CCR5 deficiency exacerbates T-cell–mediated hepatitis in mice

Abstract

Experimental T-cell–mediated hepatitis induced by concanavalin A (Con A) involves the production of different cytokines and chemokines and is characterized by leukocyte infiltration. Because the chemokine receptor CCR5 and its ligands (CCL3, CCL4, and CCL5) regulate leukocyte chemotaxis and activation, we investigated the role of CCR5 during Con A–induced liver injury. Serum levels of CCR5 ligands and their hepatic transcript levels were significantly increased after Con A injection, whereas CCR5⁺ liver mononuclear cells were recruited to the liver. CCR5-deficient (CCR5^−/−) mice disclosed increased mortality and liver injury following Con A administration compared with wild-type mice. CCR5^−/− mice also exhibited increased production of interleukin 4, tumor necrosis factor α, CCL3, CCL4, and CCL5, and a prominent liver mononuclear cell infiltrate, among which many cells were CCR1⁺. In vivo neutralization of CCR5 ligands in CCR5^−/− mice afforded a protection against hepatitis only when CCL5 was neutralized. In conclusion, CCR5 deficiency exacerbates T-cell–mediated hepatitis, and leads to increased levels of CCR5 ligands and a more pronounced liver mononuclear infiltrate, suggesting that CCR5 expression can modulate severity of immunomediated liver injury. (HEPATOLOGY 2005;42:854–862.)

A better knowledge of the basic mechanisms governing immune response in the pathogenesis of liver disease has allowed advances in the management and treatment of hepatitis.1–4 However, hepatitis remains a worldwide health problem associated with significant morbidity and mortality. Concanavalin A (Con A) administration in mice induces severe hepatitis, which is considered to be a model mimicking many aspects of human T-cell–mediated liver diseases and fulminant hepatic failure.5 Many studies have shown that cytokines play a key role in the pathogenesis of this model. Indeed, both Th-1 cytokines (tumor necrosis factor α [TNF-α], interferon γ [IFN-γ], and interleukin [IL]-12) and Th-2 cytokines (IL-4 and IL-5) mediate hepatotoxicity through proinflammatory properties.6–10 If CD4⁺ T cells, and especially natural killer (NK) T cells, play a key role in triggering liver injury, other cells such as macrophages, neutrophils, and eosinophils are also activated and participate in hepatitis.5, 8, 11, 12

Chemokines are small-protein inflammatory mediators that are classically known as chemoattractants for circulating inflammatory cells. However, these ubiquitous proteins intervene in many biological processes, including hematopoiesis, angiogenesis, and mitogenesis.13

The chemokine receptor CCR5 is a G protein–coupled receptor for the CC chemokines CCL3, CCL4, CCL5, and CCL8.14 This receptor constitutes the main coreceptor for the macrophage-tropic strains of HIV 1 and 2, which are responsible for disease transmission.15 In addition to its role in leukocyte chemotaxis, CCR5 exerts a positive regulatory effect on Th-1 differentiation in inflammatory processes.16 In mice, CCR5 is expressed in NK cells, CD4⁺ cells, CD8⁺ T cells, macrophages, and dendritic cells.17

In the liver, CCR5 is expressed at high levels in lymphocytes.18 Recent studies examining the role of a nonfunctional allele of the CCR5 gene (CCR5-Δ32) in hepatitis C have reported an association of this genetic variant with the outcome of hepatitis C virus (HCV) infection.19 Indeed, one study described milder portal inflammation but more severe fibrosis in HCV carriers homozygous for CCR5-Δ32.19 The CCR5-Δ32 mutation has also been reported to be strongly associated with primary sclerosing cholangitis and with biliary lesions following orthotopic liver transplantation.20, 21

The extent to which CCR5 contributes to the pathogenesis of T-cell–mediated hepatitis remains unclear. Therefore, the present study aimed to investigate CCR5 and its ligands in T-cell–mediated hepatitis induced by Con A administration in mice.

