Human platelet lysate‐derived extracellular vesicles enhance angiogenesis through miR‐126

Abstract Objectives Extracellular vesicles (EVs) are key biological mediators of several physiological functions within the cell microenvironment. Platelets are the most abundant source of EVs in the blood. Similarly, platelet lysate (PL), the best platelet derivative and angiogenic performer for regenerative purposes, is enriched of EVs, but their role is still too poorly discovered to be suitably exploited. Here, we explored the contribution of the EVs in PL, by investigating the angiogenic features extrapolated from that possessed by PL. Methods We tested angiogenic ability and molecular cargo in 3D bioprinted models and by RNA sequencing analysis of PL‐derived EVs. Results A subset of small vesicles is highly represented in PL. The EVs do not retain aggregation ability, preserving a low redox state in human umbilical vein endothelial cells (HUVECs) and increasing the angiogenic tubularly‐like structures in 3D endothelial bioprinted constructs. EVs resembled the miRNome profile of PL, mainly enriched with small RNAs and a high amount of miR‐126, the most abundant angiogenic miRNA in platelets. The transfer of miR‐126 by EVs in HUVEC after the in vitro inhibition of the endogenous form, restored angiogenesis, without involving VEGF as a downstream target in this system. Conclusion PL is a biological source of available EVs with angiogenic effects involving a miRNAs‐based cargo. These properties can be exploited for targeted molecular/biological manipulation of PL, by potentially developing a product exclusively manufactured of EVs.

with small RNAs and a high amount of miR-126, the most abundant angiogenic miRNA in platelets. The transfer of miR-126 by EVs in HUVEC after the in vitro inhibition of the endogenous form, restored angiogenesis, without involving VEGF as a downstream target in this system.
Conclusion: PL is a biological source of available EVs with angiogenic effects involving a miRNAs-based cargo. These properties can be exploited for targeted molecular/ biological manipulation of PL, by potentially developing a product exclusively manufactured of EVs.

| INTRODUCTION
Beyond haemostasis and thrombosis, platelets have been also described as the main regulators of angiogenesis, a key process for tissue regeneration and repair outcome of vascular insults or wound healing and based on the activation of endothelial proliferation, sprouting and organization into functional tubules. [1][2][3] The emerging role of platelets to act as inflammatory/immune effectors and to enhance angiogenesis, stems from their intrinsic physiological role to interact with the endothelium during vascular damage, preserving the integrity and vessel homeostasis. 4 Platelets exhibit a unique secretory profile of multiple combined factors with a dual pro-and anti-angiogenic role. Among them, we could list growth factors, cytokines, microRNAs, small soluble molecules and proteins, including those related to cytoskeleton, adhesion, inflammation, and extracellular matrix interaction. [5][6][7][8] This balanced combination of mediators, is mainly contained in plasma membranes delimited nanoparticles, named extracellular vesicles (EVs). These later, now conceived as signalosomes and biological vectors of heterogeneous size and composition, are released upon platelets activation, interacting with the microenvironment. 9,10 Platelets represent the most abundant source of EVs of different dimensions and quantities in the systemic circulation 11-14 depending on multiple variables including age, physiological states and lifestyle habits. 15,16 EVs mirror the haemostatic properties of platelets, 17 by exerting both anti-and pro-coagulant effects according to the subpopulation of EVs involved. 11,18,19 Hence, their circulating levels can act as predictive biomarkers of haemostatic and inflammatory disorders. 