In vitro evaluation of platelet extracellular vesicles (PEVs) for corneal endothelial regeneration

Abstract Corneal endothelial cells (CECs) slowly decrease in number with increasing age, which is a clinical issue as these cells have very limited regenerative ability. Therapeutic platelet biomaterials are increasingly used in regenerative medicine and cell therapy because of their safety, cost-effective manufacture, and global availability from collected platelet concentrates (PCs). Platelet extracellular vesicles (PEVs) are a complex mixture of potent bioactive vesicles rich in molecules believed to be instrumental in tissue repair and regeneration. In this study we investigated the feasibility of using a PEVs preparation as an innovative regenerative biotherapy for corneal endothelial dysfunction. The PEVs were isolated from clinical-grade human PC supernatants by 20,000 × g ultracentrifugation and resuspension. PEVs exhibited a regular, fairly rounded shape, with an average size of <200 nm and were present at a concentration of approximately 1011 /mL. PEVs expressed cluster of differentiation 41 (CD41) and CD61, characteristic platelets membrane markers, and CD9 and CD63. ELISA and LC-MS/MS proteomic analyses revealed that the PEVs contained mixtures of growth factors and multiple other trophic factors, as well as proteins related to extracellular exosomes with functional activities associated with cell cadherin and adherens pathways. CECs treated with PEVs showed increased viability, an enhanced wound-healing rate, stronger proliferation markers, and an improved adhesion rate. PEVs did not exert cellular toxicity as evidenced by the maintenance of cellular morphology and preservation of corneal endothelial proteins. These findings clearly support further investigations of PEV biomaterials in animal models for translation as a new CEC regeneration biotherapy.


Introduction
The gradual decrease in the number of corneal endothelial cells (CECs) with aging is a clinical problem, given their limited capacity to regenerate in vivo [1,2]. Due to a shortage of donor corneas worldwide, corneal transplantation is extremely limited for treating corneal endothelial dysfunction (CED) [3]. In addition, even the most sophisticated, focused surgical treatments are not without risk [4,5]. Transplantation of cultured CECs was proposed as an alternate therapy for CED. The transplantation of cultured human CECs (HCECs) as a sheet, with [6,7] or without [8] a carrier, and the injection of progenitor cells was investigated in preclinical studies [9,10], but an optimum adjuvant to support injected tissue-engineered HCECs is required.
Blood platelet-derived products are emerging as an attractive physiological therapy for alleviating and treating symptoms of various surface ocular diseases [11][12][13], as they contain multiple synergistic growth factors and cytokines able to promote tissue healing [14,15]. There are reports of the administration of platelet-rich plasma (PRP) into the anterior chamber to treat keratoconus hydrops [16], as an intracameral injection to treat ocular hypotension after glaucoma surgery [17], and for treating retinitis pigmentosa [18]. All studies found beneficial therapeutic results and a lack of toxicity. Besides, blood-derived biomaterial compositions share similarities in protein compositions, beneficial growth factors, vitamins, and electrolytes with tears, further supporting their use in ocular diseases [19].
Our recent study [30] showed that one particular type of human platelet lysate developed in our laboratory, and already successfully evaluated for brain administration [31,32], is a beneficial biotherapy for corneal endothelial regeneration and healing and exerts potent antioxidant activity in cellular models of oxidative stress [33]. However, utilizing a PEV preparation may emerge as being even more valuable due to the higher purity, and the potentially higher chance of standardization compared to platelet lysates due to the ability to characterize EV based on established recommentions from the International Society for Extracellular Vesicles [34]. The aim here was to establish a basis for a pioneering PEV-based biotherapy. For the first time to our knowledge, we studied how PEVs influence adhesion, proliferation, and migration of CECs using cellular models.

Collection of platelet concentrates (PCs)
Clinical-grade human PCs were obtained by apheresis from three healthy donors (Taiwan Blood Service Foundation (TBSF), Guandu, Taipei, Taiwan) according to routine donor selection, collection and processing techniques approved by the TBSF. The PCs were kept at 22 ± 2°C under mixing on a platelet agitator. When the PCs reached their expiration date (5 days after collection), they were delivered within 1 hour to Taipei Medical University (TMU, Taipei, Taiwan) in an isolated shipping box to maintain the temperature at 20 ~ 25°C. The PCs were kept on a platelet shaker at 22°C for up to 24 h with constant moderate agitation until processing, as suggested [35]. Prior to processing, a blood cell count was determined using the ABC Vet blood cell count (ABC Diagnostics, Montpellier, France).

