2022 Volume 45 Issue 2 Pages 178-183
Obesity is associated with the risk of venous thromboembolism. Thrombi are constantly formed via the coagulation cascade and degraded by the fibrinolytic system, so they tend to form in obese individuals. Adipocytes are involved in thrombus formation in obesity, but it is not clear whether bioactive factors from adipocytes directly initiate or enhance coagulation and thrombosis. In this study, we confirmed that adipocyte-derived extracellular vesicles (ADEVs) enhance procoagulant activity in vitro. ADEVs prepared from the culture supernatant of mature 3T3-L1 adipocytes shortened plasma clotting times. Moreover, the effect of ADEVs on clotting time was weakened when using plasma lacking factors of the extrinsic pathway, but not the intrinsic pathway. ADEVs contain tissue factors and phosphatidylserine, which are involved in the extrinsic pathway, and blockade of these molecules diminished the effects of ADEVs on plasma clotting time. Additionally, the effect of ADEVs on plasma clotting time was further enhanced when cells were stimulated with the proinflammatory cytokine tumor necrosis factor-α. Thus, ADEVs may be a factor in thrombus formation in obesity.
Obesity is a global health issue and increases the risk of various diseases including venous thromboembolism (VTE), coronary artery disease, stroke, diabetes mellitus, and hypertension. VTE, ischemic heart attack, and stroke are common vascular disorders,1,2) and are estimated to have an incidence of 1–2 per 1000 persons annually in the United States.3,4) Therefore, VTE is an important problem in the U.S. healthcare system. The number of VTE patients is also increasing in Japan.5)
A meta-analysis of nine studies revealed that the risk of VTE among patients with obesity was 2.33-fold that of control subjects (odds ratio, 2.33; 95% confidence interval, 1.68–3.24).6) Moreover, an increase in body mass index (BMI) is associated with the risk of VTE,7–9) implicating obesity in VTE. Thrombi are constantly formed via the coagulation cascade and degraded by the fibrinolytic system; the balance between formation and degradation is important and obese individuals tend to form thrombi, leading to VTE.
Clinical and experimental studies have investigated the mechanisms of thrombus formation in obesity. In the first mechanism, adipocytes store excess triglycerides and promote chronic inflammation by secreting proinflammatory cytokines and inducing macrophage infiltration.10) A chronic inflammatory state activates the prothrombin signaling pathway in the vascular system. Proinflammatory cytokines stimulate vascular endothelium, platelets, and other circulating vascular cells, resulting in upregulation of procoagulant factors and adhesion molecules, downregulation of anticoagulant regulatory proteins, generation of thrombin, and activation of platelets.11) The second mechanism is based on secretion of plasminogen activator inhibitor-1 (PAI-1), a serine protease inhibitor that physiologically inhibits the endogenous fibrinolytic system. Plasma PAI-1 levels are elevated in patients with obesity,12) and PAI-1 deficiency in mice abrogated obesity-induced acceleration of middle cerebral artery occlusion (a model of ischemic stroke),13) suggesting that PAI-1 promotes thrombosis in obesity. Therefore, formation of thrombus is dominant in obese individuals. However, these mechanisms are indirectly related to the coagulation cascade, and the role of coagulation factors released from adipocytes is unclear.
Extracellular vesicles (EVs) such as microparticles released from cells are elevated in patients with VTE14) and coronary artery disease,15,16) and EVs enhance or initiate procoagulant activity.15,17) Although EVs are also released by adipocytes,18) the effects of adipocyte-derived EVs (ADEVs) on coagulation are unknown. We evaluated the effects of ADEVs on procoagulant activity.
We cultured 3T3-L1 mouse embryo fibroblasts (JCBR9014; Lot 0125200, The Human Science Foundation, Osaka, Japan) in a 5% CO2 atmosphere at 37 °C in Dulbecco’s modified Eagle’s medium containing 4.5 g/L D-glucose (DMEM-high glucose; Life Technologies, Carlsbad, CA, U.S.A.) supplemented with 10% fetal bovine serum (FBS) (S1820; Biowest, France), 100 units/mL penicillin (P3032; Sigma-Aldrich/Merck KGaA, Darmstadt, Germany), and 0.1 mg/mL streptomycin (190–14342; FUJIFILM Wako, Osaka, Japan). Cells were seeded at 2.0 × 105 in six-well plates and cultured for 2 d. Next, the cells were induced to undergo adipogenic differentiation for 4 d using 0.5 mM 3-isobuthyl-1-methylxanthine (I7018; Sigma-Aldrich/Merck KGaA), 1 µM dexamethasone (D4902; Sigma-Aldrich/Merck KGaA), and 5 µg/mL bovine insulin (I5500; Sigma-Aldrich/Merck KGaA). The cells were fed with DMEM-high glucose containing 10% FBS and 5 µg/mL insulin for 2 d. At day 6 after initiation of differentiation, cells were fed DMEM-high glucose containing 10% FBS and 0.5 µg/mL insulin for 2 d. At day 8, the cells were used as 3T3-L1 adipocytes for experiments.