20 Patients with metabolic syndrome, myocardial infarction, atherosclerosis, ischemia or inflammatory diseases exhibit higher levels of circulating EVs, because of activated platelets, [21][22][23][24][25][26][27][28] suggesting their relevance to mediate pathogenetic effects beyond their physiological role. On the other hand, platelet-derived EVs have been also demonstrated to regulate angiogenesis when released at the site of endothelial sprouts 29 and secretion of VEGF, 30 or to transfer proliferative and survival biological information to the endothelium. [30][31][32][33] Plateletderived EVs can modulate the vascular tone as shown in rabbit models, 30 or even attenuate blood pressure in preeclampsia women, by stimulating the inducible nitric oxide synthase in endothelial cells, 34 38 Strong evidence of their ability to support angiogenesis has been also observed in myocardial infarction and cerebral ischemia after in vivo direct injection, when platelets, activated by thrombin, release EVs. 39,40 Moreover, the key contribution of platelet-derived EVs in supporting the angiogenic profile of cancer invasion and metastasis parallel to clinical thrombotic complications has been strongly highlighted. [41][42][43] Based on these studies, evidence that platelets and plateletderived biological products can trigger angiogenic programs in endothelial cells has encouraged a better understanding of their potential therapeutic use for those regenerative-based applications where the restoration or the enhancement of angiogenesis represents the clinical goal. Accordingly, parallel to the investigations regarding the key involvement of platelets in regulating angiogenesis, it has been demonstrated that platelet-derived clinical preparations (i.e., platelet-rich plasma and gels) are similarly able to boost and reflect the angiogenic properties of platelets. Particular attention has been dedicated to platelet lysate (PL), considered the gold preparation concentrate derived from platelets, and whose clinical efficacy is currently considered superior to other platelet-derived formulations. 44 The employment of PL, alone or even in combination with different sources of stem cells, has shown to enhance blood perfusion in peripheral artery diseases, 45 to heal difficult wounds, to sustain stromal proliferation, epithelization, angiogenesis, and to prime cardiovascular differentiation. [1][2][3][46][47][48][49] The angiogenic capacity of PL is ascribable to the plethora of highly concentrated factors in this hemoderivative. When PL is manufactured, platelets are repeatedly lysed, therefore enriching the preparations with vesicles and granules, representing a primary source of angiogenic EVs. So far, the vast majority of studies have only explored the effects of vesicles of different origins (i.e., from MSCs, fibroblasts, lymphocytes) after treatment with PL, or EVs released by intact and activated platelets. 43 Only a couple of studies have described the presence of exosomes in platelet-derived clinical formulations but as effectors of the osteogenic differentiation on MSCs or with neuroregenerative capacities. 38

| Nanotracking analysis of EVs in PL
Nanoparticles tracking analysis in terms of size distribution and concentration was performed on PL using a NanoSight NS300 instrument (Malvern Panalytical). Five 30-s videos were recorded for each sample with a camera level set at 15 of 16 and a detection threshold set between 5 and 7. The EVs concentration and size distribution were subsequently analysed with NTA 3.2 software. After incubation, the membranes were incubated with secondary antirabbit antibody (Cell Signalling; 1:10,000) and the immune complexes thus formed were detected by enhanced substrate chemiluminescence. Densitometric detection of the bands was performed by Chemidoc (Bio-Rad).

| Statistics
Statistical analysis was performed by GraphPad PRISM 5 software.
Student's t-test and one-way analysis of variance (Bonferroni correction) were used to compare the difference between the control and groups. A p < 0.05 was considered significant. Data were presented as mean ± standard error unless specified. Additional information on statistics and confidence intervals have been reported in the corresponding sections above and in figure legends.

| RESULTS
We investigated in detail the EV content and characteristics of human PL preparations. To assess the concentration and absolute size distribution, 56 Figure 1F). The characterization of EVs was further verified by cytofluorimetry. The FACS analysis confirmed the expression of CD41, the main marker of platelet origin of EVs (49.92 ± 5.22%, also known as glycoprotein IIb possessing a critical role in modulating platelet aggregation 63 ), but also the negative expression for calnexin ( Figure 1G).
