Granulocytic MDSC with Deficient CCR5 Alleviates Lipogenesis and Inflammation in Nonalcoholic Fatty Liver Disease

Abstract

C-C chemokine receptor type 5 (CCR5) positively contributes to the pathogenesis of nonalcoholic fatty liver disease (NAFLD), a common metabolic liver disease associated with chronic inflammation. CCR5 signaling also facilitates the immunosuppressive activity of a group of immature myeloid cells known as granulocytic myeloid-derived suppressor cells (g-MDSCs). While both hepatocyte and g-MDSC express CCR5, how CCR5 coordinates these two distinct cell types in the hepatic microenvironment remains largely unknown. Here, we used in vivo and ex vivo approaches to define the molecular details of how CCR5 mediates the crosstalk between hepatocytes and g-MDSCs in a mouse model of NAFLD. Global CCR5-deficient mice exhibited more severe steatosis, increased hepatic gene expression of lipogenesis, and exacerbated liver damage in diet-induced obesity. Either NAFLD or CCR5-deficiency per se is causative for the increase of g-MDSCs. Purified g-MDSCs have a higher survival rate in the fatty liver microenvironment, and blockade of CCR5 significantly decreases g-MDSCs' expression of anti-inflammatory factors. On the other hand, the null of CCR5 signaling increases hepatocytes' expression of lipogenic genes in the NAFLD microenvironment. Most importantly, inhibiting g-MDSCs' CCR5 signaling in the fatty liver microenvironment dramatically reduces STAT3 signaling, lipogenic, and pro-inflammatory gene expression in primary hepatocytes. Adoptive cell transfer experiments further demonstrate that CCR5-deficient g-MDSCs mitigate hepatic lipogenic gene expression without facilitating pro-inflammatory cytokine production and liver damage in NAFLD mice. These results suggest that targeting g-MDSCs' CCR5 signaling might serve as a potential therapeutic strategy for NAFLD.

Keywords: CCR5; NAFLD; STAT3; g-MDSC; inflammation; lipid metabolism.

Figure 1

Global CCR5 deficiency exacerbates liver damage and steatosis in NAFLD. (a) A schematic diagram of animal and dietary model. (b) Hepatic Ccr5 mRNA level was detected by real-time PCR. Relative expression was normalized to 36b4 reference gene (n = 4–6). (c) Individual liver mass and body weight were measured and the ratio was calculated (n = 6). (d) Plasma concentrations of AST, ALT, and ALP were determined (n = 6). (e) Representative liver section images of hematoxylin and eosin staining. Positions of central vein (cv) are labeled. (f) Representative liver section images of Oil-red O staining. (g) Hepatic triglyceride concentrations were determined by colorimetric assay (n = 6). Representative and quantified results (means ± SD) are shown for the indicated number of mice. **, p < 0.005; ***, p < 0.001.

Figure 2

Global CCR5 deficiency elevates the hepatic g-MDSC population in both lean and obese mice. (a) Gating strategy for immunophenotyping hepatic MDSCs via flow cytometry. Percentages of hepatic MDSCs (b) and their subsets among the hematopoietic population (c) were distinguished. Representative and quantified results (means ± SD) are shown for the indicated number of mice. *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 3

Global CCR5 deficiency causes altered hepatic STAT3 signaling and expression of cytokine genes and genes related to lipid metabolism. (a) Equal amounts of proteins from liver homogenates were subjected to SDS-PAGE followed by immunoblotting using antibodies specific for phosphorylated and total STAT3. Normalization was confirmed by β-ACTIN immunoblotting. Signal intensity was quantified for statistical analysis. The signal of phosphorylated STAT3 was normalized to total STAT3 and set the average value of WT (NCD) to 1. The levels of hepatic transcripts of genes related to g-MDSC functions (b), lipogenesis (d), and Ppar-γ with its co-activators (e) were determined by real-time PCR. Expression was normalized to the 36b4 reference gene and further normalized to the average value of the WT (NCD) group (n = 4-6). (c) A schematic diagram of targeted lipid-regulatory genes with their reported functions in NAFLD. FFA, free fatty acid; DNL, de novo lipogenesis; FAO, fatty acid oxidation. *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 4

