Coordinated Membrane Ballooning and Procoagulant Spreading in Human PlateletsClinical Perspective
Background—Platelets are central to the process of hemostasis, rapidly aggregating at sites of blood vessel injury and acting as coagulation nidus sites. On interaction with the subendothelial matrix, platelets are transformed into balloonlike structures as part of the hemostatic response. It remains unclear, however, how and why platelets generate these structures. We set out to determine the physiological relevance and cellular and molecular mechanisms underlying platelet membrane ballooning.
Methods and Results—Using 4-dimensional live-cell imaging and electron microscopy, we show that human platelets adherent to collagen are transformed into phosphatidylserine-exposing balloonlike structures with expansive macro/microvesiculate contact surfaces, by a process that we termed procoagulant spreading. We reveal that ballooning is mechanistically and structurally distinct from membrane blebbing and involves disruption to the platelet microtubule cytoskeleton and inflation through fluid entry. Unlike blebbing, procoagulant ballooning is irreversible and a consequence of Na+, Cl–, and water entry. Furthermore, membrane ballooning correlated with microparticle generation. Inhibition of Na+, Cl–, or water entry impaired ballooning, procoagulant spreading, and microparticle generation, and it also diminished local thrombin generation. Human Scott syndrome platelets, which lack expression of Ano-6, also showed a marked reduction in membrane ballooning, consistent with a role for chloride entry in the process. Finally, the blockade of water entry by acetazolamide attenuated ballooning in vitro and markedly suppressed thrombus formation in vivo in a mouse model of thrombosis.
Conclusions—Ballooning and procoagulant spreading of platelets are driven by fluid entry into the cells, and are important for the amplification of localized coagulation in thrombosis.
- blood coagulation
- blood platelets
- cell-derived microparticles
- fluorescent imaging
- membrane ballooning
Platelets play complex roles in hemostasis and arterial thrombosis, rapidly adhering to subendothelial structures and to each other to generate a platelet aggregate that is stabilized by the local production of thrombin and subsequently fibrin.1 Critical to this response is the surface exposure of aminophospholipids, particularly phosphatidylserine (PS), which promotes assembly of the tenase and prothrombinase complexes on the platelet surface. This platelet-dependent procoagulant activity therefore depends on 2 major factors: (1) the degree of PS exposure and (2) the surface area of membrane with exposed PS.
Editorial see p 1374
Clinical Perspective on p 1424
It is currently thought that a sustained rise in cytosolic Ca2+ is required for exposure of PS on the extracellular leaflet of the plasma membrane, through activation of a nonspecific phospholipid scramblase and inhibition of a PS translocase or flippase. Anoctamin-6 (gene ANO6 or TMEM16F) is identified as a key regulator of calcium-dependent PS exposure,2 and loss-of-function mutations in anoctamin-6 have been shown in 2 patients with Scott syndrome,3,4 who have aberrant calcium-dependent scramblase activity.5 However, the precise role played by anoctamin-6 is still unclear. It is possible that, like other members of the anoctamin family, it forms Ca2+-activated Cl– channels.6 Although much effort has gone into determining the molecular mechanisms regulating surface PS exposure, relatively little is known about the mechanisms by which platelet membrane surface area may be maximized. Possibly, Cl– entry may also be required for a change in membrane surface area, and this would be another distinct functional role for Cl– entry in potentiating platelet procoagulant activity.