Isolation of platelet EVs (PEVs)
PEVs were isolated under aseptic conditions as follows. Platelets were pelletized by centrifugation of apheresis PCs (at 3000 × g and 22°C for 30 min). The plasma supernatant was centrifuged at 6000 × g and 25°C for 10 min to remove residual platelets, and then at 20,000 × g and 18°C for 90 min to pelletize the PEVs. The PEV pellet was gently washed with HEPES-tyrode buffer to remove residual plasma, and the pellet was carefully resuspended using a volume of 2 mL of HEPEStyrode buffer corresponding to 1/100 th of that of the PCs. PEV samples were aliquoted and stored frozen at −80°C. Further experiments were carried out using pools of three different batches of PEVs. The preparation of PEVs is depicted in Figure 1a.

Morphology and size of PEVs by atomic force microscopy (AFM)
AFM was used to examine PEVs to reveal their three-dimensional topography. Degreased 18-mm coverslips (Marienfeld, Lauda, Germany) were coated for 1 h at 22°C on a shaker at 90 rpm with 200 µL of 0.006 mg/mL cluster of differentiation 41 (CD41; Abcam, where) in Dulbecco's phosphate-buffered saline (PBS; DPBS). PEVs were placed onto the coverslip for 1 h at 22°C and then fixed for 15 min with 300 µL of 4% paraformaldehyde (PFA). The coverslips were dried at 22°C in a sterile laminar-flow hood. PEV morphology was scanned using Nanoview1000 and an Olympus AC240TS probe (FSM, Suzhou, China). The topographic height and phase images were captured at 512 × 512 pixels with a 1-Hz scan rate and an 8.5-min scan time.

Particle concentration by a nanoparticle tracking analysis (NTA)
The amount and size distribution profile of PEVs were determined using an NTA with a 488-nm laser (Nanosight NS300, Malvern, UK). Pools of three different batches of PEVs were used for the experiment.5 μL of samples was diluted 1000fold in 0.1-µm-filtered PBS. A solution of 1000 μL was continuously injected with a syringe pump and evaluated with a Nanosight NS300 instrument. The camera level was set to 16 and the analysis detection threshold to 5. The NTA 3.4 program was used to gather three 60-s measurements and analyze them.

Size distribution by dynamic light scattering (DLS)
EV samples (800 μl) were loaded into folded capillary cells. The size distribution and zeta potential of the PEVs were measured using a Zetasizer Nano ZSP device with a 10-mW He-Ne laser operating at 633 nm and 25°C (Malvern Instrument, Worcestershire, UK).

Determination of growth factor and protein concentrations
The total protein concentration was determined by a BCA Pierce protein assay kit (Thermo Scientific, Rockford, IL, USA). Concentrations of selected growth factors including, plateletderived growth factor (PDGF)-AB, fibroblast growth factor (FGF), transforming growth factor (TGF)-β, vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), brain-derived neurotrophic factor (BDNF), hepatocyte growth factor (HGF), and insulin-like growth factor (IGF)-1 were determined using a sandwich enzyme-linked immunosorbent assay (ELISA) method (DuoSet ELISA; R&D, Minneapolis, MN, USA), as previously published [36][37][38]. PEVs samples were thawed at 37°C and centrifuged at 10,000 × g for 10 min to eliminate cell debris. Samples were then diluted 5 ~ 500-fold based on past laboratory experience. The absorbance at 450 nm was measured using Infinity M200 (Tecan, Mannedorf, Switzerland). Quantities of growth factors were estimated using standard curves, and results are expressed in ng/mL.

Profiling by a proteomics analysis
Pooled samples of PEV batches were precipitated in cold acetone overnight at a 4:1 acetone/water ratio at −20°C as described before [31]. After washing, the pellet was centrifuged at 15,000 × g for 10 min and air-dried to evaporate any remaining acetone. The pellet was then dissolved in 150 μl of 6 M urea. Equal amounts of protein samples were subsequently delivered to National Taiwan University (NTU)'s proteomics core facility for a proteomic analysis by liquid chromatography-tandem mass spectroscopy (LC-MS/MS) [30].

