Homing of inflammatory cells to the liver is key in the progression of non-alcoholic steatohepatitis (NASH). An abnormal response of CD4+ T-cells from obese mice to the chemotactic effect of CXCL12 has been reported but the mechanism involved in this process and relevance in patients are unknown. We aimed to explore the mechanism involved in the abnormal chemotaxis of CXC chemokine ligand 12 (CXCL12) in several mouse models of NASH and the relevance in the context of human non-alcoholic fatty liver disease (NAFLD). We assessed chemotactic responsiveness of CD4+ T-cells to CXCL12, the effect of AMD3100, a CXC chemokine receptor 4 (CXCR4) antagonist, in mice and lymphocytes from patients with NAFLD, and the affinity of CXCL12 for CXCR4. CXCL12-promoted migration of CD4+ T-cells from three different mouse models of NASH was increased and dependent of CXCR4. CD4+ T-cells from patients with NASH, but not from patients with pure steatosis, responded more strongly to the chemotactic effect of CXCL12, and this response was inhibited by AMD3100. Treatment with AMD3100 decreased the number of CD4+ T-cells to the liver in ob/ob mice. CXCL12 expression in the liver, CXCR4 and CXCR7 expression in CD4+ T-cells were not increased in three different mouse models of NASH. However, the affinity of CXCL12 for CXCR4 was increased in CD4+ T-cells of ob/ob mice. In conclusion, the CXCL12/CXCR4 pathway contributes in both mice and patients to the enhanced recruitment of CD4+ T-cells in NASH. An increased affinity of CXCL12 to CXCR4 rather than a higher expression of the chemokine or its receptors is involved in this process.

CLINICAL PERSPECTIVES

  • We have observed previously an increased response of CD4+ T-cells to the chemotactic effect of the chemokine CXCL12 via its receptor CXCR4 in several different mouse models of NASH.

  • In the present study we demonstrate that an increased affinity of CXCL12 to CXCR4 is involved in this process. The increase in the chemotactic effect of CXCL12 extends to patients with NASH but not to patients with pure steatosis.

  • Recruitment of inflammatory cells into the inflamed liver may be attenuated using CXCR4 antagonists, which are currently being developed.

INTRODUCTION

Non-alcoholic fatty liver disease (NAFLD) is characterized by excessive fat accumulation in the liver. Steatosis is innocuous in its pure form, as only patients presenting inflammation [i.e. non-alcoholic steatohepatitis (NASH)] develop advanced liver disease (reviewed in [1]), indicating a key role of liver inflammation (i.e. recruitment and/or activation of inflammatory cells) in the progression of NAFLD [2].

In this context, we have previously shown that liver inflammation in obese mice results from both the fatty liver having a greater potential to attract circulating lymphocytes and from lymphocytes being prone to migration to the liver, a process that was enhanced by lipopolysaccharide (LPS). We have identified that CD4+ T-cells from obese mice were more responsive to the chemotactic effect of CXC chemokine ligand (CXCL)12 [stromal cell-derived factor 1α (SDF-1α)] [3].

CXC ligand 12 (CXCL12) is constitutively expressed in various tissues, including the liver, in which it is expressed by biliary cells [3]. CXCL12 is a multifaceted chemokine playing a key role in tissue homoeostasis, immune surveillance and several pathological disorders, including cancer, autoimmunity (reviewed in [4]) and ulcerative colitis [5]. CXCL12 is the only known natural ligand of CXC receptor 4 (CXCR4) [6], which was initially identified as a co-receptor for the human immunodeficiency virus [7]. CXCR4 has also been identified as a CD14-independent lipopolysaccharide (LPS) receptor cluster [8]. CXCR7, the second CXCL12 receptor that we and others have detected in primary human leucocytes including T-cells [9], may also regulate several aspects of CXCL12 activities, including chemotaxis [10]. AMD3100, a specific competitive inhibitor of CXCR4 [11], has been shown to have antiviral activity in clinical trials [12]. In accordance with the pleiotropic role of the CXCL12/CXCR4 axis, AMD3100 also has beneficial effects in various animal models of human diseases, including inflammatory diseases, such as rheumatoid arthritis [13] and asthma [14].

The molecular mechanism involved in the abnormal response of CD4+ T-cells from obese mice to the chemotactic effect of CXCL12 remained unknown as well as the relevance in the context of human NASH. We now show herein that CXCL12/CXCR4 dysfunction extends to the context of obese patients with NASH but, most importantly, not in the context of pure steatosis. Moreover, we also demonstrate that the mechanisms accounting for the increased response of CD4+ T-cells to the chemotactic effect of CXCL12 was an abnormal high affinity of CXCL12 for its main receptor CXCR4. The recruitment of inflammatory cells into the liver was decreased by a pharmacological antagonist of CXCR4. Therefore, our study provides evidence for the involvement of the CXCL12/CXCR4 axis in the process of CD4+ T-cells recruitment in NASH.

MATERIALS AND METHODS

Mice

All mice were treated in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996). They were fed a standard chow consisting of 4.3% lipid and 70% carbohydrate, or a methionine and choline-deficient diet (MCD) consisting of 10% lipid and 67.4% carbohydrate for 2 weeks, or a high-fat diet (HFD) consisting of 60% lipid and 20% carbohydrate for 16 weeks, all supplied ad libitum.

Ten-week-old male C57BL/6 ob/ob and C57BL/6 mice were purchased from Janvier.

Mice were anaesthetized with xylazine (Rompun 2%; Centravet) and ketamine (Imalgene; Centravet) before they were killed.

Lymphocyte collection and isolation

Lymphocytes were isolated from mouse liver by a two-step perfusion procedure as described previously [3].

In the first step, circulating cells were eliminated by perfusing the liver in situ, via the portal vein with PBS supplemented with 5 mmol/l EDTA at 37°C, at a flow rate of 4 ml/min. In the second step, the liver was then excised and homogenized in 0.05% collagenase IV (Sigma–Aldrich) buffered with 0.1 mol/l Hepes for 30 min at 37°C. The resulting suspension was filtered through a filter with 70 μm pores, and the filtrate was centrifuged at 50 g at room temperature for 2 min to remove liver parenchymal cells.

Splenocytes were isolated from the spleen of mice by homogenization in PBS and passage through a 70 μm filter. Lymphocytes from spleen were purified using Lympholyte-M cell density gradient centrifugation (Cedarlane) and incubated with the following monoclonal antibodies (mAbs) (purchased from BD Biosciences): FITC-conjugated anti CD19 (clone 1D3), allophycocyanin (APC)-conjugated anti CD4 (clone RM4-5) and peridinin–chlorophyll–protein complex (PerCP)-conjugated anti CD5 (clone 53-7.3). Sorting of splenic CD19CD4+CD5+ cell subset (mainly corresponding to CD4+ T-cell population) was carried out on a FACSAria (BD Biosciences). The purity of CD4+ cells was higher than 99%, assessed by flow cytometry analysis (results not shown).

Peripheral blood mononuclear cell collection and isolation

Peripheral blood mononuclear cells (PBMCs) were obtained by Isopaque-Ficoll (Lymphoprep; Nycomed) gradient centrifugation from heparin-treated blood samples from (i) healthy blood donors (Etablissement Français du Sang); (ii) patients with NASH; and (iii) patients with obesity-related steatosis without NASH. NASH was defined by a NAFLD activity score ≥5 with a score of at least 1 for hepatocellular ballooning. Pure steatosis was defined histologically by a NAS score <5 and absence of hepatocellular ballooning, or by the absence of the metabolic syndrome, which is a strong predictor of NASH [15]. Patients were enrolled at Antoine-Béclère University hospital and gave written informed consent. The present study was approved by the ethics committee (CPP) of Ile-de-France.

Chemotaxis assays

Mouse splenocytes or PBMCs were incubated for 1 h with 10 and 25 μg/ml AMD3100 (Sigma–Aldrich), respectively. Lymphocyte chemotaxis was then evaluated with the Transwell system (Corning Costar). Cells were suspended in RPMI supplemented with 20 mmol/l Hepes and 0.5% BSA with and without AMD3100 in the upper chamber. The lower chamber contained the same medium with CXCL12, which was provided by Dr F. Baleux (Institut Pasteur, Paris, France) [16] at various concentrations or chemokine-free medium.

