Function, Proteomic and miRomic Study of a Dry Immersion Model of Microgravity implicates the Wnt/Dkk1/AXIN1 axis in Prothrombotic Platelet Phenotype

Objective: Ground based research modalities of microgravity have been proposed as innovative methods to investigate the aetiology of chronic age-related conditions such as cardiovascular disease. Dry Immersion (DI), has been effectively used to interrogate the sequelae of physical inactivity (PI) and microgravity on multiple physiological systems. Herein we look at the causa et effectus of 3-day DI on platelet phenotype, and correlate with both miRomic and biomarker expression. Approach and Results: The miRomic prole of platelets is reective of phenotype, which itself is sensitive and malleable to the exposome, undergoing responsive transitions in order to full platelets role in thrombosis and haemostasis. Heterogeneous platelet subpopulations circulate at any given time, with varying degrees of sensitivity to activation. Employing a DI model, we investigate the effect of acute PI on platelet function in 12 healthy males. 3-day DI resulted in a signicant increase in platelet count, plateletcrit, platelet adhesion, aggregation, and a modest elevation of platelet reactivity index (PRI). We identied 15 protein biomarkers whose expression levels were altered after DI. 22 “DI/PI” related miRNA were identied, of which ve have potential targets in the Wnt pathway associated with platelet biogenesis and function. These ndings are supported by increased circulating Axin1 and DKK1. Conclusions: Circulating biomarker and miRomic analysis implicates miRNA regulation of Wnt/Dkk1/AXIN1 axis in DI/PI induced primed platelet phenotype. Taken together, these ndings highlight platelets as sensitive adaptive sentinels and functional biomarkers of epigenetic drift within the cardiovascular compartment.


Introduction
Deciphering the aetiology of chronic diseases presents a continual challenge for basic, translational and clinical research. Pertinent to these endeavours is our understanding of cardiovascular disease (CVD) and associated co-morbidities. Efforts in coping with the health burden of CVD requires in depth knowledge of the causative risk factors in order to develop cost-effective preventive, management and treatment strategies. CVD risk factors can track from childhood into adulthood and are strong predictors of subclinical disease in early life [1]. Up to 80% of CVD may be prevented if modi able risk factors, e.g.
The haemostatic system is a complex ancestral pathway which is physiologically adapted to maintain haemostasis and protect vascular integrity. Risk factors perturb homeostasis, contributing to chronic One model being explored to decipher the aetiology of CVD is microgravity (µG). Gravity is a continual force impacting biology and physiology, with living organisms adapting to alterations across the gravitational continuum. Variations in gravitational levels are perceived at the molecular and cellular levels, inducing adaptive responses that in uence dynamic physiological functions. Fighting gravity requires daily physical exercise, thus exposure to µG is associated with enhanced inactivity [19]. µG provides a unique model to study the effects of global enhanced physical inactivity and deconditioning imposed on healthy subjects [17,[20][21][22][23][24][25][26][27]. Dry immersion reproduces most of the effects of microgravity, inducing rapid and profound deconditioning [20,24], similar to those observed in space ight [26,28].
To date there have been no in-depth systematic studies on the effect of DI on platelet biogenesis and concomitant phenotype. Herein, we endeavoured to elucidate the impact of physical inactivity and haemoconcentration induced by 3-Day DI on platelet phenotype as well as the cellular and molecular mechanistic alterations implicated in these adaptations.

Subjects
Twelve healthy non-athletic men aged 26 to 39 years (age 32 ± 1.4 year, weight 75 ± 2 kg, height 178 ± 2cm, BMI 23.6 ± 0.4 kg/m 2 , maximal oxygen uptake VȮ 2 max 39 ± 1.1 mL/min/kg, mean ± SEM) participated in this study. All participants were non-smokers, had no history of cardiovascular or other chronic diseases, were not taking medication prior to the experiment, and received a comprehensive clinical assessment. All subjects were informed about the experimental procedures and gave their written consent. The experimental protocol conformed to the standards set by the Declaration of Helsinki and was approved by the local Ethic Committee (CPP Sud-Ouest Outre-Mer I, France) and French Health Authorities (n° ID RCB: 2014-A 00904 − 43).