Platelets have long been reported to transform in vivo to form balloons on activation, and fibrin has been shown to fill the space between these balloon structures at wound sites.7–9 Platelet membrane ballooning has also been observed in vitro in platelets adherent to immobilized collagen.10–12 However, it is not clear whether this striking morphological change is analogous to apoptotic blebbing in other cell types.13–15 Attempts to assess ballooning in platelets have been limited by the methods of investigation, imaging resolution,10 and the fragility of the balloon structure that often results in its loss or significant deformation.11
Here, we hypothesized that platelet ballooning was important to markedly increase the surface area of exposed membrane, and exposed PS, thereby enhancing the local procoagulant response.10,16,17 We used detailed dynamic imaging approaches to visualize thrombin generation on platelet membrane surfaces and to understand the mechanisms regulating ballooning. This study revealed that the key mechanism involves fluid entry, accompanied by the genesis of a novel spread membrane structure in a process we have termed procoagulant spreading. Unlike conventional lamellipodial spreading, this form of spreading yields procoagulant surfaces and rapidly breaks up by multiple coalescences to form numerous procoagulant microvesicles. Ballooning and procoagulant spreading are therefore linked processes that are likely to contribute to hemostatic responses in vivo.
Written informed consent was obtained in accordance with the Declaration of Helsinki.
Human blood was obtained from healthy drug-free volunteers under Local Research Ethics approval (E5736). The UK Scott patient blood was obtained with NHS Research Ethics Committee approval, and has been described. This Scott patient is a compound TMEM16F heterozygote, IVS6+1G→A, resulting in exon 6 skipping. Another mutation in this patient (c.1219insT) causes premature translation termination and defective expression of TMEM16F.4,18
Details of materials used are given in the online-only Data Supplement.
Platelet-Rich Plasma Preparation
Blood drawn from healthy human volunteers was anticoagulated with 0.4% trisodium citrate and acidified with 16% acid citrate dextrose (85 mmol/L trisodium citrate, 71 mmol/L citric acid, 111 mmol/L glucose). Platelet-rich plasma was obtained by centrifugation at 180g for 17 minutes.
Washed Human Platelet Preparation
Platelet-rich plasma was centrifuged at 650g for 10 minutes in the presence of 10 μmol/L indomethacin and 0.02 U/mL apyrase, and resuspended in HEPES-Tyrode buffer modified with 0.1% (wt/vol) glucose, 10 μmol/L indomethacin, and 0.02 U/mL apyrase. Sodium- and chloride-free HEPES-Tyrode buffers were prepared by replacing Na+ and Cl– with equimolar N-methyl-d-glucamine and gluconate, respectively.
Live Cell Confocal Microscopy
Washed human platelets were preincubated (10 minutes) with calcium dye Fluo-4 AM and Alexa Fluor 568 annexin-V conjugate (1% vol/vol). Hyperosmolar Tyrode was prepared by adding 40 mmol/L sucrose to HEPES-Tyrode buffer. MatTek dishes were precoated with collagen (20 µg/mL), and aliquots of platelet suspensions were added (2×107 cells/mL), supplemented with 1 mmol/L CaCl2. Changes in relative fluorescence intensity (F/F0) over time were monitored. Details of confocal microscopy are given in the online-only Data Supplement files.
Measurement of Platelet Thrombin Generation
Platelet-rich plasma was incubated with fluorogenic thrombin substrate, Z-GGR-AMC (450 µmol/L). Platelet-rich plasma was recalcified and thrombin generation initiated with 5 pmol/L tissue factor. Thrombin substrate was measured on platelet membrane surfaces, and the traces for single platelets and platelet aggregates were converted into first-derivative curves.
Image Deconvolution and Analysis
Raw intensities from time-series (single plane over time) images were quantified after regions of interest were chosen and images corrected for background noise. For each platelet analyzed, relative fluorescence (F/F0) is reported, where F0 designates the background-subtracted fluorescence level before platelet activation. Deconvolution of Z-stack images was based on calculated point spread functions; 3-dimensional and 4-dimensional reconstruction, movie rendering, and colocalization analysis were performed by using Volocity imaging software (Perkin-Elmer, UK).
In Vitro and In Vivo Thrombosis Assays
Details of in vitro and in vivo thrombosis assays are given in the online-only Data Supplement.