Lymphocytes or PBMCs that migrated to the lower chamber during the 2-h incubation period were collected, stained with CD4 phycoerythrin (PE)-conjugated and CD3 PerCP-Cyanine5.5 (PerCP-Cy5.5)-conjugated mAbs, and counted by flow cytometry (FACS Calibur or LSR Fortessa analyser; BD Biosciences). The percentage of transmigrating cells was calculated as follows: {[(number of cells migrating to the lower chamber in response to the chemokine) − (number of cells migrating spontaneously)]/(number of cells added to the upper chamber at the start of the assay)} × 100.

Treatment of mice with AMD3100

We used AMD3100 (Sigma–Aldrich), a potent antagonist for CXCR4, by implanting Alzet® osmotic minipumps (number 2002 pumps with a delivery rate of 0.5 μl/h for 14 days) interscapularly into ob/ob and control mice. These pumps were filled with 10 mg of AMD3100 in 200 μl of PBS or with PBS alone. AMD3100 was administered for 12 days after implantation of the pumps, to ensure that sufficient AMD3100 was delivered during the treatment period.

Metabolic status of mice

Blood was collected by retro-orbital vein puncture under anaesthesia. The serum was stored at −80°C until its use for determinations of alanine and aspartate transaminases (ALT and AST, respectively) and glucose levels. Serum insulin concentration was determined by ELISA (R&D Systems).

Expression of CXCR4, CXCR7 and CXCL12

Quantification of transcripts

Total cellular RNA was isolated from freshly sorted CD4+ T-cells using the RNeasy Plus Mini kit (Qiagen). RNA was quantified using NanoDrop technology (Thermo Scientific) and integrity was assessed using a BioAnalyzer 2100 (Agilent Technologies). Samples with RNA integrity numbers superior to 7 were processed for gene expression analyses. RNA were then normalized to 1 μg/μl and reverse transcribed with poly-d(T)15 and Moloney murine leukaemia virus (MMLV) reverse transcriptase (Fisher Bioblock).

We subjected 5 μg of total RNA isolated from a frozen liver sample to treatment with deoxyribonuclease I (Invitrogen). We then used it to generate cDNA by reverse transcription with random hexamers (Roche Diagnostics) and MMLV reverse transcriptase (Invitrogen).

cDNA amplification was performed by quantitative real-time PCR on a LightCycler® 480 instrument (Roche Diagnostics) with the LightCycler® 480 SYBR Green detection kit (Roche Diagnostics).

Primer sequences

The following primer sequence were used: forward (454–473) 5′-GACCTTCAACACCCCAGCCA-3′ and reverse (689–705) 5′-CGGACGATTTCCCTCTCAGC-3′ primers for β-actin (216 bp), forward (359–378) 5′-TCATCACACTCCCCTT-CTGG-3′ and reverse (664–683) 5′-GGGTAAAGGCGGTCAC-AGAT-3′ primers for CXCR4 (286 bp), forward (414–433) 5′-GGTCAGTCTCGTGCAGCATA-3′ and reverse (547–566) 5′-GTGCCGGTGAAGTAGGTGAT-3′ primers for CXCR7 (114 bp), forward (3191–3213) 5′-GTAGTGGCTCCC-CAGGTTTGTCC-3′ and reverse (3439–3463) 5′-CGAGAG-TCCTTTTTCAGGGTCATGG-3′ primers for CXCL12 (226 bp).

PCRs

Reactions were performed using the following amplification scheme: 95°C for 10 min, and 40 cycles consisting of: 95°C for 20 s, 62°C for 15 s, and 72°C for 20 s. The dissociation curve method was used according to manufacturer's protocol (60°C–95°C) to verify the presence of a single specific PCR product. Relative quantification was performed with the standard curve method and results were expressed as CXCR4/β-actin, CXCR7/β-actin and CXCL12/β-actin ratios.

Quantification of membrane CXCR4 expression level

Levels of membrane CXCR4 expression in cells were determined by flow cytometry (FACS Calibur or LSR Fortessa analyser; BD Biosciences) with the PE-conjugated anti-mouse CXCR4 (clone 2B11) or the corresponding isotype control antibody in combination with PerCP-Cy5.5-conjugated hamster anti-CD3 (clone 2C11) and APC-conjugated rat anti-CD4 (clone RPA-T4) mAbs (BD Biosciences). CXCR4 expression in cells was determined as follows: [CXCR4 geometric mean fluorescence intensity (MFI)] − (isotype control antibody of CXCR4 geometric MFI). CXCR4 is expressed as a ratio normalized to the control group for each model. The mean of the control group corresponds to a ratio of 1.

mAbs for flow cytometry

Mouse lymphocytes were stained with PE-conjugated rat anti-CD4, FITC-conjugated rat anti-CD8, APC-conjugated rat anti-CD19 or mouse anti-NK1.1, and PerCP-Cy5.5-conjugated hamster anti-CD3 mAbs (Pharmingen). Human lymphocytes were stained with PE-conjugated mouse anti-human CD4 and PerCP-Cy5.5-conjugated mouse anti-human CD3 mAbs (Pharmingen). Lymphocyte subpopulations were analysed by four-colour flow cytometry (FACS Calibur; BD Biosciences).

Binding assay

For saturation binding experiments, splenocytes were isolated from the spleen of ob/ob and control mice. We maintained 106 cells at room temperature for 90 min in PBS supplemented with 0.2% BSA and 1 mM Hepes, with various concentrations of biotinylated CXCL12 (biot-CXCL12), as previously described [10]. Synthetic biot-CXCL12 was provided by Dr F. Baleux (Institut Pasteur). The cells were washed in ice-cold binding buffer, and incubated at 4°C with 1 μg/ml streptavidin (SAv)–PE conjugate (BD Biosciences), anti-CD4 APC-conjugated and anti-CD3 PerCP-Cy5.5-conjugated mAbs in binding buffer for 20 min. Cells were then washed once and fixed in PBS containing 2% paraformaldehyde. Results were analysed by flow cytometry (LSR Fortessa analyser; BD Biosciences). Specific binding was estimated by subtracting the non-specific binding determined in the presence of unlabelled CXCL12 at 10−6 M from the total binding. Binding parameters (KD) were determined with Prism software (GraphPad Software), using non-linear regressions applied to one-site models.

Liver cells apoptosis labelling

Parrafin-embedded liver tissues were cutted into 4-μm-thick sections. Terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling (TUNEL) experiments were realized on liver slides using a commercial kit (In situ cell death detection kit, fluorescein; Roche Diagnostics) referring to manufacturer's protocol. Slides were counterstained with Hoechst to show nuclei. Slides were scanned using NanoZoomer 2.0RS (Hamamatsu) and viewed using NDP.view2 software (Hamamatsu). Gamma were adjusted to 2.5, contrast to 200% and colour intensity of green and blue to 200%. The count of TUNEL-positive cells was determined in ten consecutive fields (×400).

Immunohistochemical staining

Immunohistochemistry was performed on the Bond® Max system (Leica Biosystems). CD3-positive cells were detected on 4 μm paraffin liver sections using monoclonal anti-CD3 antibody (clone SP7; Neomarkers) incubated for 15 min at room temperature after antigen retrieval in 1 M citrate buffer at pH 6.0 for 30 min, Bond® Polymer Refine Detection (Leica Biosystems) and 3,3′-diaminobenzidine. Slides were counterstained with Hemalun and observed using Leica DMRB microscope (Leica Microsystems). Images were obtained with camera Sony power HAD (Sony) and then digitalized using Calopix software (TRIBVN). Representative fields are shown (×1000).

Statistical analysis

All values are reported as means±S.E.M. For quantitative data, comparisons were based on non-parametric tests [Mann–Whitney; Kruskal–Wallis ANOVA post-hoc Fischer's protected least-squares difference (PLSD)] using Statview software (SAS Institute). All P values lower than 0.05 were considered significant.