Preparation of ADEVs3T3-L1 adipocytes were cultured at 37 °C for 12 h in fresh DMEM-high glucose containing 10% FBS. ADEVs were prepared from the culture supernatant by centrifugation as in Mause and Weber.19) Culture supernatant was transferred to centrifuge tubes, and dead cells and apoptotic bodies were removed by centrifugation at 2000 × g for 15 min. Next, the supernatant was centrifuged at 20000 × g for 15 min, the pellet was resuspended in 20 mM N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES) buffer (pH 7.4) containing 150 mM NaCl (suspension buffer), and the suspension was centrifuged at 20000 × g for 15 min. The pellet was collected as ADEVs, and those from 1.0 × 105 cells were resuspended in 33 µL of suspension buffer containing 0.1% bovine serum albumin (BSA) (A7030; Sigma-Aldrich/Merck KGaA). For stimulation with tumor necrosis factor-alpha (TNF-α) or lipopolysaccharide (LPS), the culture medium was replaced with fresh medium at day 8, and 10 ng/mL recombinant mouse TNF-α (201-13461; FUJIFILM Wako) or 100 ng/mL LPS from Escherichia coli 0111:B4 (L2630; Sigma-Aldrich/Merck KGaA) was added 12 h later and cells were incubated at 37 °C for 3, 6, or 12 h. The pellet was collected as TNF-α-ADEVs or LPS-ADEVs after centrifugation at 20000 × g for 15 min.
Analysis of the Effects of ADEVs on Procoagulant ActivityProcoagulant activity was measured by determining the clotting time using control human plasma (97401915; Siemens Healthcare Diagnostics, Eschborn, Germany), which contains coagulation factors and phosphatidylserine (PS). Briefly, a 90-µL sample of suspension buffer was incubated at 37 °C for 120 s, and human plasma (15 µL) was added and incubated for 30 s. Plasma clotting was initiated by the addition of 15 µL of prewarmed suspension buffer containing 50 mM CaC12 to the reaction solution, and the time to clot formation was measured using a KC4 Delta™ semi-automated coagulation analyzer (Trinity Biotech, Wicklow, Ireland). To examine the effects of ADEVs on procoagulant activity, we added 30-µL of ADEVs to the reaction solution and incubated for 30 s before adding suspension buffer containing 50 mM CaC12. To examine the involvement of the extrinsic and intrinsic pathways in ADEV-enhanced procoagulant activity, we used human plasma deficient in factor VIII (EVIII-ID, Lot. T0405; Haematologic Technologies Inc., Essex, VT, U.S.A.), IX (EIX-ID, Lot. S1214-G1; Haematologic Technologies Inc.), XI (EXI-ID, Lot. S0413; Haematologic Technologies Inc.), XII (EXII-ID, Lot. S0602-G2; Haematologic Technologies Inc.), or VII (DP030A, Lot. 070709C; Hyphen BioMed, Neuville-sur-Oise, France). To block tissue factor (TF) with an anti-TF antibody, ADEVs were incubated with 10 µg/mL anti-TF antibody (AF3178; R&D systems, Minneapolis, MN, U.S.A.) at 4 °C overnight. Next, ADEVs were centrifuged at 20000 × g for 15 min, washed with suspension buffer, and the suspension was centrifuged at 20000 × g for 15 min to collect ADEVs. Anti-goat immunoglobulin G (IgG) from goat serum (I5626, Sigma-Aldrich/Merck KGaA) was used as an isotype control antibody. To block PS with annexin V, ADEVs were incubated with 5 µL of annexin V-fluorescein isothiocyanate (FITC) (4700-100; MBL, Nagoya, Japan) in 85 µL of 10 mM HEPES buffer (pH 7.4) containing 140 mM NaCl and 2.5 mM CaC12 (annexin V-binding buffer) at room temperature for 15 min. The ADEVs were centrifuged at 20000 × g for 15 min, washed with annexin V-binding buffer, and the suspension was centrifuged at 20000 × g for 15 min. This procedure was repeated once, and the ADEVs were resuspended with suspension buffer containing 0.1% BSA.