Con A, concanavalin A; TNF-α, tumor necrosis factor α; IFN-γ, interferon γ; IL, interleukin; NK, natural killer; HCV, hepatitis C virus; WT, wild-type; mAb, mouse antibody; ALT, alanine aminotransferase; mRNA, messenger RNA; PCR, polymerase chain reaction.

Materials and Methods

Animals.

Six- to ten-week-old female C57BL/6 wild-type (WT) mice were purchased from Iffa Credo (Brussels, Belgium). Female B6;129P2 CCR5-deficient (CCR5^−/−) and B6;129PF2 (WT) mice were purchased from Jackson Laboratory (Bar Harbor, ME). Animals were maintained in our animal facilities on standard laboratory chow and received care in compliance with the national legal requirements and the National Institutes of Health guidelines.

Reagents.

Con A was purchased from Sigma-Aldrich (Bornem, Belgium). Anti-mouse CCL3, CCL4, CCL5 IgG_2A (clones 39624, 46907, and 53405, respectively), and isotype control (clone 54447) rat monoclonal antibodies were purchased from R&D Systems (Minneapolis, MN).

Experimental Protocols.

Mice were injected intravenously with Con A (in a volume of 200 μL pyrogen-free saline). Blood was obtained 2, 4, 8, and 24 hours after Con A injection. Mice were sacrificed via cervical dislocation and the livers were removed. CCL3, CCL4, and CCL5 were simultaneously or individually neutralized in vivo by injecting mice with 0.1 mg anti-CCL3, anti-CCL4, and anti-CCL5 mouse monoclonal antibody (mAb). Hepatitis severity was assessed by mortality rate, serum alanine aminotransferase (ALT), and histology 8 hours after Con A administration.

Measurements of Serum Cytokines, Chemokines, and Aminotransferases.

Concentrations of TNF-α, IFN-γ, IL-4, IL-10, CCL3, CCL4, CCL5, eotaxin, and CXCL10 were determined on serum samples via commercially available ELISA kits from R&D Systems with detection thresholds of 5.1, 2, 2, 4, 1.5, 3, 2, 3, and 2.2 pg/mL, respectively.

ALT levels (EC 2.6.1.2., normal values <35 U/L) were measured on serum samples at 30°C using commercially available kits (Boehringer Mannheim, Germany) based on methods recommended by the International Federation of Clinical Chemistry.

Liver Histology.

After excision, the livers were fixed in formaldehyde. Paraffin sections were stained with hematoxylin-eosin and examined under light microscopy. The slides were read by two of the investigators (C.M., N.N.) in a blinded manner.

Chemokine Messenger RNA Quantification via Real-Time Polymerase Chain Reaction.