RESULTS

Increase in CXCL12-promoted migration of CD4+ T-cells relies on CXCR4 in several mouse models of NASH

We first investigated the dose-dependent chemotactic response of splenic CD4+ T-cells to CXCL12. CD4+ T-cells from ob/ob mice displayed a higher migratory response to the concentrations of chemokine tested than cells from control mice. The CXCL12-induced chemotaxis of cells from both types of mice followed a typical bell-shaped dose–response, with the maximum reached at 50 nM of the chemokine. These results indicate that CXCL12 promotes the chemotaxis of CD4+ T-cells more efficiently in ob/ob mice than in control mice (Figure 1A).

CXCR4-mediated enhanced migration of CD4+ T-cells from three mouse models of NASH to CXCL12

Figure 1
CXCR4-mediated enhanced migration of CD4+ T-cells from three mouse models of NASH to CXCL12

(A) In vitro response of splenic CD4+ T-cells from ob/ob (black boxes) and control (white boxes) mice to CXCL12 and the effect of AMD3100. Graph shows means±S.E.M. for three to four mice. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Mann–Whitney test). (B) In vitro response of splenic CD4+ T-cells from HFD-fed (light grey boxes), MCD-fed (dark grey boxes), and control (white boxes) mice to CXCL12 and the effect of AMD3100 (10 μg/ml). Graph shows means±S.E.M. for four to eight mice. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Kruskal–Wallis ANOVA post-hoc Fischer's PLSD test).

Figure 1
CXCR4-mediated enhanced migration of CD4+ T-cells from three mouse models of NASH to CXCL12

(A) In vitro response of splenic CD4+ T-cells from ob/ob (black boxes) and control (white boxes) mice to CXCL12 and the effect of AMD3100. Graph shows means±S.E.M. for three to four mice. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Mann–Whitney test). (B) In vitro response of splenic CD4+ T-cells from HFD-fed (light grey boxes), MCD-fed (dark grey boxes), and control (white boxes) mice to CXCL12 and the effect of AMD3100 (10 μg/ml). Graph shows means±S.E.M. for four to eight mice. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Kruskal–Wallis ANOVA post-hoc Fischer's PLSD test).

ob/ob mice are leptin-deficient and insulin-resistant. Therefore, we studied two alternative models of steatohepatitis to decipher whether the increased chemotactic effect of CXCL12 on CD4+ T-cells involved leptin deficiency: mice fed a HFD and mice fed a MCD. CD4+ T-cells from these two mice groups also displayed enhanced migratory properties with CXCL12 (Figure 1B). It is generally assumed, although not formally demonstrated, that in vitro high concentrations of chemokine lead to increased desensitization and internalization of the cognate receptor, thus resulting in attenuated biological effect. As this physiological regulatory process is preserved in CD4+ T-cells from three different mouse models of NASH, this suggests that the enhanced migration does not result from an impaired CXCR4 inactivation [17]. Therefore, to study whether the CXCR4 receptor was involved in the enhanced CD4+ T-cells migration, we specifically blocked CXCR4 by the use of AMD3100, a specific inhibitor of CXCR4. The treatment of CD4+ T-cells from the three distinct models with AMD3100, strongly inhibited CXCL12-induced chemotaxis (Figure 1). This indicates that the higher sensitivity of CD4+ T-cells in three different mouse models of NASH mice involves the CXCR4 receptor.

Increased response of CD4+ T-cells from patients with NASH to the chemotactic effect of CXCL12

To extend our results in the context of human disease, we assessed the chemotactic effect of CXCL12 on CD4+ T-cells recovered from patients with NAFLD. Among patients with NAFLD, those with pure steatosis display no significant liver infiltration of lymphocytes, no signs of hepatocellular injury and no increase in mortality rate. In contrast, patients with NASH display liver inflammation, which may progress to cirrhosis, and have a high liver disease-related mortality rate [15,18,19]. We therefore compared the chemotactic response to CXCL12 of circulating CD4+ T-cells from patients with NASH with those of patients with pure steatosis and controls (healthy blood donors). We also investigated the impact of a blockade of the CXCL12/CXCR4 pathway using AMD3100. Characteristics of patients with NAFLD included in the present study are summarized in Table 1. Patients with NASH had a increased level of ALT and AST, whereas transaminase level was normal in patients with pure steatosis. γ-Glutamyltransferase (GGT) levels were higher in patients with NASH than in patients with pure steatosis (Table 1). The presence of NASH was shown by a liver biopsy. In transwell experiments, CD4+ T-cells from patients with NASH responded more strongly to the chemotactic effect of CXCL12 than CD4+ T-cells from patients with pure steatosis or controls (Figure 2). The response of CD4+ T-cells to CXCL12 was similar in patients with steatosis and controls. Moreover, this stronger CXCL12-mediated chemotactic response was inhibited by AMD3100 (25 μg/ml), indicating that the response of CD4+ T-cells to CXCL12 was mediated by CXCR4 (Figure 2).

CXCR4-mediated enhanced migration of CD4+ T-cells from patients with NASH to CXCL12

Figure 2
CXCR4-mediated enhanced migration of CD4+ T-cells from patients with NASH to CXCL12

In vitro response to CXCL12 of CD4+ T-cells isolated from healthy blood donors (white box), patients with pure steatosis (grey boxes) and patients with NASH (black boxes) and the effect of AMD3100. Graph shows means±S.E.M. for six healthy donors, five patients with pure steatosis and four patients with NASH. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Mann–Whitney test).

Figure 2
CXCR4-mediated enhanced migration of CD4+ T-cells from patients with NASH to CXCL12

In vitro response to CXCL12 of CD4+ T-cells isolated from healthy blood donors (white box), patients with pure steatosis (grey boxes) and patients with NASH (black boxes) and the effect of AMD3100. Graph shows means±S.E.M. for six healthy donors, five patients with pure steatosis and four patients with NASH. Results are expressed as the percentage of CD4+ T-cells migrating in response to CXCL12. Statistically significant differences are indicated (Mann–Whitney test).

Table 1
Characteristics of patients

Values are means±S.E.M. (Mann–Whitney test). NS, not significant. BMI, body mass index; GGT, γ-glutamyltransferase; HbA1c, glycated haemoglobin; HDL, high-density lipoprotein; INR, international normalized ratio; LDL, low-density lipoprotein.

Steatosis (n=5)NASH (n=4)P
Sex (% of males) 60 − 
Age (years) 42.8±2.0 55.0±4.4 NS 
Weight (kg) 122.6±10.4 95.6±8.1 NS 
Height (cm) 166±2 173±7 NS 
BMI (kg/m244.4±4.2 31.6±1.2 <0.05 
Leucocytes (g/l) 7.80±1.73 6.02±0.88 NS 
Haemoglobin (g/dl) 13.2±0.92 14.6±0.62 NS 
Platelets (g/l) 297±15 205±36 <0.05 
Blood glucose (mg/dl) 94±3 108±14 NS 
HbA1c (%) 5.5±0.1 6.1±0.5 NS 
Cholesterol (mmol/l) 5.85±0.34 4.72±0.43 NS 
Triacyglycerols (g/l) 2.30±0.59 1.69±0.33 NS 
HDL-cholesterol (mmol/l) 1.34±0.13 1.26±0.15 NS 
LDL-cholesterol (mmol/l) 3.20±0.36 2.57±0.35 NS 
ALT (IU/l) 19±3 57±2 <0.01 
AST (IU/l) 22±1 43±10 <0.01 
GGT (IU/l) 40±6 217±67 <0.05 
Total bilirubin (μmol/l) 7±1 14±4 NS 
Alkaline phosphatases (IU/l) 99±11 105±24 NS 
Prothrombin ratio (%) 100±0 79±6 <0.05 
INR 1.00±0.00 1.20±0.06 <0.05 
Steatosis (n=5)NASH (n=4)P
Sex (% of males) 60 − 
Age (years) 42.8±2.0 55.0±4.4 NS 
Weight (kg) 122.6±10.4 95.6±8.1 NS 
Height (cm) 166±2 173±7 NS 
BMI (kg/m244.4±4.2 31.6±1.2 <0.05 
Leucocytes (g/l) 7.80±1.73 6.02±0.88 NS 
Haemoglobin (g/dl) 13.2±0.92 14.6±0.62 NS 
Platelets (g/l) 297±15 205±36 <0.05 
Blood glucose (mg/dl) 94±3 108±14 NS 
HbA1c (%) 5.5±0.1 6.1±0.5 NS 
Cholesterol (mmol/l) 5.85±0.34 4.72±0.43 NS 
Triacyglycerols (g/l) 2.30±0.59 1.69±0.33 NS 
HDL-cholesterol (mmol/l) 1.34±0.13 1.26±0.15 NS 
LDL-cholesterol (mmol/l) 3.20±0.36 2.57±0.35 NS 
ALT (IU/l) 19±3 57±2 <0.01 
AST (IU/l) 22±1 43±10 <0.01 
GGT (IU/l) 40±6 217±67 <0.05 
Total bilirubin (μmol/l) 7±1 14±4 NS 
Alkaline phosphatases (IU/l) 99±11 105±24 NS 
Prothrombin ratio (%) 100±0 79±6 <0.05 
INR 1.00±0.00 1.20±0.06 <0.05 