General DI protocol
The study was conducted at the MEDES space clinic, Toulouse, France. General DI protocol is described in detail previously [29]. Brie y, protocol included three days of ambulatory baseline measurement before immersion (B-3, B-2, B-1), three days (72 hours) of dry immersion (DI1, DI2, DI3) and two days of ambulatory recovery (R0, R + 1). The subjects were asked not to exercise during the 8 days of the experiment. Two subjects in two separate baths underwent DI simultaneously. Thermoneutral water temperature (32.5-33.5°C) was continuously maintained. Light-off period was set at 23:00-07:00. Daily hygiene, weighing and some speci c measurements required extraction from the bath. During these short out-of-bath periods, subjects maintained the − 6° head-down position ("strict" protocol). Total out-of-bath supine time for the 72 h of immersion was 4.7 ± 0.16 h. Otherwise, during DI, subjects remained immersed in a supine position for all activities and were continuously observed by video monitoring. Blood pressure (BP), heart rate (HR), and body weight were measured daily at 07:00. Leg echography has been performed daily to ensure the absence of thrombophlebitis. Onset and end of immersion both occurred at 09:00, therefore morning measurements and samplings on DI1 were performed before immersion, and on day R0 -still under immersion. Water intake was ad libitum and diet was the same for all participants (standardized to body weight in energy and nutrients). Daily caloric intake was approximately 2820 kcal for baseline and recovery and 2270 kcal for the immersion period. Daily intake for sodium and potassium was approximately 3-4 g. This 3-day DI allowed for several studies on different domains performed by 8 research groups [25,27,[29][30][31][32][33]].

Blood sampling
Antecubital venous blood samples (trisodium citrate tube) were collected in the morning before breakfast, before DI (B-3; Pre), under DI on R0 (70h of DI; End) and after DI on R + 1 (22h following the accomplishment of 3-day DI; Recovery). Schematic of the blood sampling protocol is outlined in Table S1 (Supplementary Material).

Preparation of platelets
The preparation of Platelet Rich Plasma, Platelet Free Plasma and leukocyte (CD45) depletion of platelets was as previously described [34]. CD45 depletion of platelets for RNA analysis was carried out using

