Liver monocytes play a major role in the development of NASH (non-alcoholic steatohepatitis). In inflamed tissues, monocytes can differentiate in both macrophages and dendritic cells. In the present study, we investigated the role of moDCs (monocyte-derived inflammatory dendritic cells) in experimental steatohepatitis induced in C57BL/6 mice by feeding on a MCD (methionine/choline-deficient) diet. The evolution of steatohepatitis was characterized by an increase in hepatic CD45+/CD11b+ myeloid cells displaying the monocyte/macrophage marker F4-80+. In the early phases (4 weeks of treatment), Ly6Chigh/CD11b+/F4-80+ inflammatory macrophages predominated. However, their frequency did not grow further with the disease progression (8 weeks of treatment), when a 4-fold expansion of CD11b+/F4-80+ cells featuring the fractalkine receptor (CX3CR1) was evident. These CX3CR1+ cells were also characterized by the combined expression of inflammatory monocyte (Ly6C, CD11b) and dendritic cell (CD11c, MHCII) markers as well as by a sustained TNFα (tumour necrosis factor α) production, suggesting monocyte differentiation into inflammatory moDCs. The expansion of TNFα-producing CX3CR1+ moDCs was associated with an elevation in hepatic and circulating TNFα level and with the worsening of parenchymal injury. Hydrogen sulfide (H2S) has been shown to interfere with CX3CR1 up-regulation in monocyte-derived cells exposed to pro-inflammatory stimuli. Treating 4-week-MCD-fed mice with the H2S donor NaHS while continuing on the same diet prevented the accumulation of TNFα-producing CX3CR1+ moDCs without interfering with hepatic macrophage functions. Furthermore, NaHS reduced hepatic and circulating TNFα levels and ameliorated transaminase release and parenchymal injury. Altogether, these results show that inflammatory CX3CR1+ moDCs contributed in sustaining inflammation and liver injury during steatohepatitis progression.

CLINICAL PERSPECTIVES

  • In healthy livers, DCs (dendritic cells) represent a small fraction of non-parenchymal cells and have a predominantly to-lerogenic phenotype. Recent evidence indicates that NASH is associated with liver DC expansion and activation. However, the features of DCs involved in NASH progression have not been characterized in detail.

  • We have observed that DC expansion during the evolution of steatohepatitis involves a subset of cells with features of moDCs (monocyte-derived inflammatory dendritic cells) and expressing the fractalkine receptor CX3CR1. MoDCs sustain hepatic inflammation in the advanced phases of steatohepatitis by producing TNFα (tumour necrosis factor α). Interfering with CX3CR1 expression, with the H2S donor NaHS, prevents moDC differentiation and ameliorates parenchymal injury.

  • These results indicate that preventing the differentiation of CX3CR1-expressing moDCs can have application in the the-rapy of NASH.

INTRODUCTION

The development of lobular inflammation and parenchymal injury represents the key feature in the transition from NAFLD (non-alcoholic fatty liver disease) to NASH (non-alcoholic steatohepatitis) and is clinically relevant because inflammatory mechanisms are the driving forces for the disease evolution to fibrosis/cirrhosis [1]. Circulating ‘free’ (non-esterified) fatty acids, oxidative damage, endoplasmic reticulum stress and adipokine unbalances have been proposed to trigger hepatic inflammation by stimulating Kupffer cell activation [2,3]. Indeed, at the onset of NASH, Kupffer cells significantly contribute to the production of pro-inflammatory cyto/chemo-kines, which, in turn, stimulate the liver infil-tration by circulating monocytes [4,5]. These latter rapidly differentiate into M1 polarized macrophages [6] and, by interacting with activated CD4 helper T-lymphocytes and NKT (natural killer T-) cells, drive lobular inflammation [7,8]. Accordingly, Kupffer cell depletion or interference with monocyte recruitment through CCL2/CCR2 signalling prevents he-patic injury and inflammation in experimental models of NASH [4,9,10]. Furthermore, the extent of macrophage M1 response appears to modulate NASH severity among different mice strains [6].

Previous studies have revealed the existence of at least two di-stinct monocyte subsets. Inflammatory monocytes are characte-rized as Ly6Chigh/CCR2+/CX3CR1 in mice and CD14+/CD16 in humans and migrate to tissues in the early phase of the response to injury producing pro-inflammatory mediators [11,12]. A second population defined as Ly6C/CCR2/CX3CR1+ in mice and CD14/CD16+ in humans has less well-characterized functions and appears to be involved in tissue healing [11,13]. Studies using different mice models of chronic liver injury have shown that Ly6Chigh/CD11b-expressing circulating monocytes are the precursors of infiltrating macrophages in injured livers [14,15]. However, the phenotype of the monocyte-derived cells responsible for the evolution of chronic liver diseases, including NASH, are still incompletely characterized [16].

Growing evidence indicates that, under inflammatory conditions, monocytes can also differentiate into a special subset of DCs (dendritic cells), called moDCs (monocyte-derived inflammatory dendritic cells) [17]. These cells co-express both DC and monocyte/macrophage surface markers and show a high production of inflammatory mediators combined with an efficient antigen-presenting activity [17]. As a recent report has associated the development of NASH with an expansion in hepatic DCs [18], in the present study we have investigated the possible contribution of moDCs to the progression of experimental steatohepatitis. To this aim, we used steatohepatitis induced in mice by feeding on a MCD (methionine/choline-deficient) diet that, in spite of differing in its pathogenesis from human NASH, allowed us to follow hepatic chronic inflammation up to the development of overt fibrosis [19].

MATERIAL AND METHODS

Animal experimental protocol

Eight-week-old male C57BL/6 mice were purchased from Harlan-Nossan and were fed for 4 or 8 weeks on either an MCD or a control diet (Laboratorio Dottori Piccioni). In some experiments, 4-week-MCD-fed mice received NaHS (1 mg/kg of body weight) daily for a further 4 weeks while continuing on their deficient diet. The experimental protocols were approved by the Italian Ministry of Health and by the University Commission for Animal Care.

