Human Placental Mesenchymal Stem/Stromal cells (pMSCs) inhibit agonist‐induced platelet functions reducing atherosclerosis and thrombosis phenotypes

Abstract Mesenchymal stem/stromal cells isolated from human term placenta (pMSCs) have potential to treat clinically manifested inflammatory diseases. Atherosclerosis is a chronic inflammatory disease, and platelets play a contributory role towards its pathogenesis. During transplantation, MSCs interact with platelets and exert influence on their functional outcome. In this study, we investigated the consequences of interaction between pMSCs and platelets, and its impact on platelet‐mediated atherosclerosis in vitro. Human platelets were treated with various types of pMSCs either directly or with their secretome, and their effect on agonist‐mediated platelet activation and functional characteristics were evaluated. Human umbilical vein endothelial cells (HUVECs) were used as control. The impact of pMSCs treatment on platelets was evaluated by the expression of activation markers and by platelet functional analysis. A subset of pMSCs reduced agonist‐induced activation of platelets, both via direct contact and with secretome treatments. Decrease in platelet activation translated into diminished spreading, limited adhesion and minimized aggregation. In addition, pMSCs decreased oxidized LDL (ox‐LDL)‐inducedCD36‐mediated platelet activation, establishing their protective role in atherosclerosis. Gene expression and protein analysis show that pMSCs express pro‐ and anti‐thrombotic proteins, which might be responsible for the modulation of agonist‐induced platelet functions. These data suggest the therapeutic benefits of pMSCs in atherosclerosis.

haemostasis, yet from recent investigations, it has become evident that they play a central role in promoting inflammation through their interactions with leukocytes. [6][7][8] These interactions make them relevant in the pathophysiology of atherosclerosis, which is now considered an inflammatory disorder. 7,[9][10][11] Clinical studies also support an association between platelet reactivity and prognosis in patients with coronary disease. 12 In addition, numerous studies have linked platelets and hyperlipidaemia to atherothrombotic risk. 13 Drugs causing specific inhibition of platelet function are important for treatment of cardiovascular and cerebrovascular diseases. 2 Platelets express multiple receptors on its surface, which facilitate relevant physiological processes during haemostasis. [14][15][16] At the site of vascular injury, the endothelial lining and the damaged tissue generate adhesion molecules and soluble platelet agonists, such as collagen, von Willebrand factor (vWF), adenosine diphosphate (ADP) and thrombin. They interact with specific receptors on the platelet surface and initiate platelet activation followed by adhesion, activation and aggregation, thereby promoting plug formation and thrombus development. 17,18 CD36 (also called as glycoprotein IV) is a multi-ligand receptor expressed in macrophages, dendritic cells (DCs) and in platelets, which recognizes certain oxidized phospholipids. 5 Recent studies have indicated that CD36 binding to oxidized LDL (ox-LDL) results in activation of specific signalling cascade that induces platelet activation. 19,20 This binding and activation is believed to contribute to the pro-coagulant state associated with hyperlipidaemia and contribute to atherosclerosis. 21,22 Mesenchymal stem/stromal cells (MSCs) have emerged as an option in cell-based therapy for many diseases, including diabetes, cancer, CVDs and atherosclerosis. [23][24][25][26] MSCs are multipotent adult stem cells isolated from different tissues including adipose, dental pulp, bone marrow, umbilical cord and placenta. Their therapeutic potential has been attributed to their differentiation potential, immunomodulatory capabilities and tissue regeneration properties. 27 We have isolated and characterized MSCs from various parts of human term placenta (pMSCs) including, decidua basalis (DBMSCs), decidua parietals (DPMSCs) and chorionic villous (CVMSCs) regions. [28][29][30] pMSCs exhibit immunomodulatory properties, making them excellent sources for cell-based therapies against immune-mediated diseases. By secreting soluble factors, they exert immune responses on T cells, B cells, monocytes and macrophages. 31 We have previously reported the therapeutic potential of pMSCs in treating inflammatory diseases such as cancer and diabetes, 32,33 and in this study, we investigated if pMSCs modulate platelet functions and in turn play any restrictive role towards pathogenesis of atherosclerosis and thrombosis. Towards that goal, we studied the effect of pMSCs on platelet activation in presence and absence of various agonists.
The effect of pMSCs at both intracellular contact (IC) and their secretome levels (using conditioned medium (CM)) was evaluated and compared their effect to those of human umbilical vascular endothelial cells (HUVECs). In addition, we studied the effect of pMSCs on ox-LDL-induced-CD36-mediated aggregation of platelets, using the standard platelet-specific functional assays. DBMSCs and DPMSCs but not CVMSCs reduced agonist-induced activation of platelets significantly, both via the IC and with the CM treatments. However, no change was observed in the resting platelets. Decrease in platelet activation translated into their diminished spreading, limited adhesion and minimized aggregation. Most strikingly, pMSCs decreased ox-LDL-induced-CD36-mediated platelet activation significantly, as compared to resting and/or pMSCs only treated platelets. pMSCs express pro-and anti-thrombotic proteins, which may be responsible for such outcome. However, their specific roles and mechanism of action need further investigation.

