Effects of Exenatide on Blood Coagulation and Platelet Aggregation in Patients with Type 2 Diabetes

Background: To explore the effect of glucagon-like peptide-1 receptor (GLP-1R) agonist exenatide on blood coagulation function and platelet aggregation function in patients with type 2 diabetes mellitus (T2DM). Method: Thirty patients with newly diagnosed T2DM were enrolled as the case group, and 30 healthy people with matching age and sex were selected as the control group. Patients in the case group received exenatide treatment for 8 weeks. Collect the general clinical data and biochemical indicators of the patients in the case group before and after 8 weeks of exenatide treatment and the control subjects, and detect their peripheral blood platelet count (×10 12 g/L), plasma prothrombin time (PT, s), prothrombin time activity (PTA, %), activated partial thromboplastin time (APTT, s), international normalized ratio (INR), brinogen (FIB, g/L), plasma thrombin time (TT, s) , Fibrin degradation products (FDP, μg/mL), D-dimer (DD, μg/mL), nitric oxide (NO, μmol/L), CD62p (%), platelet activation complex-1 (PAC-1, %) and platelet aggregation induced by collagen, epinephrine, arachidonic acid (AA, %), and adenosine diphosphate (ADP, %). Results: There was no signicant difference in platelet count, PLT, PT, PTA, APTT, TT, INR, FDP, DD between the case group and the control group; the FIB, CD62p, PAC-1, platelet aggregation rates of the case group (EPI 87.23±6.84 , ADP 87.51±9.21, AA 90.17±3.19) is higher than normal control group (EPI 82.15±5.37, ADP 82.38±6.42, AA 83.41±6.17, P (cid:0) 0.05), NO level is lower than normal control group (68.1±14.7 vs. 79.4±11.2, P<0.05); After 8 weeks of exenatide treatment in the case group, CD62p, PAC-1, and platelet aggregation rates were lower than before (EPI: 81.62±9.02 vs. 87.23±6.84, AA: 84.62±7.12 vs. 90.17±3.19, P (cid:0) 0.05), the level of NO was higher than before (89.6±15.8 vs. 68.1±14.7); Pearson correlation analysis showed that the changes in platelet aggregation rates (Δ EPI, ΔAA) of patients in the case group after 8 weeks of exenatide treatment were positively correlated with the changes in body mass index (BMI, kg/m 2 ), waist circumference (cm), weight (kg), total cholesterol (TCH, mmol/L), triglycerides (TG, mmol/L), low-density lipoprotein (LDL-C, mmol/L), fasting plasma glucose (FPG, mmol/L), Hemoglobin A1c (HbA1c, %), CD62p, PAC-1 (P<0.05), and negatively correlated with the change of high density lipoprotein (HDL-C, mmol/L) and NO (P<0.05). Multiple linear regression analysis showed that the changes of NO, CD62p and PAC-1 were independent risk factors affecting the changes of platelet aggregation rates. is of