Inhibition of the CXCL12/CXCR4 axis decreases lymphocyte count in the liver of ob/ob mice

According to our ex vivo results on CD4+ T-cells, we therefore investigated in vivo the effect of the CXCL12/CXCR4 pathway inhibition on the liver of ob/ob mice. Ob/ob mice and controls were chronically treated with AMD3100. To this end, AMD3100 or PBS was delivered continuously via an interscapular osmotic pump, for 12 days.

Recruitment of lymphocytes into the liver was assessed quantitatively by flow cytometry. CD4+ T-cell counts were significantly lower in the livers of ob/ob mice treated with AMD3100 compared with mice treated with PBS. Similar findings were observed for CD8+ T-cells and to a lesser extent for B-cells, although this did not reach statistical significance (Figure 3A). In contrast, the number of liver lymphocytes in lean mice was not affected by AMD3100 treatment (Figure 3B). A decreased infiltration of CD3+ T-cells was also observed by immunohistochemistry in the liver of AMD3100-treated ob/ob mice (see Supplementary Figure S1 at http://www.clinsci.org/cs/128/cs1280257add.htm). These findings suggest that infiltration into the liver by CD4+ T-cells, and possibly by CD8+ T-cells and B-cells in ob/ob mice, relies at least in part on the CXCL12/CXCR4 signalling axis. In this experiment, the effect of AMD3100 was restricted to the recruitment of lymphocytes into the liver, as no significant effect on body weight, liver weight, liver/body weight ratio, blood glucose, insulin and transaminases levels (Table 2) and apoptosis of liver cells (see Supplementary Figures 2 at http://www.clinsci.org/cs/128/cs1280257add.htm) was observed.

Reduced liver count of CD4+ T-cells after AMD3100 treatment

Figure 3
Reduced liver count of CD4+ T-cells after AMD3100 treatment

Quantification, by flow cytometry, of CD4+ and CD8+ T-cells, B-cells, natural killer (NK) and natural killer T- (NKT) cells in the liver of ob/ob mice (A) treated with AMD3100 (dark grey boxes) or PBS (black boxes) and lean mice (B) treated with AMD3100 (light grey boxes) or PBS (white boxes). Results are expressed as the total number of liver lymphocytes. Graphs show means±S.E.M. for five to nine mice per group. Statistically significant differences are indicated; ns, not significant (Kruskal–Wallis ANOVA post-hoc Fischer's PLSD test).

Figure 3
Reduced liver count of CD4+ T-cells after AMD3100 treatment

Quantification, by flow cytometry, of CD4+ and CD8+ T-cells, B-cells, natural killer (NK) and natural killer T- (NKT) cells in the liver of ob/ob mice (A) treated with AMD3100 (dark grey boxes) or PBS (black boxes) and lean mice (B) treated with AMD3100 (light grey boxes) or PBS (white boxes). Results are expressed as the total number of liver lymphocytes. Graphs show means±S.E.M. for five to nine mice per group. Statistically significant differences are indicated; ns, not significant (Kruskal–Wallis ANOVA post-hoc Fischer's PLSD test).

Table 2
Characteristics of ob/ob mice and lean controls treated with PBS or AMD3100

Values are means ± S.E.M. of five to nine mice per group. AMD3100 did not affect the characteristics of the mice (Mann–Whitney test).

Control miceob/ob mice
PBSAMD3100PBSAMD3100
Body weight (g) 26.0±0.8 26.4±0.6 50.9±0.9 51.7±1.0 
Liver weight (g) 1.15±0.05 1.19±0.07 3.28±0.16 3.18±0.2 
Liver/body weight ratio (%) 4.3±0.1 4.5±0.2 6.0±0.2 6.0±0.3 
Blood glucose (mg/dl) 169±7.3 200±28.7 257±45.8 204±13.1 
Insulin (ng/ml) 1.19±0.05 0.96±0.24 5.54±1.11 4.34±1.05 
ALT (IU/l) 68±3 55±13 137±32 127±19 
AST (IU/l) 141±20 102.2±10 237±39 266±54 
Control miceob/ob mice
PBSAMD3100PBSAMD3100
Body weight (g) 26.0±0.8 26.4±0.6 50.9±0.9 51.7±1.0 
Liver weight (g) 1.15±0.05 1.19±0.07 3.28±0.16 3.18±0.2 
Liver/body weight ratio (%) 4.3±0.1 4.5±0.2 6.0±0.2 6.0±0.3 
Blood glucose (mg/dl) 169±7.3 200±28.7 257±45.8 204±13.1 
Insulin (ng/ml) 1.19±0.05 0.96±0.24 5.54±1.11 4.34±1.05 
ALT (IU/l) 68±3 55±13 137±32 127±19 
AST (IU/l) 141±20 102.2±10 237±39 266±54 

Absence of CXCR4, CXCR7 and CXCL12 up-regulation in mouse models of NASH

We investigated whether the mechanisms underlying the preferential homing of CD4+ T-cells to the liver of mouse models of NASH involved an increased expression level of CXCL12 or its receptors, namely CXCR4 and CXCR7. Therefore, we isolated CD4+ T-cells from spleens of three different mouse models of NASH and their lean controls and quantified transcripts encoding CXCR4 and CXCR7. There was no difference in CXCR4 mRNA or CXCR7 mRNA expression between splenic CD4+ T-cells (Figures 4A and 4B). We also found that CXCL12 mRNA was similarly expressed in the liver (Figure 4C). Quantification, by flow cytometry, of CXCR4 expression at the surface of splenic CD4+ T-cells showed no difference between groups (Figure 4D). However, this does not exclude the possibility of differential expression specifically in lymphocytes recruited to the liver. We therefore quantified CXCR4 membrane expression specifically in liver CD4+ T-cells, but found no increased expression in any mouse models of NASH studied (Figure 4E).

Quantification of CXCR4, CXCR7 and CXCL12 expression

Figure 4
Quantification of CXCR4, CXCR7 and CXCL12 expression

(A and B) Quantification of CXCR4 (A) and CXCR7 (B) expression by RT-PCR in splenic CD4+ T-cells of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. (C) Quantification of CXCL12 expression by RT-PCR in the livers of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. mRNA level is expressed as the ratio of CXCR4, CXCR7 or CXCL12 mRNA to β-actin mRNA. Graphs show means±S.E.M. for three to seven mice per group. No statistical differences were found (Kruskal–Wallis test). (D and E) Quantification of CXCR4 expression by flow cytometry in splenic CD4+ T-cells (D) and liver CD4+ T-cells (E) of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. Results are expressed as a ratio of geometric MFI normalized to control group. Graphs show means±S.E.M. for three to seven mice per group. No statistical differences were found (Kruskal–Wallis test).

Figure 4
Quantification of CXCR4, CXCR7 and CXCL12 expression

(A and B) Quantification of CXCR4 (A) and CXCR7 (B) expression by RT-PCR in splenic CD4+ T-cells of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. (C) Quantification of CXCL12 expression by RT-PCR in the livers of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. mRNA level is expressed as the ratio of CXCR4, CXCR7 or CXCL12 mRNA to β-actin mRNA. Graphs show means±S.E.M. for three to seven mice per group. No statistical differences were found (Kruskal–Wallis test). (D and E) Quantification of CXCR4 expression by flow cytometry in splenic CD4+ T-cells (D) and liver CD4+ T-cells (E) of control (white boxes), HFD-fed (light grey), MCD-fed (dark grey) and ob/ob (black boxes) mice. Results are expressed as a ratio of geometric MFI normalized to control group. Graphs show means±S.E.M. for three to seven mice per group. No statistical differences were found (Kruskal–Wallis test).