| Reagents and antibodies
Agonists ADP, and collagen (cat#5852), were purchased from Helena

| Collection and processing of human placentae and umbilical cord tissues
Human placentae and umbilical cord tissues were collected from healthy donors with uncomplicated gestation and with normal vaginal delivery after 38-40 weeks of gestation. Foetal viability and age of gestation were confirmed by ultrasound examinations during the gestation period. The tissues were processed within 2 h of delivery following proper experimental procedures and guidelines.

| Blood collection and platelet isolation
Human blood was collected from healthy and fasting volunteers in anticoagulant acid citrate dextrose (ACD) containing tubes. Platelets were isolated as described earlier. 34 Briefly, the tubes were centrifuged at 900 RPM for 15 min (min) at room temperature and platelet-rich plasma (PRP) was collected. Prostaglandin E 1 (PGE 1 ) at 75 nM was added to the PRP before centrifuging again for 10 min at 420g. Supernatant was discarded, and the pellet was resuspended in citrate glucose sodium chloride (CGS) solution containing PGE 1 at 75 nM final concentration. The suspension was centrifuged at 420g for 10 min, supernatant was discarded, and the platelets were resuspended in Tyrode's Buffer (138 mm NaCl, 5.3 mm KCl, 0.33 mm Na 2 HPO 4 , 0.44 mm KH 2 PO 4 , 5.5 mm glucose, pH 7.4) and centrifuged. The supernatant was discarded, and washed platelet pellets were suspended in Tyrode's buffer without PGE 1 , calcium and magnesium, and adjusted to a final concentration of 2.5 × 10 8 platelets/ ml. Washed platelets were rested at room temperature for 30 min before use.

| Platelet spreading, adhesion and light transmission aggregometry
Resting and pMSCs/HUVEC and CM-treated platelets (agonist treated or untreated) at 1 × 10 5 /ml were incubated on human fibrinogen (Enzyme Research)-coated glass coverslips (Corning), for 15 and 30 min at 37°C. Excessive protein and non-adhered platelets were washed out by rinsing. Adhered platelets were fixed with 3.7% paraformaldehyde and permeabilized with 0.01% Triton X-100.
Fixed platelets were stained for F-actin by Rhodamine Phalloidin.
The slides were visualized and photographed using a bright field microscope. We have visualized 10 slides, and 5 to 10 pictures were captured from each slide for further analysis. The surface area of the platelets was quantified using the ImageJ software ® (NIH, USA).
For adhesion assay, pMSCs/HUVEC and CM-treated and CMuntreated platelets were challenged with appropriate agonists before adding to the fibrinogen-coated 96-well plates. The wells were blocked with BSA and incubated at 37°C for 30 and 60 min. After washing the wells, platelet adhesion was quantified by acid phosphatase activity assay, and OD was measured at 405 nM, as described earlier. 36 Aggregation assays for pMSCs/HUVEC and CM-treated and CM-untreated platelets followed by challenging with different doses of agonists (collagen and ADP) were evaluated in stirring conditions in an eight channel aggregometer, PAP-8E (Bio/Data Corporation).
The samples were heated to 37°C for 5 min before subjecting to analysis. Aggregation was recorded following the degree of electrical impedance in washed and treated samples. Samples were run for 10 min, and data were recorded when the level of aggregation had reached the maximum extent.