PEV labeling and internalization by CECs
PEVs were labeled with the fluorescent dye 5-(and-6)-carboxy fluorescein diacetate succinimidyl ester (CFDA-SE; Biotium, Fremont, CA, USA) following the supplier's instructions. Briefly, 12 µl of 10 mM CFDA-SE was combined with 400 µl of a PEV suspension containing approximately 10 11 PEVs/mL (with a final CFDA-SE concentration of 20 µM) and incubated for 2 h at 37°C on a suspension mixer. Excess dye was removed by injecting a 412-µl mixture of CFDA-SE PEVs into a 5-mL Sephadex G-25 column (Sigma-Aldrich). In total, 4 mL of flowthrough fractions 2 ~ 5 (1 mL/fraction) containing the CFDA-SElabeled PEVs as detected by DLS was recovered and concentrated at 2500 × g for 2 min using an Amicon 50 K ultracentrifuge filter. Sterilized 15-mm coverslips were added to 24-well plates and coated with 1 mg/ml fibronectin for 3 h at 37°C. About 100 μg of CFDA-SE-labeled PEVs (based on BCA quantification, corresponding to approximately 30 × 10 8 PEVs) was added to 24-well plates with 50,000 B4G12 cells and incubated for 3 h. Cells were fixed with 4% paraformaldehyde (PFA) at room temperature for 15 min, permeabilized in 0.2% Triton-X 100 (Sigma-Aldrich, Darmstadt, Germany) at room temperature for 15 min, and incubated with cytoskeleton actin filament phalloidin-Alexa fluor 555 (Invitrogen) in 1% BSA for 1 h. Cells nuclei were stained with DAPI. The cellular uptake of PEVs was examined with a Leica Stellaris 8 confocal microscope (Leica, Wetzlar, Germany).

Proliferation assay and morphological assessment of PEV-treated cells
Human B4G12 cells were seeded at 2 × 10 5 cells/well in a 24-well plate with HE-SFM containing 10 ng/ml bFGF and incubated overnight at 37°C. To test the proliferation capacity of PEVs, CECs were treated for 24 or 48 h in serum-free media with or without 10 (5 × 10 8 EVs), 50 (15 × 10 8 EVs), and 100 µg (30 × 10 8 EVs) PEVs. The cell viability rate was measured by the cell counting kit-8 (CCK-8, Sigma-Aldrich). The absorbance at 450 nm was determined using a microplate reader. The number of viable cells was standardized to 100% of untreated control cells, while the number of untreated control cells was given a value of 100%. A phase-contrast microscope was utilized to observe the cell morphology (Leica).

Adhesion assay of PEV-treated cells
B4G12 cells were seeded at a density of 5000 cells/well in 24well plates and cultured in HE-SFM for 24 h or until confluent with or without 10 (5 × 10 8 EVs), 50 (15 × 10 8 EVs), and 100 µg (30 × 10 8 EVs). Cells were dissociated with 0.05% trypsin and transferred to a 24-well plate coated with 1 mg/ml fibronectin. About 50,000 cells were added to each well, along with 10, 50, and 100 µg of PEVs, and incubated for 3 h at 37°C. Cells were washed thoroughly and fixed for 15 min in 4% PFA. Cells were washed and stained with 0.5% crystal violet for 10 min at room temperature before being examined under a Leica phase-contrast microscope (Tokyo, Japan). The number of adherent cells in other wells was determined using the CCK-8 assay [39][40][41].

Proliferative biomarkers of PEV-treated cells
The Ki-67 proliferation marker was utilized to determine the proliferation of B4G12 cells. B4G12 cells were seeded at a density of 5000 cells/well in 24-well plates and incubated at 37°C overnight. The medium was changed to serum-free conditions following treatment with or without 10, 50, and 100 µg PEVs for 24 h. Cells were fixed with 4% PFA and permeabilized using 0.1% Triton X-100 at room temperature. Cells were incubated overnight at 4°C with a 1:200 dilution of AlexaFluor 488 Ki-67 (Novusbio, Colorado, USA) in blocking buffer, followed by 1 h at room temperature with a 1:500 dilution of the rabbit AlexaFluor 488 primary antibody (Abcam, Massachusetts, USA). DAPI was used to stain cell nuclei. Fluorescence was visualized using a Leica fluorescence microscope. Ki67 expression was measured in five different fields of view by ImageJ software (Scion Software, National Institutes of Health, Bethesda, MA, USA).