Background
Type 2 diabetes mellitus (T2DM) is a progressive metabolic disease characterized by chronic hyperglycemia. High glucose toxicity, lipotoxicity and insulin resistance are components of the pathophysiology of T2DM, which mainly affect the integrity of the blood vessel wall, leading to increased in ammation, endothelial dysfunction, enhanced platelet aggregation and coagulation factor dysfunction [1][2][3]. In addition to causing microvascular complications (retinopathy, nephropathy, and neuropathy), this complicated pathophysiology also leads to a 2-4 times increase in the risk of thrombosis and cardiovascular disease (CVD) in patients with T2DM [4,5]. The survey shows that CVD is the main cause of death in T2DM patients, accounting for 65% of the mortality rate in T2DM patients [6].
Therefore, the treatment of T2DM is not only to control blood sugar, but also to reduce other risk factors of CVD, such as obesity, hypertension, hyperlipidemia, and blood hypercoagulability.
Many studies have con rmed that diabetic patients are often accompanied by excessive activation of platelets, which can easily lead to thrombosis and induce adverse cardiovascular events [7]. Therefore, anti-platelet aggregation therapy has become an indispensable part of preventing cardiovascular events in T2DM. In clinical work, the method of measuring platelet aggregation function is often used to re ect the platelet activation state [8]. This method is to expose platelets to different inducers (such as ADP, collagen, epinephrine, AA) in vitro, and use light transmission aggregation measurement (LTA) to monitor platelet aggregation ability. It is often used clinically to monitor the effect of antiplatelet therapy.
Both CD62p and PAC-1 are highly sensitive and speci c markers that re ect the activation state of platelets. Among them, CD62p is a platelet activation-dependent granular membrane protein, also known as GMP-140, also known as P-Selectin, a member of the selectin family (one of the serum endothelial adhesion markers, which also includes vascular cell adhesion molecule-1 (VCAM-1), intracellular adhesion molecule-1 (ICAM-1)). PAC-1 is the GPIIb/IIIa complex brinogen receptor. Several studies have con rmed that these two indicators can be used in the monitoring and early prevention of platelet activation in hypercoagulable diseases [9].
As a new type of hypoglycemic drugs, glucagon-like peptide-1 (GLP-1) receptor agonists have received more and more attention in recent years. In addition to lowering blood sugar, these drugs also showed good effects in reducing the cardiovascular risk of T2DM [10,11]. Exenatide is a commonly used clinical GLP-1 receptors (GLP-1R) agonist. A number of clinical trials that have been completed so far have found that [12][13][14][15] exenatide can not only effectively reduce the blood glucose level of patients with T2DM, but also has obvious effects other than hypoglycemic effects, such as losing weight, lowering blood pressure, regulating blood lipids, and improving endothelial dysfunction caused by hyperglycemia and hyperlipidemia. Studies have con rmed that GLP-1R are high expressed in mouse and human platelets [16]. Cameron-Vendrig et al. [17] con rmed that the GLP-1R agonist exenatide can not only inhibit the aggregation of human and mouse platelets in vitro, but also inhibit mice arterial thrombosis in vivo. At present, the research on the effect of GLP-1R agonists on platelet activation function is mostly limited to animal experiments and in vitro experiments [16][17][18], and clinical studies on the effect of platelet activation function in T2DM patients have not been reported.
To this end, we designed and conducted a small sample control study to explore the changes of platelet aggregation and its activation markers CD62p and PAC-1 in peripheral blood of patients with T2DM after exenatide treatment, as well as the possible related factors. It aims to provide a theoretical basis for the clinical application of GLP-1R agonists and the prevention and treatment of diabetes-related cardiovascular diseases.

Study object
A total of 30 newly diagnosed T2DM patients (case group) admitted to the Department of Endocrinology of the First A liated Hospital of Anhui Medical University from October 2018 to November 2019 were selected, including 20 males and 10 females. Inclusion criteria: male or female aged 20-75 years; BMI 24 40 kg/m2; HbA1c 7.5 ~ 10%; glutamate decarboxylase antibody was negative. Received lifestyle intervention for at least 1 month; never received never received hypoglycemic drugs (including oral never received hypoglycemic drugs, GLP-1R agonists, insulin, etc.); not taking lipid-regulating drugs and antihypertensive drugs; not taking antiplatelet and anticoagulant drugs; not taking any medication that affects weight. Exclusion criteria: intolerance to GLP-1R agonists; history of blood system and chronic infectious diseases; cardiovascular and cerebrovascular events in the past 6 months; pregnant patients or recent pregnancy planners; type 2 diabetes with acute metabolic disorders; obvious abnormal heart, liver and kidney function; a history of medullary thyroid cancer; a history of gastrointestinal, pancreatic disease, and gastrointestinal surgery; received weight loss treatment in the past 6 months; secondary to other endocrine diseases and symptomatic obesity caused by other reasons; those who cannot cooperate to complete clinical research.
In addition, 30 patients with normal glucose tolerance (control group) who received physical examination at the Health Management Center of the First A liated Hospital of Anhui Medical University during the same period were selected, including 21 males and 9 females. Inclusion criteria: males or females aged 20 to 75 years; BMI 24 to 40 kg/m 2 ; normal glucose tolerance (FPG <6.1mmol/L, HbA1c<5.7%); not taking lipid-regulating drugs and antihypertensive drugs; not taking antiplatelet and anticoagulant drugs; not taking any medication that affects weight.

This study was approved by the Medical Ethics Committee of the First A liated Hospital of Anhui
Medical University, and all subjects signed the informed consent form.

