Aspirin, clopidogrel and prasugrel monotherapy in patients with type 2 diabetes mellitus: a double-blind randomised controlled trial of the effects on thrombotic markers and microRNA levels

Background Despite increased atherothrombotic risk in type 2 diabetes mellitus, (T2DM) the best preventative antithrombotic strategy remains undetermined. We defined the effects of three antiplatelet agents on functional readout and biomarker kinetics in platelet activation and coagulation in patients with T2DM. Materials and methods 56 patients with T2DM were randomised to antiplatelet monotherapy with aspirin 75 mg once daily (OD), clopidogrel 75 mg OD or prasugrel 10 mg OD during three periods of a crossover study. Platelet aggregation (PA) was determined by light-transmittance aggregometry and P-selectin expression by flow cytometry. Markers of fibrin clot dynamics, inflammation and coagulation were measured. Plasma levels of 14 miRNA were assessed by quantitative polymerase chain reactions. Results Of the 56 patients, 24 (43%) were receiving aspirin for primary prevention of ischaemic events and 32 (57%) for secondary prevention. Prasugrel was the strongest inhibitor of ADP-induced PA (mean ± SD maximum response to 20μmol/L ADP 77.6 ± 8.4% [aspirin] vs. 57.7 ± 17.6% [clopidogrel] vs. 34.1 ± 14.1% [prasugrel], p < 0.001), P-selectin expression (30 μmol/L ADP; 45.1 ± 21.4% vs. 27.1 ± 19.0% vs. 14.1 ± 14.9%, p < 0.001) and collagen-induced PA (2 μg/mL; 62.1 ± 19.4% vs. 72.3 ± 18.2% vs. 60.2 ± 18.5%, p < 0.001). Fibrin clot dynamics and levels of coagulation and inflammatory proteins were similar. Lower levels of miR-24 (p = 0.004), miR-191 (p = 0.019), miR-197 (p = 0.009) and miR-223 (p = 0.014) were demonstrated during prasugrel-therapy vs. aspirin. Circulating miR-197 was lower in those cardiovascular disease during therapy with aspirin (p = 0.039) or prasugrel (p = 0.0083). Conclusions Prasugrel monotherapy in T2DM provided potent platelet inhibition and reduced levels of a number of platelet-associated miRNAs. miR-197 is a potential marker of cardiovascular disease in this population. Clinical outcome studies investigating prasugrel monotherapy are warranted in individuals with T2DM. Trial registration EudraCT, 2009-011907-22. Registered 15 March 2010, https://www.clinicaltrialsregister.eu/ctr-search/trial/2009-011907-22/GB.

Background Thrombotic events are associated with a large burden of morbidity and mortality in the general population [1], with an even higher risk in patients with type 2 diabetes mellitus (T2DM) [2]. Enhanced platelet activation and altered fibrin clot properties are central pathological processes in the development of thrombosis in T2DM, thus increasing the risk of cardiovascular events and contributing to adverse clinical outcome following vascular ischaemia [3][4][5].
Antiplatelet drugs for the treatment and prevention of atherothrombosis have largely focussed on two pathways of platelet activation: thromboxane A 2 generation, which is blocked by the irreversible cyclo-oxygenase inhibitor aspirin, and adenosine diphosphate-(ADP-) induced amplification of platelet activation via the P2Y 12 receptor, which is irreversibly inhibited by thienopyridines such as clopidogrel and prasugrel [6]. In combination with aspirin, thienopyridines reduce the risk of thrombotic events after acute coronary syndrome (ACS) [7], but the protective effects can vary according to the agent used and the population studied. Following coronary ischaemia requiring percutaneous intervention, prasugrel has shown enhanced vascular protective properties compared with clopidogrel in T2DM patients without an increase in bleeding risk, in contrast to individuals without T2DM [8,9].
In the clopidogrel vs aspirin in patients at risk of ischemic events (CAPRIE) study, clopidogrel, used as single antiplatelet therapy (SAPT) showed better protection against vascular ischaemia compared with aspirin monotherapy in patients with T2DM [10], an effect that was even more pronounced in insulin-treated subjects [11]. In contrast to clopidogrel, prasugrel as SAPT has not been well studied and data on patients with T2DM are scarce.
Studies of aspirin for the primary prevention of cardiovascular disease in patients T2DM have been disappointing, with little or no reduction in vascular ischaemic events at the expense of a significant increase in bleeding risk [12][13][14][15][16]. Similarly, the pharmacokinetics and clinical efficacy of clopidogrel, a pro-drug, display significant variability between individuals due to differences in activity of cytochrome P450 2C19 [17]. Prasugrel, whilst also a pro-drug, is activated by a different, more predictable metabolic pathway and therefore offers better inter-individual consistency of effect [18].
In addition to the effects on platelets, aspirin and P2Y 12 inhibitors may modulate the fibrin network [19,20] and affect vascular inflammatory pathways [21,22]. In order to assess a functional readout and to gain further mechanistic insight, platelet function tests have been used to assess the response to antiplatelet agents. Micro-ribonucleic acids (miRNAs) are emerging as an adjunct to our understanding and assessment of platelet function [23][24][25]. From a translational point of view, detection of a variety of miRNAs has been linked with clinical outcomes including ischaemic heart disease [26].
Prasugrel monotherapy in individuals with T2DM may offer superior anti-thrombotic properties compared with aspirin or clopidogrel. The aim of this study was to comprehensively characterise and compare the effects of the three drugs on platelet function, fibrin network characteristics, inflammation and expression of miRNAs in a cohort of T2DM patients.

