Fractalkine is a novel human adipochemokine associated with type 2 diabetes

Abstract

Objective: Leukocyte infiltration of adipose is a critical determinant of obesity-related metabolic diseases. Fractalkine (CX3CL1) and its receptor (CX3CR1) comprise a chemokine system involved in leukocyte recruitment and adhesion in atherosclerosis, but its role in adipose inflammation and type 2 diabetes is unknown.

Research design and methods: CX3CL1 mRNA and protein were quantified in subcutaneous adipose and blood during experimental human endotoxemia and in lean and obese human adipose. CX3CL1 cellular source was probed in human adipocytes, monocytes, and macrophages, and CX3CL1-blocking antibodies were used to assess its role in monocyte-adipocyte adhesion. The association of genetic variation in CX3CR1 with metabolic traits was determined in a community-based sample. Finally, plasma CX3CL1 levels were measured in a case-control study of type 2 diabetes.

Results: Endotoxemia induced adipose CX3CL1 mRNA (32.7-fold, P < 1 × 10(-5)) and protein (43-fold, P = 0.006). Obese subjects had higher CX3CL1 levels in subcutaneous adipose compared with lean (0.420 ± 0.387 vs. 0.228 ± 0.187 ng/mL, P = 0.04). CX3CL1 was expressed and secreted by human adipocytes and stromal vascular cells. Inflammatory cytokine induction of CX3CL1 in human adipocytes (27.5-fold mRNA and threefold protein) was completely attenuated by pretreatment with a peroxisome proliferator-activated receptor-γ agonist. A putative functional nonsynonymous single nucleotide polymorphism (rs3732378) in CX3CR1 was associated with adipose and metabolic traits, and plasma CX3CL1 levels were increased in patients with type 2 diabetes vs. nondiabetics (0.506 ± 0.262 vs. 0.422 ± 0.210 ng/mL, P < 0.0001).

Conclusions: CX3CL1-CX3CR1 is a novel inflammatory adipose chemokine system that modulates monocyte adhesion to adipocytes and is associated with obesity, insulin resistance, and type 2 diabetes. These data provide support for CX3CL1 as a diagnostic and therapeutic target in cardiometabolic disease.

FIG. 1.

Adipose tissue and circulating CX3CL1 levels are increased in human endotoxemia. Adipose tissue relative mRNA expression (A) and protein levels (B) were increased by both 3 ng/kg and 0.6 ng/kg endotoxin. C: Plasma levels were also increased by both endotoxin doses. Closed circles and solid line denote 3 ng/kg endotoxin; open circles and dashed line denote 0.6 ng/kg endotoxin. Values shown are mean ± SD. *P < 0.001; **P < 0.0001 compared with pre-endotoxin (time 0) levels.

FIG. 2.

Adipose CX3CL1 is increased in obesity and is present in both the primary adipocyte fraction and the SVF. A: Obese human subjects (n = 13) had significantly higher CX3CL1 protein in subcutaneous adipose tissue than lean human subjects (n = 11). B: In obese subjects (n = 13), there was increased CX3CL1 protein in visceral compared with subcutaneous adipose tissue. In lean human adipose (n = 8), both the adipocyte and the SVFs expressed CX3CL1 mRNA (C) and protein (D). Values shown are mean ± SD. *P < 0.05 between groups.

FIG. 3.

Regulation of CX3CL1 expression in human adipocytes by inflammation and PPAR-γ. Endotoxin (100 ng/mL) upregulated CX3CL1 relative mRNA expression (A) and secreted protein (B) in fully differentiated cultured human adipocytes. Similar upregulation was seen after stimulation with 10 ng/mL TNF-α and 5 ng/mL IFN-γ in CX3CL1 mRNA (C) and protein (D); the inflammatory stimulation was ablated completely by pretreatment of adipocytes with 10 μM PPAR-γ agonist GW347845 (representative experiment with n = 3 replicates). Values shown are means ± SD. *P < 0.01; **P < 0.01; ***P < 0.0001 compared with control levels.

FIG. 4.

CX3CL1 modulates adhesion of human monocytes to primary human adipocytes. A: Mean number (± SD) of fluorescent THP-1 monocytes adherent to primary adipocytes were counted per HPF in each experimental condition. The groups designated for each block refer to blocking antibodies used in the condition. *P < 0.0001 compared with negative antibody; **P < 0.005 compared with CX3CL1 alone. Representative images for negative antibody (B), CX3CL1 antibody (C), MCP-1 antibody (D), and both CX3CL1 and MCP-1 antibodies (E) are shown. (A high-quality digital representation of this figure is available in the online issue.)

FIG. 5.

Patients with type 2 diabetes have higher levels of plasma CX3CL1 than nondiabetic subjects. Plasma CX3CL1 values were measured by ELISA in nondiabetic (n = 274) and diabetic (n = 281) subjects. Values shown are mean ± SD. *P < 0.0001 between groups.