Therefore, there was no up-regulation of the expression level of the trio CXCL12/CXCR4/CXCR7 in the three different models of NASH. Regarding CXCR4 expression level, CD4+ T-cells of mice with NASH have, at least in vitro, no potential to respond more strongly than their controls to the chemotactic effect of CXCL12.

High affinity of CXCL12 for CXCR4 in CD4+ T-cells from ob/ob mice

As the higher chemotactic response to CXCL12 of CD4+ T-cells from obese did not result from a higher expression of CXCL12, CXCR4 or CXCR7 and as CD4+ T-cells remained sensitive to the CXCR4 antagonist AMD3100, we considered a possible change in the affinity of CXCL12 for CXCR4. Therefore, we carried out binding experiments with synthetic biotin-labelled CXCL12 (biot-CXCL12) as previously described in humans [10]. The binding of biot-CXCL12 to CXCR4 in CD4+ T-cells was detected by incubation with SAv–PE conjugate and quantified by flow cytometry. The binding data indicated that the affinity of CXCL12 for CXCR4 was higher in CD4+ T-cells from ob/ob mice (apparent KD ≈ 0.02 nM) than in cells from control mice (apparent KD ≈ 0.12 nM) (Figure 5).

High affinity of CXCL12 for CXCR4 in CD4+ T-cells from obese mice

Figure 5
High affinity of CXCL12 for CXCR4 in CD4+ T-cells from obese mice

Saturation curves for the binding of biot-CXCL12 to CXCR4 in lymphocytes from ob/ob (triangles) and control (circles) mice. The data shown are duplicate determinations of the geometric MFI of SAv–PE bound to biot-CXCL12, from one experiment representative of the two independent experiments carried out.

Figure 5
High affinity of CXCL12 for CXCR4 in CD4+ T-cells from obese mice

Saturation curves for the binding of biot-CXCL12 to CXCR4 in lymphocytes from ob/ob (triangles) and control (circles) mice. The data shown are duplicate determinations of the geometric MFI of SAv–PE bound to biot-CXCL12, from one experiment representative of the two independent experiments carried out.

DISCUSSION

The presence of an inflammatory process in the liver is key in the progression of severe liver lesions in NAFLD. In mice, obesity is associated with higher levels of lymphocyte recruitment to the liver, particularly of CD4+ T-cells [3]. Here, in several different mouse models of NASH, we found that CD4+ T-cells migrated to the liver in a CXCR4/CXCL12-dependent manner. Pharmacological inhibition of the CXCR4 receptor by AMD3100 attenuated the liver accumulation of CD4+ T-cells. We extended our observations to the context of human NAFLD. The chemotactic response to CXCL12 was stronger in CD4+ T-cells from patients with NASH than in CD4+ T-cells from healthy blood donors. No such stronger response to CXCL12 was observed in patients with pure steatosis. In mice, we excluded the possible role of higher levels of CXCR4 expression. We showed that the affinity of CXCL12 to CXCR4 was stronger in CD4+ T-cells from ob/ob mice with a KD value that was 6-fold much higher than in control CD4+ T-cells. Thus, we speculate that this increased affinity is involved in the stronger chemotactic response of the cells and therefore participate to the maintenance of recruitment of inflammatory cells into the liver.

We used three different mouse models of NASH to assess CD4+ T-cells response to CXCL12. ob/ob mice are leptin-deficient, resulting in a phenotype of obesity, hepatomegaly and severe steatosis. They display insulin resistance and a low basal liver inflammation as shown by elevated levels of transaminases and lymphocyte infiltration into the liver [3]. Primary mouse lymphocytes from ob/ob mice may have altered response to inflammatory stimuli due to the impaired leptin signal [20]. Therefore, to ensure that leptin deficiency was not involved in the abnormal CXCL12 response of CD4+ T-cells from ob/ob mice, we used two other diet-induced models of NASH without leptin deficiency: mice fed a HFD and mice fed a MCD. As observed for ob/ob mice, HFD-fed mice develop steatosis, insulin resistance and liver inflammation characterized by lymphocyte infiltration but without leptin deficiency [3]. MCD-fed mice develop fatty liver and lymphocyte infiltration without obesity [21]. Overall, the results obtained in these three different models of mice reflect the phenotype of NASH in patients rather than that of pure steatosis. By using mice fed a HFD and a MCD, we clearly demonstrated that leptin deficiency was not involved in the increased response of CD4+ T-cells to CXCL12.

Several non-mutually exclusive hypotheses may be raised to explain the mechanism involved in the increased affinity of CXCR4 to CXCL12 in ob/ob mice. There may be a change in the conformation of CXCR4. Insulin resistance, which is associated with NASH, alters cholesterol metabolism and membrane-lipid composition [22], decreasing membrane lipid fluidity and deformability [23]. These modifications lead to changes in the physical properties of lipids and, consequently, to changes in the activity of some membrane proteins [24]. Patients with NASH have a specific lipidomic signature. In particular, they have high hepatic free cholesterol concentrations [25]. Membrane cholesterol has been reported to be essential for chemokine binding and for the conformational integrity and function of CXCR4 [26]. Moreover, cholesterol oxidation in T-cells induces changes in the conformation of CXCR4 associated with the inhibition of chemokine-mediated intracellular calcium mobilization and chemotaxis [27], providing evidence that cholesterol plays a critical role in CXCR4 function [26,28]. Cholesterol increase may also impact on plasma membrane domains or rafts, composed of cholesterol and saturated lipids such as sphingolipids [29], shown to facilitate the activation of T-cells [30] and to participate in intracellular signalling and inflammation processes [31]. Thus collectively, NASH-associated changes in CD4+ T-cells membrane lipid composition may contribute to the change in CXCR4 conformation and the increase in CXCL12 affinity.

CXCR7 may regulate cell migration promoted by CXCL12 [10] eventually through its capacity to recruit β-arrestin [32]. CXCR7 may also play an indirect role in CXCR4-dependent CD4+ T-cell migration either through heterodimerization with CXCR4 [33] or by acting as a decoy receptor [34], thereby modifying the concentration of CXCL12 in the liver environment. Nevertheless, in our experiments, abrogation of the CXCL12-induced chemotaxis on CD4+ T-cells by AMD3100, strongly supports a direct role of the CXCR4 receptor.

The CXCL12/CXCR4/CXCR7 axis has been reported to be involved in several liver injuries. For instance, the predominance of CXCR4 expression to CXCR7 in mouse models of chronic liver injury orients liver sinusoidal endothelial cells toward a maladaptive pro-fibrotic response [35]. The CXCL12/CXCR4 axis is also involved in hepatocellular carcinoma (HCC). Polymorphisms of CXCL12 are considered to be relevant factors underlying the occurrence and development of HCC [36]. The levels of CXCL12 and CXCR4 expression in sinusoidal endothelial cells are higher in HCC tissue than in the adjacent liver [37]. However, conflicting results have been reported, as other studies have reported no correlation between CXCL12/CXCR4 expression and HCC [38]. Obesity is associated with an increase in the risk of HCC, particularly in male patients [39]. Our results suggest that HCC in NASH patients may therefore involve the CXCL12/CXCR4 pathway, not necessarily through changes in the expression of CXCL12 and/or CXCR4, but through an increase in the chemotaxis to CXCL12, as shown here for cells from three different mouse models of NASH and patients with NASH. A similar discordance between CXCL12/CXCR4 function and expression level has also been observed in some cancer cell lines: for instance, the metastatic potency of breast cancer-derived cell lines depends on the functional association of CXCR4 with heterotrimeric G-proteins rather than on CXCR4 levels [40]. Of note, in patients with NASH, HCC may occur even in the absence of cirrhosis, suggesting a role of the inflammatory process in liver carcinogenesis [41]. Contradictory data have been obtained about the involvement of CD4+ T-cells in HCC. On the one hand, infiltration of cytotoxic CD4+ T-cells into the HCC has been associated with a good prognosis [42]. On the other hand, infiltration of CD4+ regulatory T-cells in HCC may decrease patient survival [43]. Moreover, many other CD4+ T-cells subsets may be involved: patients with HCC show high levels of circulating CD4+ T-helper 22 cells and of circulating CD4+ T-helper 17 cells [44]. Interestingly, HCC patients with CD4+ regulatory T-cells and CD4+ T-helper 17 cells infiltrating the tumour show lower survival after resection [45]. Therefore, our results suggest that inhibition of CD4+ T-cell migration into the liver may modulate the development of HCC.