Biochemical analysis

Plasma ALT and liver triacylglycerols were determined by spectrometric kits supplied by Radim and Sigma Diagnostics respec-tively. Circulating TNFα (tumour necrosis factor α) levels were evaluated using commercial ELISA kits from Peprotech.

Histology and immunohistochemistry

Steatosis and lobular inflammation were scored blind using the method of Kleiner et al. [20] in haematoxylin/eosin-stained liver sections. Hepatocyte apoptosis was detected by TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling) as reported in [7]. Collagen deposition was determined by Picro-Sirius Red staining. Liver macrophages and activated HSCs (hepatic stellate cells) were assessed in formalin-fixed sections using anti-mouse F4-80 (eBioscience) and anti-α-SMA (α-smooth muscle actin) polyclonal antibodies (Labvision, Bio-Optica) respectively in combination with a horseradish pe-roxidase polymer kit (Biocare Medical). F4-80- or α-SMA-positive cells were counted in ten different microscopic fields (magnification ×20).

mRNA extraction and real-time PCR

Liver RNA was retrotranscribed with a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Real-time PCR was performed in a Techne TC-312 termalcycler (Tecne) using TaqMan Gene Expression Master Mix and TaqMan Gene Expression probes for mouse TNFα, IL (interleukin)-12p40, CD11b, iNOS (inducible nitric oxide synthase), CX3CL1, CX3CR1, α1-procollagen, TGFβ1 (transforming growth factor β1), α-SMA and β-actin (Applied Biosystems). All samples were run in duplicate and the relative gene expression calculated as 2−ΔCT was expressed as fold increase over control samples.

Intrahepatic mononucleated cell isolation and flow cytometry analysis

Liver mononucleated cells were isolated from the livers of naive and MCD-fed mice and purified on a density gradient (Lympholyte®-M, Cedarlane Laboratories) as described in [21]. Cells were washed with Hanks medium and incubated for 30 min with decomplemented mouse serum to block unspecific immunoglobulin binding. The cells were then stained with fluorochrome-conjugated antibodies for CD45, CD11b, Ly6C, CD11c, MHCII (eBiosciences), F4-80 (Invitrogen) or CX3CR1 (R&D Systems) and analysed with a FACScalibur flow cytometer (Becton Dickinson) following prior gating for CD45 and the absence of cell aggregates. Intracellular staining for TNFα and IL-12 was performed using a specific fluorochrome-conjugated antibody (eBiosciences).

Data analysis and statistical calculations

Statistical analyses were performed using SPSS statistical software using one-way ANOVA test with Tukey's correction for multiple comparisons or Kruskal–Wallis test for non-parametric values. Significance was taken at the 5% level. Normality distribution was preliminarily assessed by the Kolmogorov–Smirnov method.

RESULTS

Feeding mice on an MCD diet for up to 8 weeks resulted in a progressive worsening of steatohepatitis as evaluated by a time-dependent increase in liver triacylglycerol accumulation, transaminase release and hepatic inflammation (Supplementary Figure S1). Although the hepatic mRNA for α1-procollagen was significantly up-regulated already after 4 weeks on the MCD diet, appreciable fibrosis, as determined by collagen staining with Picro-Sirius Red and α-SMA-positive activated HSCs, was evident in more advanced disease after 8 weeks of treatment (Supplementary Figure S1). In parallel with the progression of NASH, flow cytometry showed a time-dependent expansion of intrahepatic mononuclear cells expressing the mice monocyte/macrophage markers F4-80/CD11b+ (Figure 1). Consistently, F4-80-positive cells in liver sections increased from 71.8±23 cell/mm2 in controls up to 146.8±10.2 cells/mm2 (P<0.01) in 8-week-MCD-fed mice (Supplementary Figure S2). Previous studies have found that, in injured livers, inflamma-tory macrophages derive from Ly6C+/CD11b+ blood monocytes [16]. Accordingly, we observed that in the early phases of stea-tohepatitis (4 weeks on the MCD diet), Ly6Chigh/CD11b+/F4-80+ macrophages were prevalent. However, their frequency did not increase further in advanced disease (Figure 1) and this paralleled a decline in the mRNAs of inflammatory M1 activation markers iNOS and IL-12p40 (Supplementary Figure S1). At this stage, an elevation in the liver expression of both fractalkine (CX3CL1) and the fractalkine receptor CX3CR1 became evident (Supplementary Figure S2). CX3CR1 up-regulation largely involved monocyte-derived F4-80+ cells (Supplemen-tary Figure S2) and led to a 4-fold expansion of the pool of F4-80+/CD11b+/CX3CR1high cells (Figure 1). Further characteri-zation of these cells showed that they co-expressed CD11b and Ly6C and were largely Ly6Chigh/int (Figure 2), suggesting their origin from liver-infiltrating Ly6Chigh inflammatory monocytes rather than from the Ly6C/CCR2/CX3CR1+ monocyte subset. Moreover, these cells showed an increased expression of the DC markers CD11c and MHCII (Figure 2).

Changes in the hepatic recruitment of Ly6Chigh and CX3CR1+ monocyte-derived cell subsets during the progression of steatohepatitis

Figure 1
Changes in the hepatic recruitment of Ly6Chigh and CX3CR1+ monocyte-derived cell subsets during the progression of steatohepatitis

Mice were fed on a control (Cont) or an MCD diet over an 8-week period. The changes over time in intrahepatic CD11b+/F4-80+ monocyte/macrophages and the relative prevalence of Ly6Chigh/CD11b+/F4-80+ and CX3CR1+/CD11b+/F4-80+ cell subsets were evaluated by flow cytometry of CD45+ liver mononucleated cells isolated at different time points. The percentages indicate the proportion of cells gated as CD11b+ from four to six animals per group.

