Proteome and functional decline as platelets age in the circulation

Platelets circulate in the blood of healthy individuals for approximately 7–10 days regulated by finely balanced processes of production and destruction. As platelets are anucleate we reasoned that their protein composition would change as they age and that this change would be linked to alterations in structure and function.


| INTRODUC TI ON
Platelets circulate for approximately 10 days in healthy individuals, [1][2][3] responding rapidly as central players in hemostasis. These dynamic and metabolically active fragments contain a complex array of extracellular proteins, signaling pathways and intracellular machinery including storage granules, canalicular systems, mitochondria, and contractile proteins. Analyses of the platelet proteome have revealed a high degree of similarity in protein content among healthy individuals, with over 3000 proteins detected. 4 Being cellular fragments, platelets lack a nucleus and have only limited capacity to synthesize new proteins. However, to date studies have largely been conducted on platelets separated as a single cell type from whole blood, and so represent the collective proteome across platelets of all ages.
Newly formed platelets contain an array of the messenger ribonucleic acids (mRNA) which were present in their progenitor megakaryocyte. [5][6][7] These residual mRNAs can be used as an indicator of platelet age as they are lost from platelets as they circulate and platelets have limited capacity to generate new mRNA. [8][9][10][11][12] Numerous studies have reported that newly formed platelets, i.e. those with the highest levels of mRNA, also known as reticulated platelets or the 'immature platelet fraction', are hyper-reactive with an increased thrombotic potential. [13][14][15][16][17][18] This hyper-reactivity has been linked to a number of pathological states, including diabetes mellitus and chronic kidney disease, in which there is both increased platelet turnover and higher incidence of acute coronary syndromes associated with reduced effectiveness of standard anti-platelet therapies. [19][20][21][22][23] A hypothetical explanation for the age-related difference in reactivity is that younger platelets have a greater array of functional pathways to bring to bear to hemostatic processes; they have the full complement of proteins derived from the progenitor megakaryocytes, while these have become degraded or lost in older platelets without the general possibility for replacement. To test this idea, we developed and validated protocols to separate platelets according to circulatory age and carried out proteomic, immunofluorescence and functional assays to delineate physical and functional changes in platelets as they normally age within the circulation.

| Ethical statement: murine studies
Animal procedures were conducted under UK Home Office project license authority (PPL/8422) in accordance with "The Animals (Scientific Procedures) Act 1986", EU directive 2010/63/EU, and were subject to local approval from Queen Mary University of London and Imperial College London Ethical Review Panel.

| In vivo labelling and flow cytometric sorting of murine young and old platelets
Male C57Bl/6 wild-type mice (8-12 weeks) were purchased from Charles River UK. Anti-CD42c DyLight-x488 or x649 (Emfret) were administered (intravenous) as per supplier guidance. Briefly, anti-CD42c-x488 was injected at 0 h, followed by anti-CD42c-x649 23 h later. Blood was collected and platelet rich plasma (PRP) isolated as previously published. 24 Murine platelets were sorted using BD FACS Aria IIIu Fusion Cell Sorter (70 μm nozzle, 70 Ps; BD Bioscience); old platelets were Conclusions: Our findings demonstrate changes in protein content are linked to alterations in function as platelets age. This work delineates physical and functional changes in platelets as they age and serves as a base to examine differences associated with altered mean age of platelet populations in conditions such as immune thrombocytopenia and diabetes. • Platelet ageing is marked by a decline in protein content, particularly proteins associated with dynamic processes and associated reductions in hemostatic function • Our work highlights important changes in the platelet proteome and associated reductions in hemostatic function as platelets age within the circulation gated as CD42c-x488/CD42c-x649 dual positive, and young platelets as CD42c-x649 positive/-x488 negative events. Subsequently, platelets were pelleted at 1000 g for 10 min in the presence of prostacyclin (PGI 2 , 2 μmol/L; Tocris) and re-suspended in Qiazol (QIAGen).

