FLNA is a 280-kDa protein comprised of an N-terminal actin-binding domain followed by 24 immmunoglobulin (Ig)-like repeats of beta (β)-sheets containing seven β-strands (A-G) each ( Figure 1A). The N-terminal spectrin-related actin-binding domain contains two calponin homology domains (CH1 and CH2), and another distal secondary actin binding domain (Robertson et al., 2003; Nakamura et al., 2007). The FLNA molecule is organized into rod one containing repeats 1–15; rod two containing repeats 16–23; and two calpain-sensitive flexible loops, called hinges, separating rods one and two and repeats 23 and 24, respectively (Figure 1A). While rod one is a 58 nm extended chain, interactions between repeats 16, 18, 20 with repeats 17, 19, and 21, respectively, lead to the compact propeller-like structure of rod 2 (Nakamura et al., 2007; Ruskamo et al., 2012; Tossavainen et al., 2012). FLNA forms a V-shape upon self-dimerization at the C-terminal repeat 24 (Figure 1A), and the two actin-binding domains bind and organize actin into an orthogonal network. Rod two mediates binding to multiple proteins (Table 1) thus contributing to the major scaffolding function of FLNA.

Variants of the FLNA gene confer a group of clinical disorders collectively termed filaminopathies A (Nurden et al., 2011). The most prominent such disorder is X-linked periventricular heterotopia (PVNH), caused by a defect in neuronal migration during fetal development (Feng and Walsh, 2004; Robertson, 2005). PVNH manifests as cardiac and cerebral malformations and seizure disorders (Feng and Walsh, 2004). Frameshift, missense and nonsense mutations of FLNA underlie PVNH (Robertson, 2005). Other filaminopathies A include otopalatodigital syndrome, frontometaphyseal dysplasia and Melnick-Needles syndrome, which are characterized by skeletal dysplasia (Robertson, 2007). Multiple variants of the human FLNA gene are specifically associated with aberrant platelet function (Vassallo et al., 2020; Tanner et al., 2022). Consequently, patients with PVNH can exhibit hemorrhage, coagulopathy and thrombocytopenia (Nurden et al., 2011; Berrou et al., 2017).

The most compelling direct evidence of the critical role of FLNA in platelet function has been obtained from studies with FLNA knockout mice. Because FLNA deficiency is an embryonic lethal trait, conditional knockouts were generated where FLNA expression is specifically deleted in megakaryocytes and platelets (Falet et al., 2010; Jurak Begonja et al., 2011). These conditional knockout mice display macrothrombocytopenia, which is characterized by morphologically large platelets that circulate in low numbers (Falet et al., 2010; Jurak Begonja et al., 2011). Electron micrographs of resting platelets showed that actin filaments are dissociated from the cell membranes in FLNA-deficient platelets (Falet et al., 2010), indicating that FLNA is essential for maintaining structural continuity between the plasma membrane and the actin cytoskeleton. Importantly, FLNA-null platelets fail to assemble actin normally in response to stimulation by either thrombin or collagen-related peptide (Falet et al., 2010). In a separate study, the stability of the platelet plasma membrane was monitored (Jurak Begonja et al., 2011). After 24 h of storage, FLNA-deficient platelets exhibited increased microvesiculation relative to control platelets (Jurak Begonja et al., 2011), which further reinforces the notion that FLNA underpins membrane stability. Collectively, the available evidence clearly implicates FLNA as an essential transducer of PAR- and GPVI-driven signals to the actin cytoskeleton. Less clearly defined are the exact protein/protein interactions between FLNA and the various plasma membrane and cytoplasmic proteins that regulate the actin cytoskeleton.

