CX3CR1 differentiates F4/80low monocytes into pro-inflammatory F4/80high macrophages in the liver

The expression of chemokine receptor CX3CR1 is related to migration and signaling in cells of the monocyte-macrophage lineage. The precise roles of CX3CR1 in the liver have been investigated but not clearly elucidated. Here, we investigated the roles of CX3CR1 in hepatic macrophages and liver injury. Hepatic and splenic CX3CR1lowF4/80low monocytes and CX3CR1lowCD16− monocytes were differentiated into CX3CR1highF4/80high or CX3CR1highCD16+ macrophages by co-culture with endothelial cells. Moreover, CX3CL1 deficiency in human umbilical vein endothelial cells (HUVECs) attenuated the expression of interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), whereas recombinant CX3CL1 treatment reversed this expression in co-cultured monocytes. Upon treatment with clodronate liposome, hepatic F4/80high macrophages were successfully depleted at day 2 and recovered similarly in CX3CR1+/GFP and CX3CR1GFP/GFP mice at week 4, suggesting a CX3CR1-independent replacement. However, F4/80high macrophages of CX3CR1+/GFP showed a stronger pro-inflammatory phenotype than CX3CR1GFP/GFP mice. In clodronate-treated chimeric CX3CR1+/GFP and CX3CR1GFP/GFP mice, CX3CR1+F4/80high macrophages showed higher expression of IL-1β and TNF-α than CX3CR1−F4/80high macrophages. In alcoholic liver injury, despite the similar frequency of hepatic F4/80high macrophages, CX3CR1GFP/GFP mice showed reduced liver injury, hepatic fat accumulation, and inflammatory responses than CX3CR1+/GFP mice. Thus, CX3CR1 could be a novel therapeutic target for pro-inflammatory macrophage-mediated liver injury.

n Supplementary Fig. 2 n Supplementary Fig. 3 n Supplementary Fig. 4 n Supplementary Fig. 5 n Supplementary Fig. 6 n Supplementary Table 1 n Supplementary Table 2 n Supplementary references 2

Chronic ethanol diet
The Lieber-DeCarli liquid diet containing 5 % ethanol (Dyets Inc., Bethlehem, PA) was placed on each mouse cage for ethanol diet feeding. The liquid diet was kept in a bottle and replaced daily to prevent evaporation of ethanol. At week 8, mouse liver tissue and blood were collected for the evaluation of liver injury. Sectioned liver tissues were subjected to hematoxylin and eosin (H & E) staining, immunohistochemistry, and Oil Red O staining. Serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and triglyceride were assessed using kits purchased from IDEXX Laboratories (Westbrook, ME).

Generation of chimeric mouse
Chimeric mice were generated as previously described 1 . Briefly, antibiotic-containing water was fed to recipient mice for 7 days before radiation treatment. Next, the mice were irradiated twice at a dose of 475 Rad at 3-h intervals. Six hours after the first radiation, total BMCs (3 × 10 6 cells) from donor mice were injected via the tail vein. To deplete the recipient's Kupffer cells, clodronate liposome treatment was performed 2 days before BMC transplantation as previously described 2 .

Liver non-parenchymal cell isolation
Livers from mice were homogenized and filtered through a 70-μm nylon cell strainer (BD Bioscience, San Jose, CA). After removing debris, hepatic non-parenchymal cells were collected and suspended in 40% Percoll. The cell suspension was gently overlaid onto 70% Percoll and centrifuged at 4°C for 30 min at 120 xg. Liver mononuclear cells (MNCs) were collected from the interface. Then, liver MNCs were resuspended in fresh PBS after RBC lysis. The cell suspension was filtered through a 70-μm cell strainer and centrifuged at 42 xg for 5 min at room temperature to remove hepatocytes.
The supernatant was transferred to a new tube and centrifuged at 400 g for 10 min at 4°C. The pellet was resuspended in 6 ml of 11.5 % Optiprep (Sigma-Aldrich, St. Louis, MO, USA), loaded carefully onto 6 ml of 20 % Optiprep, and centrifuged at 1,800 xg 3 for 17 min at 4°C. The cellular fraction in the interface between 11.5% and 20% Optiprep was gently collected 3 .