Figure 1
Changes in the hepatic recruitment of Ly6Chigh and CX3CR1+ monocyte-derived cell subsets during the progression of steatohepatitis

Mice were fed on a control (Cont) or an MCD diet over an 8-week period. The changes over time in intrahepatic CD11b+/F4-80+ monocyte/macrophages and the relative prevalence of Ly6Chigh/CD11b+/F4-80+ and CX3CR1+/CD11b+/F4-80+ cell subsets were evaluated by flow cytometry of CD45+ liver mononucleated cells isolated at different time points. The percentages indicate the proportion of cells gated as CD11b+ from four to six animals per group.

CX3CR1-positive monocyte-derived cells associated with steatohepatitis show features of monocyte-derived DCs

Figure 2
CX3CR1-positive monocyte-derived cells associated with steatohepatitis show features of monocyte-derived DCs

Mice were fed on a control (Cont) or an MCD diet for 8 weeks and liver CD45+ mononucleated cells were analysed by flow cytometry. (A) F4-80+/CX3CR1+ cells were characterized for the relative distribution of inflammatory monocyte markers CD11b and Ly6C as well as for the expression of DC markers CD11c and MHCII. Grey lines refers to isotypic controls. Results are from one experiment representative of three. (B) Distribution of Ly6C expression among hepatic F4-80+/CX3CR1+ cells. Results are from one experiment representative of three. (C) Liver CD45+ mononucleated cells were evaluated for the relative prevalence of CD11chigh/MHCII+ DCs. The percentages indicate the proportion of cells gated as CD45+, with quantitative evaluation from four animals per group.

Figure 2
CX3CR1-positive monocyte-derived cells associated with steatohepatitis show features of monocyte-derived DCs

Mice were fed on a control (Cont) or an MCD diet for 8 weeks and liver CD45+ mononucleated cells were analysed by flow cytometry. (A) F4-80+/CX3CR1+ cells were characterized for the relative distribution of inflammatory monocyte markers CD11b and Ly6C as well as for the expression of DC markers CD11c and MHCII. Grey lines refers to isotypic controls. Results are from one experiment representative of three. (B) Distribution of Ly6C expression among hepatic F4-80+/CX3CR1+ cells. Results are from one experiment representative of three. (C) Liver CD45+ mononucleated cells were evaluated for the relative prevalence of CD11chigh/MHCII+ DCs. The percentages indicate the proportion of cells gated as CD45+, with quantitative evaluation from four animals per group.

In line with previous observations implicating DCs in NASH [18], we observed that the progression of steatohepatitis was characterized by an appreciable increase in the fraction of he-patic CD11chigh/MHCII+ DCs (Figure 2). Such an expansion involved a pool of cells that were F4-80high and Ly6Chigh/int and expressed CX3CR1 (Figure 3), Interestingly, CX3CR1high cells were particularly evident within the CD11chigh/F4-80+ DC pool (Figure 3). On the other hand, CD11c+/MHCII+/B220+ plasmocytoid DCs and CD11c+/MHCII+/CD8a+ lymphocytoid DCs were significantly decreased (Supplementary Figure S3). Altogether, these data suggested that DC expansion occurring during the progression of steatohepatitis involved CX3CR1high moDCs [17].

DC expansion during the evolution of steatohepatitis involved a pool of CX3CR1-positive cells with features of monocyte-derived DCs

Figure 3
DC expansion during the evolution of steatohepatitis involved a pool of CX3CR1-positive cells with features of monocyte-derived DCs

Mice were fed on a control or an MCD diet for 8 weeks and liver CD45+ mononucleated cells were analysed by flow cytometry. (A) CD11chigh/MHCII+ hepatic DCs were characterized for the expression of inflammatory monocyte markers F4-80 and Ly6C and for that of CX3CR1. The percentages indicate the proportion of cells gated as CD11chigh/MHCII+, with quantitative evaluation from four animals per group. (B) Distribution of CX3CR1high-expressing cells among CD11chigh/F4-80+ and CD11clow/F4-80+ cells obtained from the livers of 8-week-MCD-fed mice. The percentages indicate the proportion of cells gated. One experiment representative of three.

Figure 3
DC expansion during the evolution of steatohepatitis involved a pool of CX3CR1-positive cells with features of monocyte-derived DCs

Mice were fed on a control or an MCD diet for 8 weeks and liver CD45+ mononucleated cells were analysed by flow cytometry. (A) CD11chigh/MHCII+ hepatic DCs were characterized for the expression of inflammatory monocyte markers F4-80 and Ly6C and for that of CX3CR1. The percentages indicate the proportion of cells gated as CD11chigh/MHCII+, with quantitative evaluation from four animals per group. (B) Distribution of CX3CR1high-expressing cells among CD11chigh/F4-80+ and CD11clow/F4-80+ cells obtained from the livers of 8-week-MCD-fed mice. The percentages indicate the proportion of cells gated. One experiment representative of three.

M1 activation of hepatic macrophages has been shown to be an important factor in driving hepatic inflammation in NASH through the production of TNFα and other pro-inflammatory mediators [4,6]. However, as mentioned above, advanced steatohe-patitis was characterized by a lowering of M1 activation markers as compared with early NASH (Supplementary Figure S1). The intracellular TNFα content of Ly6Chigh/CD11b+/F4-80+ macrophages also peaked in early NASH and subsequently decreased in more advanced disease (Supplementary Figure S4). Yet, a steady elevation in both the hepatic mRNA and serum levels of TNFα was evident during NASH progression (Figure 4), and the individual levels of circulating TNFα positively correlated with transaminase release (r=0.82; P=0.035). An elevated production of inflammatory mediators, including TNFα, is a feature of moDCs [17]. In line with this, we observed that CD11chigh/F4-80+ moDCs had an enhanced expression of TNFα which specifi-cally involved CX3CR1+ cells (Figure 4). Furthermore, the prevalence of CX3CR1+/TNFα+ cells increased in the livers of 8-week-MCD-fed mice (Figure 4), suggesting that CX3CR1+ moDCs might sustain hepatic TNFα production during the progression of steatohepatitis.