Study design
We performed a single-centre, double-blind, crossover, randomised controlled trial of patients with a confirmed diagnosis of T2DM. Patients were eligible to participate if they were aged 18-75 years, already on treatment with aspirin 75 mg once-daily (OD) and able to give informed consent. Eligible participants receiving aspirin 75 mg OD were randomised 1:1 to one of two medication sequences ( Fig. 1). All patients continued aspirin 75 mg OD for an initial lead-in period of 14 days. One half then received clopidogrel 75 mg OD for 28 days then prasugrel 10 mg OD for 28 days, whilst the other half received prasugrel 10 mg OD for 28 days followed by clopidogrel 75 mg OD Fig. 1 Design of the study. mg milligrams, OD once daily, R point of randomisation for 28 days. Aspirin was discontinued after study completion if it was deemed clinically unnecessary. Randomisation was performed by shuffled sealed opaque envelopes prepared by a member of pharmacy staff. Both patients and investigators were blind to treatment allocation until all study data was collected. A range of pharmacodynamic measurements were made at the end of each treatment period. The study was approved by the United Kingdom National Health Service Research Ethics Service (reference 09/H1307/110). Written consent was obtained from participants before any study activities took place.
Exclusion criteria included: any type of diabetes other than T2DM; any coagulation disorder; neoplastic disease; history of ACS within 3 months of enrolment; history of stroke or transient ischaemic attack; history of deep venous thrombosis or pulmonary embolism; treatment with oral anticoagulant or non-steroidal antiinflammatory drugs; abnormal liver enzyme tests defined as alanine transferase > threefold upper limit of normal; any previous or current upper gastrointestinal pathology; weight < 60 kg; and women of child-bearing age and refusing to use contraception.

Blood samples
Venous blood samples were collected by venepuncture, using syringe and 18G needle, and anticoagulated with trisodium citrate dihydrate 3.13%. Platelet-rich plasma was prepared by centrifugation for 10 min at 200×g and platelet-poor plasma was prepared by further centrifugation for 10 min at 1500×g. All analyses described below were undertaken by individuals blinded to the sample details and type of antiplatelet treatment.

Platelet P-selectin expression
Surface expression of platelet P-selectin after stimulation with 0.3, 1, 3, 10 & 30 μmol/L ADP was quantified using flow cytometry and expressed as percentage positive events, as previously described [28].

Fibrin clot turbidimetric analysis
High-throughput turbidimetric analysis was performed as previously described and validated [29][30][31][32]. Briefly, citrated platelet-poor plasma samples were mixed with standard lysis and activation mixes to form acellular clots. Serial absorbance was measured using an automated plate reader during clot formation until lysis was achieved. Variables recorded were lag time, representing the period from the addition of clot activation mix to the start of clot formation (a measure of clotting tendency), final clot turbidity (maximum absorbance, a representation of fibre thickness and clot density), and lysis time (time from full clot formation to 50% lysis, a measure of fibrinolysis potential).

RNA isolation and miRNA quantification
Quantification of miRNAs with known relevance to platelet function and cardiovascular disease (miRs 21, 24, 27b, 28, 93, 122, 126, 150, 191, 197, 223, 320, 451a and 486) was performed in platelet-poor plasma samples using reverse transcription quantitative polymerase chain reaction (RT-qPCR), as previously described [24,26]. Total RNA was isolated using the miRNeasy Mini kit (Qiagen, Hilden, Germany). In brief, 100 µL of plasma were combined with 694.75 µL of QIAzol, 4 µL of diluted Caenorhabditis elegans miR-39-3p (cel-miR-39) spike-in and 1.25 µL carrier MS2. Following a brief incubation at ambient temperature, 140 µL of chloroform were added and the solution was mixed vigorously. Samples were then centrifuged at 13,500 relative centrifugal force (rcf ) for 15 min at 4 °C. The upper aqueous phase was carefully transferred to a new tube and 1.5 volumes of ethanol were added. The samples were then applied directly to columns and washed according to the company's protocol. Total RNA was eluted with 35 µL of nuclease-free water. A fixed volume of 3 μL of the 35μL RNA eluate was used as input for reverse transcription (RT) reactions. MiRNAs were reverse-transcribed using Mega Cel-miR-39 was used for normalisation and as a quality control. Quantification results were calibrated with pooled RNA from 50 samples. Quantification cycle (Cq) values > 32 were considered to fall below the limit of detection. Relative quantification was performed with Microsoft Excel, version 15.32 for Mac using the 2 (−∆∆Cq) method [35].

