Adiponectin inhibits chemokine-induced cell migration via direct interaction with platelet factor 4

Abstract

Despite adiponectin’s recognized anti-inflammatory properties, its impact on cardiovascular homeostasis involves poorly defined mechanisms. We investigated the effect of adiponectin on chemokine-induced cell migration and their potential intermolecular interactions. Our findings revealed that cell migration induced by recombinant PF4, MCP-1, or RANTES in HL-60 cells was significantly inhibited by pre-treating cells with adiponectin. Surface plasmon resonance analysis and molecular docking analysis indicated that only PF4 binds to adiponectin with a higher affinity of adiponectin to the PF4 binding site respectively. These results suggest that adiponectin’s atheroprotective functions may be mediated by its ability to reduce PF4-induced monocyte migration through direct interaction.

Introduction

Chemokines, or chemotactic cytokines, are small secreted proteins that play a crucial role in mediating cell recruitment to sites of infection or inflammation by interacting with their specific receptors. These proteins are essential for immune system homeostasis and are involved in both protective and destructive immune and inflammatory responses.⁽¹⁾ In the context of obesity, circulating levels and tissue contents of various chemokines are significantly increased, promoting the influx of inflammatory leukocytes, particularly pro-inflammatory monocytes, into hypertrophic adipose tissue.^(1,2) For example, monocyte chemoattractant protein-1 (MCP-1), generated by macrophages and endothelial cells, is upregulated in obesity and contributes to macrophage polarization and monocyte recruitment in adipose tissue, leading to insulin resistance and hepatic steatosis.⁽³⁾ Other chemokines, such as regulated on activation, normal T cell expressed and secreted (RANTES) and platelet factor 4 (PF4), also participate in the inflammation of obese adipose tissue. RANTES recruits blood monocytes and exacerbates obesity-associated insulin resistance,⁽⁴⁾ while PF4, predominantly derived from platelets, targets various blood cells and its receptor CXCR3 is upregulated in obesity, modulating obesity-induced adipose inflammation and insulin resistance.⁽⁵⁾

On the other hand, adiponectin, a 30 kDa protein secreted by adipocytes, circulates in the plasma as monomers or higher-order complexes, including trimeric, hexameric, and oligomeric forms. Additionally, full-length adiponectin multimers can be cleaved to release globular domain fragment which is known to have a potent metabolic effect.⁽⁶⁾ During obesity, adiponectin receptors expression and its plasma levels are reduced, and such reduction is considered to be the leading cause of obesity-linked complications including diabetes, and atherosclerosis.⁽⁷⁾

Studies by us and others have demonstrated that adiponectin binds to several growth factors and modulate their biological activity.^(8,9) However, to our knowledge, there is no report about the physical interaction of adiponectin with chemokines. Therefore, the present study aimed to explore the potential effect of adiponectin on chemokine-induced cell migration, as well as their intermolecular interactions.

Materials and Methods

HL-60 cell culture and cell chemotaxis assay

Human promyelocytic leukemia, HL-60 cells were cultured in RPMI1640 medium (Wako Pure Chemicals, Osaka, Japan) containing 10% fetal calf serum (FCS; Trace Scientific Ltd., Melbourne, Australia), 100 ‍U/ml penicillin, and 100 ‍μg/ml streptomycin under a humidified atmosphere of 5% CO₂ at 37°C. Chemotaxis assay of HL-60 cells were performed using the QCM^TM chemotaxis 96-well cell migration assay kit (Chemicon, Temecula, CA). HL-60 cells were cultured in FCS-free RPMI1640 medium for 16 ‍h prior to the cell migration assay. The following proteins including, full-length adiponectin (Biovendor Laboratory Medicine Inc., Brno, Czech Republic), globular adiponectin (Wako), PF4 (PeproTech EC, London, UK), MCP-1 (Sigma-Aldrich Fine Chemicals, St. Louis, MO), and RANTES (R&D Systems, Minneapolis, MN) diluted in FCS free, RPMI medium at different concentrations were added to the bottom wells of the chamber. Cells suspended in FCS-free RPMI medium (1 × 10⁶/ml) were added to the top wells of the chamber and allowed to migrate for 30 ‍min at 37°C in a 5% CO₂ atmosphere. The number of HL-60 cells that migrated to the bottom chamber of each well were determined using the hemocytometer.

