Keywords
platelet - thrombus - cytoskeleton - lamellipodia - actin
Introduction
Platelets are anucleated blood cells and play an invaluable role in hemostasis. After
vascular injury, platelets arrest on exposed subendothelial components, become activated,
and form a hemostatic plug which is essential to prevent excessive blood loss. However,
uncontrolled thrombus formation in diseased vessels may lead to irreversible vessel
occlusion and infarction of vital organs. Platelet adhesion and thrombus formation
is a multistep process mediated by multiple receptor–ligand interactions.[1] The initial capture of flowing platelets occurs via the interaction of the platelet
glycoprotein (GP) Ib–V–IX complex with immobilized von Willebrand factor on subendothelial
collagen. This interaction allows the binding of the receptor GPVI to collagen, thereby
initiating platelet activation.[2] The conformational shift of β1 and β3 integrins to a high-affinity state finally
enables firm platelet adhesion to the subendothelial components and platelet aggregation.
As a result, local production of thrombin and the release as well as the synthesis
of soluble mediators reinforce and sustain platelet activation. Besides enhancing
platelet activation, thrombin also catalyzes the production of activated coagulation
factors and converts fibrinogen to fibrin. The fibrin network covers thrombi and determines
the thrombus microelasticity.[3]
Platelet Cytoskeleton and Biomechanics
Platelet Cytoskeleton and Biomechanics
The cytoskeleton plays a major role in platelet function as it maintains the discoid
shape of circulating platelets and dynamically rearranges in response to activation.
The platelet cytoskeleton consists of the proteins spectrin, myosin, tubulin, and
actin as well as several cytoskeletal-regulatory proteins ([Fig. 1A]).[4]
[5] Spectrins (approximately 2,000 molecules in a single platelet) form heterodimers
(α/β subunits) and assemble into tetramers of 200 nm lengths. The spectrin-based membrane
cytoskeleton interconnects with actin filaments and supports the maintenance of the
platelet shape.[6]
[7] Microtubules are dynamic, polarized structures and their assembly is regulated by
multiple microtubule-associated proteins. Approximately 250,000 tubulin dimers (α/β
subunits) are present in a single platelet with up to 55% organized in microtubule
filaments. In resting platelets, the microtubules form approximately 8 to 12 characteristic,
ring-shaped microtubule coils ([Fig. 1B]), the so-called marginal band, which was described to be maintained by the antagonistic
actions of the microtubule-interacting proteins dynein and kinesin.[8]
[9] Activated platelets rapidly undergo shape change from discoid to spherical as a
result of marginal band coiling due to altered dynein/kinesin activity.[10]
[11] Dmitrieff et al examined using in silico simulations how the marginal band elasticity
and cortical tension determine platelet morphology, and found that when cortical tension
increases faster than cross-linkers can unbind, the marginal band will coil at the
onset of platelet activation.[12] The cross-linked filamentous network of actin cytoskeleton is one of the major structural
components of the cell. A single platelet contains approximately 2 million actin monomers
(G-actin), which is approximately 15% of the total protein content. A total of 40%
of the G-actin is polymerized into 2,000 to 5,000 actin filaments (F-actin) building
a mechanically rigid cytoplasmic actin network ([Fig. 1B]).[13]
[14] During platelet activation, the actin cytoskeleton undergoes rapid reorganization
regulated by actin-binding proteins and confers multiple functions, including force
generation and mechanotransduction. The list of actin-associated and -regulating proteins
has grown at a steady pace, thereby increasing our knowledge about the role of the
cytoskeleton machinery in platelet function.[15]
Fig. 1 Platelet cytoskeleton. (A) Scheme of the cytoskeleton of a resting platelet. Created with BioRender.com. (B) Platinum-replica electron microscopy image of the cytoskeleton of a resting mouse
platelet on poly-L-lysine. Lower image: pseudo-colored zoom-in image. The arrow head points to a clathrin-coated pit.