Immunoblot Analysis with Blue Native-Polyacrylamide Gel ElectrophoresisADEVs from culture supernatant of 3T3-L1 adipocytes (2 × 106 cells) were resuspended in sample buffer containing 100 mM bis-Tris (pH 7.0), 10% (v/v) glycerol, 50 mM 6-aminocaproic acid, 0.5% (w/v) CBB G-250, and 1% sodium dodecyl sulfate (SDS), and incubated for 20 min at 50 °C. The cathode chamber was filled with native polyacrylamide gel electrophoresis (PAGE) Cathode Buffer (50 mM Tricine/NaOH, 15 mM bis-Tris–HCl [pH 7.0]), and the anode chamber was filled with native PAGE Anode Buffer (50 mM bis-Tris–HCl [pH 7.0]). After pre-running on 5–20% Tris-Glycine gels (E-T520L; ATTO, Tokyo, Japan) at 100 V for 60 min, native PAGE Cathode Buffer was replaced with native PAGE Cathode Buffer containing 0.02% CBB G-250. Next, the sample and native molecular weight markers were resolved by running at 100 V for 60 min (4 °C). Native PAGE Cathode Buffer was replaced with native PAGE Cathode Buffer containing 0.002% CBB G-250, and the gel was run at 150 V for 60 min (4 °C).
The resolved proteins were transferred onto a polyvinylidene difluoride membrane (Immobilon-P; Merck Millipore, Billerica, MA, U.S.A.), which was blocked for 1 h with 1% skimmed milk (Morinaga, Tokyo, Japan) in Tris-buffered saline. TF protein was probed using a specific antibody (AF3178) and HRP-linked anti-goat IgG (A5420; Sigma-Aldrich/Merck KGaA) and visualized by enhanced chemiluminescence using ImmunoStar LD reagent (FUJIFILM Wako). Analysis was performed on an Image Station 4000 MM (Kodak, Rochester, NY, U.S.A.), and quantitated using ImageJ software (National Institutes of Health, Bethesda, MD, U.S.A.).
We prepared ADEVs from the culture supernatant of 3T3-L1 adipocytes as shown in Fig. 1A. The pellet after centrifugation at 20000 × g for 15 min (ppt ①) shortened the clotting time (Fig. 1B); i.e., ppt ① enhanced the procoagulant activity of control human plasma. The other pellets (ppt ② to ⑥) did not affect the clotting time (Fig. 1B). ADEVs that precipitated at 20000 × g19) were responsible for the enhanced procoagulant activity.
(A) Preparation of ADEVs by centrifugation. Culture supernatant of 3T3-L1 adipocytes was centrifuged as indicated. (B) Effects of pellets on clotting time. Values are means ± standard deviation (S.D.) (n = 6). Asterisk indicates significant differences (* p < 0.05, vs. ppt minus) calculated by Student’s t-test.
Coagulation is initiated by the intrinsic or extrinsic pathway followed by the common pathway (Fig. 2A, left panel). The intrinsic pathway involves factors XII, XI, IX, VIII, and PS. The extrinsic pathway involves factors VII, TF, and PS. Because EVs exert a procoagulant effect via the extrinsic pathway,20) we ascertained whether ADEVs enhanced procoagulant activity via the extrinsic pathway. First, we confirmed the involvement of the extrinsic pathway by using plasma deficient in factors of the intrinsic or extrinsic pathway. The effect of ADEVs on the plasma clotting time was unchanged by use of plasma deficient in factors of the intrinsic pathway (Fig. 2A, right). In contrast, the effect of ADEVs on the clotting time was weakened by use of plasma lacking factor VII of the extrinsic pathway.
(A) Effects of ADEVs on procoagulant activity were measured by determining clotting time using control human plasma or plasma deficient in factors of the intrinsic and extrinsic pathway. Values are means ± S.D. (n = 3). (B) Detection of TF in ADEVs by BN-PAGE using a specific antibody (left panel). Effect of pretreatment with an anti-TF antibody on ADEV-enhanced procoagulant activity (right panel). TF antibody (10 µg/mL) was added to a suspension of ADEVs, and the suspension was incubated overnight at 4 °C. The plasma clotting time was measured. Values are means ± S.D. (n = 3). (C) Effect of pretreatment with annexin V on ADEV-enhanced procoagulant activity. Before measuring the plasma clotting time, ADEVs were incubated with annexin V for 15 min at room temperature. Values are means ± S.D. (n = 3). Asterisk indicates significant differences (* p < 0.05) calculated by Student’s t-test.
Next, we examined the involvement of TF using an anti-native TF antibody (Fig. 2B, left panel). The effect of ADEVs on plasma clotting time was weakened by pretreatment with an anti-TF antibody (Fig. 2B, right panel). Moreover, because PS is present on the surface of EVs and is required for the activation of TF,21) we examined the involvement of PS in the effect of ADEV on the plasma clotting time. The effect of ADEVs on plasma clotting time was weakened by pretreatment with annexin V, which binds PS (Fig. 2C). Therefore, ADEVs enhanced procoagulant activity via the extrinsic pathway.