The initial platelet adhesion to the damaged vascular wall requires the binding of von Willebrand factor (vWF) to the GPIb-IX-V receptor complex at the platelet surface (Varga-Szabo et al., 2008a). This quadripartite complex consists of GPIbα, GPIbβ, GPIX, and GPV (Zhang et al., 2022); notably, FLNA constitutively binds GPIbα, thus anchoring the receptor complex to the actin cytoskeleton (Nakamura et al., 2006). Specifically, the hydrophobic region of FLNA Ig repeat 17 binds the cytoplasmic tail of GPIbα at amino acids 563–571 (Meyer et al., 1997; Cranmer et al., 2005; Nakamura et al., 2006). Multiple evidences support the functional importance of the FLNA-GPIbα interaction. For example, FLNA-deficient platelets exhibit abnormal surface expression and distribution of GPIbα (Falet et al., 2010). Moreover, transgenic mice expressing a mutant, non-FLNA-binding GPIbα produce platelets that fragment under shear stress (Cranmer et al., 2011). Similarly, Chinese hamster ovary (CHO) cells transfected with a non-FLNA-binding GPIbα failed to generate contractile forces on VWF substrates (Feghhi et al., 2016). Conversely, a gain-of-function mutation in FLNA Ig repeat 16 was proposed to promote the GPIbα-FLNA interaction and platelet adhesion on a VWF surface (Berrou et al., 2013). Combined with data suggesting that FLNA is required for normal GPIbα trafficking to the cell surface (Kanaji et al., 2012), it is therefore clear that the GPIbα-FLNA association is of crucial importance in the context of platelet adhesion and hemostasis. A recently published report implicates the GPIbα-FLNA interaction as a determinant in GPIbα receptor shedding, which diminishes the function of stored platelets (Zhou et al., 2022). The authors report that GPIbα receptor shedding is directly related to the stability of the actin cytoskeleton and the integrity of GPIbα-FLNA binding (Zhou et al., 2022), thus further underscoring the role of FLNA in platelet signal transduction.

Integrins are heterodimeric, transmembrane receptors that adopt a folded, closed conformation in the resting platelet. Integrins can be activated via their cytoplasmic domains (termed “inside-out signaling”) (Huang et al., 2019) or by ligand binding to their extracellular domain (“outside-in” signaling) (Shattil and Newman, 2004). FLNA interacts with integrin beta (β) subunits. This interaction requires FLNA Ig repeat 21 and a region located between two endocytic NPxY/F motifs on β subunits that also interact with talin-1 (Kiema et al., 2006) and kindlin-3 (Yates et al., 2012).

Megakaryocytes and platelets express both β1 and β3 integrins: the collagen receptor α2β1, the fibronectin receptor α5β1, the laminin receptor α6β1, the fibrinogen receptor αIIbβ3, and the vitronectin receptor αVβ3 [recently reviewed in (Yang et al., 2022)]. While the β3 subunit does not have an optimal FLNA-binding motif (unlike the β1 subunit), most studies on platelets have focused on the interaction between FLNA and αIIbβ3 and its role in fibrinogen binding and platelet aggregation. As the fibrinogen-binding receptor (Bennett et al., 2009; Nieswandt et al., 2009; Jackson and Schoenwaelder, 2010), the glycoprotein IIb/IIIa (αIIbβ3 integrin) receptor is critical for normal platelet function and transmits signals to/from the platelet actin cytoskeleton (Morse et al., 2014). Two cytoplasmic proteins, talin (Petrich et al., 2007a; Petrich et al., 2007b) and kindlin-3 (Moser et al., 2008) are identified as integrin “activators” that bind the cytoplasmic domain of the β3 integrin subunit, triggering conformational changes that expose the integrin’s extracellular ligand binding site (Ginsberg, 2014) to promote integrin activation (Ma et al., 2007). Notably, the Ig repeat 21 of FLNA also binds the β3 integrin at amino acids 747–755, thus competing with talin and kindlin-3 (Kiema et al., 2006; Liu et al., 2015; Rosa et al., 2019).