Stimulation of HUVECs with TNF-α
HUVECs were used in passages 4 to 6 and cultured in EGM-2 SingleQuots ® medium (Lonza, Walkersville, MD). HUVECs were seeded with a density of 1 × 10 5 cells/cm 2 in 24 well plates. After 24 h, cells were stimulated with 1 ng/ml TNF-α for another 24 h. Absence of mycoplasma contamination was confirmed using e-Myco VALiD Mycoplasma PCR detection kit (iNtRON Biotechnology, Seongnam, Korea) using input genomic DNAs in HUVECs.

Co-culture of mouse and human monocytes with mouse LSECs and human umbilical vein endothelial cells (HUVECs)
To differentiate macrophages from monocytes, 2 × 10 5 monocytes of liver or spleen were co-plated onto 5 × 10 5 of LSECs. The latter cells had been plated 3 h previously.
The cells were co-cultured in Roswell Park Memorial Institute (RPMI)-1640 containing 10% fetal bovine serum (FBS) for 3, 6, 12, or 24 h. Whole cells were detached from the plate at each time point using trypsin-EDTA (Invitrogen, Carlsbad, CA) and the monocytes were separated from LSECs using magnetic-activated cell sorting. Isolated monocytes were analyzed by flow cytometry. In some experiments, 5 × 10 5 whole liver non-parenchymal cells were cultured at each time point. Human monocytes were isolated from peripheral blood mononuclear cells (PBMCs) using MagniSort TM human CD14 positive selection kit (Invitrogen) following the manufacturer's protocol. Human monocytes (5 ×10 5 ) were co-cultured with 5 × 10 5 HUVECs that had been plated 3 h previously. The co-culture was carried out in EGM-2 SingleQuots ® medium for 12, 24, and 48 h. The monocytes were subjected to flow cytometry after detachment using trypsin-EDTA.

Isolation of splenic F4/80 low CD11b + monocytes
The spleen was removed and minced into small pieces, which were homogenized and filtered through a 70-μm cell strainer. The cell suspension was centrifuged at 400 xg, 4°C for 5 min. The supernatant was removed and pellets were treated with 5 ml RBC 4 lysis buffer (Biolegend, San Diego, CA) for 5 min. The cells were suspended in PBS, centrifuged at 400 xg, 4°C for 10 min, and the splenocytes were collected. F4/80 low CD11b + monocytes were isolated using FACS Aria III device (BD Bioscience, San Jose, CA).

Silencing of CX3CL1 in HUVECs and treatment with recombinant CX3CL1
HUVECs were stably transfected with small interfering RNA (siRNA) for CX3CL1 using

cDNA synthesis and qRT-PCR
Sorted or collected cells and liver tissue were treated with TRIzol (Invitrogen) for extraction of total RNA. cDNA was synthesized from RNA with amfiRivert cDNA synthesis master mix (GenDEPOT, Barker, TX) following the manufacturer's protocol.
qRT-PCR was performed with SYBR Green real-time PCR master mix (Toyobo, Osaka, Japan) using the CFX96 system (Bio-Rad, Hercules, CA). Supplementary Table 1 and 2 list the PCR primers. The expression of each gene was compared with those of βactin and GAPDH and analyzed using ΔΔCt values.

Blood biochemistry
Serum was collected and assayed for ALT, AST and triglyceride (TG) using kits purchased from IDEXX Laboratories (ME, USA).

Gene expression omnibus dataset
To analyze the expression of genes in mouse livers with alcoholic steatohepatitis and alcoholic hepatitis, gene expression data were obtained from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database 6 (accession number GSE97234) and previously published high-throughput sequencing data 4 . As previously described 5 , Upper Quartile normalization was performed to normalize all samples of read counts per gene/transcript, and Partek Gene Specific Analysis method was used to analyze differential expression. For each comparison, differentially expressed genes (DEGs) were defined as a fold-change >1.0 in either direction and a false discovery rate of 0.05. Among DEGs, heatmaps for Cx3cr1, Cx3cl1, Il1b, and Tnf were generated using GraphPad Prism (version 7.00; GraphPad Software Inc., La Jolla, CA).