Molecular interaction analysis using surface plasmon resonance (SPR) method

The protein-protein interactions between chemotactic factors with adiponectin were examined using a methodology previously described in our published work.⁽⁹⁾ Briefly, full-length adiponectin as a ligand was immobilized at a concentration of 10 ‍μg/ml onto the carboxymethylated dextran surface of the CM5 sensor chip. The adiponectin-chip surface was then perfused with HBS-EP buffer (10 ‍mM HEPES, 150 ‍mM NaCl and 0.005% Surfactant P20, pH ‍7.4) at 37°C, followed by exposure to increasing concentrations of PF4, MCP-1, or RANTES dissolved in buffer at a flow rate of 20 ‍μl/min for 105 ‍s. Following the addition of each analyte, dissociation was evaluated by passing the buffer alone over the chip for 120 ‍s. Based on the dissociation constant (Kd) (s–1) and association constant (Ka) (M–1s–1) obtained, the binding constant KD (M) was calculated by dividing Kd by Ka.

Molecular docking assessment

The three-dimensional structures of human adiponectin, PF4, MCP1, and RANTES were sourced from the UniProt database (https://www.uniprot.org/). To explore the interactions between these proteins, molecular docking was performed using the ClusPro server (https://cluspro.org/help.php). In this study, adiponectin was selected as the ligand, while PF4, MCP1, and RANTES acted as individual receptors. The interaction profiles between adiponectin and each of the receptor proteins were carefully analyzed using the PDPsum server (https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/Generate.html). The results of these analyses were then visualized using BIOVIA Discovery Studio software.

Statistical analysis

The results are expressed as mean ± SE. One-way analysis of variance (ANOVA) followed by the Tukey–Kramer post-hoc test was utilized for chemotaxis assay. A p value of less than 0.05 was considered statistically significant.

Results

Chemotaxis assay was conducted in vitro using HL-60 cells. As presented in Fig. 1, there was no difference in the number of chemotactic cells between FCS-free medium used as a negative control and the FCS-free medium supplemented with either full-length or globular adiponectin. In contrast, supplementation of the medium with any of three chemotactic factors, dose-dependently increased the number of migrated HL-60 cells compared to the control. Nevertheless, the simultaneous addition of either full-length or globular adiponectin to the chemotactic-supplemented medium exhibited a tendency to reduce cell migration. Notably, both forms of adiponectin significantly suppressed PF4-induced chemotaxis.

Fig. 1.

Effect of adiponectin on chemotaxis of HL-60 cells induced by PF4, MCP-1, or RANTES. HL-60 cells were placed on the upper side of the top chemotaxis analysis chambers, and cells that migrated to the lower side of the chamber was measured in medium supplemented with either full-length adiponectin (fADP, 0.75 ‍μg/ml) or globular adiponectin (gADP, 0.5 ‍μg/ml) and (A) PF4 (1 and 10 ‍μg/ml), (B) MCP-1 (0.1, 1, and 10 ‍μg/ml), or (C) RANTES (1 and 10 ‍μg/ml) (n = 4). Values represent mean ± SE. * indicates significant difference (p<0.05) for FCS-free medium. ^# indicates significant difference (p<0.05) for medium with the same concentration of chemotactic factor and without adiponectin.

Next, we examined the interaction between adiponectin and the three chemotactic factors using SPR method. The adiponectin used in this study includes trimeric, hexameric and multimeric form. Platelet-derived growth factor-BB (PDGF-BB) was loaded as a positive control analyte based on our recent finding.⁽⁹⁾ As shown in Fig. 2A, infusion of increasing doses of PF4 on the surface of immobilized full-length adiponectin led to an increase in RU in a dose-dependent manner, reflecting their ligand-binding, while stopping this infusion resulted in a drop in RU, an indication of their separation from the ligand. Kinetics analysis showed that PF4 bound selectively to full-length adiponectin with binding constant (K_D) of 160 nM, which is slightly higher to those of PDGF-BB (25 nM) (Fig. 2D). On the other hand, there was no detectable interactions between adiponectin and MCP-1 or RANTES (Fig. 2B–D).

Fig. 2.

Molecular interactions between adiponectin and chemotactic factors. (A–C) Increasing concentrations of (A) PF4 (0.5, 1, 2.5, 5, and 10 ‍mg/ml), (B) MCP-1 (0.5, 1, 2.5, 5, and 10 ‍mg/ml), and (C) RANTES (0.5, 1, 2.5, 5, and 10 ‍mg/ml) were perfused on the surface of full-length adiponectin (fAPDN; 10 ‍μg/ml)-immobilized BIAcore sensor chip. Respective overlay plots of sensorgram are shown. (D) Summary of analyte binding to adiponectin.

The molecular docking interactions of adiponectin, as a ligand, with PF4, MCP1, and RANTES are illustrated in Fig. 3 and summarized in Supplemental Table 1–3*. Adiponectin demonstrates the strongest interaction with the PF4 binding site, forming 12 hydrogen bonds, 4 Pi interactions, and 1 salt bridge, with a binding free energy of −766.2 (Fig. 3A and Supplemental Table 1*). For the MCP1 binding site, adiponectin exhibits a binding free energy of −584.7, forming 19 hydrogen bonds and 4 Pi interactions (Fig. 3B and Supplemental Table 2*). Interaction with the RANTES binding site involves 9 hydrogen bonds, 8 Pi interactions, and 2 salt bridges, with a binding free energy of −737.2 (Fig. 3C and Supplemental Table 3*).