The progression of steatohepatitis is characterized by the increase in CX3CR1-positive monocyte-derived DCs producing TNFα

Figure 4
The progression of steatohepatitis is characterized by the increase in CX3CR1-positive monocyte-derived DCs producing TNFα

Mice were fed on a control (Cont/Con) or an MCD diet for up to 8 weeks. (A and B) Hepatic mRNA expression and circulating levels of TNFα were evaluated in control and MCD-fed mice. Boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. Results are from six to eight animals per group. (C) TNFα expression by CD11chigh/F4-80+ and CD11clow/F4-80+ cells was evaluated by flow cytometry along with (D) the differential TNFα expression by F4-80+ positive or negative for CX3CR1 and (E) the proportion of TNFα-producing CX3CR1+ cells in control and NASH livers. The percentages indicate the proportion of cells gated in the areas indicated by the arrows. Grey lines refers to isotypic controls. Quantitative data were from four to five animals per group.

Figure 4
The progression of steatohepatitis is characterized by the increase in CX3CR1-positive monocyte-derived DCs producing TNFα

Mice were fed on a control (Cont/Con) or an MCD diet for up to 8 weeks. (A and B) Hepatic mRNA expression and circulating levels of TNFα were evaluated in control and MCD-fed mice. Boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. Results are from six to eight animals per group. (C) TNFα expression by CD11chigh/F4-80+ and CD11clow/F4-80+ cells was evaluated by flow cytometry along with (D) the differential TNFα expression by F4-80+ positive or negative for CX3CR1 and (E) the proportion of TNFα-producing CX3CR1+ cells in control and NASH livers. The percentages indicate the proportion of cells gated in the areas indicated by the arrows. Grey lines refers to isotypic controls. Quantitative data were from four to five animals per group.

Previous studies have shown that genetic and pharmacological interference with CX3CR1 ameliorates the evolution of atherosclerotic plaques [2224]. In this context, hydrogen sulfide (H2S) has been reported to improve atherosclerosis by preventing the up-regulation of CX3CL1/CX3CR1 in monocyte/macrophages exposed to pro-inflammatory stimuli [24]. As CX3CR1-expressing DCs are already present in healthy livers (Figure 3), in subsequent experiments, we sought to investi-gate whether treatment of mice with the H2S donor NaHS might selectively influence the development of CX3CR1+ moDCs in MCD-induced steatohepatitis. Preliminary analysis showed that chronic administration of NaHS (1 mg/kg of body weight) did not influence transaminase release and hepatic inflammation markers in control mice (results not shown). In subsequent experiments, mice fed for 4 weeks on the MCD diet received daily NaHS while continuing on the diet up to the eighth week. In these ani-mals, we observed that NaHS ameliorated CX3CL1 and CX3CR1 mRNA up-regulation (Figure 5), without interfering with that of CCL2, CCR2 or CD11b (Supplementary Figure S5). NaHS supplementation did not modify the hepatic pools of inflamma-tory macrophages and of DCs (Supplementary Figure S5), but halved CX3CR1 expression in F4-80+ or CD11chigh cells (Fi-gure 5). In particular, NaHS treatment selectively reduced the fraction of CX3CR1high/F4-80+/CD11chigh moDCs (Figure 5). Furthermore, NaHS decreased intracellular TNFα levels as well as the fraction of TNFα-producing cells (Figure 6). In line with this, hepatic TNFα mRNA and circulating TNFα levels were lowered in NaHS-supplemented mice (Figure 6). NaHS treatment did not appreciably influence the histopathological scores of steatosis (2.3±0.8 compared with 1.8±0.4; P=0.1) and lo-bular inflammation (1.7±0.6 compared with 1.6±0.5; P=0.8). However, it significantly reduced the number of necrotic foci and apoptotic cells (Figure 6) and also prevented further ele-vation of transaminase release in the animals maintained on the MCD diet (Figure 6), indicating that the sustained production of TNFα by CX3CR1+ moDCs contributed to hepatocellular injury in advanced NASH. Although TNFα-producing DCs have also been implicated in promoting hepatic fibrosis [25], in our hands, treatment of mice with NaHS did not appreciably affected α1-procollagen, α-SMA and TGFβ1 mRNAs as well as collagen staining with Picro-Sirius Red (Supplementary Figure S6).

Treatment of mice with the H2S donor NaHS reduces hepatic CX3CL1 expression and CX3CR1-positive monocyte-derived DCs associated with the progression of steatohepatitis

Figure 5
Treatment of mice with the H2S donor NaHS reduces hepatic CX3CL1 expression and CX3CR1-positive monocyte-derived DCs associated with the progression of steatohepatitis

Mice were fed on an MCD diet for 8 weeks. NaHS (1 mg/kg of body weight) was administered to MCD-fed mice starting from the fourth week of treatment. (A) The hepatic expression of CX3CL1 and CX3CR1 was evaluated by reverse transcription–PCR in mRNA extracted from control or MCD-fed mice with or without NaHS supplementation. Results are expressed as fold increases over control (Cont) values after normalization to the β-actin gene and are from six to nine animals per group; boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. (B) The effect of NaHS supplementation on CX3CR1 expression by F4-80+ cells and CD11chigh cells was evaluated by flow cytometry. Isotypic controls are shown as broken lines. (C) Changes in the distribution of CX3CR1high moDCs following NaHS supplementation of MCD-fed mice. The percentages indicate the proportion of cells gated as F4-80+/CD11chigh. Results are from three or four animals per group.

Figure 5
Treatment of mice with the H2S donor NaHS reduces hepatic CX3CL1 expression and CX3CR1-positive monocyte-derived DCs associated with the progression of steatohepatitis

Mice were fed on an MCD diet for 8 weeks. NaHS (1 mg/kg of body weight) was administered to MCD-fed mice starting from the fourth week of treatment. (A) The hepatic expression of CX3CL1 and CX3CR1 was evaluated by reverse transcription–PCR in mRNA extracted from control or MCD-fed mice with or without NaHS supplementation. Results are expressed as fold increases over control (Cont) values after normalization to the β-actin gene and are from six to nine animals per group; boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. (B) The effect of NaHS supplementation on CX3CR1 expression by F4-80+ cells and CD11chigh cells was evaluated by flow cytometry. Isotypic controls are shown as broken lines. (C) Changes in the distribution of CX3CR1high moDCs following NaHS supplementation of MCD-fed mice. The percentages indicate the proportion of cells gated as F4-80+/CD11chigh. Results are from three or four animals per group.