Data were analyzed using GraphPad Prism 6 (San Diego, CA) and presented as interleaved box plots with whiskers showing minimum to maximum values and interquartile ranges. We determined statistical significance by the Friedman test, followed by the Dunn multiple comparison test or by Wilcoxon signed rank test. P<0.05 (*) or P<0.01 (**) was considered significant.
Spatiotemporal Dynamics of Platelet Membrane Ballooning and PS Exposure
On interaction with the subendothelial matrix, platelets are transformed into balloonlike structures as part of the hemostatic response.7–10 To obtain more temporal and dynamic insight into platelet ballooning, we used phase-contrast and confocal live cell imaging microscopy techniques to study platelet adhesion and ballooning on a collagen-coated surface. In the early phases of contact with collagen, platelets formed small, retractable membrane blebs (Figure 1A, images 2–7, blue arrows). However, in many platelets, one of these blebs can swell to become a balloon (Figure 1A, image 10, yellow arrow), which ranged between 1 and 6 µm in diameter and typically did not retract (Figure 1A, Movie I in the online-only Data Supplement). We identified 3 distinct phases leading to balloon formation, which we termed phases Ph1, Ph2 and Ph3 (Figure 1B). Ph1 was characterized by blebbing (membrane protrusion ≤1 µm; Figure 1A and 1B, images 1–5). During Ph2, the membrane of nonretracted blebs rapidly expanded (images 6–8) and in Ph3, the expansion plateaued (images 9 and 10). The initiation of ballooning was typically by 5 minutes after adhesion of the platelet to collagen fiber, and progressed rapidly to Ph3 by a further 5 minutes. Importantly, we observed platelet ballooning in vivo, in a mouse thrombus formation model (Figure IA in the online-only Data Supplement), consistent with previous observations.7–9 Ballooning also occurred in human thrombi, as shown in in vitro flow studies (Figure IB and IC in the online-only Data Supplement). Interestingly there was no ballooning in platelets adherent to von Willebrand factor in the presence of botrocetin, suggesting agonist specificity of the response.
A 3-dimensional reconstruction showed that the platelet body and the balloon membrane bound labeled annexin-V, demonstrating PS exposure and generation of a procoagulant surface5,19–21 (Figure 1C, Movie II in the online-only Data Supplement). Annexin-V binding was characterized by 2 stages: Initially, a low level (F/F0=1.3–1.5) of annexin-V accumulated on the membrane of the platelet body alone (Figure 1D, images 1–5), followed by annexin-V binding to the balloon as well (Figure 1D, images 7–10, Movie III in the online-only Data Supplement). Importantly, PS externalization did not temporally correspond with membrane ballooning (Figure 1E and 1F), suggesting a distinction between the mechanisms underlying these 2 events. Furthermore, integrin αIIbβ3 was activated at an early time point before ballooning, but was sustained and localized just to the platelet body (Figure 1G and H).
Thrombin Is Generated on the Ballooned Surface of Human Platelets Adherent to Collagen
To demonstrate that ballooned platelets support a procoagulant response, we visualized the formation of thrombin using the fluorogenic thrombin substrate Z-Gly-Gly-Arg aminomethyl coumarin (ZGGR-AMC, Figure 2Ai, cyan blue). Three-dimensional reconstruction further confirmed that thrombin was generated all over the balloon and platelet body (Figure 2Aii). Figure 2Aiii shows that thrombin is generated within 300 to 450 s after platelet adhesion to collagen. This strongly correlated with balloon formation and annexin-V binding by single platelets (compare with Figure 1B, 1E, and 1F).
Platelets from patients with Scott syndrome show significantly impaired PS exposure and procoagulant activity.18 Here, we show that platelets from the UK patient with Scott syndrome have a marked defect in balloon formation in comparison with control platelets (Figure 2B, see arrows). Platelet ballooning may therefore be important for the procoagulant response, by increasing the available surface area for recruitment of the tenase and prothrombinase complexes (Figure II in the online-only Data Supplement).