Study methods
Individuals included in the case group were supervised by professionals in the elds of diet, exercise, and blood sugar monitoring throughout the study. Also, they were commenced on exenatide treatment (5 μg, sc, bid). Four weeks later, the dose of exenatide was increased to 10 μg (sc, bid), and the treatment continued for another four weeks, with a total of 8 weeks of exenatide treatment. One telephone follow-up was conducted every week while an outpatient follow-up was scheduled monthly. If nausea, vomiting, or other digestive symptoms occurred during the higher-dose exenatide treatment (10 μg, sc, bid), a reduced dose of exenatide (5 μg, sc, bid) should be administered till the higher-dose treatment was resumed a week later. If the patient responded poorly to an increased dose of exenatide, he/she should continue the lower-dose treatment till the end of the study. Should serious hyperglycemia (FPG levels at two different time points both exceeding 13.9 mmol/L) develop during the study, the patient was approved to withdraw from the study upon completion of relevant examinations.
Ulnar vein blood was drawn after fasting for at least 8 hours. Blood samples were collected, processed, and analyzed as required. A platelet aggregation system (Helena Laboratories, US) was used by a professional to determine the rates of platelet aggregation (to epinephrine (EPI), collagen, arachidonic acid (AA), and adenosine diphosphate (ADP)) pre-and post-exenatide treatment with light transmittance aggregometry (LTA). Three-color ow cytometry (FACSCalibur Flow Cytometer, BD, US) was performed for detection of platelet CD62p and PAC-1 (represented by the percentage of positive platelet for each activation marker) in whole-blood samples at the commencement and the end of exenatide treatment. an automatic biochemical analyzer (COBAS 8000, Roche, Switzerland) was utilized to measure the levels of fasting plasma glucose (FPG), total cholesterol (TCH), triglyceride (TG), low-density lipoprotein (LDL-C), and high-density lipoprotein (HDL-C) at Week 0 and at Week 8. An automatic coagulation analyzer (AYL-4-013, Stago, France) was used for determination of prothrombin time (PT), international normalized ratio (INR), brinogen (FIB), activated partial thromboplastin time (APTT), brin degradation products (FDP), and d-dimer. An automatic hematology analyzer (XE-2100, Sysmex, Japan) was applied to the calculation of blood platelet counts (BPCs) at Week 0 and at Week 8. A Glycosylated hemoglobin analyzer (AYFY24319, BIO-RAD, Japan) was utilized to quantify glycosylated hemoglobin A1c (HbA1c) pre-and post-exenatide treatment. FPG was measured by the hexokinase method. TG and TCH were analyzed using enzymatic colorimetric methods. High-density lipoprotein cholesterol (HDL-C) was measured by direct methods. Low-density lipoprotein cholesterol (LDL-C) values were calculated using equations. Measurements of PT, FIB, and APTT were performed in the magnetic bead-based settings. Values of INR and PT activity were calculated directly. FDP and d-dimer were detected via immunoturbidimetry. PLT was analyzed using optical methods. HbA1c values were obtained from high-pressure liquid chromatography (HPLC). Nitrogen monoxide (NO) test: 2.0 mL of venous blood was collected, centrifuged and stored at -70°C for future use. After all blood samples were collected, the nitrate reductase assay (Nanjing Jiancheng Bioengineering Institute) was performed as instructed by the manufacturer. The concentration of serum NO was measured based on the optical density (OD) at 550 nm. Each patient's body height (m), body weight (kg), and waist were measured to calculate his/her body mass index (BMI) value according to the following equation: BMI = body weight (kg) / body height (m) 2 . Differences (Δ) pre-and postexenatide treatment were calculated using the values of the above-mentioned clinical indicators measured at Week 0 and at Week 8.

Statistical methods
The SPSS 21.0 statistical software was used to process the data, and the measurement data conforming to the normal distribution were expressed as mean ± standard deviation (x̄ ± s). The comparison of variables between the case group and the control group by independent sample t test and 2 test; the comparison of variables before and after treatment in the case group by paired t test. Pearson correlation analysis explored the correlation between the changes in platelet aggregation function before and after exenatide treatment and the changes in other variables. Multiple linear regression analysis was used to explored the independent in uencing factors of platelet aggregation function before and after exenatide treatment. Test level α=0.05, two-sided P<0.05 indicates that the difference is statistically signi cant.