Accumulation of lymphocytes, including CD4+ T-cells, into the liver is reversed by the inhibition of CXCL12 binding to CXCR4 using pharmacological inhibitor AMD3100. The decrease of inflammatory cells into the liver was not associated with a decrease of transaminases level or a correction of insulin resistance. This discrepancy between the strong effect of AMD3100 on CD4+ T-cells of mice and patients with NASH and the absence of decrease of transaminases highlights a limited effect of AMD3100 on obesity-associated liver injury. This may be related to the persistence of other associated mechanisms involved in liver inflammation and/or to a too short duration of AMD3100 treatment.

Inhibitors of the CXCL12/CXCR4 pathway are currently being developed for the long-term therapeutic use such as treatment of cancers [46] and lupus nephritis, in which recruitment of leucocytes to the kidney involves the CXCL12/CXCR4 axis [47]. Nevertheless, as CXCL12/CXCR4 has numerous physiological functions, it is difficult to predict whether the long-term administration of a pharmacological inhibitor would improve liver inflammation without adverse effects [4]. Interestingly, a recent Phase 1 clinical trial of long-term low-dose of AMD3100 has provided preliminary evidence for safety [48].

In conclusion, we have identified the CXCL12/CXCR4 axis in CD4+ T-cells as an actor on the recruitment into the liver of CD4+ T-cells in obesity. The increase in the affinity of CXCL12 for CXCR4 is likely to be involved in this process, which may be attenuated by pharmacological targeting.

Abbreviations

     
  • ALT

    alanine aminotransferase

  •  
  • APC

    allophycocyanin

  •  
  • AST

    aspartate aminotransferase

  •  
  • CXCL

    CXC chemokine ligand

  •  
  • CXCR

    CXC chemokine receptor

  •  
  • biot-CXCL12

    biotinylated CXCL12

  •  
  • HCC

    hepatocellular carcinoma

  •  
  • HFD

    high-fat diet

  •  
  • mAb

    monoclonal antibody

  •  
  • MCD

    methionine and choline-deficient diet

  •  
  • MFI

    mean fluorescence intensity

  •  
  • MMLV

    Moloney murine leukaemia virus

  •  
  • NAFLD

    non-alcoholic fatty liver disease

  •  
  • NASH

    non-alcoholic steatohepatitis

  •  
  • PBMC

    peripheral blood mononuclear cell

  •  
  • PE

    phycoerythrin

  •  
  • PerCP

    peridinin–chlorophyll–protein complex

  •  
  • PLSD

    protected least-squares difference

  •  
  • SaV

    streptavidin

  •  
  • SDF-1α

    stromal cell-derived factor 1α

  •  
  • TUNEL

    terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling

AUTHOR CONTRIBUTION

Hedia Boujedidi contributed to the acquisition of data, analysis and interpretation of data and drafting of the manuscript. Olivier Robert helped with the acquisition of data, analysis and interpretation of data and drafting of the manuscript. Alexandre Bignon performed acquisition of data, analysis and interpretation of data and drafting of the manuscript. Anne-Marie Cassard-Doulcier contributed with the drafting and critical revision of the manuscript for important intellectual content. Marie-Laure Renoud, Helene Gary-Gouy, Patrice Hémon, Hugo Tharinger and Sophie Prévot helped with the technical support. Francoise Bachelerie performed drafting and critical revision of the manuscript for important intellectual content. Sylvie Naveau helped with the statistical analysis and critical revision of the manuscript for important intellectual content. Dominique Emilie contributed with study concept and design. Karl Balabanian helped with the study concept and design, drafting and critical revision of the manuscript for important intellectual content. Gabriel Perlemuter performed the study concept and design, study supervision, analysis and interpretation of data, statistical analysis, drafting of the manuscript and also obtained funding.

We thank L. Bouchet-Delbos (Univ. Paris-Sud, Laboratoire “Cytokines, Chimiokines et Immunopathologie”, INSERM U996, Clamart, France) and J. Taieb (Hôpital Antoine Béclère, Service de Biochimie, Clamart, France) for technical assistance. We thank Dr F. Baleux (Unité de Chimie Organique, Institut Pasteur, Paris, France) for providing us with CXCL12 proteins.

FUNDING

This work was supported by the INSERM, Paris-South University (Univ. Paris-Sud); the Agence Nationale de la Recherche (ANR) [grant number 2010 JCJC 1104 01 (to K.B.)]; the European Union FP6 (INNOCHEM) [grant number LSHB-CT-2005-518167] (to K.B. and G.P.); Cardiovasculaire, obésité, rein, diabète, domaines d'intérêt majeur (CORDDIM) (to G.P.); the Société Nationale Française de Gastroentérologie (SNFGE) (to G.P.); the Association Française pour l’Etude du Foie (AFEF) (to G.P.); and the Ministère de l’Enseignement Supérieur et de la Recherche (MESR) (to H.B., O.R. and A.B.). K.B., F.B. and A.B. are members of the Laboratory of Excellence LERMIT supported by a grant from the ANR [grant number ANR-10-LABX-33] under the programme “Investissements d’Avenir” [grant number ANR-11-IDEX-0003-01].

References

References
1
Perlemuter
 
G.
Bigorgne
 
A.
Cassard-Doulcier
 
A. M.
Naveau
 
S.
 
Nonalcoholic fatty liver disease: from pathogenesis to patient care
Nat. Clin. Pract. Endocrinol. Metab.
2007
, vol. 
3
 (pg. 
458
-
469
)
[PubMed]
2
Neuschwander-Tetri
 
B. A.
 
Hepatic lipotoxicity and the pathogenesis of nonalcoholic steatohepatitis: the central role of nontriglyceride fatty acid metabolites
Hepatology
2010
, vol. 
52
 (pg. 
774
-
788
)
[PubMed]
3
Bigorgne
 
A. E.
Bouchet-Delbos
 
L.
Naveau
 
S.
Dagher
 
I.
Prevot
 
S.
Durand-Gasselin
 
I.
Couderc
 
J.
Valet
 
P.
Emilie
 
D.
Perlemuter
 
G.
 
Obesity-induced lymphocyte hyperresponsiveness to chemokines: a new mechanism of Fatty liver inflammation in obese mice
Gastroenterology
2008
, vol. 
134
 (pg. 
1459
-
1469
)
[PubMed]
4
Karin
 
N.
 
The multiple faces of CXCL12 (SDF-1alpha) in the regulation of immunity during health and disease
J. Leukoc. Biol.
2010
, vol. 
88
 (pg. 
463
-
473
)
[PubMed]
5
Uo
 
M.
Hisamatsu
 
T.
Miyoshi
 
J.
Kaito
 
D.
Yoneno
 
K.
Kitazume
 
M. T.
Mori
 
M.
Sugita
 
A.
Koganei
 
K.
Matsuoka
 
K.
Kanai
 
T.
Hibi
 
T.
 
Mucosal CXCR4+ IgG plasma cells contribute to the pathogenesis of human ulcerative colitis through FcgammaR-mediated CD14 macrophage activation
Gut.
2012
, vol. 
62
 (pg. 
1734
-
1744
)
[PubMed]
6
Bachelerie
 
F.
Ben-Baruch
 
A.
Burkhardt
 
A. M.
Combadiere
 
C.
Farber
 
J. M.
Graham
 
G. J.
Horuk
 
R.
Sparre-Ulrich
 
A. H.
Locati
 
M.
Luster
 
A. D.
, et al 
Update on the extended family of chemokine receptors and introducing a new nomenclature for atypical chemokine receptors
Pharmacol. Rev.
2013
, vol. 
66
 (pg. 
1
-
79
)
[PubMed]
7
Feng
 
Y.
Broder
 
C. C.
Kennedy
 
P. E.
Berger
 
E. A.
 
HIV-1 entry cofactor: functional cDNA cloning of a seven-transmembrane, G protein-coupled receptor
Science
1996
, vol. 
272
 (pg. 
872
-
877
)
[PubMed]
8
Triantafilou
 
K.
Triantafilou
 
M.
Dedrick
 
R. L.
 
A CD14-independent LPS receptor cluster
Nat. Immunol.
2001
, vol. 
2
 (pg. 
338
-
345
)
[PubMed]
9
Infantino
 
S.
Moepps
 
B.
Thelen
 
M.
 