CCL3, CCL4, and CCL5 messenger RNA (mRNA) were measured with real-time polymerase chain reaction (PCR) as recently described.22 Liver specimens were frozen in liquid nitrogen after collection. Frozen liver samples were then homogenized in the lysis solution with a MagNalyser (Roche Diagnostics, Brussels, Belgium) with one run of 50 seconds at 6,500 rpm. Total RNA was then extracted via the Tripure procedure (Roche Diagnostics). Reverse transcription was performed as follows: 9 μL H₂O containing 1 μg total RNA was mixed with 4 μL oligo-dT primer (0.1 μg/μL) and incubated at 65°C for 5 minutes. Samples were chilled on ice, and 7 μL RT mix containing the following components were added: 4 μL Transcriptor 5× (Roche Diagnostics) buffer; 2 μL dNTP (deoxyribonucleoside triphosphate) mix (10 mmol/L each); 0.5 μL porcine RNase inhibitor (31.75 U/mL) (Amersham Biosciences, Roosendaal, the Netherlands); 0.5 μL Transcriptor Reverse Transcriptase (20 U/μL) (Roche Diagnostics). The mixture was then incubated for 1 hour at 42°C and then 15 minutes at 70°C. Quantitative PCR was performed with real-time fluorogenic PCR. Amplification of complementary DNA was performed with forward and reverse specific primers, and fluorogenic probes were used for the detection of amplified products (Eurogentec, Seraing, Belgium). Amplification was performed on a LightCycler (Roche Diagnostics). A total of 45 cycles was performed. β-Actin was used as the reference housekeeping gene. The copy number was calculated as previously described.10 Primers and probes were designed using Primer3 software (Whitehead Institute for Biomedical Research, Cambridge, MA). The sequence of primers and probes for the real-time PCR reactions was: β-actin sense, 5′-TCC-TGA-GCG-CAA-GTA-CTC-TGT-3′; β-actin antisense, 5′-CTG-ATC-CAC-ATC-TGC-TGG-AAG-3′; β-actin probe, 5′-(6-Fam)ATC-GGT-GGC-TCC-ATC-CTG-GC-3′ (Tamra) (phosphate); CCL3 sense, 5′-AAG-TCT-TCT-CAG-CGC-CAT-ATG-3′; CCL3 antisense, 5′-GTG-GAA-TCT-TCC-GGC-TGT-AG-3′; CCL3 probe, 5′-(6-Fam)CCC-GAC-TGC-CTG-CTG-CTT-CTC-3′ (Tamra) (phosphate); CCL4 sense, 5′-TTC-TGT-GCT-CCA-GGG-TTC-TC-3′; CCL4 antisense, 5′-CGG-GAG-GTG-TAA-GAG-AAA-CAG-3′; CCL4 probe, 5′-(6-Fam)CCA-ATG-GGC-TCT-GAC-CCT-CCC-3′ (Tamra) (phosphate); CCL5 sense, 5′-ATC-TTG-CAG-TCG-TGT-TTG-TCA-3′; CCL5 antisense, 5′-TTC-TTG-AAC-CCA-CTT-CTT-CTC-TG-3′; and CCL5 probe, 5′-(6-Fam)CCG-CCA-AGT-GTG-TGC-CAA-CC-3′ (Tamra) (phosphate).

Flow Cytometry Analysis.

To obtain liver mononuclear cells, livers were removed and perfused through the portal vein with 5 mL Hank's balanced salt solution (HBSS) containing 100 U/mL collagenase (Gestimed, Brussels, Belgium). Livers were minced with a sterile razor blade and then digested for an additional 50 minutes with 5 mL of a 400 U/mL collagenase solution at 37°C. Organs were teased apart, and the resulting mixture was passed through a 75-μm nylon filter. Cells in suspension were then separated by density centrifugation as previously described.23 Isolated hepatic mononuclear cells were suspended in PBS/0.5% bovine serum albumin. Nonspecific binding of mAbs to Fc receptors was blocked by preincubating cells with culture supernates of anti-FcγRII/III mAb (2.4G2). FITC-conjugated anti-CD3 (clone 145-2C11), anti-CD4 (clone GK1.5), anti-CD8 (clone 53-6.7), anti-CD11b (clone M1/70), anti-NK1.1 (clone PK136), and phycoerythrin-conjugated anti-CCR5 (clone C34-3448), anti-NK1.1 (clone PK136) (BD Biosciences, Erembodegem, Belgium) were used to identify cell populations in the liver through fluorescent-activated cell sorting (FACScalibur; Becton Dickinson, Mountain View, CA) using CellQuest software (Becton Dickinson). To identify CCR1⁺ mononuclear cells recruited to the liver after Con A injection, hepatic mononuclear cells were first fixed and permeabilized using a cytofix-cytoperm plus kit (BD Biosciences) and then incubated with a phycoerythrin-labeled polyclonal antibody against the cytoplasmic tail of murine CCR1 (Santa Cruz Biotechnology, Santa Cruz, CA).

Statistical Analysis.

Results are expressed as the mean ± SEM. Statistical comparisons were made using a two-tailed Mann-Whitney U test. The log rank test was used to compare survival rates. Analyses were performed using SPSS 11.0 software (SPSS Inc., Chicago, IL).

Results

CCR5 Ligand Expression During Con A–Induced Hepatitis.