Wound-healing scratch assay of PEV-treated cells
Human B4G12 cells were cultivated at 2 × 10 5 cells/well in a 24well plate in the presence of HE-SFM with 10 ng/ml bFGF and incubated at 37°C overnight. Cells were scraped with a 200-µL pipette tip and rinsed three times in PBS. The medium was changed to serum-free conditions after treatment with and without 100 µg PEVs (5 × 10 8 EVs). Cells were viewed at 0, 6, 12, and 18 h under a microscope, and photos were taken [42]. The enhancement of wound closure from five different photos was examined using ImageJ software (NIH) [42]. The wound-healing index, i.e., the percentage difference between the original wound area and the end wound area, was used to express the results.

Statistical analysis
All tests were carried out in three separate independent experiments. Data are presented as the mean and standard deviation (SD). For the statistical analysis, GraphPad PRISM was utilized (vers. 7.0, California, USA). A one-way analysis of variance (ANOVA) was used to establish significance, followed by Tukey's post-hoc analysis. Student's t-test was used to determine the significance of the differences in the wound-healing scratch assay. We utilized a statistically significant p value of 0.05 (* p < .05, ** p < .01, *** p < .001, **** p < .0001).

Blood cell count
Platelet counts in the initial PCs ranged from 420 to 540 × 10 3 cells/mm 3 comparable to 2.5 to 3.6 x 10 11 /unit. The residual red blood cell (RBC) <0.6 x 10 6 /mm 3 and white blood cell (WBC) <0.7 x 10 3 /mm 3 values were also below the detection thresholds.

Biophysical and functional characterization of PEVs
We characterized PEV products obtained by 20,000 × g ultracentrifugation of PC supernatants using four methods: AFM, NTA, DLS, and Western blotting utilizing standard PEV markers. In both two-and three-dimensional topographic AFM pictures, AFM revealed that PEVs had a regular, spherical form, with size heterogeneity. PEVs had a diameter of 84 ~ 150 nm (Figure 1b). NTA revealed a homogeneous population of nanoparticles in the PEV preparation with a mode size of 128 nm and a concentration of about 10 11 particle/ mL (Figure 1c). DLS showing an average diameter of approximately 202 nm (Figure 1d). The NTA and DLS data did not reveal the presence of micrometer-size events, indicating that intact cells and large cell debris were removed during the PEV preparation process. The EV markers of CD9 and CD63 and the platelet membrane markers of CD41 and CD61 were clearly expressed in PEVs according to the Western blot analysis (Figure 1f-g and supplementary file 2).

Growth factor measurements, proteomic profiling, and biochemical analysis
Using the BCA assay, the average total protein concentration of the PEV preparations was found to be 3 ~ 4 mg/mL. According to ELISA results, PEVs contained a considerable number of growth factors (PDGF, TGF-β, BDNF, EGF, FGF, VEGF, IGF, and HGF;  Table 1). Cellular components and functional annotations were discovered using a DAVID Bioinformatics Resources 6.8 GO enrichment analysis. Extracellular exosomes (terminology equivalent to EVs in this context) were the most abundant proteome component found in PEVs, according to the cellular component analysis (Figure 2b). Signal transduction, cell adhesion, immunological processes, complement activation, metabolic processes, cell migration regulation, and wound healing were among the biological processes involved (Figure 2c). The most highly enriched cluster found in the functional annotation clustering analysis was related to cell adherens junction and cell adhesion (Figure 2d). Supplementary Table 2 shows the biochemical analysis composition of PEVs.

PEVs were internalized by CECs
To monitor their uptake by B4G12 cells, PEVs were first labeled with a green fluorescent CFDA-SE dye. Labeled PEVs were then added to B4G12 cells for 3 h. The labeled PEVs were discovered in the perinuclear region of B4G12 cells (Figure 3), consistent with PEV internalization by B4G12 cells.

PEVs promote CEC proliferation and migration
The proliferation rate of human B4G12 cells was measured to assess the influence of PEVs on CEC proliferation. B4G12 cells were treated with 10, 50, or 100 μg PEVs or kept in their normal culture conditions without PEV exposure (control). PEV-induced cell proliferation was observed using a phase-contrast microscope. At 24 and 48 h, proliferation rates of human B4G12 cells treated with 50 and 100 μg PEVs had significantly increased in a dose-dependent manner (p < .05, Figure 4a,b).
Using a scratch test, we evaluated the capacity of PEVs to improve wound closure in human B4G12 cells. B4G12 cells were scraped with a 200-μL pipette tip when they had reached confluence. After being thoroughly cleaned, cells were cultured in serum-free medium with or without 100 μg PEVs (30 × 10 8 EVs). Wound closure was assessed at 0, 6, 12, and 18 h after scratching to determine the wound-closure rate. When comparing the control and 100 μg PEVs, results showed that 100 μg PEVs significantly improved endothelial wound closure in B4G12 cells after 12 and 18 h (p < .05, p < .05, respectively; Figure 4c,d).