Expression and regulation of the orphan receptor RDC1 and its putative ligand in human dendritic and B cells
J. Immunol.
2006
, vol. 
176
 (pg. 
2197
-
2207
)
[PubMed]
10
Balabanian
 
K.
Lagane
 
B.
Infantino
 
S.
Chow
 
K. Y.
Harriague
 
J.
Moepps
 
B.
Arenzana-Seisdedos
 
F.
Thelen
 
M.
Bachelerie
 
F.
 
The chemokine SDF-1/CXCL12 binds to and signals through the orphan receptor RDC1 in T lymphocytes
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
35760
-
35766
)
[PubMed]
11
Schols
 
D.
Este
 
J. A.
Henson
 
G.
De Clercq
 
E.
 
Bicyclams, a class of potent anti-HIV agents, are targeted at the HIV coreceptor fusin/CXCR-4
Antiviral Res.
1997
, vol. 
35
 (pg. 
147
-
156
)
[PubMed]
12
Hendrix
 
C. W.
Collier
 
A. C.
Lederman
 
M. M.
Schols
 
D.
Pollard
 
R. B.
Brown
 
S.
Jackson
 
J. B.
Coombs
 
R. W.
Glesby
 
M. J.
Flexner
 
C. W.
Bridger
 
G. J.
, et al 
Safety, pharmacokinetics, and antiviral activity of AMD3100, a selective CXCR4 receptor inhibitor, in HIV-1 infection
J. Acquir Immune Defic. Syndr.
2004
, vol. 
37
 (pg. 
1253
-
1262
)
[PubMed]
13
Matthys
 
P.
Hatse
 
S.
Vermeire
 
K.
Wuyts
 
A.
Bridger
 
G.
Henson
 
G. W.
De Clercq
 
E.
Billiau
 
A.
Schols
 
D.
 
AMD3100, a potent and specific antagonist of the stromal cell-derived factor-1 chemokine receptor CXCR4, inhibits autoimmune joint inflammation in IFN-gamma receptor-deficient mice
J. Immunol.
2001
, vol. 
167
 (pg. 
4686
-
4692
)
[PubMed]
14
Lukacs
 
N. W.
Berlin
 
A.
Schols
 
D.
Skerlj
 
R. T.
Bridger
 
G. J.
 
AMD3100, a CxCR4 antagonist, attenuates allergic lung inflammation and airway hyperreactivity
Am. J. Pathol.
2002
, vol. 
160
 (pg. 
1353
-
1360
)
[PubMed]
15
Chalasani
 
N.
Younossi
 
Z.
Lavine
 
J. E.
Diehl
 
A. M.
Brunt
 
E. M.
Cusi
 
K.
Charlton
 
M.
Sanyal
 
A. J.
 
The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Gastroenterological Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology
Gastroenterology
2012
, vol. 
142
 (pg. 
1592
-
1609
)
[PubMed]
16
Crump
 
M. P.
Gong
 
J. H.
Loetscher
 
P.
Rajarathnam
 
K.
Amara
 
A.
Arenzana-Seisdedos
 
F.
Virelizier
 
J. L.
Baggiolini
 
M.
Sykes
 
B. D.
Clark-Lewis
 
I.
 
Solution structure and basis for functional activity of stromal cell-derived factor-1; dissociation of CXCR4 activation from binding and inhibition of HIV-1
Embo J.
1997
, vol. 
16
 (pg. 
6996
-
7007
)
[PubMed]
17
Balabanian
 
K.
Brotin
 
E.
Biajoux
 
V.
Bouchet-Delbos
 
L.
Lainey
 
E.
Fenneteau
 
O.
Bonnet
 
D.
Fiette
 
L.
Emilie
 
D.
Bachelerie
 
F.
 
Proper desensitization of CXCR4 is required for lymphocyte development and peripheral compartmentalization in mice
Blood.
2012
, vol. 
119
 (pg. 
5722
-
5730
)
[PubMed]
18
Ekstedt
 
M.
Franzen
 
L. E.
Mathiesen
 
U. L.
Thorelius
 
L.
Holmqvist
 
M.
Bodemar
 
G.
Kechagias
 
S.
 
Long-term follow-up of patients with NAFLD and elevated liver enzymes
Hepatology
2006
, vol. 
44
 (pg. 
865
-
873
)
[PubMed]
19
Dunn
 
W.
Xu
 
R.
Wingard
 
D. L.
Rogers
 
C.
Angulo
 
P.
Younossi
 
Z. M.
Schwimmer
 
J. B.
 
Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study
Am. J. Gastroenterol.
2008
, vol. 
103
 (pg. 
2263
-
2271
)
[PubMed]
20
De Rosa
 
V.
Procaccini
 
C.
Cali
 
G.
Pirozzi
 
G.
Fontana
 
S.
Zappacosta
 
S.
La Cava
 
A.
Matarese
 
G.
 
A key role of leptin in the control of regulatory T cell proliferation
Immunity.
2007
, vol. 
26
 (pg. 
241
-
255
)
[PubMed]
21
Itagaki
 
H.
Shimizu
 
K.
Morikawa
 
S.
Ogawa
 
K.
Ezaki
 
T.
 
Morphological and functional characterization of non-alcoholic fatty liver disease induced by a methionine-choline-deficient diet in C57BL/6 mice
Int. J. Clin. Exp. Pathol.
2013
, vol. 
6
 (pg. 
2683
-
2696
)
[PubMed]
22
Younsi
 
M.
Quilliot
 
D.
Al-Makdissy
 
N.
Delbachian
 
I.
Drouin
 
P.
Donner
 
M.
Ziegler
 
O.
 
Erythrocyte membrane phospholipid composition is related to hyperinsulinemia in obese nondiabetic women: effects of weight loss
Metabolism.
2002
, vol. 
51
 (pg. 
1261
-
1268
)
[PubMed]
23
Sola
 
E.
Vaya
 
A.
Santaolaria
 
M. L.
Hernandez-Mijares
 
A.
Reganon
 
E.
Vila
 
V.
Martinez-Sales
 
V.
Corella
 
D.
 
Erythrocyte deformability in obesity measured by ektacytometric techniques
Clin. Hemorheol. Microcirc.
2007
, vol. 
37
 (pg. 
219
-
227
)
[PubMed]
24
Escriba
 
P. V.
 
Membrane-lipid therapy: a new approach in molecular medicine
Trends Mol. Med.
2006
, vol. 
12
 (pg. 
34
-
43
)
[PubMed]
25
Puri
 
P.
Baillie
 
R. A.
Wiest
 
M. M.
Mirshahi
 
F.
Choudhury
 
J.
Cheung
 
O.
Sargeant
 
C.
Contos
 
M. J.
Sanyal
 
A. J.
 
A lipidomic analysis of nonalcoholic fatty liver disease
Hepatology
2007
, vol. 
46
 (pg. 
1081
-
1090
)
[PubMed]
26
Nguyen
 
D. H.
Taub
 
D.
 
CXCR4 function requires membrane cholesterol: implications for HIV infection
J. Immunol.
2002
, vol. 
168
 (pg. 
4121
-
4126
)
[PubMed]
27
Nguyen
 
D. H.
Taub
 
D. D.
 
Inhibition of chemokine receptor function by membrane cholesterol oxidation
Exp. Cell Res.
2003
, vol. 
291
 (pg. 
36
-
45
)
[PubMed]
28
Ablan
 
S.
Rawat
 
S. S.
Viard
 
M.
Wang
 
J. M.
Puri
 
A.
Blumenthal
 
R.
 