Ballooned and Procoagulant-Spread Platelets
With the use of 4-dimensional imaging, we observed 4 distinct platelet phenotypes adherent to collagen (Figure 3A): (1) conventionally spread nonballooned platelets (annexin-V–), (2) ballooned and procoagulant-spread platelets (BAPS, annexin-V+), (3) ballooned nonspread platelets (BNS, annexin-V+), and (4) nonballooned and nonspread platelets. The nonballooned and nonspread platelets phenotype was typically annexin-V– but could be induced to expose PS with prolonged stimulation. Procoagulant spreading, in BAPS platelets, was only clearly identifiable by visualizing annexin-V just above (between 0 and 0.75 μm) the collagen-coated surface (Figure 3A and 3B, in red; Movie IV in the online-only Data Supplement). The area covered by procoagulant-spread platelets is generally much greater than by conventionally spread platelets and can extend to around 20 μm (Figure 3A). BAPS platelets constitute a distinct subpopulation of adherent platelets, and this report is the first characterization of these structures.
Ballooning and Procoagulant Spreading Are Synchronized Events
Four-dimensional live-cell imaging showed balloon formation started within 4 minutes after platelet adhesion and reached a maximal diameter by 9 to 10 minutes, as visualized by annexin-V binding (Figure 4A and 4B). Procoagulant membrane spreading followed membrane ballooning after a 1- to 2-minute delay (Figure 4A through 4Ci) and both were temporally correlated (Figure 4Ci, Movies V and VI in the online-only Data Supplement). Only a small proportion of platelets (≈10%) ballooned without procoagulant-spreading (BNS; Figure 4Cii), and we did not observe procoagulant spreading without ballooning. These data suggest that ballooning is likely to be required for procoagulant spreading, but not vice versa. In this experiment, after adhesion of platelets to collagen for 1 hour, the mean relative proportions of platelets were conventionally spread nonballooned platelets (27.4%), BNS (10.4%), BAPS (50.0%), nonballooned and nonspread platelets AnxV– (5.9%), and nonballooned and nonspread platelets AnxV+ (6.3%; Figure 4Cii).
BAPS platelets showed a punctate or cobblestone annexin-V staining pattern (Figure 4A and 4B). In addition, BAPS platelets released microvesicles in a time-dependent manner (Figure III in the online-only Data Supplement), and scanning electron microscopy showed the cobblestone appearance to be microvesicles formed from BAPS platelets after 1 hour of adhesion (Figure 4D through 4F). This was clearly distinct from the conventionally spread nonballooned platelets platelet (Figure 4E). We therefore suggest that procoagulant spreading forms the basis for microvesicle formation and release on adhesion of platelets to surfaces.
Inhibition of Actomyosin Promotes Membrane Ballooning While Blocking Procoagulant Spreading
To determine the role of the actin cytoskeleton in ballooning and procoagulant spreading, platelets were incubated with modulators of actin polymerization or myosin motor activity. Figure 5A shows representative figures of annexin-V binding to platelets adhered to collagen in the presence of these inhibitors. Blebbistatin, which blocks myosin-II ATPase activity,22 significantly promoted the progression of platelets ballooning and increased balloon diameter at Ph3 (Figure 5Aii, Figure III in the online-only Data Supplement). Similar results were found with platelets pretreated with cytochalasin-D or Y27632. However, while enhancing ballooning, blebbistatin, cytochalasin-D, and Y27632 inhibited spreading, and therefore formation of the BAPS morphology, so that platelets were restricted to the BNS phenotype (Figure 5Aii). Platelets treated with jasplakinolide, to promote actin polymerization,23 did not show any membrane ballooning or spreading. The data indicate that actin polymerization and myosin contraction negatively control membrane ballooning. Inhibition of procoagulant spreading by using cytochalasin D also resulted in diminished thrombin generation (Figure 5B).