Statistical analysis
The three treatments were compared by repeated measures ANOVA with Greenhouse-Geisser correction. The primary endpoints of the study were platelet aggregation responses to ADP, collagen and AA; platelet P-selectin expression and fibrin clot dynamics. Other analyses were exploratory. For variables with a significant difference (p < 0.05) between the treatments, pairwise comparisons were performed using Bonferroni correction. SPSS statistics v25 (IBM software) was used for these analyses and graphical representations generated using GraphPad PRISM v7. Correlation and subgroup analyses were performed using RStudio v1.1.456: adjustment for multiple comparisons was not made for these as they were exploratory and intended for hypothesis generation.
A total of 56 patients were needed to detect a 10% difference in platelet aggregation response to various agonists comparing the different therapies at p < 0.05 and 90% power, based on the assumption that the common standard deviation of the response variable is 16%. The study also had the power to detect a 7% difference in clot final turbidity based on the assumption that standard deviation for the response variable is 11% (p < 0.05, 90% power).

Recruitment and participant characteristics
Of 310 patients who were approached, 64 were enrolled and 56 completed the study (Additional file 1: Figure S1). Baseline characteristics are shown in Table 1. Comparing platelet aggregation responses when receiving standard-of-care aspirin 75 mg OD to those at the end of the study aspirin period, there were no differences (Additional file 1: Table S1).

Coagulation factors and inflammatory markers
No significant differences in fibrinogen, circulating leukocyte count, CRP or complement C3 were observed between the treatments (all p > 0.05, Table 2).

Correlation between miRNA detection and other markers
Platelet  Table S3). No significant correlations were observed between ADP-or collagen-induced platelet aggregation and circulating miRNA levels (Additional file 1: Table S3). Of the fibrin clot parameters studied, there was a significant positive correlation between final clot turbidity    Table 3). Subgroup analysis of levels miR-21, miR-24, miR-191, miR-197 and miR-223 by HRPR status revealed no significant differences between the groups (Additional file 1: Table S4).

miR-126
Although levels of miR-126 have been linked to platelet and endothelial function in the general population [25], T2DM may be associated with lower detectable quantities [39]. P-selectin expression in response to ADP stimulation showed positive correlation with miR-126 (R = 0.27, p = 0.0004) (Additional file 1: Table S3, Additional file 1: Figure S4). On the other hand, we failed to show significant differences in quantification of miR-126 between the treatments and there was no evidence of a significant correlation between aggregation responses and miR-126 (Additional file 1: Table S2).

Effect of presence or absence of cardiovascular disease
Subgroup analysis by presence (n = 32) or absence (n = 24) of a history of macrovascular atheromatous disease (history of coronary artery disease, cerebrovascular ischaemia or peripheral arterial disease) at enrolment revealed no significant differences between the subgroups in markers of platelet aggregation during each of the three treatment periods (Additional file 1: Table S5).