Platelet-free plasma isolation
Platelet free plasma (PFP) was isolated for microvesicle work. In order to generate PFP, a double centrifugation method was employed. Blood was drawn using a 21G needle into a sodium citrate vacutainer (0.32% v/v nal concentration). As standard procedure, the rst 3mls of blood was discarded to avoid contamination from tissue damage due to venepuncture. The blood sample was mixed by gentle inversions to ensure even distribution of the anticoagulant followed by centrifugation at 1550xg for 20 minutes at room temperature (20-22 o C) to pellet the cells. The supernatant PFP was carefully aspirated leaving a layer of approximately 0.5 cm undisturbed on top of the cells. The collected PFP was centrifuged again at 13,000xg for 2 minutes. The PFP was then collected, leaving 20% of the sample at the bottom of the tube. The PFP was separated in 250 µl aliquots and stored at -80°C until further analysis, at which point it was thawed on ice.
Proseek® multiplex CVD II 96x96 & Proseek® multiplex in ammation I 96x96 are high-throughput multiplex immunoassays, each enabling analysis of 92 CVD-or in ammation-related protein biomarkers using 1µL of sample and across samples simultaneously. This high level of multiplexing is achieved by proximity extension assay (PEA) technology. A pair of oligonucleotide-labelled antibodies speci c for each biomarker are allowed to pair-wise bind to each target protein in the sample allowing for detection and quanti cation by real time PCR. Proseek assays were performed by Olink Bioscience (Upsala, Sweden) to evaluate the expression of 152 CVD and in ammatory biomarkers in the PFP samples. Brie y, 1µl of each sample or negative control was incubated with the conjugated antibodies at 4°C overnight. The PEA mixture was added and the products were extended and pre-ampli ed using PCR (ABI 2720 Thermal cycler, Life Technologies). The detection reagent was added to 2.8 µl of the extended and pre-ampli ed product, mixed and then loaded into the Fluidigm Gene Expression 96x96 Dynamic arrays (Fluidigm Corporation). The chip was primed using Fluidigm IFC controller HX and loaded into a Fluidigm Biomarker system. Detection and sample analysis was performed using the Fluidigm® BioMark™ HD System. This PCR platform enables simultaneous detection of 96 analytes in 96 samples creating 9,216 data points from a single run.
Proseek Multiplex Immunoassay: Data Analysis Raw data was analysed using Fluidigm PCR software as per manufacturers guidelines. Brie y, generated Cq values and data was normalized using the extension control and a background value. The data used for statistical analysis was expressed on a log2 scale, where a high value corresponded to a high protein expression. The limit of detection (mean negative control plus 3 x standard deviation) was determined per biomarker per sample. The data was normalised and analysed using GenEx software (MultiD, Gothenburg, Sweden). All statistical analyses (dynamic principal component analysis and one-way ANOVA) were performed on normalized data. miRNA isolation and ampli cation All RNA procedures were undertaken using the strictest sterile techniques using appropriate RNase free consumables and reagents. The mirVana® RNA extraction kit was used for isolation and puri cation of total RNA from platelet samples as per manufacturer's instructions. Platelets were prepared as previously outlined and pelleted by centrifugation at 2000xg for 12 minutes at RT. 400µl of total lysis binding solution was added to the cells at the commencement of the protocol. The total RNA was then eluted into a fresh collection tube by centrifugation at 10,000xg for 30 seconds using 100 µl of elution buffer, which was pre-heated to 95°C. The RNA was analysed on a NanoDrop® Spectrophotometer and samples stored at -80°C.
Platelet microRNA and pro ling Platelet miRNA pro les from 8 out of the 12 subjects (due to sample limitations) were assessed at the Pre-DI and End-DI time points (B-3 and R0). Total (non-normalized) leukocyte-depleted platelet RNA (ranging from 3.5 to 7.0 ng) was extracted from equal volumes of platelet rich plasma using the miRVANA RNA extraction kit. The miRNA pro le was determined by RT-qPCR using Applied Biosystems OpenArray® plate technology on the QuantStudio™ 12K Flex Real-Time System. For a complete miRNA pro le, 754 human miRNAs were quanti ed. miRNA pro ling was carried out using the Applied Biosystems® TaqMan® Low Density Array (TLDA) Human miRNA A (v2.0) and B (v3.0) cards set. For analysis on the TLDA cards, total RNA was rstly extracted from platelets using the Ambion™ mirVana® miRNA isolation kit as per manufacturer's instructions. Single stranded cDNA was synthesised from total platelet RNA using the Applied Biosystems TaqMan® miRNA Reverse Transcription (RT) Kit. For a full miRNA pro le two RT reactions were needed incorporating primers for both pool A and B miRNA panels. The RT reaction had a nal volume of 7.5 µl and contained: 3 µl (1-350 ng) total RNA and 4.5 µl of RT master mix. As the total RNA yield from platelets is generally lower than nucleated cells (less than 350 ng), a preampli cation step was carried out prior to committing the cDNA to the TaqMan miRNA arrays to uniformly pre-amplify desired cDNA prior to quanti cation with the TLDA cards. The sample was diluted with 75 µl of 0.1x TE buffer (pH 8) and used immediately for array analysis or stored for up at -80°C for future use. A 7900HT PCR system was used for initial miRNA pro ling and a QuantStudio™ 12K Flex Real-Time PCR system was used for large scale pro ling. DNA polymerase from the TaqMan® Universal PCR Master Mix ampli es the target cDNA using sequence-speci c primers and a probe on the TaqMan microRNA array. 100 µl of the master mix was dispensed into each chamber of the array. The card was centrifuged twice for 1 minute at 1000 x g to ll each of the 384 wells of the card in an Eppendorf 5810R centrifuge. OpenArray® plate technology on the QuantStudio™ 12K Flex Real-Time System (Paris, France) was used for large scale miRNA pro ling.
Bioinformatics analysis of platelet microRNA data Bioinformatics methods were used to extrapolate biological meaning from miRNA that were signi cantly up or downregulated after the DI. First, microRNA target prediction was performed using online software tools Targetscan (http://www.targetscan.org/vert_71/) and microRNA.org. Following this, involvement of these targets in pathways of interest from the Kyoto Encyclopaedia of Genes and Genomes (KEGG) was analysed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (http://david.niaid.nih.gov).
Platelet function assay with Impact-R cone and plate analyser The Impact-R device tests platelet function in anticoagulated whole blood under near physiological conditions. An image analyser measures the adhered platelets and results are expressed as a percentage of the well surface covered (SC%) as an index of platelet adhesion, and average size of the aggregates (AS µm 2 ) as an index of aggregation. Blood samples drawn into sodium citrate vacutainers were analysed 1 hour post-draw as per manufacturers recommendations. Blood samples were mixed for 1 minute at 10 rpm prior to loading on the apparatus. A 130µl aliquot of the blood sample was applied to the centre of the well followed immediately by the bell housing and cone on top. The selected programme was started (arterial shear rate of 1800 per second for 2 minutes). Platelet were stained with 500µl of May Grunwald stain solution for 1 minute. Images were then captured at random. Seven images were captured in total. The software analysed the captured images by eliminating the four least readable images and calculating the average of the remaining three. Results were expressed as SC and AS with a visual and graphical result provided for each test.