Probing live platelets with AlexaFluor-350 phalloidin showed staining in the spread membranes of BAPS platelets (Figure 5C), indicating that their membrane integrity was likely to be compromised and that the balloons were actin-rich (see also Movie VII in the online-only Data Supplement). In contrast, conventionally spread nonballooned platelets were phalloidin negative.
Platelet Ballooning Involves Microtubule Disruption at the Exit Point of the Protrusion
Transmission electron microscopy images of platelets in the expansion phase (Ph2) of ballooning revealed a neck structure that delineated the balloon from the main platelet body (Figure IVA in the online-only Data Supplement, yellow arrows), beneath which lay a disrupted microtubule ring structure, shown in detail by TEM tomography (Figure IVAiii in the online-only Data Supplement, green arrows; Movie VIII in the online-only Data Supplement). Significantly, there was no microtubule architecture present in Ph3 balloons (Figure IVB in the online-only Data Supplement). In the later phases of balloon formation, the integrity of the membrane becomes compromised, because it becomes leaky to 2 low-molecular-weight dyes, calcein and propidium iodide (data not shown).
Platelet Ballooning and Procoagulant Spreading Requires Salt and Water Entry
The rapid growth of the platelet balloon suggested that it may be driven by fluid entry. Our data showed a 1.6±0.02–fold increase in [Cl–]cyt after 5 minutes of contact with collagen (Figure 6A and 6B), which was absent in Cl–-free medium (not shown). In similar experiments, we recorded transient Na+ entry during platelet adherence to collagen (Figure 6C and 6D), which was absent in Na+-free medium (not shown). The percentage of adhered platelets that ballooned (Figure 6E) and underwent procoagulant spreading (BAPS, Figure 6F) was significantly attenuated in media lacking Cl– or Na+ ions. Interestingly, the peak change and half-life of annexin-V binding were significantly reduced under these conditions (Figure III in the online-only Data Supplement). It is likely that Cl– entry was substantially mediated by calcium-activated chloride channels (CaCC), because the inhibitor of these channels, CaCCinh-A01, induced similar effects to removal of extracellular Cl– (Figure 6E and 6F).
To assess whether water entry was required for ballooning, we increased extracellular osmolality, using sucrose, by 40 mmol/L, which significantly attenuated both ballooning (Figure 7A) and procoagulant spreading (BAPS, Figure 7B), but not blebbing (Figure 7A). Furthermore, mean balloon diameter was attenuated in the small proportion of platelets able to balloon (Figure IIIA in the online-only Data Supplement). Acetazolamide has been shown to block aquaporin water channels,24 and it induced a characteristic retraction of the ballooning membrane (Figure 7A) and attenuated membrane accumulation of annexin-V (Movie IX in the online-only Data Supplement). Importantly, there was also a marked reduction in formation of BAPS platelets (Figure 7B), suggesting a requirement for water entry in procoagulant spreading.
In addition, acetazolamide and hyperosmotic challenge significantly reduced PS exposure and thrombin generation on the platelet surface (Figures 7C and 7D, Figure IIIB in the online-only Data Supplement, and Movies IX and X in the online-only Data Supplement). The mean intensity of thrombin substrate per unit membrane area was similar in BAPS and BNS platelets (Figure 7E), suggesting that the greater thrombin generation seen in controls was attributable to increased membrane surface area provided by ballooning and procoagulant spreading.
Blocking Water Entry and Ballooning Significantly Reduces In Vivo Thrombus Formation
Mice were treated by bolus intravenous administration of acetazolamide (7 mg/kg), and carotid artery damage was induced by application of FeCl3 (10% vol/vol). Accumulation of platelets, labeled with DyLight 488-antiGp1bα, was visualized by video epifluorescence microscopy. Whereas control mice showed rapid and sustained accumulation of platelets in a growing and eventually occlusive thrombus, mice treated with acetazolamide were markedly spared from this event (Figure 8). Balloon-enhanced local generation of thrombin will therefore likely form a positive feedback system to further activate platelets and promote coagulation.