PEVs promotes the Ki-67 CEC proliferation marker
To further confirm the proliferation assay, we performed Ki-67 immunofluorescence(IF) labeling using B4G12 cells (green) (Figure 5a). We discovered that after 24 h of treatment with PEVs, percentages of Ki67 expression in B4G12 cells treated with 50 and 100 μg PEVs were significantly increased compared to the control group (p < .01, p < .001, respectively; Figure 5b). This finding is in line with the CCK8 experiment, which found that PEVs increased B4G12 cell growth. The cell viability rate is given as a percent of the control (100%). Representative image of B4G12 cell wound healing at various time periods following scratching after treatment with and without 100 µg PEVs (30 × 10 8 EVs) (C). ImageJ software was used to quantify the wound closure rate (D). * p < .05, ** p < .01 vs. the control.

PEVs enhanced the CEC adhesion rate
A cell-adhesion test was performed to see whether PEVs promoted CEC adhesion. Figure 6a shows sample photos of adherent B4G12 cells after crystal violet staining with different dosages of PEVs. After 3 h of incubation, PEVs at doses of 50 and 100 μg had significantly increased B4G12 cell attachment compared to the control (p < .05, p < .01, respectively; Figure 6b).

Effect of PEVs on corneal endothelial protein expression
We tested the influence of PEVs on CEC markers in human B4G12 cells. After CECs were treated with PEVs for 48 h, we assessed their protein expression using IF labeling and Western blotting. B4G12 cells treated with 100 μg PEVs showed a hexagonal form identical to the control group, according to IF labeling. Levels of ZO-1, a corneal endothelial protein marker, and N-cadherin were found to be similar in PEV-treated cells, indicating that corneal cell integrity had been preserved. Furthermore, IF staining revealed that PEVs did not produce cell toxicity, as seen by the maintenance of ZO-1, N-cadherin, and Na + /K + -ATPase protein expressions (Figure 7a-c). Compared to control cells, ZO-1 expression was maintained in B4G12 cells treated with 10, 50, and 100 μg PEVs, as shown by Western blotting. In B4G12 cells treated with PEVs, N-cadherin and Na + /K + -ATPase expressions were significantly higher than in control cells. As seen in Figure 7 d and e, PEVs exhibited no cytotoxic effects compared to control cells.