Due to their high intracellular actin content and surface receptor density, platelets
are considered as ideal cells to mechanosense their environment and transduce cellular
signals in response to a myriad of physicochemical and biomechanical stimuli ([Fig. 2]).[16] Within this context, the platelet actomyosin complex plays a central role during
retraction (i.e., volume shrinkage) of newly formed blood clots by compaction of the
fibrin network.[17] Platelets achieve this by mechanically pulling fibrin transversely as well as along
their longitudinal axes.[18] Although the detailed biophysical mechanisms underlying platelet mechanobiology
are only poorly explored, recent use of force-sensitive biophysical tools has unraveled
the link between platelet mechanotransduction and cytoskeleton-dependent biomechanics
at different length scales. For example, with regard to nanomechanics of single-platelet
contraction, Myers et al developed a high-throughput hydrogel based platelet-contraction
cytometer that uses displacement of fibrinogen microdots and showed that Wiskott–Aldrich
syndrome and May–Hegglin Disorder platelets generate reduced contractile forces compared
with control platelets.[19] Similarly, microclot elastometry was used to assess platelet-dependent clot contraction
in bulk using shear gradients. It was shown that under these conditions platelet forces
are significantly reduced when blood samples are treated among others with inhibitors
of myosin, GPIb–IX–V, and integrin αIIbβ3.[20] Implementation of most of these biophysical approaches has been limited to the investigation
of adherent platelet biomechanics, since assessment of intrinsic biomechanical properties
of nonadherent single platelets in their native environment has remained challenging.
To specifically address this, high-throughput real-time deformability cytometry (RT-DC)
has been established[21] and is meanwhile used to study deformability of platelets from genetically modified
mice. It was demonstrated that platelets from coactosin-like 1 (Cotl1, an F-actin
binding protein) deficient mice (Pf4-Cre/LoxP) are smaller in size but exhibit a softer
biomechanical phenotype (i.e., are more deformable) than their wild-type counterparts.
Cotl1 mutant platelets also showed markedly reduced adhesion and thrombus formation
on von Willebrand factor at high shear, which demonstrated that Cotl1 is essential
for GPIb-mediated platelet tethering/adhesion and downstream mechanotransduction.[22] Furthermore, it was recently shown that RT-DC-based label-free mechanophenotyping
can be also used for quality assessment of cellular blood products, including platelets.[23] For more detailed insights on biophysical methods, which have been applied in platelet
biology, the reader is referred to Ciciliano et al,[24] Feghhi et al,[25] and to the recent reviews from Zaninetti et al[26] and Sachs et al.[27]
Fig. 2 Simplified schematic representation of dynamic platelet–extracellular matrix
and platelet–platelet interactions governing mechanosensing and thrombus biomechanics. Under hydrodynamic shear after initial platelet tethering, cellular activation and
inside-out upregulation of integrin affinity are essential for firm platelet adhesion.
Downstream mechanotransductional signaling occurs through coupling of integrins via
talin to F-actin generating forces. Activated platelets spread through filopodial
extensions with crosslinked F-actin bundles. Released second wave mediators amplify
integrin activation on adherent platelets and mediate thrombus growth by activating
additional platelets (not depicted). Those platelets are arrested by the activated
primary layer of platelets through multiple receptor–ligand interactions (e.g., integrin–fibrinogen).
These interactions further amplify signal mechanotransduction at the platelet–platelet
interface under hydrodynamic shear and drive thrombus growth and clot compaction.
Created with BioRender.com.
Platelet Spreading under Static Conditions
Platelet Spreading under Static Conditions
Upon platelet activation, a rapid reorganization of the cytoskeleton takes place,
which results in platelet shape change and contributes to platelet function. Platelets
start to spread under static in vitro conditions when they bind to immobilized adhesive
proteins such as fibrinogen. This assay is frequently performed to assess platelet
integrin outside-in signaling and the capacity to rearrange the cytoskeletal components.