Acute hepatitis was induced in C57BL/6 WT mice via an intravenous 0.3-mg Con A injection (ALT serum levels at 8 hours: 1,738 ± 815 U/L [mean ± SEM of 8 mice]). Serum levels of CCR5 ligands were determined during the course of liver injury. Significantly increased serum levels of CCL3, CCL4, and CCL5 were detected 2, 4, and 8 hours after Con A administration (P < .05). By 24 hours after Con A injection, serum levels of these chemokines had returned to basal levels (Fig. 1A). In parallel, a significant increase in hepatic CCL3 and CCL4 mRNA expression was observed 2, 4, and 8 hours after Con A injection (P < .05), whereas hepatic CCL5 mRNA expression was significantly increased 8 hours after Con A injection only (P < .05) (Fig. 1B).

Figure 1. Serum (protein) and hepatic (transcript) levels of CCR5 ligands. (A) Serum levels of CCL3, CCL4, and CCL5 were measured via ELISA 0, 2, 4, 8, and 24 hours after Con A administration (n = 4 mice per group and per time point). (B) Hepatic CCL3, CCL4, and CCL5 mRNA expressions were measured via quantitative PCR 0, 2, 4, 8, and 24 hours after Con A administration. Data are expressed as the mean ± SEM. (n = 4 mice per group and per time point). *P < .05 versus control. mRNA, messenger RNA; Con A, concanavalin A.

Con A–Induced Hepatitis Is Characterized by the Recruitment of CCR5⁺ Mononuclear Cells to the Liver.

The percentage of CCR5-expressing mononuclear cells recruited to the liver was determined via flow cytometry analysis. A significant increase in hepatic CCR5⁺ mononuclear cells was observed 8 and 24 hours after Con A administration (P < .05) (Fig. 2A). This observation prompted us to analyze liver cell populations showing increased CCR5 expression. Liver CD4⁺, CD11b⁺, and NK1.1⁺ cells all disclosed a significant increase in CCR5 expression at 8 and 24 hours. Liver CD8⁺ cells showed only a marginal increase in CCR5 expression that was significant at 24 hours after Con A injection (Fig. 2B).

Figure 2. Liver recruitment of CCR5⁺ mononuclear cells. At different time points after Con A administration, mice were killed and mononuclear cells were isolated from the livers. (A) CCR5 expression of liver mononuclear cells was measured via flow cytometry analysis at 0, 2, 8, and 24 hours after Con A administration (n = 4 mice per group and per time point). (B) CCR5 expression on different liver mononuclear cell populations was measured via flow cytometry analysis at 0, 8, and 24 hours after Con A administration. Data are expressed as the mean ± SEM (n = 4 mice per group and per time point) *P < .05 versus control. Con A, concanavalin A.

CCR5-Deficient Mice Develop More Severe Liver Injury.

The role of CCR5 in mediating Con A–induced liver injury was addressed using CCR5-deficient (CCR5^−/−) mice. Animal survival, ALT serum levels, and liver histology were compared in CCR5^−/− and control WT mice. After a 0.2-mg Con A injection, the mortality rate reached 80% in CCR5^−/− mice, whereas all of the WT mice survived (P = .01) (Fig. 3A). A lower dose of Con A was then used to induce nonlethal hepatitis in CCR5^−/− mice. Whereas WT mice only slightly increased their serum ALT levels following 0.1-mg Con A administration, CCR5^−/− mice developed severe hepatitis as demonstrated by a striking elevation in serum ALT levels at 8 hours (2,476.5 ± 1,160.8 vs. 59.3 ± 8.8 U/L in CCR5^−/− vs. WT mice; P < .01) (Fig. 3B). The difference in ALT serum levels was confirmed by histopathological changes. Livers from CCR5^−/− mice exhibited lymphoid and neutrophilic inflammatory infiltrates, numerous apoptotic bodies, and widespread hepatocellular necrosis in the intermediate zone (zone 2) of the liver lobule. In contrast, minimal necrotic changes surrounded by some neutrophils were rarely observed in WT mice (Fig. 4).

Figure 3. Mortality and serum ALT levels in WT and CCR5-deficient mice. (A) Mortality was studied after 0.2-mg Con A injection (n = 5 mice per group; P = .01 at 8 hours). (B) CCR5^−/− and WT mice were injected intravenously with 0.1 mg Con A, and serum ALT levels were measured 8 and 24 hours afterward. Results are from two independent experiments with 5 to 6 mice per group in each experiment. Results are expressed as the mean ± SEM. **P < .01. WT, wild-type; Con A, concanavalin A; ALT, alanine aminotransferase.