Platelets are the surveillance cells of the vascular system, detecting vessel damage events and acting as the early and rapid response system to address the damage they recognize. They rapidly adhere to subendothelial matrix collagen, recruit more platelets to form an aggregate, and stabilize the structure by initiating blood coagulation. This procoagulant activity depends on the surface exposure of negatively charged aminophospholipids, particularly phosphatidylserine (PS), which binds the tenase and prothrombinase complexes. The dynamics of platelet membrane transformation therefore underpin the coagulant response. Previous studies have shown that platelets are able to undergo both membrane blebbing and ballooning on activation,7–11 and that platelet balloon formation is part of the normal hemostatic response. In particular, the studies of Wester et al7,8 used a template bleeding technique to induce acute wounds to human skin, and excised the wounds at various time periods after wounding. By using histological and electron microscopic approaches they showed the existence of platelet balloons within the structure of the associated thrombus. Importantly, they showed that fibrin was deposited in the margins between platelet balloons, suggesting a functional link between balloon formation and coagulation in vivo. Consistent with this, we demonstrated that activated factors V and X are bound to ballooned platelet membranes (Figure II in the online-only Data Supplement). We have determined the molecular regulation of these events, and revealed in real time the formation of a novel BAPS platelet phenotype in a process we termed procoagulant spreading. The balloon and procoagulant-spread structures are functionally linked and both contribute to the local generation of thrombin. BAPS platelets break into a multitude of procoagulant microvesicles, and this process may form a major route for the generation of platelet-derived microvesicles which contribute to local hemostasis, but have potential for pathological roles in various cardiovascular diseases.25
Although previous studies had described platelet membrane ballooning in some detail,7,8,10,11,26 this is the first report of coordinated membrane ballooning and procoagulant spreading. Platelets with characteristics similar to the ballooned platelets seen here have been reported, particularly COATED and SCIP platelets.18,26 However, there are clear differences, because platelets in this study showed sustained integrin αIIbβ3 activation (Figure 1G and 1H and Figure II in the online-only Data Supplement) and transient increase in cytosolic calcium (data not shown). Furthermore, data in Figures II and IVb in the online-only Data Supplement show that the balloon retains some of the molecular and subcellular components of the platelet from which it was derived, including various surface receptor markers and mitochondria.
The timing and localization of PS exposure in platelets indicates that it does not appear until ballooning is well under way, suggesting that PS exposure is not a requirement for ballooning. Rather, platelet balloons result from physical disruption to the circumferential microtubule, accompanied by an increase in internal hydrostatic pressure provided by a coordinated Na+, Cl–, and water entry mechanism, which inflates the balloon. By contrast, membrane blebbing analogous to that seen in cancerous cells is reversible and occurs independent of fluid entry mechanisms.27 It is possible that Na+ may enter through nonselective cation channels, such as TRPC628, or by Na+/Ca2+ exchange. For Cl–, anoctamin (TMEM16) genes are key components of CaCCs.29–31 Anoctamin-6 is a CaCC6,32 which functionally couples to TRPC channels.33 Salt may therefore enter through regulated pathways, providing an osmotic drive for water entry and ballooning. The role of calcium in these events is pivotal, because blockade of calcium entry or release from intracellular stores abolished ballooning and procoagulant spreading, but not lamellipodial spreading (data not shown). A rise in cytosolic free Ca2+ may therefore trigger CaCC opening6 and allow chloride entry.34 Interestingly, the rise in cytosolic Na+ and Cl– ions concentrations are predominantly in the platelet cell body, rather than the balloon (Figure 6). A possible explanation is that the major volume change occurs in the balloon, and not in the cell body. Likely, all increases in ion concentration in the balloon are continuously being diluted through entry of water. Related to this, we have never observed balloons to continue to grow or to burst, but always observe them reaching a plateau size. We suggest that this is likely to be a product of this transient increase in ion permeability leading to a self-limiting water entry, constraining the balloon to a steady-state sustained volume.