Moreover, static platelet spreading assays have often been used to test the functional
properties of in vitro produced platelets. Four distinct stages can be defined during
platelet spreading. (1) Platelets start to adhere to the surface, (2) followed by
the formation of filopodia, (3) a phase in which filopodia and lamellipodia can be
observed, and (4) until only lamellipodia are formed. Filopodia are finger-like protrusions
with parallel actin filaments ([Fig. 3A]), whereas lamellipodia are flat, undulating cellular protrusions with orthogonally
arrayed short actin filaments ([Fig. 3B]).[28] Platelets are considered as fully spread when a circumferential zone of orthogonally
arrayed short actin filaments within lamellipodia has been formed ([Fig. 3C]).
Fig. 3 Platelet shapes under static and dynamic conditions. Cytoskeletal rearrangement of mouse platelets on fibrinogen under static conditions.
(A) Platelet filopodia formation. The arrow head points to filopodium with parallel bundles of actin filaments. (B) Platelet lamellipodium formation. The arrows point to the circumferential zone of orthogonally arrayed short filaments. (C) Zoom-in of the peripheral zone of fully spread platelet in (B). White arrow head: long actin filaments; dashed white arrow: microtubules; white arrow: zone of short actin filaments. (D) Platelet with filopodium-like structures (dashed white arrows) when in direct contact to collagen fibers (black arrow head) at the bottom of a thrombus. (E) Scanning electron microscopy image of platelets in a thrombus. (F) Schematic illustration of the platelet shape in a collagen-induced thrombus. Created
with BioRender.com.
Initial studies in platelets and other cell types have shown that rapid actin polymerization
of new filaments at the leading edge of lamellipodia is mediated by Ca2+-gelsolin, which is crucial for actin filament uncapping and severing, and by the
small GTPase Rac, which activates the Arp2/3 complex.[29] The Arp2/3 complex is an important actin filament nucleator and essential to create
branched actin filament networks required for formation of lamellipodia. Arp2/3 is
activated by nucleation-promoting factors, the best characterized being Wiskott–Aldrich
syndrome protein (WASp) and the WASp-family verprolin-homologous protein (WAVE)-regulatory
complex (WRC).[30] Platelet nodule formation, an F-actin structure present early during spreading and
suggested to be important for adhesion processes, was shown to be dependent on the
activity of WASp.[31] However, branching of actin filaments remained functional in fully spread platelets
in the absence of WASp.[32] Interestingly, the WRC subunit cytoplasmic fragile X mental retardation 1–interacting
protein 1 (Cyfip1) has recently been shown to be crucial for formation of branched
actin filaments in platelet lamellipodia.[33] The inactive WRC is activated upon interaction of the subunit Cyfip1 with the small
GTPase Rac1.[34] Subsequently, the active WRC interacts with Arp2/3, and mediates formation of branched
actin filaments. Similar to Rac1-deficient[35]
[36] and Arp2/3-deficient platelets,[37] Cyfip1-deficient platelets are unable to reorganize the actin filaments into short,
branched actin filaments and therefore do not fully spread.[33] Taken together, the Rac1-WAVE/Cyfip1-Arp2/3 pathway in platelets plays a crucial
role for branching of actin filaments and for lamellipodium formation.
Platelet Shape in Thrombus Formation and Stability
Platelet Shape in Thrombus Formation and Stability
Blood platelets play a key role in the formation of hemostatic plugs and obstructive
thrombi and therefore have to rearrange their cytoskeleton and change their shape.
There has been controversial discussion in the literature on the importance of platelet
lamellipodial structures in thrombus formation and stability. While lamellipodia can
be easily observed under in vitro conditions, it has been challenging to study lamellipodium
formation under dynamic in vivo conditions. However, studies on mouse models with
deficiencies in the Rac-Cyfip1/WAVE-Arp2/3 complex signaling pathway have contributed
to a better understanding on the role of lamellipodia in thrombus formation.
In 2005, McCarty et al investigated the role of the small GTPase Rac in platelet lamellipodium
formation. For this, they generated mice with deletion of Rac1, Rac2, and Rac1/Rac2
in the hematopoietic system.[35] Platelets were allowed to spread on fibrinogen-, collagen-, or laminin-coated surfaces.