| Ethical statement: human studies
All studies were conducted according to the principles of the

| Flow cytometric measurement of activation markers pre-and post-sorting
Pre-or post-sorted platelets were diluted in annexin binding buffer

| Protein extraction and proteomic analysis of sorted human platelets
Protein was extracted from sorted platelets (65 million per subpopulation) using lysis buffer (100 mM TRIS pH 7.5, 2% sodium dodecyl sulfate (SDS), 1 cOmplete™ Mini Protease Inhibitor Cocktail Tablet per 10ml; all Sigma) and protein content determined by Nanodrop Spectrophotometer ND-1000 (Thermo Fisher Scientific).
Proteomic analysis was performed on 30 µg of total protein in each subpopulation with tryptic digests obtained using the Filter Aided Sample Preparation protocol using 30k filter units (Microcon YM-30, Millipore) and sequencing grade trypsin (Trypsin Gold, Promega) with an enzyme to protein ratio of 1:50. 25 The concentration of tryptic peptides was estimated by UV spectrometer at 280 nm, and 10 µg peptides were used for mass spectrometry analysis. Sample preparation, LC-MS/MS analysis and data acquisition was performed as previously described. 26 Raw mass spectrometry data files were processed in the  27 Only proteins present in at least 50% of the samples in at least one group of differently aged platelets were considered identified. The false discovery rate (FDR), determined by reverse database searching, was set to 0.01 for the peptides and proteins.
Proteins found to be differentially expressed between groups (p < .05) were subjected to pathway mapping analysis and were dis-
Analysis was conducted using Zen Software and ImageJ; spreading stage was quantified as; i: adherent, not spread; ii: filopodia; iii: lamellipodia; iv: fully spread.

LC-MS/MS was performed at the National Institute of Environmental
Health Science, North Carolina, USA as described previously. 6

| Statistical analyses
Data are expressed as mean ± SEM. Graphs and statistical analysis were generated using GraphPad Prism 8 (GraphPad Software Inc.). Statistical analyses were performed with a paired t-test or a one-way ANOVA with Tukey's post-test for multiple comparisons.
Correlations were assessed by simple linear regression. Significance was defined as p < .05.

| Platelets can be separated by circulatory age using cell sorting based upon thiazole orange fluorescence
To validate assays to use in samples from healthy human individuals, we began with experiments in mice in which we could employ in vivo labelling. Temporal in-vivo antibody labelling in mice and cell sorting followed by qRT-PCR demonstrated that the newest circulating platelets (<24 h old), had significantly lower mean cycle threshold (Ct) values for ITGA2B, PF4 and TUBB1 than old platelets (2-5 days). This is consistent with the loss of megakaryocytic mRNAs as platelets age and demonstrates that the levels of these megakaryocytic mRNAs are indicators of platelet circulatory age ( Figure 1A, p < .05, n = 4). Next, as we can- 30.4 ± 2.8; young vs. intermediate platelets, p < .05 for all, n = 3). There were highly significant correlations between TO-determined platelet age and log2 fold differences in mRNA, as determined from Ct values, for ITGA2B (r 2 0.53, p < .03), PF4 (r 2 0.55, p < .02) and TUBB1 (r 2 0.65, p < .008) ( Figure 1D). The levels of a further ten mRNAs relevant to platelet function were also noted to decline strongly, although for these Ct values >40 in old platelets precluded full analyses (Table S1).
From these data we can conclude that our TO staining and sorting protocols allow the separation of human platelets from healthy volunteers on the basis of circulatory age as confirmed by decline in megakaryocytic mRNAs. Importantly, activation status did not change during the staining and sorting protocol ( Figure S2) indicating that platelets were not adversely activated by these processes.
Notably these correlations are incompatible with the notion that mRNA and protein vary by platelet size rather than age, as protein varies across the TO-defined populations linearly and mRNA by powers of 2. As further confirmation, immunofluorescence revealed there was no difference in the cross-sectional area of our sorted young and old platelets (young platelets, 8.9 ± 0.2 µm 2 ; old platelets, 8.5 ± 0.5 µm 2 ; Figure 2C, p > .05,

| Proteomic analysis identified differences in proteins affecting fundamental biological processes
Proteomic analysis identified 583 proteins within the sorted platelets (Table S2) of which 94 proteins were significantly modulated among the three subpopulations (p < .05; Table S3). Targeted analysis between young and old platelets identified relative differences in the levels of 78 proteins (Table 1, Figure 3A).
Ingenuity Pathway Analysis predicted an association between the 78 altered proteins and 28 biological processes and functions ( Figure 3B). Twenty-two of these processes were predicted as higher in young platelets including hemostasis, calcium flux, as well as transmembrane potential of mitochondria. The remaining six functions were predicted as higher in old platelets including apoptosis and senescence ( Figure 3B).