Figure 4. Liver histology. CCR5^−/− and WT mice were injected intravenously with 0.1 mg Con A and were sacrificed 8 hours later. Liver sections were stained with hematoxylin-eosin (original magnification ×200). (A) Widespread hepatocellular necrosis and inflammatory infiltrate were observed in CCR5^−/− mice. (B) Livers from WT mice only exhibited minimal necrotic changes surrounded by neutrophils (arrows).

CCR5 Deficiency Affects Cytokine and Chemokine Production.

Serum levels of cytokines that were described to be critically involved in the pathogenesis of Con A–induced liver injury were compared in CCR5^−/− and WT mice (Fig. 5A). The peak of IL-4 and TNF-α secretion was significantly increased in CCR5^−/− mice as compared with WT mice (P < .05). In the same line, a trend toward increased IFN-γ serum levels, although not significant, was observed in CCR5^−/− mice. Serum levels of CCR5 ligands were strongly increased in CCR5^−/− mice when compared with WT mice (Fig. 5A). This increased production was observed at 2 hours after injection for all CCR5 ligands (P < .01) and also at 8 hours after injection for CCL5 (P < .01). In parallel, hepatic CCL3, CCL4, and CCL5 mRNA levels were all significantly increased in CCR5^−/− mice when compared with WT mice 2 hours after Con A administration (Fig. 5B). In contrast, the serum levels of two other chemokines—eotaxin and CXCL10 (C-X-C chemokine ligand 10)—were not different between both groups of mice (Fig. 5A).

Figure 5. Production of cytokines and chemokines in CCR5-deficient and control mice. CCR5^−/− and WT mice were injected intravenously with 0.1 mg Con A. Blood and liver tissues were obtained at different time points. (A) Serum levels of IL-4, TNF-α, IFN-γ, CCL3, CCL4, CCL5, CXCL10, and eotaxin were measured via ELISA. (B) Expression of hepatic CCL3, CCL4, and CCL5 mRNA was measured via quantitative PCR. Results are expressed as the mean ± SEM (n = 7 to 9 mice per group and per time point). *P < .05. **P < .01. IL, interleukin; WT, wild-type; TNF-α, tumor necrosis factor α; IFN-γ interferon γ; Con A, concanavalin A; mRNA, messenger RNA.

Liver Mononuclear Cell Populations in CCR5-Deficient and Control Mice.

Increased serum levels of CCL3, CCL4, and CCL5 observed in CCR5^−/− mice, in comparison with WT mice, prompted us to compare liver mononuclear cell populations between both groups of mice. Eight hours after 0.1-mg Con A injection, a significant increase in liver mononuclear cells was observed in CCR5^−/− mice compared with WT mice (P < .05). When looking at liver mononuclear cell populations, the increase concerned CD4⁺, CD8⁺, CD11b⁺, and CD3⁺NK1.1⁺ cells (Fig. 6A).

Figure 6. Liver mononuclear cell populations in CCR5-deficient and control mice. CCR5^−/− and WT mice were injected intravenously with 0.1 mg Con A and were sacrificed 8 hours later. (A) Absolute number of liver CD4⁺, CD8⁺, CD11b⁺, and CD3⁺ NK1.1⁺ cells. (B) CCR1 expression in liver mononuclear cells was determined via flow cytometry analysis. Results are expressed as the mean ± SEM (n = 4 mice per group). *P < .05. WT, wild-type; Con A, concanavalin A.

Because CCL3 and CCL5 also bind CCR1, we analyzed hepatic CCR1⁺ mononuclear cells in CCR5^−/− and WT mice. Eight hours after 0.1-mg Con A injection, the percentage of CCR1⁺ mononuclear cells recruited to the liver was determined via flow cytometry analysis. A significant increase in the percentage and absolute number of hepatic CCR1⁺ mononuclear cells was observed in CCR5^−/− mice 8 hours following Con A administration (P < .05), whereas the recruitment of CCR1⁺ mononuclear cells into the liver was not increased in WT mice (Fig. 6B).