To discriminate the biological effects of EVs from the whole PL, we isolated the EVs according to methodological standardized guidelines by high-speed ultracentrifugation. 64,65 Afterwards, we investigated whether EVs may convey haemostatic properties, such as aggregation and pro-coagulant abilities, which are two key physiological properties exerted by platelets but also reported for EVs. 66 We stimulated platelets of healthy subjects with increasing percentages of EVs (5%, 10% and 20%). PL (10% and 20%) and collagen were used as biological and positive references, respectively. In addition, the quantification of the soluble fragment 1 + 2 of prothrombin (F1 + 2) was employed to test the coagulation property of EVs. Results showed that neither the increasing concentration of EVs nor PL were able to induce aggregation compared to collagen ( Figure 2A,B). A similar amount of F1 + 2 among samples was detected (comparable to physiological soluble levels in the human plasma), with no statistically significant differences ( Figure 2C).
As one of the most significant bioactive properties of PL is the ability to induce angiogenesis, 2,46-49 we investigated the contribution of EVs to the angiogenesis stimuli mediated by PL. We isolated and labelled the EVs with the green fluorescent dye CFSE. Afterwards, HUVECs were stimulated for 24 h with the EV preparation (10% vol/vol, corresponding to the same PL volume in percentage routinely employed in cell culture [47][48][49]51 ). HUVECs were able to uptake EVs, as demonstrated by the presence of green fluorescent dots visible in the cytoplasm ( Figure 3A). When HUVECs were subjected to the in vitro angiogenesis Matrigel assay at increasing concentrations of EVs (5%, 10%, 20%), we found that 10% EVs was the optimal percentage to significantly enhance the number of closed loops ( Figure  Several studies have demonstrated the modulation of the redox status in cells exposed to intact platelet-derived EVs. 67 Thus, we investigated the levels of hydrogen peroxide in the conditioned media of HUVECs collected after 24 h of treatment with EVs. Results showed a lower release of hydrogen peroxide after treatment with 10% EVs compared to PL (p = 0.03; Figure 4A). We observed that the treatment with all percentages of EVs were able to maintain very low amounts of H 2 O 2 in the media as both controls (EBM and EGM-2). However, the 10% EVs reveals as the optimal anti-oxidant stimulation respect to 10% PL (p < 0.05). This result was also coherent with the lowest expression level of the NADPH isoform Nox4 (the main and specific isoform responsible for the direct production of hydrogen peroxide by endothelial cells 52,53,68-70 ) after stimulation with 10% EVs among the three concentrations of EVs (p < 0.05; Figure 4B). The Nox4 mRNA levels in presence of 10% EVs were similarly downregulated as PL and EGM-2 with respect to EBM (p < 0.05).
Some key functions of platelets, such as aggregation, activation and angiogenesis, are known to be mediated by miRNAs released by platelets in response to a wide range of stimuli, both physiological and pathological. 71,72 This ability can also be mediated by EVs, since they are known to transfer information to target cells through miRNAs, 73 and therefore to determine diverse biological effects in relation to the cargo within the vesicles. With these premises, we hypothesized the presence of miRNAs in PL-based formulations and assessed this by analysing the miRNA profile of two different batches of PL for a total of four replicates. Results showed that the majority of the small RNA content in PL is represented by miRNAs (43%), followed by Y RNAs (17%), anti-sense RNAs (10%), and lincRNAs (8%) ( Figure 5A).
A miscellaneous group is also represented (22%). After applying a cutoff of >10 copies in all analysed batches of PL (average count among the four PL samples; Table 1), the identified miRNAs clustered into three macrogroups based on their expression levels (low, medium and high) when analysed by heatmap with hierarchical clustering ( Figure 5B).  As a further selective step, we set a threshold for miRNAs with over 1000 reads, thus obtaining a shortlist of 39 miRNAs (Table 1) After intersecting each GO term with the top 39 miRNA shortlist, we extrapolated potential eligible candidates for the abovementioned roles. We found that the highest overlap was with the angiogenesis gene list as displayed in the function-expression interaction network that we generated by software ( Figure 5C). Afterwards, we compared the results and shortlisted the main group of 31 miRNAs and a further subset of 11 miRNAs correlating with two and all five GO categories, respectively, where three miRNAs (hsa-miR-320a, hsa-miR-25 and hsa-miR-126) were selected to quantitatively validate the seq data by real-time PCR (Table 2). Notably, miR-126 is a key regulator of angiogenesis and is known as the angio-miRNA and one of the most abundant and specific miR to endothelial cells, human platelets and Note: Specifically, the selection here reported was made by including only those miRNAs with average counts >10 copies. The miRNome was performed on four replicates of PL batches. The first 39 miRNAs with a threshold of over 1000 reads are highlighted in red. Abbreviation: PL, platelet lysate.