VASP/P2Y 12 phosphorylation assay
Flow cytometry experiments in this thesis were performed on a Accuri C6 (BD Biosciences). To ensure day-to-day sample reproducibility, all cytometers were calibrated daily. To determine the VASP phosphorylation state of whole blood, an adapted standardized ow cytometric assay (BioCytex, France) was employed [35]. The VASP-P analysis was performed within 4h after blood collection as per manufacturer's instructions.

Microvesicle quanti cation and analysis
The NanoSight NS300 and Syringe Pump were used to quantify microvesicles (exosomes and microparticles) in PFP samples. Nanoparticle tracking analysis technology (NTA) used in this device combines the properties of light scattering and Brownian motion to attain measurements including concentration and size distribution of particles in a liquid suspension (Methods; Supplementary Material).

Statistical analysis
Results are expressed as mean ± SEM. Statistical comparisons were performed using a variety of tests, depending on the experimental procedure. Main tests included independent t-tests, paired samples t-tests, repeated measures ANOVA and one-way ANOVA/one-way ANCOVA. Pearson product coe cient was used to examine relationships between variables. Statistical signi cance was set at a level of 0.05. SPSS v19 statistical package was used to analyse results.

Effect of dry immersion on physiological and haematological characteristics
The overview of physiological responses to this 3-day dry immersion protocol has previously been reported [29]. Overall, DI was well tolerated, with no dropouts. Moderate back pain was reported at the beginning, as it is usually seen in dry immersion protocols. HR and BP remained within normal limits throughout the protocol. HR rose slightly at END-DI (57 ± 3 bpm) and RECOVERY-DI (57 ± 2 bpm) vs. PRE-DI (53 ± 2 bpm). SBP and DBP were not signi cantly modi ed ( Figure S1; supplementary material). Body weight during immersion decreased by approximately 1-2kg ( Figure S1). Daily echo-doppler of the lower limbs did not reveal venous problems/thrombosis. Estimated plasma volume showed a 14 ± 2% decrease at END-DI vs. PRE-DI (p < 0.0001). At RECOVERY-DI plasma volume did not signi cantly differ from PRE-DI (+ 4 ± 2%, p = 0.103). Changes in haematological parameters in response to DI are shown in Fig. 1.
There was a signi cant increase in WBC, RBC concentration, HGB and HCT at the end time point. There was a signi cant decrease in WBC, RBC concentration and HCT between post and recovery, almost to basal levels.

Effect of dry immersion on platelet indices
There were no changes in platelet large cell ratio, mean platelet volume and platelet distribution width between any time points. There was a signi cant increase in platelet count and plateletcrit from PRE-DI to END-DI and subsequently, a signi cant decrease in these parameters from END-DI time point to RECOVERY-DI (Fig. 2).

Effect of dry immersion on platelet VASP phosphorylation
To assess if basal platelet VASP phosphorylation was affected by physical inactivity, the standardised P2Y 12 /VASP kit was used in a ow cytometry assay. The platelet reactivity index (PRI) represents changes in VASP phosphorylation. There was a small insigni cant increase in PRI post DI ( Figure S2). There were individual uctuations in VASP phosphorylation during the DI, possibly re ective of interindividual variability due to genetic architecture.