Result
Comparison of clinical data and biochemical indicators between case group and control group Compared with the control group, the levels of TG, FPG, HbA1c, FIB, CD62p and PAC-1 before treatment in the case group and the platelet aggregation rates induced by epinephrine, ADP and AA in case group were signi cantly higher than those in the control group (P < 0.05), the levels of HDL-C and NO were lower than those of the control group (P < 0.05), and there was no signi cant difference in other clinical indicators between the two groups (P > 0.05), Table 1  Comparison of clinical indicators after 0 weeks and 8 weeks of exenatide treatment in case group subjects After 8 weeks of exenatide treatment, the levels of weight, BMI, waist, SBP, TCH, TG, LDL-C, FPG, HbA1c, FIB, CD62p and PAC-1, as well as platelet aggregation rates induced by epinephrine and AA were signi cantly decreased in case group compared with 0 weeks (P < 0.05), the levels of HDL-C and NO were signi cantly increased (P < 0.05), while DBP, PT, PTA, APTT, INR, TT, FDP, DD and platelet aggregation rates (collagen and ADP as inducers) had no signi cant differences before and after treatment (P > 0.05), Table 3, Table 4, Fig. 1.   Pearson correlation analysis showed that the Δ platelet aggregation rates (ΔEPI, ΔAA) before and after 8 weeks of exenatide treatment in the case group were positively correlated with ΔBMI, Δ weight, Δ waist, ΔTCH, ΔTG, ΔLDL-C ΔFPG, ΔHbA1c, ΔCD62p and ΔPAC-1, and negatively correlated with ΔHDL-C and ΔNO (P < 0.05), and no signi cant correlation with ΔSBP and ΔFIB (P > 0.05), Fig. 3.
Factors affecting platelet aggregation before and after treatment with exenatide According to the results of Pearson correlation analysis, ΔBMI, Δ waist, Δ weight, ΔTCH, ΔTG, ΔLDL-C, ΔHDL-C, ΔFPG, ΔHbA1c, ΔNO, ΔCD62p, and ΔPAC-1 were included in the multiple linear regression analysis as independent variables, and Δ platelet aggregation rates (ΔEPI, ΔAA) as the dependent variable. The results show that ΔNO, ΔCD62p and ΔPAC-1 are independent risk factors affecting Δ platelet aggregation rates (ΔEPI ΔAA) (P < 0.05), Table 5. Annotate: Δ Weight: change in weight before and after exenatide treatment Δ Waist: change in waist before and after exenatide treatment ΔBMI change in body mass index before and after exenatide treatment ΔTCH change in total cholesterol before and after exenatide treatment ΔTG: change in triglycerides before and after exenatide treatment ΔLDL-C: change in low density lipoprotein cholesterol before and after exenatide treatment ΔHDL-C: change in high density lipoprotein cholesterol before and after exenatide treatment ΔFPG: change in fasting plasma glucose before and after exenatide treatment ΔHbA1c: change in hemoglobin A1c before and after exenatide treatment ΔNO: change in nitric oxide before and after exenatide treatment; ΔPAC-1: change in platelet activation complex-1 before and after exenatide treatment ΔEPI changes in platelet aggregation rate with epinephrine as inducer before and after exenatide treatment ΔAA: changes in platelet aggregation rate with arachidonic acid as inducer before and after exenatide treatment. CI: con dence interval