Role of CCR5 Ligands in the Pathogenesis of Liver Injury in CCR5-Deficient Mice.

The striking increase in CCL3, CCL4, and CCL5 serum levels prompted us to neutralize these chemokines in CCR5^−/− mice. Liver injury was dramatically reduced in CCR5^−/− mice after neutralization of all three ligands (ALT serum levels 8 hours after Con A administration: 190 versus 1,236 U/L in anti-CCL3, anti-CCL4, and anti-CCL5 versus isotype control mAb pretreatment; P < .05), suggesting that increased levels of CCR5 ligands are involved in exacerbating liver damage in CCR5^−/− mice (Fig. 7A).

Figure 7. Effect of neutralization of CCR5 ligands on liver injury in CCR5-deficient mice. Individual neutralization of CCR5 ligands was performed 1 hour before intravenous 0.1-mg Con A administration via intravenous injection of anti(α)-CCL3, anti(α)-CCL4, anti(α)-CCL5, or isotype control mAb (100 μg of each). Simultaneous neutralization of CCR5 ligands was performed via intravenous injection of 100 μg of anti(α)-CCL3, anti(α)-CCL4, and anti(α)-CCL5, or 300 μg isotype control mAb. (A) Serum ALT levels were measured 8 hours after Con A administration. Results are expressed as the percentage compared with control (n = 6 mice per group). (B) Liver mononuclear cell infiltration was measured via flow cytometry analysis 8 hours after Con A administration (n = 6 mice per group). *P < .05. **P < .01. ALT, alanine aminotransferase; ctrl, control; mAb, monoclonal antibody.

To further investigate the role of each chemokine on liver injury following Con A administration, we neutralized each chemokine individually with monoclonal antibodies. Liver injury was reduced by CCL5 neutralization, but no significant difference was noted following CCL3 or CCL4 neutralization (ALT serum levels 8 hours after Con A injection: 3,508 vs. 2,728 vs. 1,629 vs. 3,546 U/L in anti-CCL3 vs. anti-CCL4 vs. anti-CCL5 vs. isotype control mAb pretreatment; P < .05 between anti-CCL5 and isotype control mAb) (Fig. 7A). In parallel, a less important infiltration of the liver by mononuclear cells was observed following CCL4 and particularly CCL5 neutralization (Fig. 7B).

Discussion

During the last decade, several chemokines and their receptors, including CCR5, have been described to be potential actors involved in the pathogenesis of liver diseases.13, 19, 24 However, the lack of interventional studies in T-cell–mediated liver disease models means that the extent to which CCR5 contributes to the pathogenesis of human T-cell–mediated liver diseases remains largely unknown. Con A–induced hepatitis in mice serves as a prototypic model for T-cell–mediated hepatitis.5–12 In the present study, we show that the expression of CCR5 and its ligands is increased in the liver during Con A–induced hepatitis. More importantly, CCR5 deficiency exacerbates hepatic injury and is characterized by an increased number of inflammatory cells in the liver, a striking increase in chemokines that normally bind CCR5, and an increased production of deleterious cytokines. This suggests an immunomodulatory and anti-inflammatory role of CCR5 in this model of acute hepatitis.

CCR5 is a G protein–coupled receptor for several chemokines of the CC family: CCL3, CCL4, CCL5, and CCL8.14 Chemokines of the CC class are of particular interest because they attract and activate macrophages and T lymphocytes. In accordance with a previous study,24 we have identified an early intrahepatic production of CCL3 and CCL4 during Con A–induced hepatitis. In comparison with CCL3 and CCL4, CCL5 mRNA expression is delayed after Con A administration and is also accompanied by a prolonged rise in serum levels. Thus, CCR5 ligands are all upregulated, but with distinct kinetics after Con A administration.

CCR5 is expressed by T cells, NK cells, macrophages, and dendritic cells.17, 25 In this study, we observed an increased recruitment of CCR5-expressing mononuclear cells into the liver, which peaked between 8 and 24 hours after Con A injection and concerned CD4⁺, CD11b⁺, and NK1.1⁺ liver cells and, to a lesser degree, CD8⁺ cells.