Effect of dry immersion on protein biomarker expression
Protein biomarker expression of platelet poor plasma (PPP) was assessed using the Cardiovascular and In ammatory protein biomarker panels (Olink Bioscience). The levels of 15 proteins differed signi cantly between time points (Table 1). Results are expressed as normalised protein expression (NPX) on a log2 scale. Therefore, a normalised increase of 1 is equal to a two-fold increase in protein amount. For 15 proteins out of 131 detected, expression levels differed signi cantly between different stages of the DI protocol. Seven of these proteins affected by DI are related to platelet function, namely Heat shock protein 27 (HSP27), Lectin-like oxidised LDL receptor (LOX-1), NF-Kappa-B essential modulator (NEMO), Proto-oncogene tyrosine protein kinase (SRC), Dickkopf-related protein (DKK1), Axin1, SIRT2, and IL-6. The changes in biomarker expression levels are displayed in Figs. 4 and 5. Table 1 Protein biomarkers which were differentially expressed after the DI. The column on the left displays differentially expressed proteins from the CVD panel and the column on the right displays differentially expressed proteins from the in ammation panel. Bioinformatic analysis of differentially expressed miRNA Bioinformatic interrogation was employed to ascribe biological relevance of the signi cantly differentially expressed miRNA END-DI. In silico analysis was used to determine putative targets of these identi ed miRNA of interest. A comprehensive approach was employed using complementary online software tools including Targetscan and microRNA.org. Subsequent pathway analysis was undertaken based on the Kyoto Encyclopaedia of Genes and Genomes (KEGG) and accessed using the DAVID bioinformatics database. A table was constructed to illustrate the number of potential genes the miRNA could target and the corresponding KEGG pathways common to a number of differentially expressed miRNA which are involved in platelet function (Table S2 -Wnt signalling, cytoskeletal dynamics, ECM-Receptor interactions and Toll like receptor signalling).

KEGG pathway analysis
Maps of key pathways involved in platelet function and activation were downloaded from the KEGG database using DAVID. Pathways were chosen based on their involvement in platelet function and signalling, speci cally in adhesion and aggregation, but also their in ammatory potential. Included in these were the Wnt signalling pathway, regulation of actin cytoskeleton, ECM interaction and Toll-like receptor pathway, of which the Wnt signalling pathway appeared repeatedly in both the up and downregulated miRNA targets. Genes that were potential targets for multiple miRNAs were circled red, genes that were targets for a single miRNA were circled yellow. Potential miRNA targets of genes involved in the Wnt signalling pathway are shown in Figure S6, potential miRNA targets of genes involved in the reorganisation of the actin cytoskeleton pathway are shown in Figure S7 and potential miRNA targets of genes involved in the ECM receptor-interaction pathway are shown in Figure S8.
miRNA targets of the Wnt signalling pathway The Wnt signalling pathway involved the largest number of differentially regulated miRNA post DI, as shown previously in Table 2. The protein biomarkers Axin1 and DKK1 were also differentially expressed END-DI. We used online databases to determine additional miRNA which could target Axin1 and DKK1 ( Fig. 7A & 7B). The combined action of multiple down or upregulated miRNA could have affected the gene and subsequent protein expression of Axin1 and DKK1. distinguished by a reduced capability for exercise, orthostatic intolerance and tachycardia [24]. Initial reactions to DI occur in the rst 12 hours and are caused by immediate modi cations in body uid distribution and a removal of support structure [20][21][22][23].
In the 1980's, Kirichenko et al. demonstrated that 7 days of DI induced a signi cant increase in platelet count and hyper-coagulopathic changes to platelet hemostasis [37,38]. A 2011 study examined the effects of simulated microgravity on the miRNA pro le of human lymphoblastic cells using a high aspect ratio vessel to model microgravity in space [39]. More recently, 42 miRNAs from cultured human blood lymphocytes from 12 healthy subjects were differentially expressed in microgravity stimulated cells compared to static cells, with resultant mRNA gene targets involved in in ammatory and apoptotic responses [40]. Malkani et al. elucidated the role of circulating microRNAs as both a potential biomarker for health risks associated with space ight and a countermeasure to mitigate the damage caused to the body by the space environment [41].