Discussion
Despite the fact that all subjects in this study were included in the category of overweight or obesity, in the case group, the accumulation of cardiovascular risk factors in T2DM patients is more obvious, including a signi cant increase in the blood sugar, blood pressure, blood fat, and FIB levels, which indicated that compared with those with normal glucose tolerance in the control group, the T2DM patients appeared to be at higher risk of developing CVD. Apart from this, the levels of FIB and platelet activation markers, and the rates of platelet aggregation in the case group were signi cantly higher than in the control group, indicating a relatively high degree of functional activation of platelets, a dramatic enhancement of aggregation function, and a prothrombotic state in the T2DM patients.
The molecular mechanism of increased platelet aggregation and adhesion in T2DM patients is not yet fully understood. Existing studies have found that: 1) insulin resistance becomes less effective in inhibiting hyperfunction of platelets [19]; 2) in a high-glucose environment, the number of glycosylated products increases due to non-enzymatic glycosylation of the platelet membrane, which consequently reduces the membrane uidity, promotes platelet aggregation and adhesion and improves the sensitivity of platelets to pro-aggregation substances [20,21]; 3) an increase of P2Y12 on platelet in patients with diabetes mellitus (DM) can induce protein kinase A (PKA)-mediated vasodilation, leading to reduced cyclic adenosine monophosphate (cAMP)-dependent phosphorylation of vasodilator-stimulated phosphoproteins (VASP-Ps) and lower bioavailability of NO in endothelial cells [20,22]. Additionally, hypertriglyceridemia and very-low-density lipoprotein (VLDL) can trigger platelet hyperfunction in DM patients through interaction between apolipoprotein E (apoE) and platelet LDL receptor [21]. A study [23] revealed that in T2DM, the increase in the level of FIB in the peripheral blood of patients with T2DM may be caused by the combined effect of in ammatory cytokines and insulin on the liver, resulting in increased liver synthesis.
As is well known, the potential cardiovascular protective activity of GLP-1R agonists is attributed to the anti-hypoglycemic effects of these agents, which help improve insulin resistance, aid weight loss, lower blood pressure, modify lipid distribution and directly act on the myocardium and blood vessel endothelium. The study by Martinez et al. [24] showed an average weight loss of 3.9 kg after six months of exenatide treatment. In a 30-week randomized, double-blind, controlled clinical trial, it was found that exenatide could signi cantly reduce SBP in patients with newly diagnosed T2DM, while DBP was also slightly reduced [25]. Sun et al. [26] carried out a meta-analysis of 13 studies (GLP-1R agonist treatment requires at least a 6-week duration) and pointed out that GLP-1R agonists could moderately reduce the LDL-C, TCH, and TG levels. In this study, it was found that following the 8-week exenatide treatment, the T2DM patients lost 4.8 kg, the SBP level was reduced by 5.5 mmHg, and the DBP level was slightly decreased,however, the differences pre-and post-treatment were not statistically signi cant. Likewise, blood lipids were also improved after 8 weeks of exenatide treatment, manifested by reduced TCH, TG, and LDL-C levels and an increase in HDL-C, which was basically consistent with the aforementioned ndings. Presently, it is widely accepted that GLP-1R agonist-mediated SBP reduction is probably associated with weight loss, improved endothelial function, natriuresis, and relaxation of renal vascular smooth muscles [27][28][29].
In fact, in addition to the above-mentioned bene cial effects of reducing the risk for CVD, it has become a popular interest of research to investigate the effects of GLP-1R agonists on the platelet aggregation/activation function [17,18]. Hemostatic time, clotting time, platelet activating markers, and the rate of platelet aggregation serve common indicators for the evaluation of platelet function [30][31][32].
The rate of platelet aggregation is considered as an important indicator for the platelet aggregation function, which plays an important role in the prevention, treatment, and monitoring of thrombosis. A range of speci c markers have been released in the process of platelet activation, especially CD62p and PAC-1 [33]. CD62p is a transmembrane protein stored in α granules of platelets and Weibel-Palade bodies of endothelial cells. Under normal circumstances, CD62p only has a relatively low level of endothelial surface expression. In response to stimulation, the surface expression of platelets increases sharply to mediate the adhesion function of activated endothelial cells, mononuclear cells, and neutrophils using the lectin-like, N-terminal domain. These activated cells not only promote brin deposition but also play a role in in ammatory response and thrombosis. Activated platelets treated with anti-CD62p antibodies no longer have adhesive attraction to each other, which shows that as one of the important markers of platelet activation, CD62p is able to mediate the adhesion between activated platelets or between endothelial cells and leukocytes [23]. PAC-1, as a platelet membrane glycoprotein IIb/IIIa complex and abundant platelet surface glycoprotein, consists of binding sites speci c to attachment proteins such as brinogen, bronectin, and von Willebrand factor (vWF). Under normal conditions, PAC-1 exists as a monomer without ligand-binding ability. Upon activation of platelets, the PAC-1 receptor reveals its ligand-binding mechanism and binds to speci c attachment proteins, which promotes platelet-brinogen-platelet aggregation and eventually contributes to platelet thrombus formation. Apart from this, PAC-1 also plays a role in intracellular signal transduction and thus it is useful for direct detection of activated platelets [34] In terms of coagulation function, this study found that the anticoagulant indexes PT, APTT, INR, TT, PTA, FDP, and DD did not signi cantly change after 8 weeks of exenatide treatment, and the FIB in T2DM patients after 8 weeks of exenatide treatment was signi cantly decreased compared with that before treatment. As far as we know, this nding is the rst to be reported. It is known that FIB is a protein with coagulation function synthesized in the liver, and is the coagulation factor with the highest content in plasma. As a marker of thrombosis and in ammation, FIB is an important reaction substrate for thrombosis, participates in the key steps of thrombosis, and plays a very important role in the pathogenesis of cardiovascular diseases. In this study, it was found that after 8 weeks of exenatide treatment, plasma FIB levels in T2DM patients were signi cantly reduced, but the relevant mechanism is still unclear, which may be related to exenatide inhibiting in ammation and improving insulin resistance [23].