Treatment of mice with the H2S donor NaHS reduces hepatic TNFα production and improves hepatic injury during the progression of steatohepatitis

Figure 6
Treatment of mice with the H2S donor NaHS reduces hepatic TNFα production and improves hepatic injury during the progression of steatohepatitis

Mice were fed on an MCD diet for 8 weeks. NaHS (1 mg/kg of body weight) was administered to MCD-fed mice starting from the fourth week of treatment. (A) TNFα expression and the relative prevalence of liver F4-80+/TNFα+ cells was evaluated by flow cytometry. Results are from three or four animals per group. Isotypic controls are shown as broken lines. (B) Hepatic TNFα mRNA and circulating TNFα levels were evaluated in MCD-fed mice with or without NaHS supplementation. Liver TNFα mRNA levels were measured by reverse transcription–PCR and expressed as fold increase over control values after normalization to the β-actin gene. Circulating TNFα levels were determined by ELISA in the same animals. Results are from six to nine animals per group; boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. (C and D) Liver histology was evaluated in haematoxylin/eosin-stained sections from control or MCD-fed animals (magnification ×200). Necro-inflammatory foci and apoptotic cells were counted as described in [17]. (E) Liver damage was assessed by circulating alanine aminotransferase (ALT) release.

Figure 6
Treatment of mice with the H2S donor NaHS reduces hepatic TNFα production and improves hepatic injury during the progression of steatohepatitis

Mice were fed on an MCD diet for 8 weeks. NaHS (1 mg/kg of body weight) was administered to MCD-fed mice starting from the fourth week of treatment. (A) TNFα expression and the relative prevalence of liver F4-80+/TNFα+ cells was evaluated by flow cytometry. Results are from three or four animals per group. Isotypic controls are shown as broken lines. (B) Hepatic TNFα mRNA and circulating TNFα levels were evaluated in MCD-fed mice with or without NaHS supplementation. Liver TNFα mRNA levels were measured by reverse transcription–PCR and expressed as fold increase over control values after normalization to the β-actin gene. Circulating TNFα levels were determined by ELISA in the same animals. Results are from six to nine animals per group; boxes include the values within the 25th and 75th percentiles, whereas the horizontal bars represent the medians. The extremities of the vertical bars (10th–90th percentile) comprise 80% of the values. (C and D) Liver histology was evaluated in haematoxylin/eosin-stained sections from control or MCD-fed animals (magnification ×200). Necro-inflammatory foci and apoptotic cells were counted as described in [17]. (E) Liver damage was assessed by circulating alanine aminotransferase (ALT) release.

DISCUSSION

DCs are a heterogeneous population of specialized bone-marrow-derived cells mainly involved in antigen presentation to lymphocytes [26]. In healthy livers, DCs represent a small fraction of non-parenchymal cells and have a predominant to-lerogenic phenotype [27], but a dramatic DC expansion occurs in chronic liver disease in combination with a stimulation of their antigen-presenting activity and the release of pro-inflammatory cytokines [28]. Although most DCs derive from common bone marrow precursors, growing evidence indicates that, under inflammatory conditions, infiltrating monocytes can differentiate into moDCs that co-express both DC and monocyte/macrophage surface markers and produce large amounts of inflammatory mediators [17,29].

Experiments in a mouse model of NASH have found that he-patic DCs expand and mature in the early phases of the disease and acquire the capacity to specifically stimulate CD4+ T-cells [18]. Such an activation is likely to be related to the stimulation of both humoral and cellular immune responses that also characterizes the evolution of NASH [7]. However, the specific features of NASH-associated DCs have not been investigated in detail. The results of the present study support the involvement of DCs in NASH by showing that, with the progression of stea-tohepatitis, DC expansion involves a subset of cells featuring monocyte markers (F4-80 and Ly6C) along with CX3CR1 expression and TNFα production. On this basis, we propose that, during the evo-lution of NASH, a subset of Ly6Chigh monocytes might acquire CX3CR1 and differentiate to TNFα-producing moDCs. Suppor-ting this view, Barlic et al. [30] have observed that, during the development of atherosclerosis, oxidized lipids stimulate human monocytes to switch from CCR2 to CX3CR1 expression. Furthermore, studies have shown that Ly6Chigh monocytes differentiate to Ly6C+/CD11b+/CX3CR1+ moDCs in the intestinal lamina propria. This process is greatly enhanced during gut inflammation and CX3CR1+ moDCs exacerbate colitis by secreting TNFα [31,32].

H2S is increasingly recognized as an endogenous mediator exerting anti-inflammatory and cytoprotective activity in se-veral tissue including the gastrointestinal tract [33,34]. Zhang et al. [24] have shown that, in either RAW 246.7 cells or mice peritoneal macrophages, H2S selectively antagonizes CX3CR1 expression induced by LPS (lipopolysaccharide) or interferon-γ by signalling through the transcription factor NF-κB (nuclear factor κB) and PPARγ (peroxisome-proliferator-activated receptor γ). They also demonstrate that, by interfering with the CX3CL1/CX3CR1 dyad, supplementation of mice with the H2S donor NaHS reduces the development of atherosclerotic plaques [24]. Mice fed on a high-fat diet show a reduced hepatic H2S production [35], whereas H2S supplementation ameliorates oxidative stress and hepatic inflammation in mice with MCD-induced NASH [36]. In our hands, NaHS does not have a generalized anti-inflammatory action, but specifically interferes with the up-regulation of CX3CL1/CX3CR1 dyad associated with the progression of steatohepatitis. Furthermore, NaHS selectively blocks the development of TNFα-producing CX3CR1high moDCs, indi-cating that CX3CL1/CX3CR1 signalling might have an important role in the differentiation of Ly6chigh inflammatory monocytes to moDCs. NaHS treatment also prevents further elevation of transaminase release in the animals maintained on the MCD diet, indicating that CX3CR1+ moDCs can contribute to stea-tohepatitis by sustaining hepatic TNFα production. This is in line with the observation that TNFα-producing DCs sustain hepatic inflammation in mice treated with the hepatotoxic agent thioacetamide [25].

