showed significantly reduced levels of transaminases, minimal macroscopic lesions, and less necrotic areas (Fig. 7E,F), as compared to mice treated with ConA and receiving PBS.

Figure 7.

IL-25 protects mice against ConA-induced FH. Mice were given IP IL-25 or vehicle 1 hour before treatment with ConA. Control mice received PBS. (A) Serum levels of ALT and AST were measured 6, 12, and 24 hours after ConA treatment in six mice per group. Data indicate the mean ± standard error of the mean (SEM) of all mice (N = 6 per group); *P < 0.05; **P < 0.01. (B) Representative photomicrographs (original magnification: ×100) showing H&E-stained frozen liver sections of mice sacrificed at the indicated time points after ConA administration. Images are representative of six mice/group per each time point. (C) Representative IF photomicrographs (original magnification: ×100) showing frozen liver sections stained with anti-GR1 (upper panels) and -CD11b (bottom panels). Images are representative of six mice euthanized 24 hours after ConA treatment. (D) Representative dot plots showing HMNCs, isolated from livers of mice treated, as described in (A), and analyzed for GR1 and CD11b by FCM. Numbers in quadrants indicate percentages of GR1/CD11b+ cells. Dot plots are representative of six experiments in which HMNCs were isolated from livers of six mice per group, euthanized 12 hours after ConA treatment. A representative dot plot showing IgG control staining is also shown. (E) IL-25 is therapeutic in mice with ConA-induced FH. Mice were treated with ConA and, 6 hours later, with IL-25 or vehicle. Serum levels of ALT were measured 12, 24, 36, and 48 hours after ConA treatment. Data show mean ± SEM of five mice per group; *P < 0.05; **P < 0.01; ***P < 0.001. (F) Upper panel: representative pictures of livers explanted from mice treated as described above and sacrificed 72 hours after ConA treatment. Images are representative of livers explanted in five mice in each group. Lower panel: representative photomicrographs of H&E-stained frozen liver sections of mice treated as described above and euthanized 72 hours after ConA treatment (original magnification: ×100; insets, magnification: ×200). This analysis was performed in five mice per group.

IL-25 Expression Is Reduced in Patients With Acute Fulminant Hepatitis

In a final set of studies, we analyzed IL-25 protein expression in paraffin-embedded liver sections of patients with FH and controls by confocal IF. IL-25-positive cells were clearly evident in control livers, particularly in hepatocytes, but staining was markedly reduced in liver sections from patients with FH (Fig. 8), thus confirming that acute hepatocyte damage is associated with decreased production of IL-25.

Figure 8.

IL-25 expression is down-regulated during human FH. Representative photomicrographs (original magnification: ×100 and ×200) showing IL-25 immunostaining in liver paraffin-embedded specimens from five controls and five patients with FH. Isotype control staining is also shown. Upper right insets show higher magnification images (×400).

Discussion

IL-25 (also known as IL-17E), a member of the IL-17 cytokine gene family, is made by several immune and nonimmune cell types and plays a critical role in expansion of Th2 cell responses and negative regulation of both Th1 and Th17 immunity.[8, 9, 13, 14] Deregulation of IL-25 production has been described in many inflammatory disorders and is supposed to contribute to the progression of the pathology.[8, 9, 13, 14] For example, high IL-25 sustains inflammation in airways of patients with asthma, whereas defective IL-25 synthesis helps perpetuate chronic inflammation in the gut of patients with inflammatory bowel disease.[9, 26] This later finding fits with the demonstration that IL-25 delivers negative signals to macrophages and DCs with the downstream effect of suppressing detrimental inflammatory responses in the gut.[9] The data in the present study expand on these data and indicate that IL-25 is produced in both human and mouse liver. Despite its ability to amplify Th2 cell programs, IL-25 does not polarize Th cell responses along the Th2 pathway.[8, 26] Therefore, it is not surprising that expression of IL-25 in the uninjured liver was associated with no induction of Th2 cytokines. In contrast, data of the present work support the notion that IL-25 constitutively synthesized by epithelial cells plays a key role in maintaining tissue homeostasis and limiting detrimental immune responses.[21, 27] Consistently, IL-25 synthesis was markedly reduced during acute and severe liver damage. The decreased synthesis of IL-25 in livers of mice with FH was paralleled by enhanced synthesis of IL-6 and no significant change in AFP, suggesting that decline in IL-25 synthesis is not secondary to exhaustion of cytokine production. Factor(s)/mechanism(s) involved in down-regulation of IL-25 during FH remain unknown, even though cytokines produced during liver damage could negatively regulate IL-25 expression. One such cytokine could be TNF-α because it is overproduced during FH,[28] and we have previously shown that TNF-α inhibits IL-25 production in the gut.[18]