The role of cholesterol and sphingolipids in chemokine receptor function and HIV-1 envelope glycoprotein-mediated fusion
Virol. J.
2006
, vol. 
3
 pg. 
104
 
[PubMed]
29
Brown
 
D. A.
London
 
E.
 
Structure and function of sphingolipid- and cholesterol-rich membrane rafts
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
17221
-
17224
)
[PubMed]
30
Harder
 
T.
 
Lipid raft domains and protein networks in T-cell receptor signal transduction
Curr. Opin. Immunol.
2004
, vol. 
16
 (pg. 
353
-
359
)
[PubMed]
31
Luo
 
C.
Wang
 
K.
Liu de
 
Q.
Li
 
Y.
Zhao
 
Q. S.
 
The functional roles of lipid rafts in T cell activation, immune diseases and HIV infection and prevention
Cell Mol. Immunol.
2008
, vol. 
5
 (pg. 
1
-
7
)
[PubMed]
32
Rajagopal
 
S.
Kim
 
J.
Ahn
 
S.
Craig
 
S.
Lam
 
C. M.
Gerard
 
N. P.
Gerard
 
C.
Lefkowitz
 
R. J.
 
Beta-arrestin- but not G protein-mediated signaling by the “decoy” receptor CXCR7
Proc. Natl. Acad. Sci. U. S. A.
2010
, vol. 
107
 (pg. 
628
-
632
)
[PubMed]
33
Levoye
 
A.
Balabanian
 
K.
Baleux
 
F.
Bachelerie
 
F.
Lagane
 
B.
 
CXCR7 heterodimerizes with CXCR4 and regulates CXCL12-mediated G protein signaling
Blood
2009
, vol. 
113
 (pg. 
6085
-
6093
)
[PubMed]
34
Naumann
 
U.
Cameroni
 
E.
Pruenster
 
M.
Mahabaleshwar
 
H.
Raz
 
E.
Zerwes
 
H. G.
Rot
 
A.
Thelen
 
M.
 
CXCR7 functions as a scavenger for CXCL12 and CXCL11
PLoS One.
2010
, vol. 
5
 pg. 
e9175
 
[PubMed]
35
Ding
 
B. S.
Cao
 
Z.
Lis
 
R.
Nolan
 
D. J.
Guo
 
P.
Simons
 
M.
Penfold
 
M. E.
Shido
 
K.
Rabbany
 
S. Y.
Rafii
 
S.
 
Divergent angiocrine signals from vascular niche balance liver regeneration and fibrosis
Nature
2013
, vol. 
505
 (pg. 
97
-
102
)
[PubMed]
36
Chang
 
C. C.
Chen
 
S. C.
Hsieh
 
Y. H.
Chen
 
Y. C.
Chen
 
T. Y.
Chu
 
Y. H.
Ma
 
HJ
Chou
 
M. C.
Tsai
 
H. T.
Yang
 
S. F.
 
Stromal cell-derived factor-1 but not its receptor, CXCR4, gene variants increase susceptibility and pathological development of hepatocellular carcinoma
Clin. Chem. Lab. Med.
2009
, vol. 
47
 (pg. 
412
-
418
)
[PubMed]
37
Li
 
W.
Gomez
 
E.
Zhang
 
Z.
 
Immunohistochemical expression of stromal cell-derived factor-1 (SDF-1) and CXCR4 ligand receptor system in hepatocellular carcinoma
J. Exp. Clin. Cancer Res.
2007
, vol. 
26
 (pg. 
527
-
533
)
[PubMed]
38
Shibuta
 
K.
Mori
 
M.
Shimoda
 
K.
Inoue
 
H.
Mitra
 
P.
Barnard
 
G. F.
 
Regional expression of CXCL12/CXCR4 in liver and hepatocellular carcinoma and cell-cycle variation during in vitro differentiation
Jpn J. Cancer Res.
2002
, vol. 
93
 (pg. 
789
-
797
)
[PubMed]
39
Paradis
 
V.
Zalinski
 
S.
Chelbi
 
E.
Guedj
 
N.
Degos
 
F.
Vilgrain
 
V.
Bedossa
 
P.
Belghiti
 
J.
 
Hepatocellular carcinomas in patients with metabolic syndrome often develop without significant liver fibrosis: a pathological analysis
Hepatology
2009
, vol. 
49
 (pg. 
851
-
859
)
[PubMed]
40
Holland
 
J. D.
Kochetkova
 
M.
Akekawatchai
 
C.
Dottore
 
M.
Lopez
 
A.
Mc Coll
 
S. R.
 
Differential functional activation of chemokine receptor CXCR4 is mediated by G proteins in breast cancer cells
Cancer Res.
2006
, vol. 
66
 (pg. 
4117
-
4124
)
[PubMed]
41
Ertle
 
J.
Dechene
 
A.
Sowa
 
J. P.
Penndorf
 
V.
Herzer
 
K.
Kaiser
 
G.
Schlaak
 
J. F.
Gerken
 
G.
Syn
 
W. K.
Canbay
 
A.
 
Non-alcoholic fatty liver disease progresses to hepatocellular carcinoma in the absence of apparent cirrhosis
Int. J. Cancer.
2011
, vol. 
128
 (pg. 
2436
-
2443
)
[PubMed]
42
Fu
 
J.
Zhang
 
Z.
Zhou
 
L.
Qi
 
Z.
Xing
 
S.
Lv
 
J.
Shi
 
J.
Fu
 
B.
Liu
 
Z.
Zhang
 
J. Y.
, et al 
Impairment of CD4+ cytotoxic T cells predicts poor survival and high recurrence rates in patients with hepatocellular carcinoma
Hepatology
2013
, vol. 
58
 (pg. 
139
-
149
)
[PubMed]
43
Pedroza-Gonzalez
 
A.
Verhoef
 
C.
Ijzermans
 
J. N.
Peppelenbosch
 
M. P.
Kwekkeboom
 
J.
Verheij
 
J.
Janssen
 
H. L.
Sprengers
 
D.
 
Activated tumor-infiltrating CD4+ regulatory T cells restrain antitumor immunity in patients with primary or metastatic liver cancer
Hepatology
2013
, vol. 
57
 (pg. 
183
-
194
)
[PubMed]
44
Qin
 
S.
Ma
 
S.
Huang
 
X.
Lu
 
D.
Zhou
 
Y.
Jiang
 
H.
 
Th22 cells are associated with hepatocellular carcinoma development and progression
Chin. J. Cancer Res.
2014
, vol. 
26
 (pg. 
135
-
141
)
[PubMed]
45
Huang
 
Y.
Wang
 
F.
Wang
 
Y.
Zhu
 
Z.
Gao
 
Y.
Ma
 
Z.
Xu
 
R.
Du
 
Z.
 
Intrahepatic interleukin-17+ T cells and FoxP3+ regulatory T cells cooperate to promote development and affect the prognosis of hepatocellular carcinoma
J. Gastroenterol. Hepatol.
2014
, vol. 
29
 (pg. 
851
-
859
)
[PubMed]
46
Burger
 
J. A.
Peled
 
A.
 
CXCR4 antagonists: targeting the microenvironment in leukemia and other cancers
Leukemia
2009
, vol. 
23
 (pg. 
43
-
52
)
[PubMed]
47
Kyttaris
 
V. C.
Tsokos
 
G. C.
 
Targeting lymphocyte signaling pathways as a therapeutic approach to systemic lupus erythematosus
Curr. Opin. Rheumatol.
2011
, vol. 
23
 (pg. 
449
-
453
)
[PubMed]
48
Mc Dermott
 
D. H.
Liu
 
Q.
Velez
 
D.
Lopez
 
L.
Anaya-O’Brien
 
S.
Ulrick
 
J.
Kwatemaa
 
N.
Starling
 
J.
Fleisher
 
T. A.
Priel
 
D. A.
, et al 
A phase 1 clinical trial of long-term, low-dose treatment of WHIM syndrome with the CXCR4 antagonist plerixafor
Blood
2014
, vol. 
123
 (pg. 
2308
-
2316
)
[PubMed]

Author notes

1

These authors contributed equally to this work.

2

In memoriam.

Supplementary data