For methods of isolation and culture of pMSCs and HUVECs, collection of conditioned media (CM), real-time PCR (RT-PCR), immunoblotting
and statistical analysis, please refer to the Supplementary Methods.

| Dose response for CM treatment on platelet activation
Agonists (Collagen, U46619 and ADP) were added to the pMSCs CM-treated platelets at different doses (collagen (1 μg/ml and 2.5 μg/ml), ADP (1 μM and 5 μM) and U46619 at 10 μM) and incubated for staggered time points ranging from 2-10 min. As shown in Figure 1, the expression of activation marker P-selectin and PAC1 decreased robustly in a dose-dependent manner, ranging from 5% to 20% as compared to the collagen-treated (CM-untreated) control. This change in expression levels was significant (p < 0.05) for 10% and 20% compared to the 5% treatment in CM-DBMSCs  In resting (CM and agonist untreated) platelets, the expression levels of both P-selectin and PAC1 were significantly low (p < 0.05) as compared to the agonist treated and activated platelets.

| pMSCs partially inhibit agonist-induced platelet activation
DBMSCs and DPMSCs inhibited activation of collagen-treated platelets at all cellular ratios tested in the co-culture experiments. As shown in Figure 2A and B, there was a significant (p < 0.05) decrease in activation at the 1:2 and 1:5 (platelet to MSCs) ratio, as compared to untreated control. However, as observed for CM-CVMSCs, the cellular F I G U R E 1 Conditioned media of pMSCs inhibit platelet activation: Washed platelets incubated with conditioned medium (CM) at 5%, 10% and 20% concentration and then induced with collagen. Flow cytometry was performed to assess the expression of activation markers, P-selectin (CD62P) and PAC1. As 2.5 × 10 6 cells are an agreed dose necessary for successful transplantation in clinical settings, 37 we used these cell numbers against specific number of platelets to achieve a ratio of 1:5. As shown in Figure 3, the expression levels of both P-selectin and PAC1 were significantly reduced (p < 0.05) after treatment with DBMSCs as well as with the DPMSCs, whereas CVMSCs have no effect on collageninduced platelet activation. HUVECs decreased expression levels of both the P-selectin and the PAC1 in collagen-treated platelets. This phenotype was not observed in resting platelets, irrespective of their treatment with cells or untreated controls ( Figure 3A and B). dose-dependent effect pertaining to reduced platelet activation after their induction with ADP. However, there was no effect on P-selectin or PAC1 expression in washed platelets with or without treatment with pMSCs or HUVECs ( Figure S3B-D). Although there was a decrease in platelet adhesion with CVMSCstreated platelets, yet the data were not statistically significant ( Figure 4C). Figure 4D