Fig. 3.

Molecular docking interactions of adiponectin with chemotactic factors. (A) Interaction of adiponectin with PF4 by binding free energy of −766.2 ‍kcal/mol. (B) Interaction of adiponectin with MCP-1 by binding free energy of −584.7 ‍kcal/mol. (C) Interaction of adiponectin with RANTES by binding free energy of −737.2 ‍kcal/mol. Details of molecular interactions of adiponectin with PF4, MCP-1, and RANTES are presented in Supplemental Table 1–3*, respectively.

Discussion

Chemokines are released in response to pro-inflammatory signals, attracting immune cells like neutrophils, lymphocytes, and monocytes to sites of inflammation or injury. Persistent infiltration of activated macrophages and T cells into atherosclerotic lesions significantly contributes to atherosclerosis progression.⁽¹⁰⁾ Although adiponectin does not affect cell proliferation, it plays an atheroprotective role by inhibiting the proliferation and migration of vascular smooth muscle cells induced by several growth factors.⁽⁸⁾ In line with this, we demonstrated that both globular and full-length adiponectin forms reduce HL-60 cell migration induced by PF4, MCP-1, and RANTES, chemokines essential for monocyte attraction and key players in atherosclerosis development.

Previous studies reported that adiponectin may have anti-inflammatory mechanisms independent of the suppression of NF-κB signaling in target cells through methods that do not involve the stimulation of intracellular signaling. One mechanism is that adiponectin affects biological activity of chemokines and growth factor through physical interaction and thereby abolish their effect at their perspective pre-receptor level.^(8,11,12) We previously showed that adiponectin binds to and modulate the biological activity of nerve growth factor β.⁽⁹⁾ In the present study, intermolecular interaction analysis by SPR method revealed that only PF4 binds to adiponectin with a binding constant of 160 nM, while molecular docking analysis indicated a higher affinity of adiponectin for PF4 binding sites compared to those of MCP-1 or RANTES. While our results show that adiponectin does not directly bind to MCP-1 or RANTES, it may still inhibit their chemotaxis activity through indirect mechanisms beyond physical interaction. For instance, adiponectin has been shown to modulate macrophage function and polarization, which could affect the cellular response to these chemokines.⁽¹³⁾ Additionally, adiponectin can suppress the expression of adhesion molecules and reduce monocyte attachment to endothelial cells, potentially interfering with the chemotactic process initiated by MCP-1 and RANTES.^(14,15)

Platelet dysfunction and hyperactivity is a typical complication in obese individuals and believed to be key factors in the initiation and progression of atherothrombosis in obese patients.⁽¹⁶⁾ A study on rats fed a high-fat diet showed that plasma level of PF4 increased significantly along with cholesterol and triglyceride levels.⁽¹⁷⁾ Moreover, markers of platelet activation, such as P-selectin and sCD40L, have been reported at higher levels in obese subjects.^(18,19) It’s noteworthy that adiponectin reduces platelet aggregation and sCD40L release from platelets.⁽²⁰⁾ Moreover, an accelerated thrombus formation has been reported in adiponectin knockout male mice.⁽²¹⁾ In line with these observations, our in vitro binding and migration assays suggest that adiponectin might regulate the development of atherosclerosis by controlling the bioavailability of PF4, hindering their binding to the CXCR3 receptor, and thus attenuating their chemotactic actions. However, further research is necessary to elucidate the details of the underlying mechanisms.

In summary, considering the opposite roles of adiponectin and chemokine on the cell’s recruitment, it is highly likely that an imbalance exists between adiponectin and chemokines in obesity, and that alteration may contribute to the development of systemic inflammation, and vascular complications associated with obesity. Our findings give a new insight to the potential significance of the interaction between adiponectin and PF4 as an additional mechanism contributing to the anti-atherogenic properties of adiponectin. Additional research is required to comprehensively understand the interplay between adiponectin and PF4 in the regulation of chemotaxis.

Author Contributions

ME: data curation, formal analysis, investigation, writing – original draft preparation, and writing – review & editing. SG: data curation, formal analysis, and investigation. AE-F: data curation and formal analysis. YO-O: data curation, formal analysis, and writing – review & editing. KK: conceptualization, funding acquisition, project administration, supervision, and writing – review & editing. All authors have read and approved the final manuscript.

Funding

This work was supported by a JSPS KAKENHI grant number 19H03116 to KK.

Data Availability

Research data is available upon request.

Conflict of Interest

No potential conflicts of interest were disclosed.