Rac1- and Rac1/Rac2 double-deficient platelets extended filopodia, but failed to generate
lamellipodia. In contrast, Rac2-deficient platelets were comparable to controls. These
results demonstrated that Rac1, but not Rac2, is essential for platelet lamellipodium
formation. Further, it was revealed that only Rac1 is required for stable thrombus
formation under shear flow both ex vivo and in vivo.[35] In this study, the question was raised whether the increased embolization of Rac1-deficient
platelets can be attributed to the defect in lamellipodium formation or to a defect
in collagen-induced GPVI signaling. In a later study, Pleines et al followed up on
this and demonstrated that Rac1 deficiency in platelets leads to a specific GPVI-dependent
phospholipase Cγ2 activation defect, resulting in impaired platelet adhesion and defective
thrombus formation on collagen under flow, which could be rescued in a co-infusion
ex vivo assay using the second wave platelet activators ADP and thromboxane A2.[36] Based on the results of this study, it was suggested that the insufficient GPVI-mediated
platelet activation and release of secondarily acting agonists rather than the inability
to form lamellipodia is the major cause for the reduced thrombus stability of Rac1−/−
platelets. However, the role of platelet lamellipodia in this process, especially
in vivo, was still not sufficiently answered.
To investigate the contribution of the Arp2/3 complex to platelet function, Arpc2fl/fl, PF4-Cre
mice (deletion of the p34 subunit encoded by the Arpc2 gene) with a markedly decreased expression of the Arp2/3 complex were analyzed.[37] These mutant mice displayed a microthrombocytopenia and their platelets generated
filopodia but were unable to form lamellipodia on a fibrinogen matrix, demonstrating
an essential role of Arp2/3-dependent actin filament branching in lamellipodium formation.
Arpc2fl/fl, PF4-Cre
mice were subjected to the tail clip bleeding assay, which allows the determination
of the hemostatic function, and to the FeCl3-induced carotid artery thrombosis assay. Intriguingly, no significant differences
were found regarding hemostatic function and thrombus formation.[37] However, deletion of the Arp2/3 complex subunit 2 (Arpc2) was not completely successful
in platelets (>95% reduction), which could potentially rescue platelet function. Thus,
the role of lamellipodia in this process was still unclear and direct proof was lacking.
Recently, we analyzed platelet-specific Cyfip1 knockout mice to investigate the role
of Cyfip1 in platelet function.[33] This mutant mouse line displayed a normal platelet count and a slight reduction
in platelet size. However, lamellipodium formation of mutant platelets was abrogated
on all adhesive matrices, such as fibrinogen, collagen IV, collagen-related peptide,
and laminin. Activation of Cyfip1-deficient platelets was only moderately reduced
but a selective GPVI-signaling defect could be excluded.[33] In contrast, the mouse models of Rac1 and Arp2/3 deficiency described above displayed
severe defects in GPVI signaling[36] and in platelet production as well as incomplete knockout efficiency,[35] respectively. Thus, without having those side effects, the Cyfip1 knockout mouse
line represents a unique mouse model to study the role of lamellipodia in the hemostatic
function and thrombus formation.[33] Bleeding times after tail-tip amputation and thrombus growth in different experimentally
induced arterial thrombosis models were comparable between control and Cyfip1-deficient
mice showing that lamellipodial structures are not required for classical hemostatic
function and pathological thrombus formation. Interestingly, investigation of wild-type
platelet morphology revealed that in general platelets with circumferential lamellipodia,
equivalent to fully spread platelets in the static spreading assay, could not be observed
under flow ex vivo. Platelets under those dynamic conditions were flattened and contained
filopodial structures and parallel actin bundles when bound to collagen fibers at
the bottom of the thrombus ([Fig. 3D]).[33] Moreover, it was observed that platelets in the thrombus shell ex vivo only formed
filopodia, whereas plate-like protrusions were absent ([Fig. 3E]). Similarly, analysis of the platelet shape in the thrombus shell after experimentally
induced in vivo thrombus formation revealed that platelets do not form lamellipodia.