| DBMSCs inhibit ox-LDL-induced-CD36mediated platelet activation
As shown in Figure [38][39][40] Therapeutic potential of pMSCs to treat several inflammatory diseases, such as diabetes and cancer, was reported earlier. 28,32,[38][39][40][41] Since atherosclerosis is also an inflammatory disease, 3  In order to assess their impact on platelet activation, we investigated the effects of pMSCs on platelet physiology by performing cell-tocell contact treatment assays followed by evaluation of platelet activation markers. Using various clinically relevant number of stem cells versus the platelets, different ratios were tested to examine the effect of pMSCs on platelet activation. At a ratio of 1:1, 1:2 and 1:5, DBMSCs and DPMSCs reduced, the agonist (collagen and thromboxane A2) induced activation of platelets in a dose-dependent manner.
However, it has been shown that a minimum dose of 2.5 × 10 6 CD34 + cells was necessary for successful engraftment for neutrophils and platelets, yet reinfusion of 5.0 × 10 6 CD34 + cells resulted in prompt engraftment and was agreed to be the preferred target. 37 However, as observed for the CM, the inhibitory effect of cell-to-cell contact was observed for DBMSC and DPMSC only and was limited to collagen and thromboxane A2 agonists only. CVMSCs as compared to HUVECs did not show any effect on collagen and thromboxane A2treated platelets. These data confirm the outcome observed for CM, and the surface expression may be somewhat similar to the secretome of pMSCs. We have previously reported that pMSCs show surface expression of several important molecules which could inhibit the agonist-induced activation of platelets in a co-culture setup. 38,39 In order to assess the effect of pMSCs on platelet hyperactivity post-agonist treatment, we performed platelet functional analyses and confirmed the inhibitory role of pMSCs and their secreted products. Several parameters that describe platelet activation in diseased setup at the site of injury, or after agonist induction, include their interaction to other cells (leukocytes or endothelial cells), their adhesion to artificial or natural surfaces, the receptor conformational change, changes in their actin cytoskeleton, secretion of molecules from the granules and most importantly the platelet-platelet aggregation. 42 In classical platelet activation pathways, agonists, such as collagen, thrombin or ADP, interact with platelet surface receptors and initiate intracellular signalling events resulting in activation of integrins, which culminate at shape change, secretion of platelet granule contents and platelet aggregation. 43 With regard to these functional outcomes after agonist and/or pMSCs treatment, we report that barring CVMSs, both DBMSCs and the DPMSCs, increased platelet spreading, decreased platelet adhesion and platelet aggregation significantly, as compared to the HUVEC controls. These data suggest that DBMSCs and DPMSCs, but not the CVMSCs, possess anti-thrombotic activities. This differential outcome may be related to the particular niche from which each cell type was isolated. However, further investigation is required to decipher the mechanism of difference in the outcome of the pMSCs isolated from the same source, the placenta. As expected, these inhibitory phenotypes were not observed in platelet-rich plasma (PRP) or washed platelets, treated with DBMSCs or DPMSCs and not treated with any agonist. These data negate the role of any platelet-specific intrinsic factor responsible for inhibition of platelet activity, as observed in pMSCs-treated and agonist-induced platelets. However, a moderate increase in activation of resting platelets while treating F I G U R E 7 DBMSCs inhibit platelet activation induced by ox-LDL and collagen: (A) Washed platelets were initially treated with DBMSCs for 10 min and subsequently incubated with 50 μM ox-LDL (as per manufacturer's instructions) for 30 min. Collagen at 2.5 µg was added under stirring conditions, and aggregation was recorded. In comparison, other experimental conditions such as platelets treated with collagen and ox-LDL, platelets treated with collagen only and platelets treated with DBMSCs and induced with collagen were included in the experiments. (B) Final percentage aggregation of platelets with different treatments is shown as a bar graph with ±SE from 3 independent experiments. *p ≤ 0.05 them with CVMSCs may prove beneficial in other haemostasis setup and further suggest the specific role of each cell type studied. It further proves that CVMSCs therapy may not be beneficial in certain diseases associated with a risk of thromboembolism. These results are in agreement with the previously reported studies performed on platelets using CD133 + bone marrow, lipoaspirate and cord blood stem cells. 44,45 As previously reported, pMSCs display immunomodulatory properties making them useful therapeutic agents for various diseases. 31  It has been demonstrated that ox-LDL binds to platelets via CD36 and induces platelet activation. 21 Since platelet hyperactivity is as- is expressed on platelets in addition to other mammalian cell types including the macrophages, DCs, adipocytes. 48 It binds to ox-LDL and activates a specific signalling cascade that induces platelet activation. 5,49 It has also been demonstrated that CD36 contributes to thrombus formation in response to vascular injury in mice confirming the role of this pathway in platelet activation and promoting atherosclerosis. 43 Results obtained in our study provide strong support to the hypothesis that pMSCs treatment not only inhibits the agonist-induced platelet activation but also significantly inhibits the lipid-induced platelet aggregation, making them a potential candidate for atherosclerosis therapy.

| CON CLUS ION
Our studies primarily demonstrate that pMSCs provide a protective role in agonist-induced platelet activation, inhibiting thrombosis and atherosclerosis. These preliminary results depict that pMSCs not only inhibit agonist-induced platelet activation but DBMSCs also inhibit ox-LDL-induced-CD36-mediated platelet aggregation providing a direct link between pMSCs therapy, hyperlipidaemia and atherosclerosis. Taken together, our study provides strong evidence that pMSCs-based therapy has a potential to be used against atherosclerosis. However, a comprehensive plan is imminent to understand the inhibitory mechanism and to confirm their therapeutic potential in animal models, before applying them in clinical trials.

ACK N OWLED G M ENTS
We are thankful to the volunteers, mothers and nurses at the Department of Obstetrics and Gynecology of King Abdul Aziz Medical City, Riyadh for providing us the placentae.

CO N FLI C T O F I NTE R E S T
Authors declare that there is no conflict of interest.

The Institutional Review Board (IRB) at King Abdullah International
Medical Research Centre (KAIMRC), Saudi Arabia, approved this study and the consent form. Placentae were obtained after delivery from uncomplicated human pregnancies at 38-40 gestational weeks, after signing a consent form.

CO N S E NT FO R PU B LI C ATI O N
All authors agree to publish this manuscript.

AUTH O R D I S CLOS U R E S TATE M E NT
The authors declare no competing financial interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated during this study are included in this manuscript.