Scott syndrome is extremely rare and patients show a bleeding disorder associated with defective expression of TMEM16F and a resultant defect in microvesicle formation and exposure of PS.5,35 Our observations show that platelet ballooning is also markedly impaired in this syndrome (Figure 1C). It is possible that TMEM16F provides the critical mechanism driving ballooning, through regulated ion influx followed by water, and that the membrane stretching lowers the activation energy required for scrambling of membrane phospholipids and movement of inner leaflet PS to the outer leaflet.
Importantly, this study has revealed a novel phenotype that we have termed BAPS platelets. These structures break up to form procoagulant microvesicles and thereby increase the PS-exposing membrane surface area for procoagulant activity. That it had been previously missed in the literature was likely the result of the extremely thin and particulate nature of the spread membrane, which is only visible when platelets are stained with labeled annexin-V, under physiological concentrations of extracellular calcium and monitored over varying z-heights. There was a direct link between ballooning and procoagulant spreading because inhibition of ballooning by CaCCinh-A01 (Figure 6E) or jasplakinolide or acetazolamide (Figures 5 and 7) also blocked procoagulant spreading and microparticle generation (Figure III in the online-only Data Supplement). Agents that inhibit procoagulant-spreading independent of ballooning, such as blebbistatin, cytochalasin-D, or Y27632, also largely inhibited microparticle formation (Figure III in the online-only Data Supplement). It is therefore likely that all these events are functionally linked, and sequential, from ballooning to procoagulant spreading to microparticle formation (Movies V and VI in the online-only Data Supplement). It is interesting to note that endothelial PS exposure has recently been shown to play a major role in thrombin generation.36 However, it is possible that procoagulant spreading may also allow platelets to contribute to coagulation over a much wider surface area than just the platelet cell body, which may explain any apparent discrepancy with Ivanciu et al.36
Our data would also suggest that, as the hydrostatic pressure in the platelet cell body rises, it is able to sustain this increase in pressure because of the intact cytoskeleton of the cell. However, once a weakness in the cytoskeleton develops, the membrane rapidly inflates to generate the balloon. This leaves further questions, such as the molecular nature of the ion and water channels, and whether disruption of the cytoskeleton is a tightly coordinated event. The fact, however, that acetazolamide is able to markedly diminish in vitro ballooning, and also act to potently inhibit thrombus formation in vivo, indicates the importance of this event, and suggests potentially novel ways to control thrombus formation pharmacologically.
In conclusion, this study has uncovered the molecular mechanisms that control dramatic platelet membrane ballooning and revealed a novel procoagulant-spread membrane structure on platelet activation by collagen. The events are mechanistically coupled and are likely to amplify the procoagulant responses at wound sites. They also suggest a route by which platelets generate microparticles on contact with collagen. The mechanism of membrane ballooning shown here, which involves salt and water entry into the cells, may lead to new therapeutic directions for the control of thrombus formation in vivo.
We acknowledge the MRC and the Wolfson Foundation for funding the University of Bristol’s Bioimaging Facility. We thank Alan Leard, Katy Jepson, and Judith Mantell of the Wolfson Bioimaging Facility for their assistance. We thank Dr Nicholas Timpson (School of Social and Community Medicine) and Dr Peter Brennan (School of Physiology and Pharmacology) of the University of Bristol, for their help with the statistical review.
Sources of Funding
This work was supported by the British Heart Foundation (RG/10/006/28299; PG/12/25/29488; FS/11/62/28934; PG/10/100/28658; FS/12/22/29510), Netherlands Heart Foundation (2011T6 to JMEMC), ZonMW (MKMD 114021004 to JMEMC and JWMH), and United Kingdom National Institute for Health Research (II-LB-0313-20003).
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.114.015036/-/DC1.
- Received December 19, 2014.
- Accepted July 30, 2015.