In terms of platelet activation status, this study found that after 8 weeks of exenatide treatment in T2DM patients (newly diagnosed overweight or obese), platelet activation markers represented by CD62P and PAC-1, and platelet aggregation rates induced by epinephrine and AA were signi cantly reduced. Pearson and multiple linear regression correlation analysis showed that after exenatide treatment, the Δ platelet aggregation rates (ΔEPI, ΔAA) were signi cantly positively correlated with ΔCD62p and ΔPAC-1, and ΔCD62p and ΔPAC-1 were independent factors affecting Δ platelet aggregation rates, indicating that Exenatide can signi cantly inhibit the activation of platelets in patients with T2DM, thereby decreasing the ability of peripheral blood platelets to aggregate. It is worth mentioning that in our study, no signi cant changes in the platelet aggregation rate with ADP and collagen as inducers were observed after 8 weeks of exenatide treatment. This suggested that there may be differences in the detection results of platelet aggregation function with different inducers in T2DM patients, and the application of single inducer may lead to bias in the study results, while the detection combined with multiple inducers can better guide clinical practice. Some studies have found that when the antiplatelet aggregation function is detected by turbidimetric method, the decrease of platelet aggregation rate is not obvious at the early stage, but it will decrease signi cantly over time [35]. We speculated that the platelet aggregation rate with ADP and collagen as inducers in this study might signi cantly decrease with the further extension of exenatide treatment, which requires further research to con rm.
In order to further understand the other in uencing factors of Δ platelet aggregation rate after treatment with GLP-1R agonist exenatide, we also explored the correlation between Δ platelet aggregation rates and cardiovascular risk factors in T2DM patients, and the results showed that Δ platelet aggregation rates (ΔEPI, ΔAA) are positively correlated with ΔBMI, Δ waist, weight, ΔTCH ΔTG ΔLDL-C ΔFPG and HbA1c, negatively correlated with ΔHDL-C and ΔNO, and have no correlation with ΔSBP and ΔFIB. Among them, ΔNO is an independent risk factor that affects the platelet aggregation rates, suggesting that the effect of exenatide on improving platelet function may be related to weight loss, lipid regulation, blood sugar reduction, and increase of NO concentration. Simeone et al. [36] randomly assigned 62 patients (obese subjects with prediabetes or early T2DM) to the liraglutide group and the lifestyle intervention group according to 1:1, and the two groups were monitored until the weight loss goal (-7% of initial body weight) was achieved. The results showed that U-11-dehydro-TXB2 (a metabolite of thromboxane that re ects platelet activation) was signi cantly reduced after achievement of the weight loss target regardless of the intervention arm, suggesting that the inhibition of platelet activation by GLP-1R agonists is related to weight loss. Dyslipidemia is closely related to platelet activation [37]. Studies have shown that dyslipidemia, especially hypertriglyceridemia, is directly related to platelet function [38]. Ebara et al. [7] believe that oxidized HDL is negatively related to blood coagulation and brinolysis in patients with T2DM. In addition, many studies have con rmed that hyperglycemia can have a signi cant adverse effect on platelet function [20,21]. In addition to decreasing the uidity of the platelet membrane and promoting platelet aggregation, coagulation factors and PAI-1 (plasminogen activator inhibitor-1) will also increase under the condition of hyperglycemia, breaking the balance of coagulation and brinolysis, and promoting thrombosis [39]. Besides improvements in blood sugar control and weight loss, the inhibitory effect of GLP-1R agonists on platelets has been veri ed in animal models [17,18] and healthy volunteers [40]. Therefore, GLP-1R agonists have the potential, at least in theory, to regulate platelet activation directly through the effect on platelet function and indirectly through the control of body weight and metabolism. The above arguments further support the results of this study.
NO is considered to be the most important vasodilator factor secreted by vascular endothelial cells, which can relax vascular smooth muscle and dilate blood vessels. Liraglutide has been reported to inhibit platelet activation in animal models [17] and healthy volunteers [40] by increasing the concentration of NO. In vitro studies we have also con rmed that liraglutide can increase the expression of endothelial nitric oxide synthase (eNOS), and reduce the expression of inducible nitric oxide synthase (iNOS) at the levels of transcription and translation by inhibiting nuclear factor kappa B (NF-p65) phosphorylation to improve endothelial cell function, and prevent diabetic atherosclerosis [41]. Similarly, Jia and colleagues also con rmed that GLP-1R agonists regulate platelet activity by inducing eNOS-dependent mechanisms, increase the bioavailability of NO in vascular endothelial cells, and improve vascular function [18]. The above study results all suggest that GLP-1R agonists reduce platelet activation and inhibit thrombosis by enhancing the production and utilization of NO, which is basically consistent with the results of this study.
In fact, the molecular mechanism of the effect of GLP-1R agonists on platelet aggregation/activation function has not yet been elucidated. It is currently recognized that the adhesion/aggregation process of platelets is regulated by the balance between procoagulant and anti-aggregation circulatory agents (such as NO). NO stimulates soluble guanylate cyclase (sGC) in platelets, activates cyclic guanosine monophosphate (cGMP) and protein kinase G (PKG), and then inhibits platelet activation through various pathways. In addition, cGMP can indirectly increase cellular cAMP levels by inhibiting 3'phosphodiesterase to activate PKA [42] and induce eNOS activity [43]. The synergistic effect of cGMP and cAMP inhibit platelet aggregation [44,45]. Secondly, cGMP can also inhibit the activation of phosphoinositide3-kinase (PI3K) [46], causing the activation of the GP IIb-IIIa brinogen receptor [47]. In addition to the cGMP-dependent pathway described above, there is evidence that NO can also regulate platelet function independently of cGMP without being affected by sGC [48][49][50]. Therefore, any change in eNOS activity or/and NO bioavailability by GLP-1R agonists is a key factor in determining platelet function [18,51].