ADEVs from Adipocytes Stimulated with TNF-α Strongly Enhanced Procoagulant ActivityBecause adipocytes secrete substances in response to various stimuli, we also investigated the relationships between inflammatory stimuli, such as TNF-α, and enhancement of procoagulant activity by ADEVs. The effects of TNF-α-ADEVs from adipocytes stimulated with TNF-α on clotting time were stronger than those from non-stimulated cells (Fig. 3). To evaluate the effects of stimuli other than cytokines, cells were stimulated with LPS; LPS-ADEVs had no effect on clotting time. These results suggest that specific inflammatory stimuli trigger the effect of ADEVs on procoagulant activity.
3T3-L1 adipocytes were stimulated with 10 ng/mL TNF-α or 100 ng/mL LPS for the indicated periods, and the effects of TNF-α-ADEVs and LPS-ADEVs on clotting time were measured. Values are means ± S.D. (n = 3). Asterisk indicates significant differences (* p < 0.05, vs. none) calculated by Student’s t-test.
EVs secreted by a variety of cell types (including platelets, monocytes, endothelial cells, and cancer cells) exhibit procoagulant activity. Although adipocytes also release EVs, the effects of ADEVs on procoagulant activity are unclear. Here, we demonstrated that ADEVs that precipitate at 20000 × g enhance procoagulant activity via the extrinsic pathway, and we also found that TF and PS in ADEVs are involved in these effects.
TF must be converted from an inactive to an active conformation to exert its procoagulant effect.22,23) Several factors activate TF, including monomerization, disulfide bond formation, phosphatidylserine exposure outside the membrane, and lipid rafts.24) The ADEVs contained monomeric and dimeric TF (Fig. 2B). Additionally, annexin V binding indicated that PS is exposed on the exterior of ADEVs. These results suggest that ADEVs contain active TF. Therefore, ADEVs supply active TF and increased active TF enhances procoagulant activity via the extrinsic pathway.
Are ADEVs involved in thrombosis in vivo? Because adipocytes exist outside blood vessels, to accelerate blood coagulation ADEVs must be transported into the circulation. EVs carrying tumor-associated antigen are present in the serum of patients with ovarian carcinoma.25) Moreover, tumor-derived EVs are elevated in plasma of patients with gastric cancer.26) Because EVs derived from cells outside blood vessels are present in blood, ADEVs may reside in blood vessels in vivo. Moreover, EVs are potential biomarkers for thrombosis in cancer.27) In this study, we measured the effects of ADEVs from 1 × 105 cells on procoagulant activity in 120 µL of buffer. Assuming a blood volume of 5 L, a similar response will occur when ADEVs from approximately 4 × 109 cells are present in the blood. In vivo, there is clearance of EVs, and only a small percentage of ADEVs can be transferred to the blood. However, the number of adipocytes in the body has been reported as 4–8 × 1010 cells.28) We also measured the procoagulant activity of ADEVs using 10-fold diluted plasma. Therefore, ADEVs may be present in blood vessels and accelerate thrombosis in vivo.
Obesity is thought to be a risk factor for thrombosis. TNF-α secretion is increased in obese individuals, and our findings showed that TNF-α stimulation, but not LPS stimulation, enhanced the effects of ADEVs on plasma clotting time. Moreover, the number of EVs secreted from TNF-α-stimulated cells was higher than that from unstimulated cells (Supplementary Fig. S1) and monomeric TF was increased in ADEVs from cells stimulated with TNF-α (Supplementary Fig. S2). Therefore, an increase in the number of ADEVs and amount of monomeric TF may be involved in the effects of ADEVs on plasma clotting time enhanced by TNF-α stimulation. The procoagulant activity of ADEVs released from adipocytes in obese subjects may be stronger than that in normal subjects, and ADEVs are likely involved in the development of thrombosis in obese subjects.
This is the first study to demonstrate that adipocyte-derived EVs enhance procoagulant activity, which was promoted by stimulation of cells with TNF-α. Further analysis of the mechanism by which TNF-α promotes the ADEV-enhanced procoagulant activity will provide insight into the roles of adipocytes in thrombosis in patients with obesity.
We thank Y. Hayashi (Teikyo University), R. Suzuki (Teikyo University), and T. Maruyama (Teikyo University). This work was supported in part by ACRO Incubation Grants from Teikyo University.
The authors declare no conflict of interest.
This article contains supplementary materials.