Haematological parameters
Both space ight and its proxies result in an initial plasma volume decrease (approximately 10-15 %) after which it remains stable [22,42]. We noted a 14% decrease in plasma volume post DI, with return to baseline at the recovery time point. Blood viscosity usually increases due to the decrease in plasma volume, with a corresponding decrease in RBC is required to maintain blood viscosity [43]. However, decreases in RBC due to altered erythropoiesis takes time, so haemoconcentration was still present at the end of immersion (END-DI), re ected by an increase in RBC. There were also signi cant increases in Hb and HCT during the immersion, although these values remained within the normal limits for healthy individuals. These ndings were as expected, as lower plasma volume would result in increased concentrations of RBCs. In general, real and simulated microgravity results in changes to physical properties of RBCs [20,44]. Both a 7-day DI [45] and 5-day DI [46] resulted in an increase in RBCs and altered morphological composition of red blood in healthy males. Navasiolava et al. reported a signi cant increase in RBC, Hb and HCT after seven days of DI, which returned to normal after recovery [23]. Unlike our study, they did not observe a signi cant difference in WBC. However, Bedendeeva et al. noted a 40% increase in leukocytes after DI [47].

Platelet indices
The effects of DI on platelet indices were investigated as primary indicators of platelet phenotype, and also as surrogate markers of altered megakaryocytopoiesis (Fig. 2). There was an increase in platelet count (PLT) between PRE-DI to END-DI time points, probably re ecting the loss of plasma volume during the DI. There was a signi cant decrease in PLT between END-DI and RECOVERY-DI suggesting that platelet count returned to pre-immersion levels once subjects began to resume ambulatory activity. Navasiolava et al. found no signi cant difference in PLT after 7-days DI [22]. Other studies have indicated a signi cant increase in PLT after 7-day DI [37,38]. There were no changes in MPV, PDW or PLCR after three days of DI. There was a signi cant increase in plateletcrit from PRE-DI to END-DI and a signi cant decrease in plateletcrit from END-DI to RECOVERY-DI. Plateletcrit appears to be one of the more sensitive platelet indices markers. Overall, the combined increases in PLT and PCT observed at the time-point END-DI may indicate altered megakaryocytopoiesis.

Platelet function
The effect of PI on platelet function was quanti ed using the Impact R cone and plate analyser. We observed a signi cant increase in platelet adhesion from PRE-DI to END-DI suggesting stronger plateletsurface interactions in response to physical inactivity and DI. There was a signi cant decrease in SC% from END-DI to RECOVERY-DI indicating platelet adhesion levels had returned to their basal state. There was a signi cant increase in AS from PRE-DI to END-DI suggesting that PI/DI results in elevated platelet aggregation with increased activation of GPαIIbβ3 and increased a nity for brinogen binding. Platelet aggregation decreased slightly from END-DI to RECOVERY-DI. The reduction in blood volume as a consequence of reduced total body water in the body during microgravity has been suggested as a factor for thrombotic tendencies, which could have been a determinant of increased platelet activation in this study [48].
PI causes endothelial dysfunction as well as increasing soluble von Willebrand factor (vWF), tissue factor (TF) expression and endothelial microparticle production, facilitating platelet activation under disturbed ow, inferring that altered shear stress may not activate platelets directly [49]. Haemoconcentration, body uid redistribution, hypodynamia experienced during DI may increase the risk of thrombosis and could also have contributed to enhanced platelet adhesion and aggregation observed after the DI [20,23]. The increase in blood viscosity after 24 hours of DI may have been a contributory factor to the platelet hyperaggregability [50]. The use of a viscometer and in parallel with platelet function tests could provide informative data [51]. Interestingly, Kuzichkin et al. observed an increase in plasma brinogen concentration after short-term space ights and 7-day DI [52]. Assessment of brinogen levels and αIIbβ3 activation levels would provide an insight into the increased platelet aggregation END-DI [53].
Platelet VASP phosphorylation VASP is an intracellular regulator of actin dynamics in platelets and plays a key role in regulating platelet adhesion and aggregation [54]. VASP is phosphorylated by cAMP-and cGMP-regulated protein kinases and re ects with inhibition of platelet activation, inhibition of αIIbβ3 and a restriction of VASP to bind to F-actin [55]. Consequently, decreased VASP phosphorylation can result in platelet hyperreactivity. We observed a minor non-signi cant increase in the average PRI (5%) at END-DI suggesting a reduction in VASP phosphorylation and increase in platelet activation ( Figure S2). Assinger et al. showed that VASP phosphorylation at basal levels on the Ser239 residue was signi cantly reduced in smokers [56].
Employing both picomolar and nanomolar concentrations of PGE 1, smokers demonstrated a reduced VASP phosphorylation correlating with increased P-selectin expression. However, at maximal PGE 1 concentrations traditionally used for in vitro assays, no difference between smokers and non-smokers was noted. As the VASP/P2Y 12 kit uses a maximal effective doses of PGE 1 , the PRI was similar (i.e. no statistical difference) between the smokers and the non-smokers. This experimental observation highlights the importance of employing a range of PGE 1 concentrations during assay development and emphasises the assessment of VASP phosphorylation in the presence of submaximal quantities of PGE 1 could be more physiologically informative and relevant. Physical inactivity could produce similar responses and require the same level of investigation.