- © 2015 American Heart Association, Inc.
- Kunzelmann K,
- Nilius B,
- Owsianik G,
- Schreiber R,
- Ousingsawat J,
- Sirianant L,
- Wanitchakool P,
- Bevers EM,
- Heemskerk JW
- Castoldi E,
- Collins PW,
- Williamson PL,
- Bevers EM
- Tian Y,
- Schreiber R,
- Kunzelmann K
- Heemskerk JW,
- Vuist WM,
- Feijge MA,
- Reutelingsperger CP,
- Lindhout T
- Jackson SP,
- Schoenwaelder SM
- Abaeva AA,
- Canault M,
- Kotova YN,
- Obydennyy SI,
- Yakimenko AO,
- Podoplelova NA,
- Kolyadko VN,
- Chambost H,
- Mazurov AV,
- Ataullakhanov FI,
- Nurden AT,
- Alessi MC,
- Panteleev MA
- van Kruchten R,
- Mattheij NJ,
- Saunders C,
- Feijge MA,
- Swieringa F,
- Wolfs JL,
- Collins PW,
- Heemskerk JW,
- Bevers EM
- López JA,
- Andrews RK,
- Afshar-Kharghan V,
- Berndt MC
- van de Vijver E,
- De Cuyper IM,
- Gerrits AJ,
- Verhoeven AJ,
- Seeger K,
- Gutiérrez L,
- van den Berg TK,
- Kuijpers TW
- Michelson AD
- Kovács M,
- Tóth J,
- Hetényi C,
- Málnási-Csizmadia A,
- Sellers JR
- Gavin RH
- Holzinger A
- Kulkarni S,
- Jackson SP
- Fackler OT,
- Grosse R
- Harper MT,
- Londoño JE,
- Quick K,
- Londoño JC,
- Flockerzi V,
- Philipp SE,
- Birnbaumer L,
- Freichel M,
- Poole AW
- Caputo A,
- Caci E,
- Ferrera L,
- Pedemonte N,
- Barsanti C,
- Sondo E,
- Pfeffer U,
- Ravazzolo R,
- Zegarra-Moran O,
- Galietta LJ
- Grubb S,
- Poulsen KA,
- Juul CA,
- Kyed T,
- Klausen TK,
- Larsen EH,
- Hoffmann EK
- Paradiso AM,
- Ribeiro CM,
- Boucher RC
- Ivanciu L,
- Krishnaswamy S,
- Camire RM
Platelets play a central role in hemostasis and arterial thrombosis, in particular, being the critical cells that cause arterial blockage in coronary artery disease. They rapidly form platelet aggregates at sites of blood vessel damage, such as occurs at sites of atherosclerotic plaques, and these aggregates are stabilized by the local production of thrombin and the generation of a fibrin clot. Platelets coordinate the localized clotting process by exposing phosphatidylserine on their cell surface, allowing the recruitment of the tenase and prothrombinase complexes to form a procoagulant surface. The procoagulant activity therefore depends on the surface area of membrane with exposed phosphatidylserine. We showed here that this is enhanced substantially by the formation of balloon structures, with subsequent procoagulant spreading of platelets over the adherent surface. Bleeding defects may thus be attributable to aberrant procoagulant membrane dynamics as exemplified by Scott patients. Furthermore, platelet membrane ballooning and procoagulant spreading is driven by a coordinated system of salt and water entry, which may be modified pharmacologically. Acetazolamide, used clinically for several indications, has actions that include the blockade of water entry into cells. We showed that acetazolamide not only substantially impaired ballooning and thrombin generation in vitro, but also markedly reduced thrombus formation in an in vivo model of thrombosis in mice. Drugs that modify water entry into platelets may therefore represent novel antithrombotic therapies for the management of coronary artery disease and the prevention of stroke. Furthermore, we provide added rationale for monitoring patient bleeding risk when acetazolamide is coadministered with anticoagulants.