Informed consent
Informed consent was obtained from all individual participants included in the study.

Data Availability
The data sets used to support the ndings of this study are available from the corresponding author upon request.

Declarations of interest
The authors declare no con ict of interest.

Authors' contributions
Yaqin Zhang and Ruofei Chen conceived and designed the study, analyzed the data, and wrote the paper; Yangyang Jia collected data with Yaqin Zhang; Mingwei Chen and Zongwen Shuai discussed results and advised during the completion of the study.
All authors read and approved the nal manuscript. Changes in platelet aggregation rates after 8 weeks of exenatide treatment Annotate: EPI-pre platelet aggregation rate before exenatide treatment (epinephrine) EPI-post platelet aggregation rate after exenatide treatment (epinephrine) ADP-pre platelet aggregation rate before exenatide treatment (adenosine diphosphate) ADP-post platelet aggregation rate after exenatide treatment (adenosine diphosphate) COLL-pre platelet aggregation rate before exenatide treatment (collagen) COLL-post platelet aggregation rate after exenatide treatment (collagen) AA-pre platelet aggregation rate before exenatide treatment (arachidonic acid) AA-post platelet aggregation rate after exenatide treatment (arachidonic acid).

Figure 3
Correlation between changes in platelet aggregation rates (ΔEPI ΔAA) and changes in other clinical parameters (r) Annotate: ΔBMI change in body mass index before and after exenatide treatment ΔSBP: change of systolic blood pressure before and after exenatide treatment Δ Weight change in weight before and after exenatide treatment Δ Waist change in waist before and after exenatide treatment ΔTCH change in total cholesterol before and after exenatide treatment ΔTG change in triglycerides before and after exenatide treatment ΔLDL-C change in low density lipoprotein cholesterol before and after exenatide treatment ΔHDL-C change in high density lipoprotein cholesterol before and after exenatide treatment ΔFPG change in fasting plasma glucose before and after exenatide treatment ΔHbA1c change in hemoglobin A1c before and after exenatide treatment ΔNO: change in nitric oxide before and after exenatide treatment; ΔFIB change in brinogen before and after exenatide treatment ΔPAC-1: change in platelet activation complex-1 before and after exenatide treatment ΔEPI changes in