Because IL-25 targets many immune cells (e.g., macrophages and T cells), which have been involved in the pathogenesis of FH,[1, 2] we next explored the role of this cytokine in acute liver damage. Using two well-established models of FH in mice by activating liver macrophages and T cells by systemic administration of D-Gal/LPS or ConA, respectively, we showed that a single dose of IL-25 was sufficient to prevent liver damage in both models, and this effect was associated with a marked inhibition of pathogenic cytokines in the liver. IL-25 did not directly prevent AMD/TNF-α-induced apoptosis of cultured hepatocytes, suggesting that the IL-25-mediated protective effect against D-Gal/LPS-driven hepatocyte apoptosis is probably secondary to reduced production of apoptotic inducers, such as TNF-α. Interestingly, IL-25 was also therapeutic in the ConA-induced FH model. Whereas this study was ongoing, Meng et al. showed that IL-25 protects mice from bile duct ligation-induced liver fibrosis.[29] Overall, these data strengthened the importance of the cytokine in the negative control of pathogenic cell responses in the liver.

To dissect the mechanism(s) whereby IL-25 counter-regulates inflammatory reactions in the liver, we next performed a detailed analysis of immune cells infiltrating the liver of mice with FH either treated or not with IL-25. Whereas IL-25 by itself was not able to modify the type of cell infiltrate in livers of mice without damage, pretreatment of animals with IL-25 before administration of D-Gal/LPS caused a significant increase in the numbers of cells expressing GR1 and CD11b. These cells, termed MDSCs, are induced in various inflammatory diseases, where they contribute to restrain immune cell activation and favor the resolution of detrimental immune reactions.[30-32] The demonstration that mice with D-Gal/LPS-induced liver damage contained more GR1- and CD11b-positive cells than control mice is not surprising, because it has been reported that inflammation is required for induction of MDSC.[33] MDSCs sorted from IL-25-treated mice given D-Gal/LPS inhibited in vitro proliferation of activated T cells, thus confirming their suppressive nature. Unfortunately, there is no procedure that allows us to selectively deplete MDSCs and test the hypothesis that IL-25-induced MDSC mediates the anti-inflammatory effect of this cytokine. To circumvent these difficulties, we used an alternative approach and evaluated the effect of depletion of GR1 cells on the effect of IL-25 on the D-Gal/LPS-induced FH. Depletion of GR1 cells from mice abolished the IL-25-mediated protection against D-Gal/LPS-induced liver damage. However, we would like to point out that the anti-GR1 Ab we used in the in vivo studies can deplete both MDSCs and neutrophils, so we cannot exclude the possibility that neutrophils are also involved in the anti-inflammatory action of IL-25.

The exact mechanism by which IL-25 promotes accumulation of MDSCs in livers of mice with FH remains to be ascertained. It is unlikely that IL-25 converts tissue-resident cells into MDSCs because the accumulation of MDSCs in livers of IL-25-treated mice was evident at early time points after cytokine administration (i.e., 6 hours). It is more plausible that IL-25 increases the recruitment of MDSCs from the periphery during FH. This hypothesis is supported by the demonstration that livers of mice with hepatitis given IL-25 overproduce CCL17 and that MDSCs isolated from livers of IL-25-treated mice express CCR4, the CCL17 receptor.[34] In line with these findings is the demonstration that IL-25R-deficient, allergen-sensitized mice express low amounts of CCL17 in lung.[35] We have attempted to prove the role of CCL17 in IL-25-mediated accumulation of MDSCs in the liver by injecting mice with a neutralizing CCL17 Ab. However, our preliminary data indicate that mice given anti-CCL17 still accumulate MDSCs into the liver after IL-25 administration. However, we do not know whether this later finding is the result of our inability to fully inhibit CCL17 activity with the neutralizing Ab or reflects the action of other MDSC-attracting chemokines, which were up-regulated in mice given anti-CCL17.

The ability of exogenous IL-25 to induce activity of GR1/CD11b-positive cells has also been recently described in lung, where these cells exacerbate, rather than inhibit, asthmatic allergic reactions.[35] Taken together, these findings are consistent with the demonstration that both IL-25 and MDSCs may have a dual role in the control of inflammatory processes.

In conclusion, our study is the first to show that IL-25 expression is down-regulated in the liver of both humans and mice with FH and that IL-25 exerts both preventive and therapeutic effects in murine models of acute liver damage, raising the possibility that IL-25-based therapies could advance the way we manage patients with this disorder.

Ancillary