Two previous papers have reported that CX3CR1 genetic deficiency exacerbates hepatic injury and fibrosis induced by chronic CCl4 treatment and bile duct ligation [37,38]. In particular, Karlmark et al. [38] have shown that CX3CR1 is required for the survival and the differentiation of liver-infiltrating macrophages and can limit their M1 polarization. In the same vein, unspecific he-patic DC destruction exacerbates acetaminophen hepatotoxicity [39] and unexpectedly worsens the evolution of experimental NASH [18]. At present, there is no explanation for the discrepancies between these findings and the results of the present study. We have observed that in the livers of control animals, ∼30% of DCs constitutively express CX3CR1. Liver DCs are known to have a predominant immune-suppressive activity [27] and CX3CR1+ DCs also share these properties [40]. Thus it is possible that genetic CX3CR1 deficiency or hepatic DC ablation might enhance damage-associated inflammation by affecting a population of tolerogenic DCs and that this latter effect might overcome the protection given by preventing moDC differentiation.

We are well aware that steatohepatitis induced by the MCD diet does not reproduce important features of the human NASH such as obesity and insulin resistance [19]. However, the difficulties of differentiating the role of moDCs in both clinical and experimental settings justify the use of this model as, in contrast with other experimental protocols, it causes extensive stea-tohepatitis. Nonetheless, further studies are required to better define the relative contribution of constitutive compared with monocyte-derived CX3CR1-expressing DCs during the evolution of chronic liver diseases.

In conclusion, the results of the present study indicate that the evolution of steatohepatitis involves the emergence of CX3CR1-expressing moDCs that can sustain hepatic inflammation in advanced disease through TNFα production. Our results also show that interference with CX3CR1 up-regulation prevents the diffe-rentiation of moDCs, pointing to CX3CR1 as a possible target for the therapy of NASH.

AUTHOR CONTRIBUTION

Salvatore Sutti and Irene Locatelli designed the study and performed the experiments. Aastha Jindal, Stefania Bruzzi and Marco Vacchiano performed the experiments. Cristina Bozzola evaluated the pathology. Emanuele Albano contributed to the study design, and supervised the study and the preparation of the paper.

FUNDING

This work was supported by the Fondazione Cariplo, Milan, Italy [grant number 2011-0470].

Abbreviations

     
  • DC

    dendritic cell

  •  
  • HSC

    hepatic stellate cell

  •  
  • IL

    interleukin

  •  
  • iNOS

    inducible nitric oxide synthase

  •  
  • MCD

    methionine/choline-deficient

  •  
  • moDC

    monocyte-derived inflammatory dendritic cell

  •  
  • NAFLD

    non-alcoholic fatty liver disease

  •  
  • NASH

    non-alcoholic steatohepatitis

  •  
  • α-SMA

    α-smooth muscle actin

  •  
  • TGFβ1

    transforming growth factor β1

  •  
  • TNFα

    tumour necrosis factor α

References

References
1
Lazo
 
M.
Clark
 
J.M.
 
The epidemiology of nonalcoholic fatty liver disease: a global perspective
Semin. Liver Dis.
2008
, vol. 
28
 (pg. 
339
-
350
)
[PubMed]
2
Marra
 
F.
Gastaldelli
 
A.
Svegliati Baroni
 
G.
Tell
 
G.
Tiribelli
 
C.
 
Molecular basis and mechanisms of progression of non-alcoholic steatohepatitis
Trends Mol. Med.
2008
, vol. 
14
 (pg. 
72
-
81
)
[PubMed]
3
Tilg
 
H.
Moschen
 
A.R.
 
Evolution of inflammation in non-alcoholic fatty liver disease: the multiple parallel hits hypothesis
Hepatology
2010
, vol. 
52
 (pg. 
1836
-
1846
)
[PubMed]
4
Tosello-Trampont
 
A.C.
Landes
 
S.G.
Nguyen
 
V.
Novobrantseva
 
T.I.
Hahn
 
Y.S.
 
Kupffer cells trigger nonalcoholic steatohepatitis development in diet-induced mouse model through tumor necrosis factor-α production
J. Biol. Chem.
2012
, vol. 
287
 (pg. 
40161
-
40172
)
[PubMed]
5
Leroux
 
A.
Ferrere
 
G.
Godie
 
V.
Cailleux
 
F.
Renoud
 
M.L.
Gaudin
 
F.
Naveau
 
S.
Prévot
 
S.
Makhzami
 
S.
Perlemuter
 
G.
Cassard-Doulcier
 
A.M.
 
Toxic lipids stored by Kupffer cells correlates with their pro-inflammatory phenotype at an early stage of steatohepatitis
J. Hepatol.
2012
, vol. 
57
 (pg. 
141
-
149
)
[PubMed]
6
Maina
 
V.
Sutti
 
S.
Locatelli
 
I.
Vidali
 
M.
Mombello
 
C.
Bozzola
 
C.
Albano
 
E.
 
Bias in macrophage activation pattern influences non-alcoholic steatohepatitis (NASH) in mice
Clin. Sci.
2012
, vol. 
122
 (pg. 
545
-
553
)
[PubMed]
7
Sutti
 
S.
Jindal
 
A.
Locatelli
 
I.
Vacchiano
 
M.
Gigliotti
 
L.
Bozzola
 
C.
Albano
 
E.
 