| Old platelets have reduced mitochondrial number and activity
Proteomic analysis demonstrated a relative reduction in the levels of key mitochondrial proteins in old platelets, notably citrate synthase and ADP/ATP translocase 2 (Table 1)

| Old platelets have an altered cytoskeleton resulting in reduced platelet spreading
Proteomics also indicated a reduction in the amount of a number of cytoskeletal binding proteins suggesting age related alterations to cytoskeletal structure (Table 1). Immunofluorescence demonstrated a significant decrease in the fluorescence intensity 1364 ± 139AU; Figure 5D-F; p < .05, n = 5) in young vs. old platelets, respectively. Furthermore, western blotting confirmed the reduction in the expression of α-tubulin and β-actin ( Figure S3A-B).
Notably, the reduction in cytoskeletal proteins was not accompanied by a change in the cross-sectional area of resting platelets ( Figure 2C).
Categorization of spreading stage demonstrated that old platelets were able to form filopodia but unlike young platelets did not transition into lamellipodia and fully spread ( Figure 5J). Furthermore, the reduction in spreading was accompanied by a decrease in the percentage of platelets forming actin nodules (63 ± 1% vs. 24 ± 1%; young vs. old platelets; p < .05, n = 4; Figure S3C-E).

| Platelet ageing is associated with alterations in intracellular protein components
Proteomics also identified a relative increase in the levels of a small number of circulating proteins as platelets age (complement C5, C1s and C4a, and fibrinogen alpha chain; Table 1). Microscopy confirmed an increased abundance of complement C4 and fibrinogen in old platelets (respectively, 661 ± 150 AU vs. 1519 ± 181 AU and 580 ± 93 AU vs. 961 ± 179 AU, in young vs. old platelets; Figure 6C,F; p < .05, n = 4) localized both at the cell periphery and within the cytoplasm consistent with internalization ( Figure 6A-F).

| Platelet ageing is associated with a reduction in activation and secretory pathways
Imaging flow cytometry of fabricated mixed populations of 50% young and 50% old platelets demonstrated that young platelets responded more rapidly to stimulation, forming the core of aggregates with old platelets binding to the periphery ( Figure 7A). Young platelets were present in 95 ± 2% of aggregates compared with 42 ± 6% of old platelets ( Figure 7B; p < .05, n = 3). Notably, 57 ± 6% of aggregates were composed exclusively of young platelets but only 4 ± 2% were composed exclusively of old platelets ( Figure 7C;