Protein biomarker expression
We also examined the effect of PI on platelet poor plasma (PPP) protein biomarkers. The expression pro les of these proteins were analysed at PRE-DI, END-DI and RECOVERY-DI. 131 out of 157 of the proteins were detected in all samples, with 15 proteins signi cantly ΔDE between different stages of the DI (Table 1). Key proteins affected by physical inactivity included Heat shock protein 27 (HSP27), Lectinlike oxidised LDL receptor (LOX-1), NF-Kappa-B essential modulator (NEMO), Proto-oncogene tyrosine protein kinase (SRC) and Dickkopf-related protein (DKK1) (Fig. 4).
Numerous environmental and physiological stressors mediate the expression of heat shock proteins (HSPs). HSPs have been identi ed in atherosclerosis [57] and post exercise [58]. HSP-27 was signi cantly elevated END-DI suggesting a stress response to the effects of acute physical inactivity. HSP27 has been proposed as a key player in actin polymerisation during platelet shape change, and phosphorylated HSP27 is released from human platelets upon collagen activation, and is associated with the acceleration of platelet aggregation [59]. A number of in ammatory proteins were also differentially regulated END-DI (Fig. 5). Axin-1 is a key member of the Wnt signalling pathway, acting as a scaffold protein and a negative regulator of the Wnt signalling pathway [65]. As Wnt signalling negatively regulates platelet function and modulates the major platelet receptor GPαIIbβ3, an increase in Axin-1 levels could also have contributed to platelet adhesion and aggregation levels in this study [66][67][68].
IL-6, a pleiotropic in ammatory cytokine, was signi cantly increased END-DI. IL-6 has been adversely linked with sedentary time in a large study of > 500 participants aged ~ 63 years and at high risk for T2DM [69] and is elevated after physical inactivity in our study. While platelets do not express IL-6, it can affect platelet activation [70]. IL-6 is also a potent thrombopoietic factor promoting maturation of human megakaryocytes in vitro [71]. A number of in ammatory and CVD proteins were ΔDE after our DI, suggesting that extensive physical inactivity affects both immune and platelet function. Future proteomic studies would prove insightful.