Thus, it can be concluded that platelets use only filopodial structures to adhere
to collagen fibers and pull on fibrin for clot retraction. Taken together, these results
demonstrate that (1) changes in platelet morphology are profoundly different between
static and dynamic conditions and (2) that lamellipodium formation is not required
for formation and stabilization of a physiological hemostatic plug or pathological
thrombus ([Fig. 3F]).
Discussion
Blood platelets are structurally simple cells with a high load of actin, which can
be quickly rearranged upon stimulation. Activated platelets are able to generate contractile
forces and pull at fibrin fibers resulting in reduction of the clot volume.[24] To achieve this, platelets require functional surface receptors, signaling pathways,
and contractile proteins. However, while the role of platelets in preventing bleeding
is well characterized from a biological perspective, the mechanobiological aspects
are still only poorly understood. Therefore, additional studies are needed to better
understand and quantitatively describe the biomechanical aspects of platelets not
only in hemostasis and thrombosis, but also in other platelet-mediated processes.
It is envisioned that a comprehensive analysis of platelet cytoskeleton-dependent
biomechanical properties using cutting-edge biophysical tools will have diagnostic
and prognostic applications: potentially, biophysical characteristics of platelets
may be used for the prediction of bleeding risk, risk stratification depending on
type of cytoskeletal defect, and thus may also be helpful in determination of clinical
need for prophylactic or therapeutic drug intervention and platelet transfusions.[26]
Platelets can extend two different types of cellular protrusions, namely filopodia
and lamellipodia, in a static spreading assay. It has been generally believed that
platelet filopodial structures are important for sensing the extracellular environment,
whereas lamellipodial structures are rather important for sealing the wound and propagating
thrombus formation. However, controversial data have been published on the role of
lamellipodia in the context of hemostasis and thrombosis.[35]
[36]
[37] Recent data provide compelling evidence that platelets, which are unable to fully
spread and adopt the “fried egg” shape, can perfectly form hemostatic plugs and occlusive
thrombi ex vivo and in vivo.[33] Most likely, the short time window between single platelet adhesion and initiation
of thrombus formation does not allow extensive cytoskeletal rearrangement to generate
lamellipodia.
However, it has been reported that other platelet shapes play a role in hemostasis.
It was shown that collagen-adherent platelets can transform into phosphatidylserine-exposing
balloon-like structures, and thereby contribute to thrombin generation and localized
coagulation.[38] Cyfip1-deficient platelets were able to regulate coagulant activity.[33] This shows that the process of platelet membrane ballooning does not require normal
lamellipodium formation.
But what is then the role of lamellipodia in platelet function? It is well known that
these structures are important for migration of other cell types.[39] Gaertner et al showed that platelets migrate to sites of infection to help trap
bacteria and clear the vascular surface.[40] Inhibition of actin polymerization and branching stopped platelet migration in vitro
indicating that migration depends on branched lamellipodial actin networks.[40] The same group also shows that haptotaxis plays a critical role in vascular surveillance
during inflammation and infection in vivo. Lamellipodium-dependent migrating platelets
scan the vascular endothelium for inflammatory microlesions and plug them, thereby
preventing inflammatory microbleeds and suppressing bacterial dissemination.[41] Interestingly, Gupta et al[42] observed that platelets are also important to maintain the vascular homeostasis
without an injury or inflammation. Absence of platelets, GPVI, PLCγ2, or dense granules
increased extravasation of 40-kDa dextran from capillaries and postcapillary venules.
Although the authors currently do not favor the hypothesis that platelets directly
block local mini-leaks, it is tempting to speculate that platelet shape change might
contribute to this process.
In general, further studies will be required to understand more about lamellipodial
structures in platelet-related processes. It is also important to identify novel molecular
regulators of lamellipodium formation. Is it possible to support migration of platelets
to sites of inflamed endothelium? We are just about to start understanding the contribution
of platelet structures and forces in platelet function.