Discussion
Patients with T2DM present particular challenges with regards to atherothrombotic protection. There is no current clear strategy of antiplatelet therapy for primary prevention in those with T2DM. For example, whilst some evidence suggests that aspirin therapy targeted by assessment of cardiovascular and bleeding risk may be of benefit [40], guidelines remain conflicted on the extent to which, if at all, aspirin therapy should be recommended in this situation [41,42]. P2Y 12 inhibitors are alternative antiplatelet agents to aspirin with the potential benefits of prolonged presence in the plasma that might overcome reduced aspirin effect due to high platelet turnover and avoiding gastric erosion [43]. Three orally-active drugs, clopidogrel, prasugrel and ticagrelor are commonly-available [6]. Monotherapy with clopidogrel offers only modest clinical benefits over aspirin in the setting of secondary prevention, but these may be amplified in those with T2DM [11]. Furthermore, the newer P2Y 12 inhibitors ticagrelor and prasugrel are more potent and consistent in effect than clopidogrel [44]. Certainly, when given in combination with aspirin, ticagrelor and prasugrel offer net clinical benefit after ACS [6]. Additionally, there is evidence that potent P2Y 12 inhibition may offer benefits in  Relative quantification of circulating miR-197 in participants with and without a history of cardiovascular disease during the three treatment periods. Large dots and lines represent mean ± SD. p values were generated by t-tests patients with T2DM and complications such as lower extremity arterial disease, in which ticagrelor improves microvascular flow, for example [45]. Although there has been concern over greater bleeding risk increasing the potency of P2Y 12 inhibition, recent evidence suggests, for example, ticagrelor may have similar safety to clopidogrel in patients at high risk of bleeding, such as those who are elderly with ST-elevation myocardial infarction [46], and low-dose prasugrel appears to be of comparable safety to clopidogrel when used in triple therapy (aspirin, P2Y 12 inhibitor and anticoagulant) for patients with atrial fibrillation undergoing percutaneous coronary intervention [47].
In this study, we have comprehensively characterised the effects of aspirin, clopidogrel and prasugrel when given as SAPT to patients with T2DM. Prasugrel provided the strongest effect on ADP-induced platelet aggregation, as did aspirin on AA-induced aggregation. Prasugrel also provided greater and more consistent inhibition of ADP-induced platelet aggregation when compared to clopidogrel, in agreement with previous studies on populations with and without diabetes receiving dual antiplatelet therapy [44,48] and of studies of prasugrel vs. clopidogrel loading doses when given as SAPT [49]. In concert with this was the fact that there was a large reduction in the proportion of patients with HRPR when receiving clopidogrel vs. aspirin, and again when receiving prasugrel vs. either comparator. P-selectin expression after stimulation with ADP followed a similar pattern to the aggregation responses. Collagen-induced platelet aggregation, perhaps representing the best global assessment of effects on platelet macroaggregation, was more strongly inhibited by prasugrel than aspirin and clopidogrel, suggesting that prasugrel acts as the most potent antiplatelet drug of the three when used as monotherapy in patients with T2DM.
For the first time, we studied a panel of miRNAs in patients with diabetes receiving three different antiplatelet agents, showing that potent P2Y 12 inhibition with prasugrel reduced detectable levels of miR-24, miR-191, miR-197 and miR-223 compared to aspirin. Although there were no significant differences between aspirin and clopidogrel nor clopidogrel and prasugrel, there did appear to be similar trends; thienopyridines reduced miRNA levels, an effect that was most pronounced with prasugrel. The known actions and associations of miRNAs are broad-ranging, but circulating levels of miR-21, miR-24, miR-197 and miR-223 are most strongly associated with platelets and platelet microparticles, along with miR-126 [26]. These, as well as miR-191, have been shown to be reduced by antiplatelet therapy in healthy volunteers, and some also in patients with cardiovascular disease [24,25]. Our study, in individuals with T2DM, supports the association between plasma levels of these miRNAs and platelet function.
Whilst there were positive correlations between miRNA levels and ADP-induced platelet P-selectin expression, there was no evidence of a significant correlation with ADP or collagen-induced aggregation, and we observed negative correlations with AA-induced aggregation in some cases. P-selectin expression, which is known to be reduced by P2Y 12 inhibitors but not aspirin [28], occurs when alpha granules fuse with the cell membrane upon platelet activation. These data suggest that plasma levels of miRNAs reflect primarily the tendency for platelet degranulation and indeed it has previously been shown that miRNAs may be involved in regulating degranulation [50]. This has not been reported before in individuals with diabetes receiving antiplatelet therapy and future research is required to understand the specific role of each of the miRNAs, which may help with risk stratification and/or uncover alternative therapeutic targets to control platelet activation in this population.
We saw lower plasma quantities of miR-197 in T2DM patients with known cardiovascular disease compared to those without, which is consistent with the previous finding in a general population cohort suggesting that lower miR-197 might be associated with increased risk of myocardial infarction [26,51]. In contrast to previous findings in the general population, we did not see evidence of an association between elevated levels of miR-126 and presence of cardiovascular disease. Notably, however, studies have suggested that T2DM itself is associated with lower detectable quantities of miR-126 and therefore its prognostic significance as a vascular marker is potentially lost in the presence of diabetes and this may further explain the failure to demonstrate a treatment effect [39].

Conclusion
In summary, our data suggest that prasugrel monotherapy is superior to either aspirin or clopidogrel in inhibiting platelet function in diabetes. From a translational perspective these findings could have the potential to be implemented in personalised treatment options for patients with T2DM and cardiovascular disease. Moreover, our miRNA results indicate that assessing response to antiplatelet therapy does not necessarily require fresh blood samples and tests can be conducted on miRNA measurements as biomarkers from stored acellular plasma samples. miRNA measurements may provide a platform to identify patients at greater risk of ischaemic heart disease relating to their platelet function. Based on these data and acknowledging that platelet activation is the central process in the development of atherothrombosis, a trial assessing the effects on clinical outcomes of prasugrel monotherapy may be warranted for the primary or secondary prevention of ischaemic heart disease in patients with T2DM. Also, further analysing the role of miRNA in predicting vascular outcome in individuals with diabetes may offer a tool to measure the clinical efficacy of antiplatelet agents.