Discussion
Clinicians and researchers in the field of ophthalmology have long worked on developing tissue engineering-based biotherapies that use cell carriers for cornea transplantation, as an alternative or adjuvant to corneal transplantation [43], as well as topical treatment for in situ repair. Recently, there has been a rise in interest in researching the use of clinical-grade, blood-derived products for direct in situ cell and tissue regeneration, notably for treating ocular surface problems [15]. Numerous studies in alternative regenerative medicine are now focusing on applications of EVs from various cells [44] and PEVs in particular [24]. PEV preparations may be more effective than platelet lysates due to their superior purity and safety, as well as a greater probability of standardization due to the removal of the free protein components that are present in PRP or platelet lysates. Serum and PRPderived microvesicles were shown to improve keratinocyte wound healing in vitro [45], chondroregeneration in vitro [46], cutaneous wound healing in diabetic rats [47], in vitro and in vivo bone regeneration [48], and muscle damage recovery in clinical settings [49]. In vitro [50] and in vivo [51], platelet microparticles induce neurogenesis. To the present, however, no research on utilizing PEVs for corneal endothelium regeneration treatment has been published, justifying our study.
An in-depth investigation of the composition of our PEV preparations using an ELISA and LC-MS/MS-based proteomics revealed the presence of different growth factors known to be favorable for ocular regeneration. Our PEVs were comprised of PDGF, EGF, TGF-β, HGF, and IGF, at doses in the 1 ~ 100-ng/ml range, which are critical for the cornea's regeneration capabilities [80]. Among those, PDGF, BDNF, and TGF-β are growth factors particularly abundant in platelets, and therefore their relative abundance in PEVs is expected, while the relatively low content in IGF reflects the fact that this growth factor is largely present in plasma, which is removed during the preparation of this PEV preparation [27,32]. We used DAVID Bioinformatics Resources proteomics tool to learn about the functions and pathways associated with PEV protein contents. The presence of PEV constituents were implicated in numerous biological processes, such as signal transduction, cell adhesion, immunological processes, and complement activation, as validated by a gene ontology (GO) analysis. The most enriched cluster found by a functional annotation clustering analysis was linked to cell adherens junctions and cell adhesion, both of which are important in the cornea endothelial repair mechanism [52,53].
Based on the DAVID cellular component analyses, proteins associated with extracellular exosomes (or EVs) were confirmed to be the most abundant components in PEVs. This proteomics finding confirmed the EV characterization results, which showed that our isolation approach created well isolated EVs. Our PEVs ranged in size from 128 to 202 nm according to the NTA data and expressed the CD9 and CD63 EV markers in accordance with recommendations of the International Society of Extracellular Vesicles [54]. The characteristics (number, size, marker expression) of the PEVs preparations were very consistent with those prepared in our laboratory for other studies.
In this study, we first examined the possible toxicity and survival of CECs after PEV therapy. According to our findings, PEVs induced no cellular toxicity, improved the proliferation rate and migratory ability of CECs in a wound-healing assay, and expressed higher levels of the Ki-67 proliferation marker. It was interesting to note that the regeneration ability of the corneal endothelium attributed to PEVs was dose-dependent. As a result, we confirmed that the various trophic factor combinations contained within PEVs are likely to contribute to corneal regeneration repair. Furthermore, PEVs stained with CFDA-SE were taken up by CECs after 3 h of incubation, supporting the direct effect of PEVs on CEC proliferation and their migratory ability.
One of the common applications of PEVs (with different preparation methods) has been in treating injuries and wounds. The woundhealing process, in particular, was associated with increases in fibroblast and keratinocyte migration and proliferation in vitro [46,47]. These effects can be attributed to the cargo delivered by PEVs, which contain PDGF, FGF-2, TGF, and VEGF [47]. Our study found that PEVs significantly improved wound migration of CECs. Previous research on PEVs (prepared in various ways) demonstrated that PEVs enhance cellular angiogenic mobilization and migration. Several investigations [55][56][57] reported vaso-regeneration and the preservation of arterial integrity following damage. In vitro and in vivo studies later demonstrated increased vascular endothelial cell recruitment and adhesion, and a regenerative effect [55]. Another study found that PEVs influenced endothelial permeability [58]. The role of platelet exosomes in bone regeneration was studied using in vitro and in vivo models of osteonecrosis. Those investigations showed that PRP exosomes can promote proliferation while preventing apoptosis, resulting in bone healing [59]. In more-thorough studies, PEVs were also linked to neuro-regenerative responses. Platelet microparticles induce brain stem cell proliferation and neurogenesis in vitro, which was connected to numerous growth factors contained in PEVs [50]. An in vivo investigation using rats that examined a motor impairment test showed improved neurological competence after an ischemic stroke [51].
PEVs were discovered to improve CEC adherence. This finding can be attributed to the adhesion process in platelet activity, which assists in wound healing [60]. A prior study that used a platelet lysate gel discovered that it was advantageous for cell adhesion in cellular tissue engineering [61]. Several materials were investigated to promote corneal adhesion, which in turn improves corneal wound healing [62][63][64]. In cornea endothelial regenerative therapy, adhesion ability is essential as it may contribute to the practical use of tissue-engineered human corneal endothelial cell lines [65].