Adaptive immune responses triggered by oxidative stress contribute to hepatic inflammation in NASH
Hepatology
2014
, vol. 
59
 (pg. 
886
-
897
)
[PubMed]
8
Wehr
 
A.
Baeck
 
C.
Heymann
 
F.
Niemietz
 
P.M.
Hammerich
 
L.
Martin
 
C.
Zimmermann
 
H.W.
Pack
 
O.
Gassler
 
N.
Hittatiya
 
K.
, et al 
Chemokine receptor CXCR6-dependent hepatic NK T cell accumulation promotes inflammation and liver fibrosis
J. Immunol.
2013
, vol. 
190
 (pg. 
5226
-
5236
)
[PubMed]
9
Baeck
 
C.
Wehr
 
A.
Karlmark
 
K.R.
Heymann
 
F.
Vucur
 
M.
Gassler
 
N.
Huss
 
S.
Klussmann
 
S.
Eulberg
 
D.
Luedde
 
T.
, et al 
Pharmacological inhibition of the chemokine CCL2 (MCP-1) diminishes liver macrophage infiltration and steatohepatitis in chronic hepatic injury
Gut
2012
, vol. 
61
 (pg. 
416
-
426
)
[PubMed]
10
Miura
 
K.
Yang
 
L.
van Rooijen
 
N.
Ohnishi
 
H.
Seki
 
E.
 
Hepatic recruitment of macrophages promotes nonalcoholic steatohepatitis through CCR2
Am. J. Physiol. Gastrointest. Liver Physiol.
2012
, vol. 
302
 (pg. 
G1310
-
G1321
)
[PubMed]
11
Ziegler-Heitbrock
 
L.
 
Monocyte subsets in man and other species
Cell. Immunol.
2014
, vol. 
289
 (pg. 
135
-
139
)
[PubMed]
12
Murray
 
P.J.
Wynn
 
T.A.
 
Protective and pathogenetic functions of macrophage subsets
Nat. Rev. Immunol.
2011
, vol. 
11
 (pg. 
723
-
737
)
[PubMed]
13
Karlmark
 
K.R.
Weiskirchen
 
R.
Zimmermann
 
H.W.
Gassler
 
N.
Ginhoux
 
F.
Weber
 
C.
Merad
 
M.
Luedde
 
T.
Trautwein
 
C.
Tacke
 
F.
 
Hepatic recruitment of the inflammatory Gr1+ monocyte subset upon liver injury promotes hepatic fibrosis
Hepatology
2009
, vol. 
50
 (pg. 
261
-
274
)
[PubMed]
14
Baeck
 
C.
Wei
 
X.
Bartneck
 
M.
Fech
 
V.
Heymann
 
F.
Gassler
 
N.
Hittatiya
 
K.
Eulberg
 
D.
Luedde
 
T.
Trautwein
 
C.
Tacke
 
F.
 
Pharmacological inhibition of the chemokine C-C motif chemokine ligand 2 (monocyte chemoattractant protein 1) accelerates liver fibrosis regression by suppressing Ly-6C+ macrophage infiltration in mice
Hepatology
2014
, vol. 
59
 (pg. 
1060
-
1072
)
[PubMed]
15
Ehling
 
J.
Bartneck
 
M.
Wei
 
X.
Gremse
 
F.
Fech
 
V.
Möckel
 
D.
Baeck
 
C.
Hittatiya
 
K.
Eulberg
 
D.
Luedde
 
T.
, et al 
CCL2-dependent infiltrating macrophages promote angiogenesis in progressive liver fibrosis
Gut
2014
, vol. 
63
 (pg. 
1960
-
1971
)
[PubMed]
16
Tacke
 
F.
Zimmermann
 
H.W.
 
Macrophage heterogeneity in liver injury and fibrosis
J. Hepatol.
2014
, vol. 
60
 (pg. 
1090
-
1096
)
[PubMed]
17
Domininguez
 
P.M.
Ardavin
 
C.
 
Differentiation and function of mouse monocyte-derived dendritic cells in steady state and inflammation
Immunol. Rev.
2010
, vol. 
234
 (pg. 
90
-
104
)
[PubMed]
18
Henning
 
J.R.
Graffeo
 
C.S.
Rehman
 
A.
Fallon
 
N.C.
Zambirinis
 
C.P.
Ochi
 
A.
Barilla
 
R.
Jamal
 
M.
Deutsch
 
M.
Greco
 
S.
, et al 
Dendritic cells limit fibroinflammatory injury in nonalcoholic steatohepatitis in mice
Hepatol.
2013
, vol. 
58
 (pg. 
589
-
602
)
19
Larter
 
C.Z.
Yeh
 
M.M.
 
Animal model of NASH: getting both pathology and metabolic contest right
J. Gastroenterol. Hepatol.
2008
, vol. 
23
 (pg. 
1635
-
1648
)
[PubMed]
20
Kleiner
 
D.E.
Brunt
 
E.M.
Van Natta
 
M.
Behling
 
C.
Contos
 
M.J.
Cummings
 
O.W.
Ferrell
 
L.D.
Liu
 
Y.C.
Torbenson
 
M.S.
, et al 
Design and validation of a histological scoring system for nonalcoholic fatty liver disease
Hepatology
2005
, vol. 
41
 (pg. 
1313
-
1321
)
[PubMed]
21
Crispe
 
I.N.
 
Isolation of mouse intrahepatic lymphocytes
Curr. Protoc. Immunol.
1997
, vol. 
Chapter 3
 (pg. 
3.21.1
-
321.8
)
22
Lesnik
 
P.I.
Haskell
 
C.A.
Charo
 
I.F.
 
Decreased atherosclerosis in CX3CR1−/− mice reveals a role for fractalkine in atherogenesis
J. Clin. Invest.
2003
, vol. 
111
 (pg. 
333
-
340
)
[PubMed]
23
Poupel
 
L.
Boissonnas
 
A.
Hermand
 
P.
Dorgham
 
K.
Guyon
 
E.
Auvynet
 
C.
Charles
 
F.S.
Lesnik
 
P.
Deterre
 
P.
Combadiere
 
C.
 
Pharmacological inhibition of the chemokine receptor, CX3CR1, reduces atherosclerosis in mice
Arterioscler. Thromb. Vasc. Biol.
2013
, vol. 
33
 (pg. 
2297
-
2305
)
[PubMed]
24
Zhang
 
H.
Guo
 
C.
Wu
 
D.
Zhang
 
A.
Gu
 
T.
Wang
 
L.
Wang
 
C.
 