The finding that CCR5^−/− mice exhibit more inflammation in their liver is intriguing. Indeed, one could expect a defect in cell recruitment in the absence of a chemokine receptor, as previously described for CCR126 and CCR2.27 However, the absence of macrophage or T-cell recruitment defect has already been reported in CCR5^−/− mice.28 Moreover, recent studies in other experimental models of disease using CCR5^−/− mice support an immunoregulatory role of CCR5. In a model of influenza A virus, CCR5^−/− mice developed increased pulmonary inflammation and mortality.29 CCR5^−/− mice also proved to have enhanced delayed-type hypersensitivity28 and increased T-cell–dependent immune responses in a graft-versus-host disease mouse model,30, 31 and in Mycobacterium tuberculosis infection.32 To our knowledge, there is no evidence in the literature that CCR5-positive cells possess a regulatory function; rather, CCR5 ligation has a Th-1–type profile.16 In the present work, increased inflammatory infiltrate in CCR5^−/− mice concerned CD4⁺ lymphocytes, macrophages, and NKT cells. Because all these cell populations are involved in the pathogenesis of Con A–induced liver injury,5, 8, 11 a more prominent inflammatory infiltrate may be responsible for exacerbating liver damage in CCR5^−/− mice.

To further unravel the mechanisms involved in the worsening of liver injury observed in CCR5^−/− mice, we first studied the cytokines that play a pivotal role in the pathogenesis of Con A–induced liver injury. Cytokines such as IFN-γ, TNF-α, and IL-4 have all been found to be proinflammatory in Con A–induced hepatitis because mice pretreated with anticytokine antibody or cytokine-deficient mice proved to be resistant to Con A–induced liver injury.6–8 TNF-α produced by Kupffer cells is reported to play a crucial role in Con A–induced liver damage because attenuated liver damage observed after depletion of Kupffer cells is accompanied by a decreased TNF-α production.11

Several reports have also shown that IL-4 plays an important role in T-cell–mediated hepatitis.8, 33, 34 Indeed, IL-4 activates NKT cells in an autocrine manner, leading to Fas ligand expression on their surface, which induces apoptosis of hepatocytes.8 Therefore, our observation of increased TNF-α and IL-4 production may partly explain enhanced hepatic injury observed in CCR5^−/− mice.

Secondly, a surprising major increase in CCL3, CCL4, and CCL5 serum levels is observed after Con A administration in CCR5^−/− mice. If CCL4 acts only through CCR5 ligation, CCL3 and CCL5 are able to bind other receptors: CCL3 binds CCR1, and CCL5 exerts its effects through CCR1 and CCR3 ligation.13, 25 We demonstrate that in the absence of CCR5, many CCR1-positive cells are recruited into the liver of CCR5^−/− mice. One explanation could be that increased levels of CCL3 and CCL5 attract more CCR1⁺ cells in the livers of CCR5^−/− mice. Moreover, CCR1 has been recently shown as an important chemokine receptor involved in the pathogenesis of Con A–induced hepatitis.35 Interestingly, methionylated RANTES, a dual CCR1/CCR5 peptide antagonist, decreases hepatitis severity after Con A administration by inhibiting liver recruitment of CD4⁺ T cells, suggesting that CCR1 is a major chemotactic signal for lymphocytes in this model of liver disease.35 Our results also suggest that in the absence of CCR5, increased inflammatory cell infiltration occurs through other chemokine receptor signaling, underlining the complex interplay between redundant chemokines and the chemokine receptor system allowing compensatory mechanisms.

Because CCR5 has been shown to act as a death receptor and NKT cells have been shown to rapidly decline after Con A injection as a result of apoptosis,36–38 we investigated the hypothesis that a more pronounced mononuclear cell infiltration in the livers of CCR5^−/− mice might be related to reduced cell death after Con A stimulation. However, flow cytometry analysis using annexin V staining did not reveal any difference regarding apoptosis in NK1.1⁺ cells between CCR5^−/− and control mice (data not shown).