Next, we validated the role of miR-126 in the biological effects of PL and EVs on HUVECs, by comparing the sole stimuli that were the media supplemented with 10% PL or 5%, 10% and 20% EVs (with volumes and dilutions adjusted to correspond to 10% PL). Although an increasing but not statistically significant trend of miR-126 was observed among the percentages of EVs, real-time PCR testing confirmed the equivalent content of miR-126 in all media recipes ( Figure 6A). The treatment with 10% EVs was the only able to significantly upregulate the levels of intracellular miR-126 in endothelial cells compared to the negative control ( Figure 6B; p < 0.05). We  Our results demonstrated that EVs after being internalized by endothelial cells, positively enhances angiogenesis by fostering endothelial tubule-like networks in a complex 3D microenvironment, similarly to PL. This phenomenon is in line with those describing the angiogenic effects of EVs derived from circulating intact platelets, 82,83 or from other non-platelet cell types. 84,85 Although PL also contains a plethora of several soluble mediators with angiogenic function, 2,[46][47][48][49] it is conceivable that EVs implement this property that PL normally possesses. Notably, 10% of EVs were revealed as the optimal condition in culture, showing a non-canonical dose-response of EVs without additional effects at a higher percentage in line with the variable biological effects of EVs already verified. 86 The 10% EVs might represent a sort of 'balanced' amount. We have already experienced that the 10% PL itself is the optimal percentage for angiogenic assay also in presence of inhibition of specific soluble factors within the preparation. Below or above this threshold angiogenic effects are not optimal. 46 Nevertheless, some biological differences exist between PL and EVs: these latter preserve the physiological levels of hydrogen peroxide (<100 μM) 87  The angio-miRNA miR-126 is one of the most abundant miRNA expressed in platelets. 103,104 Sharing with miR-320 (that we also found as highly represented) the unique expression also in endothelial cells, miR-126 is able to downregulate adhesion molecules (e.g., VCAM-1) upon the influence of specific cytokines (i.e., VEGF), therefore contributing to endothelial migration, proliferation, activation and vascular inflammation. 75 Exosomes enriched in miR-126 are strictly correlated with protection from ischemic events 105 and atherosclerosis progression. 106 Changes in circulating levels of miR-126 have been described in patients with acute ischemic stroke, 107 coronary artery disease or type 2 diabetes. 108  Our data confirm that miR-126 of platelet origin plays a key role in the angiogenic homeostasis of endothelial cells. Accordingly, results highlight that HUVECs increase intracellular levels of miR-126 upon stimulation with EVs of PL origin, by adding exogenous miR126 by EV transfer when the endogenous miR-126 is silenced. Thus, our results demonstrate that a fraction of the angiogenic effect induced by the whole PL preparation is directly ascribable to the EV cargo, specifically to platelet-derived miR-126.
This study has some limitations. Although we found that VEGF is a target of miR-126, we could not observe any modulatory effect in our system, suggesting that alternative mechanisms are needed to be verified. Only a few of them have been already described. For instance, the DNA methyltransferase, playing a role in hypoxia tolerance, has been found as a target of miR-126 contained in exosomes. 112 Further mechanisms can coexist, including the reduction of cell apoptosis, 105  For instance, the exosomal derived-miR25 has been found to promote angiogenesis, vascular permeabilization, metastatic niches in cancer and involvment in cardiovascular disorders. 116,117 To date, the individual contribution of EVs within PL has not been fully elucidated in terms of regenerative angiogenesis. Certainly, the methodology to manufacture PL severely impacts the quantity and the quality of EVs within the formulations and in particular that employed to concentrate, lyse or activate platelets 118