Platelet poor plasma microvesicles
We investigated the effect of physical inactivity on the number and size of MVs in PPP. Figure  Platelet microRNA (miRNA) To further understand the signi cance of PI on the molecular mechanisms which underpin platelet phenotype, we investigated miRNA pro les PRE-DI and END-DI, to determine if the miRNome for platelets was altered by physical inactivity. In brief, miRNA are short (18-24) nucleotide long non-coding RNA molecules. They regulate gene expression by hybridising to the 3' UTR of mRNA. The existence and functionality miRNA in the anucleate human platelets has been described and constitute 80% of all small RNAs in platelets [72], expressing relatively high quantities of miRNA compared to their nucleated counterparts [73]. Platelet miRNA levels have been demonstrated to be associated with phenotype [74].
Platelets expressed a total average of 436 miRNA PRE-DI and 438 END-DI, in line to previous similar published studies [72,73]. We identi ed 22 miRNA which were signi cantly up or downregulated at END-DI (Fig. 6). 10 miRNA were signi cantly downregulated and 12 miRNA which were signi cantly upregulated at END-DI. Interestingly, miR-374a family has been demonstrated to regulate Wnt/beta-catenin signalling [75].
As platelets do not express the miRNA nuclear machinery Drosha and DGCR8, the de novo synthesis of new miRNA in platelets is negligible. Upregulation of miRNA in response to physical inactivity could be derived from the processing of pre-miRNA to mature miRNA or as a re ection of increased levels miRNA in their megakaryocyte precursor. As platelets can release MV-containing miRNA upon activation, downregulated miRNA in this study may re ect this process. These ndings suggest that platelet miRNAs correlate platelet activation in vitro and may have great potential as biomarkers of PI.

Bioinformatics of differentially expressed DI regulated miRNA
The fold change, number of conserved targets, and number of pathways the predicted genes are linked to are included in Table S2 for each ΔDE miRNA. Table 2 shows the miRNA potentially involved in the pathways of interest and whether these speci c miRNAs were up or downregulated with DI. Pathways were chosen based on their involvement in platelet function and signalling.
Of major interest from the ndings of this study was the predicted involvement of ΔDE miRNA on the Wnt signalling pathway, as both DKK1 and Axin-1 proteins were differentially expressed END-DI, and are involved in the Wnt pathway. DKK1 was one of the signi cantly upregulated proteins identi ed from the biomarker panel END-DI and we therefore sought to identify ΔDE miRNA that could target DKK1 (Fig. 7A). A number of these miRNA were downregulated in our study, suggesting that a simultaneous downregulation of multiple miRNA targeting DKK1 may act together increasing DKK1 expression. miR-302a, the most down regulated miRNA, targets DKK1. The increase in DKK1 could negatively regulate the Wnt signalling pathway and ultimately, elevated platelet adhesion levels.
Axin-1, a key mediator of the Wnt/β-cat pathway, was identi ed as one of the proteins that was upregulated after the DI [77]. Interestingly, two downregulated miRNAs, miR-203 and miR-200a were identi ed as potential regulators of Axin-1. Additionally, a number of other miRNA predicted to target Axin-1 were also downregulated as shown in Fig. 7B. The simultaneous downregulation of multiple miRNA targeting Axin-1 may explain the observed increase in Axin-1 and result in reduced Wnt signalling in platelets with elevated platelet adhesion and aggregation.
Evidence of physical activity-speci c microRNA signatures have seeded the notion that there must also be physical inactivity-speci c miRNA pro les [78]. Epigenetic variation could be a potential mechanism allowing for independent or perhaps synergistic effects of physical inactivity on platelet function. Hibler et al. recently described indications for epigenetic variation and miRNA expression as a link between physical activity and sedentary lifestyle [79]. We hypothesised that an epigenetic adaptation to physical inactivity is also present inducing epigenetic changes in megakaryocytopoiesis and altered platelet phenotypes.
Conclusion 3-day DI induced rapidly reversible shift to primed platelet phenotype in healthy men, re ected by increased adhesion and aggregation. We have identi ed 15 ΔDE protein biomarkers associated with µG/PI. Their expression trends could be of importance for developing 'biosignatures' of physical inactivity. We also identi ed 22 ΔDE platelet miRNA END-DI. These PI related miRNA have potential targets involved in pathways associated with platelet function. It is evident from this study that the identi cation of unique signatures of several miRNA, rather than a single miRNA in isolation may enhance diagnostic accuracy [80]. The canonical Wnt signalling pathway may signify a novel endogenous mechanism for regulating platelet activity in response to physical inactivity. Collectively, our results provide evidence for the early and robust deleterious impact of reduced daily activity on platelet function and phenotype, highlighting the vulnerability of the vasculature to a sedentary lifestyle. It also highlights the importance of physical activity and exercise medicine.