The mechanisms through which CCR5-binding chemokines are increased in CCR5^−/− mice was not demonstrated, but increased levels are apparently specific for CCR5 ligands. Indeed, serum levels of two other chemokines, CXCL10 and eotaxin, both of which have been described to be upregulated after Con A administration,10, 36 were not different between both groups of mice in our study. We also observed that serum and hepatic transcript levels of CCR5 ligands are increased in CCR5^−/− mice compared with control mice, suggesting the loss of a negative feedback loop in the absence of CCR5. Moreover, a consumption/scavenging phenomenon has recently been described for CCL2/CCR2 in a model of lung inflammation.40 In this line, an absence of consumption of CCR5 ligands by cells expressing their receptor in CCR5^−/− mice could also cause the observed differences. Finally, the unexpected finding of increased CCR5 ligand expression in CCR5^−/− mice has also been described in other disease models.29, 32, 41

We also demonstrated that increased serum levels of CCR5 ligands in CCR5^−/− mice play a pathological role during Con A hepatitis, because their neutralization dramatically reduces the extent of hepatocellular damage in these mice. Interestingly, when each chemokine was neutralized individually with a monoclonal antibody, only CCL5 neutralization could protect against liver injury in CCR5^−/− mice. Moreover, this protective effect was accompanied by a decrease in liver mononuclear cell infiltration, indicating that increased production of CCL5 in the absence of CCR5 is partly responsible for liver inflammation and necrosis. Interestingly, in an experimental model of neurological inflammation, antibody targeting of CCL5 resulted in diminished leukocyte infiltration and neurological damage.42 The protection induced by CCL5 neutralization on liver injury seems milder than following simultaneous CCL3, CCL4, and CCL5 neutralization, and can probably be explained by the fact that chemokines form a redundant system in which the blocking of several components is more efficient than individual neutralization.

In humans, a 32-bp deletion in the coding region of the CCR5 gene, termed CCR5-Δ32, leads to a frame shift in the open reading frame of CCR5. Patients homozygous for this mutation cannot express CCR5 on the cell surface, whereas heterozygosity results in decreased expression of the functional CCR5 protein. The CCR5-Δ32 mutation has gained clinical interest because it confers protection against infection with HIV.15 More recently, CCR5-Δ32 mutation was found to influence disease susceptibility and severity in patients who have liver disease. Studies in chronic viral hepatitis due to HCV have suggested that CCR5 deficiency may affect the outcome of HCV infection, but contradictory results have been published.19, 43 Hellier et al.19 reported an association between the CCR5-Δ32 allele and decreased portal inflammation. Surprisingly, CCR5-Δ32 homozygotes also had more advanced liver fibrosis. However, the role of CCR5 in chronic HCV infection has yet to be clarified, because no association could be found between the CCR5-Δ32 allele and clinical parameters of HCV infection in another study.43 The expression of CCR5 seems to play a role in autoimmune liver disease and in liver transplantation. First, the allele frequency of CCR5-Δ32 is higher in patients with sclerosing cholangitis and is especially high in patients suffering from severe liver disease.20 Secondly, the CCR5-Δ32 mutation has also been reported to be a risk factor for the development of ischemic-type biliary lesions after liver transplantation and to be associated with a decrease in survival.21 Thus, several data corroborate our findings of a protective role of CCR5 in liver pathology and suggest that CCR5 expression in liver cells can influence disease susceptibility and severity in patients who have liver disease. Consequently, the occurrence of hepatotoxicity during treatment with CCR5 antagonists—which are promising new drugs for the treatment of HIV-infected patients—should be carefully considered.

In conclusion, experimental T-cell–mediated liver injury is severely worsened in the absence of CCR5. CCR5 deficiency induces an increase in the production of CCL5, thereby enhancing the recruitment of inflammatory cells into the liver and the production of proinflammatory cytokines. The present study highlights the role of CCR5 in immune response as a brake to limit liver inflammation.

Acknowledgements

The authors thank Dr. Alain Le Moine for helpful discussions and Dr. Marianna Arvanitakis for reviewing the manuscript.