CCR5 signaling modulates the anti-inflammatory profile of g-MDSC in the hepatic microenvironment. (a) A schematic diagram of preparing liver-conditioned M199 culture medium (CM). (b) Primary g-MDSCs were isolated from WT bone marrow cells by FACS sorting. (c) Isolated g-MDSCs were pretreated with 5 μM Maraviroc (MVC) (+) or vehicle control (−) for 1 h and then stimulated with indicated liver-CM for 18 h. Cells were also treated with vehicle control and M199 medium as a negative control. Dead cells were quantitated by flow cytometry after fixable viability staining with FVS780. Additionally, g-MDSCs’ transcript levels of genes related to STAT3 activation (d) and anti-inflammation (e) were determined by real-time PCR. Expression was normalized to the 36b4 reference gene and further normalized to the average value of the first group (n = 3). (f) Primary WT g-MDSCs were pretreated with 5 μM MVC (+) or vehicle control (−) for 1 h and then stimulated with indicated liver-CM for 10 min. The levels of phosphorylated and total STAT3 were determined by immunoblotting. The signal of phosphorylated STAT3 was normalized to total STAT3 and set the first group as 1 (n = 3). *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 5

CCR5 regulates STAT3 signaling in primary hepatocytes. (a) Microscopy images of primary WT hepatocytes were taken before and after plating on a collagen-coated culture dish for 24 h. Subsequent experiments were performed in primary hepatocytes 24 h after plating. (b) Primary hepatocytes were pretreated with 5 μM MVC, 10 μM Stattic or vehicle control (−) for 1 h, and then either left unstimulated (−) or stimulated with 10 ng/mL of CCL5 for 18 h. Total lysates were analyzed by immunoblotting to determine the status of STAT3 expression and phosphorylation. The amount of total STAT3 was normalized to β-ACTIN and set the first group as 1 (n = 3). The levels of transcripts from genes related to hepatic lipogenesis (c) and Ppar-γ with its co-activators (d) were determined by real-time PCR. Expression was normalized to the 36b4 reference gene and further normalized to the average value of the first group (n = 3). *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 6

Blockade of g-MDSCs’ CCR5 signaling reduces STAT3 signaling and lipogenic gene expression in hepatocytes under the fatty liver microenvironment. (a) A schematic diagram of preparing g-MDSC-conditioned culture medium (doubly conditioned medium; dCM). (b) Primary WT hepatocytes were pretreated with 5 μM MVC (+) or vehicle control (−) for 1 h and then stimulated with liver-CM or g-MDSC-dCM for 18 h. Total lysates were analyzed by immunoblotting to determine the status of STAT3 expression. The amount of total STAT3 was normalized to β-ACTIN and set the first group as 1 (n = 3). The levels of transcripts from genes related to hepatic lipogenesis (c), Ppar-γ with its co-activators (d), and pro-inflammatory cytokines (e) were determined by real-time PCR. Expression was normalized to the 36b4 reference gene and further normalized to the average value of the first group (n = 3). *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 7

CCR5-deficient g-MDSCs reduce hepatic STAT3 activation and lipogenic gene expression in NAFLD. (a) A schematic diagram of the animal model for g-MDSC adoptive transfer. Obese WT mice were retro-orbitally injected with PBS, WT g-MDSCs, or R5-KO g-MDSCs three times before being sacrificed by 24 weeks of HFD. (b) The liver-to-body weight ratio was measured and calculated. (c) The amount of hepatic g-MDSCs was quantitated by flow cytometry and shown as percentages in the hematopoietic population as described for Figure 2. (d) The status of STAT3 phosphorylation was determined by immunoblotting of liver homogenates. The signal of phosphorylated STAT3 was normalized to total STAT3 and set the average value of the PBS-injected group to 1. Hepatic transcript levels of genes related to lipogenesis (e), Ppar-γ with its co-activators (f), and pro-inflammatory cytokines (g) were determined by real-time PCR. Expression was normalized to the 36b4 reference gene and further normalized to the average value of the PBS-injected group. (h) Plasma concentrations of AST and ALT were analyzed (n = 2-5). *, p < 0.05; **, p < 0.005; ***, p < 0.001.

Figure 8

A schematic diagram of key conclusions. Results from our mechanistic studies of NAFLD in WT (gray area) and CCR5-deficient (pink area) backgrounds are illustrated in the proposed diagram.