PEV therapy did not generate cellular damage, as evidenced by the maintenance of the corneal protein expressions of ZO-1 and N-cadherin, and of Na + /K + -ATPase according to Western blotting and IF labeling. Junctional ZO-1 and N-cadherin adhesion proteins perform critical roles in cell integrity, cell connection stabilization, and cell morphology maintenance [52]. Furthermore, the cornea's transparency is principally determined by sodium bicarbonate ion transport, which is mediated by the Na + /K + -ATPase pump [66]. Our findings demonstrated that PEV treatment could sustain the expressions of ZO-1, N-cadherin, and Na + /K + -ATPase, implying that PEVs were not harmful to corneal cells. The Western blot analysis revealed that PEVs increased the expression of the adhesion protein N-cadherin. These data are relevant to the proteomic results of the DAVID bioinformatic functional annotation clustering analysis that the cell adhesion and cell cadherin pathways were the most highly enriched clusters.
Results of our research demonstrated the possibility of employing a new blood-derived product to regenerate the corneal endothelium. One value of our study is to have used clinical-grade PCs collected from healthy donors as starting materials, which are of a morestandardized quality than PRP, which is known to exhibit variable blood cell counts depending upon the production devices, in particular. Indeed, allogeneic PC produced by blood establishments under good manufacturing practices represent a biological material meeting established quality specifications for platelet count and residual red blood cells and white blood cells, thereby attenuating the level of variability [67]. Several production variables may, however, potentially affect the characteristics of the PEV recovered from allogeneic PCs. Platelet apheresis procedures have been shown to impact the number of PEVs generated by platelets [68]. The shelf-life of the source PC varies between 5 and 7 days depending on legislations potentially impacting the extent of platelet storage lesions [69] and the number and protein content of EVs released by platelets during storage [69,70]. The use and type of platelet additive solutions to substitute partially for the plasma in which platelets are suspended [71] can influence the content and type of EVs released by platelets [72]. The implementation of a pathogen reduction treatment, combined or not with a phase of adsorption to remove unwanted chemicals, can influence the miRNA content of EV. Last but not least, the introduction of cold storage of PC to decrease the risk of bacterial contamination and extend shelf-life could profoundly influence EV release by platelets [73,74]. Therefore, as is the case for PRP for regenerative medicine [75,76] and allogeneic human platelet lysates for cell therapy [77,78], monitoring process variables and controlling characteristics will be essential to ensure consistent clinical outcomes of PEVs.
The PEV production process uses PC supernatants, making it compatible with the use of pelletized platelets to produce lysates for other therapeutic applications, like treating disorders of the central nervous system [31,32], thereby optimizing blood resources and contributing to the United Nations' Sustainable Development Goals (SDGs) at the global level, including lowand middle-income countries. We have also recently shown that various PEV preparations from outdated PCs maintain functional regenerative capacity in neuronal cell models [79]. However, systematic studies are needed to understand whether PEVs from fresh or outdated PC are equally effective in various regenerative medicine applications.
One relevant limitation of our study is that we did not compare the functional activity of these PEVs to that of a platelet lysate. This is due to several reasons. First, we wanted to prove first the functional activity of the PEV in the particular CEC model used. Second, the PEVs used in this study, by contrast to our previous recent work [79], were isolated from the plasma supernatant of a platelet concentrate, not from a platelet lysate, therefore constituting a stand-alone product of its own. Nevertheless, experiments comparing a full platelet lysate (which contains both free proteins and PEV), isolated PEVs, and PEV-depleted platelet lysate would be valuable to discriminate the respective role of free trophic factors and PEVs in the repair of the damaged corneal endothelium. In a very recent study using a dedicated heat-treated platelet lysate for brain administration, we have found that the PEV depletion of this material by nanofiltration on 19-nm filter, did not significantly affect the neuroprotective and antiinflammatory activities in cellular and animal models [81]. This suggested that for this specific platelet lysate and that particular application, the PEV did not exert a major contribution to functionality.
In conclusion, the PEVs in this study were successfully manufactured, and their non-toxicity was established. Our study showed that PEVs could be effective for corneal endothelial regeneration, due to increased support forCEC proliferation, migration, and adhesion. It should be examined whether PEVs can be employed topically or as a component of a biomaterial for pre-established endothelial keratoplasty to improve success rates. These data strongly imply that PEVs should be further studied in preclinical animal models as a promising CEC regeneration biotherapy.

Disclosure statement
No potential conflict of interest was reported by the author(s)

Funding
The study was supported in part by a Taipei Medical University-Taipei Medical University Hospital research grant (108TMU-TMUH -03) from Taipei Medical University Hospital and the Ministry of Science and Technology of Taiwan (MOST 107-2314-B-038-032-MY2). The funding bodies had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Authors' contributions
RW, TB, and TJW conceived the experimental design, discussed the results, and wrote the manuscript; RW acquired the data and performed the analysis. YWW, LD, and DYL provided scientific and technical assistance. All authors approved the final article.

Ethical approval and consent to participate
The Institutional Review Board of Taipei Medical University approved this study (TMU-JIRB no. 201802052)