Hydrogen sulfide inhibits the development of atherosclerosis with suppressing CX3CR1 and CX3CL1 expression
PLoS One
2012
, vol. 
7
 pg. 
e41147
 
[PubMed]
25
Connolly
 
M.K.
Bedrosian
 
A.S.
Mallen-St Clair
 
J.
Mitchell
 
A.P.
Ibrahim
 
J.
Stroud
 
A.
Pachter
 
H.L.
Bar-Sagi
 
D.
Frey
 
A.B.
Miller
 
G.
 
In liver fibrosis, dendritic cells govern hepatic inflammation in mice via TNF-α
J. Clin. Invest.
2009
, vol. 
119
 (pg. 
3213
-
3225
)
[PubMed]
26
Hashimoto
 
D.
Miller
 
J.
Merad
 
M.
 
Dendritic cell and macrophage heterogeneity in vivo
Immunity
2011
, vol. 
35
 (pg. 
323
-
335
)
[PubMed]
27
Crispe
 
I.N.
 
Immune tolerance in liver disease
Hepatology
2014
, vol. 
60
 (pg. 
2109
-
2117
)
[PubMed]
28
Rahman
 
A.H.
Aloman
 
C.
 
Dendritic cells and liver fibrosis
Biochim. Biophys. Acta
2013
, vol. 
1832
 (pg. 
998
-
1004
)
[PubMed]
29
Mildner
 
A.
Yona
 
S.
Jung
 
S.
 
A close counter of the third kind: monocyte-derived cells
Adv. Immunol.
2013
, vol. 
120
 (pg. 
69
-
103
)
[PubMed]
30
Barlic
 
J.
Zhang
 
Y.
Foley
 
J.F.
Murphy
 
P.M.
 
Oxidized lipid-driven chemokine receptor switch, CCR2 to CX3CR1, mediates adhesion of human macrophages to coronary artery smooth muscle cells through a peroxisome proliferator-activated receptor γ-dependent pathway
Circulation
2006
, vol. 
114
 (pg. 
807
-
819
)
[PubMed]
31
Varol
 
C.
Vallon-Eberhard
 
A.
Elinav
 
E.
Aychek
 
T.
Shapira
 
Y.
Luche
 
H.
Fehling
 
H.J.
Hardt
 
W.D.
Shakhar
 
G.
Jung
 
S.
 
Intestinal lamina propria dendritic cell subsets have different origin and functions
Immunity
2009
, vol. 
31
 (pg. 
502
-
512
)
[PubMed]
32
Rivollier
 
A.
He
 
J.
Kole
 
A.
Valatas
 
V.
Kelsall
 
B.L.
 
Inflammation switches the differentiation program of Ly6Chi monocytes from antiinflammatory macrophages to inflammatory dendritic cells in the colon
J. Exp. Med.
2012
, vol. 
209
 (pg. 
139
-
155
)
[PubMed]
33
Wallace
 
J.L.
Ferraz
 
J.G.
Muscara
 
M.N.
 
Hydrogen sulfide: an endogenous mediator of resolution of inflammation and injury
Antioxid. Redox Signal.
2012
, vol. 
17
 (pg. 
58
-
67
)
[PubMed]
34
Chan
 
M.V.
Wallace
 
J.L.
 
Hydrogen sulfide-based therapeutics and gastrointestinal diseases: translating physiology to treatments
Am. J. Physiol. Gastrointest. Liver Physiol.
2013
, vol. 
305
 (pg. 
G467
-
G473
)
[PubMed]
35
Peh
 
M.T.
Anwar
 
A.B.
Ng
 
D.S.
Atan
 
M.S.
Kumar
 
S.D.
Moore
 
P.K.
 
Effect of feeding a high fat diet on hydrogen sulfide (H2S) metabolism in the mouse
Nitric Oxide
2014
, vol. 
41
 (pg. 
138
-
145
)
[PubMed]
36
Luo
 
Z.L.
Tang
 
L.J.
Wang
 
T.
Dai
 
R.W.
Ren
 
J.D.
Cheng
 
L.
Xiang
 
K.
Tian
 
F.Z.
 
Effects of treatment with hydrogen sulfide on methionine choline deficient diet-induced non-alcoholic steatohepatitis in rats
J. Gastroenterol. Hepatol.
2014
, vol. 
29
 (pg. 
215
-
222
)
[PubMed]
37
Aoyama
 
T.
Inokuchi
 
S.
Brenner
 
D.A.
Seki
 
E.
 
CX3CL1–CX3CR1 interaction prevents carbon tetrachloride-induced liver inflammation and fibrosis in mice
Hepatology
2010
, vol. 
52
 (pg. 
1390
-
1400
)
[PubMed]
38
Karlmark
 
K.R.
Zimmermann
 
H.W.
Roderburg
 
C.
Gassler
 
N.
Wasmuth
 
H.E.
Luedde
 
T.
Trautwein
 
C.
Tacke
 
F.
 
The fractalkine receptor CX3CR1 protects against liver fibrosis by controlling differentiation and survival of infiltrating hepatic monocytes
Hepatology
2010
, vol. 
52
 (pg. 
1769
-
1782
)
[PubMed]
39
Connolly
 
M.K.
Ayo
 
D.
Malhotra
 
A.
 
Dendritic cell depletion exacerbates acetaminophen hepatotoxicity
Hepatology
2011
, vol. 
54
 (pg. 
959
-
968
)
[PubMed]
40
Liao
 
G.
van Driel
 
B.
Magelky
 
E.
O'Keeffe
 
M.S.
de Waal Malefyt
 
R.
Engel
 
P.
Herzog
 
R.W.
Mizoguchi
 
E.
Bhan
 
A.K.
Terhorst
 
C.
 
Glucocorticoid-induced TNF receptor family-related protein ligand regulates the migration of monocytes to the inflamed intestine
FASEB J.
2014
, vol. 
28
 (pg. 
474
-
484
)
[PubMed]

Author notes

1

These authors equally contributed to this work.

Supplementary data