| DISCUSS ION
Changes in the age profile of platelet populations have been associated with many disease states and increased risk of thrombotic events.
However, to date there has been very limited systematic investigation into the changes that platelets undergo as they age naturally. To address this need we have isolated, phenotyped and functionally characterized platelets as they age within the healthy human circulation.
To separate human platelets by age, we used TO fluorescence intensity as a surrogate marker of mRNA content, corroborating the relevance of this to platelet ageing by determining the levels of particular megakaryocyte-derived platelet-specific mRNAs with qRT-PCR.
This approach was validated in mice using temporal antibody labelling and is supported by recent research using nucleic acid dye staining with TO and SYTO13 to sort platelet subpopulations for transcriptomic analyses. 29,30 We found highly significant correlations between the TOdetermined platelet age and platelet levels of mRNA for ITGA2B, PF4, TUBB1, and between TO-determined platelet age and total platelet protein; the former being a log2 function and the latter a linear relationship.
It was particularly notable that as platelets aged there was a marked reduction in total protein content, such that old platelets had lost almost half of the protein present in young platelets. As platelets are anucleate, with a limited translational capacity, this is perhaps unsurprising and could be explained by protein degradation due to basal cellular processes and release of proteins, perhaps encapsulated within platelet microvesicles. 7,31,32,33,34 While our data show a large general loss of proteins, this does not mean that individual proteins are not more subtly regulated. Platelets have the capacity to replenish their protein levels through endocytosis via the open canalicular system, as well as through limited protein synthesis. 35,36 The latter of these processes can be engaged in response to stimulation, suggesting a controlled mechanism of synthesis to support the activation process, that might help counter the general protein loss. 32,37 Given the rate of mRNA decay in platelets, ongoing protein synthesis may perhaps occur primarily within the first few days of platelet lifespan. 9 We identified 583 proteins in our proteomic analysis and established significant differences in the levels of 78 proteins between young and old platelets. These differences were determined against the general loss of proteins noted above and so demonstrate particular variations rather than general changes. There were 64 proteins that compared to general proteins were significantly lower in old platelets than young platelets, indicating accelerated loss during ageing. A further 14 proteins were proportionately more abundant in old platelets, indicating selective retention, synthesis, or uptake from the circulation. Among the proteins which demonstrated accelerated loss in old platelets there was a particular signal for those involved in dynamic processes, such as in mitochondria (12 proteins) and the cytoskeleton (6 proteins). 4,38 Ingenuity pathway analyses of the significantly modulated proteins predicted strong associations between older platelet age and reductions in cell activation and calcium flux, cytoskeletal organization, microtubule dynamics, formation of filaments and lamellipodia, binding of platelets and hemostasis, which we substantiated in the functional studies discussed below. The pathway analyses also associated old platelets to reduced cell viability and increased necrosis, apoptosis, and senescence. Previous research into platelet ageing has produced contradictory evidence as to whether platelet size changes with age, with some groups suggesting platelet size to be independent of age whilst others report an association between mean platelet volume and thrombotic risk. [60][61][62][63][64][65][66] Indeed, the use of TO to identify platelets containing elevated levels of mRNA as 'young' has been questioned on the basis that larger platelets may uptake more dye and so skew subsequent analysis. 67 However, this cannot explain our data in healthy individuals as we noted a log2 relationship between particular megakaryocytic mRNAs and TO-determined platelet age; i.e. platelets would have to vary in size by around 32-64 fold for such a relationship to hold. Furthermore, we found no differences in the cross-sectional areas of young and old platelets. We have not, however, examined associations between platelet age and size in any pathological conditions in which different relationships may well exist. Interestingly, it was hypothesized some 40 years ago that buoyant density could be an indicator of platelet age, with the observation that high density platelets are more metabolically active. 68 This idea is supported by our data in which younger platelets have higher protein content and contain more mitochondria consistent with higher density and metabolic activity. Consistent with this, our data also support the related and much more recent suggestion that mitochondria number is a better indicator of platelet age than platelet size. 69 The discussion above has focused upon proteins that are decreased at accelerated rates as platelet age, however our proteomic analysis also demonstrated significant increases in the relative levels of some other proteins. These can largely be categorized as circulating proteins, including complement proteins and fibrinogen, which are most likely endocytosed as platelets circulate. Supporting this notion, research has implicated anionic phospholipids, such as phosphatidylserine, as promotors of complement protein activation. 70,71 Conversely, recent transcriptomic research has indicated a relative increase in complement C5 transcripts in old platelets, so perhaps there is synthesis of complement proteins within platelets. 30 In conclusion, our work demonstrates changes in total and relative protein content as platelets age and associated alterations in hemostatic function. We propose that young platelets are rapid hemostatic responders, due to their more robust cytoskeleton and higher mitochondria number, forming the core of aggregates and recruiting older platelets to the periphery. Old platelets have blunted hemostatic responses, accumulate circulating proteins, and bear indicators of cells marked for clearance from the circulation. Our research provides a detailed characterization of protein and functional changes as platelets normally age within the circulation providing key information for studies of platelet function in health and disease.

| Limitations
Research into platelet subpopulations is still in its relative infancy and much of the work here has been performed at the technical boundaries. In particular, to minimise activation during the sorting process but also supply responsive platelets, samples were sorted at a slow flow rate for a limited time, restricting the number of platelets available for study. As a result of lower numbers of platelets used for proteomic analysis, the total number of proteins detected within these samples is lower than previously published literature.
Furthermore, technical constraints on obtaining defined platelet pellets after cell sorting, meant that we were unable to perform high resolution electron microscopy to establish changes in granule number and determine alterations in other structural components.

ACK N OWLED G EM ENTS
The authors acknowledge the support the Flow Cytometry Core Facilities at the Blizard Institute and Charterhouse Square, Queen Mary University of London and the UCD Conway Institute mass spectrometry facilities.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N
HEA designed the research, performed the assays, and collected data, analyzed and interpreted data, performed statistical analy-