The prevailing theory regarding integrin activation is that FLNA binding to the β3 integrin serves primarily to retain the latter in a resting state (Liu et al., 2015), and that the dissociation of FLNA from the integrin promotes talin and kindlin binding to, and activation of, the integrin (Ithychanda et al., 2009). This contention is partially supported by a report of increased β3 integrin function conferred by a FLNA variant near the C-terminus (repeat 24) carried by a human subject (Berrou et al., 2017). The patient’s platelets exhibited increased aggregation, secretion, and αIIbβ3 integrin activity, as well as an increased association between talin and the β3 subunit (Berrou et al., 2017). The same research group recently generated a knock-in mouse that recapitulates the FLNA mutation; platelets from this knock-in mouse essentially replicate the gain-of-function phenotype observed in the human subject (Adam et al., 2022). These data clearly point to the functional importance of the FLNA-αIIbβ3 integrin association with regards to hemostasis and thrombosis. However, it should be noted that platelets completely devoid of FLNA exhibit comparable, yet not elevated, αIIbβ3 integrin activity relative to controls (Falet et al., 2010). This finding suggests that the regulation of integrin activity is perhaps more complex than is currently known, and that further research is required to fully validate the hypothesized role of FLNA as a strict suppressor of integrin activation.

FLNA interacts with the adaptor protein PACSIN2, a member of the Bin/amphiphysin/Rvs (BAR) family of proteins that bind and tubulate membranes (Begonja et al., 2015). PACSIN2 has been implicated in receptor internalization, caveolae biogenesis, endosomal trafficking, and cell adhesion, spreading, and migration. In platelets, PACSIN2 colocalizes with GPIbα in membrane invaginations reminiscent of the open canalicular system (OCS), the membrane reservoir for platelet spreading and channels for granule secretion following platelet activation (Heijnen and van der Sluijs, 2015). In megakaryocytes, PACSIN2 colocalizes with the initiating demarcation membrane system (DMS), the highly invaginated membrane system that provides membrane for future platelets. This interaction requires FLNA Ig repeat 20 and the tip of PACSIN2 F-BAR domain, and regulates membrane tubulation in vitro, in platelets, and in megakaryocytes (Begonja et al., 2015). Single nucleotide polymorphisms in PACSIN2 have been associated with platelet count and size (Astle et al., 2016; Eicher et al., 2016; Chen et al., 2020; Vuckovic et al., 2020).

In addition to directly binding GPIbα and integrin αIIbβ3, FLNA also transduces signals from other platelet receptors. For example, thrombin is a serine protease and a potent physiological agonist that activates platelets at concentrations as low as 0.1 nM (Greco et al., 1995; Michelson, 2013). Thrombin signals via PARs (Vu et al., 1991), which are seven-transmembrane G-protein coupled receptors (GPCRs) (Michelson, 2013). Other platelet GPCRs include the P2Y₁₂ receptor, which recognizes ADP, and the thromboxane receptor (TP receptor), which recognizes thromboxane A2 (Jurk and Kehrel, 2005; Dowal and Flaumenhaft, 2010). As mentioned in Section 1 above, PAR-driven signaling is thought to elicit changes in the actin cytoskeleton via G_q, PLCβ, rise in [Ca²⁺]_i and PKC activation. However, a model has been proposed in which Ig repeat 21 of FLNA binds GPCRs with high affinity (Tirupula et al., 2015). Although this model has not been specifically validated in platelets, it does raise the interesting possibility that FLNA modulates thrombin-induced shape change via direct bridging of PARs with the actin cytoskeleton.

The rise in intracellular calcium ([Ca²⁺]_i) following platelet activation is essential for actin assembly. This is exemplified by data indicating that platelets fail to spread in the presence of the Ca²⁺-chelating agent BAPTA/AM (Mazharian et al., 2007). Accordingly, the [Ca²⁺]_i rise is necessary for complete platelet aggregation and thrombus formation (Jardin et al., 2007; Varga-Szabo et al., 2008b). Interestingly, FLNA was recently found to regulate Ca²⁺ signaling in platelets. Lopez et al. reported that stromal interaction molecule 1 (STIM1), which serves as a calcium-sensing molecule at the endoplasmic reticulum (ER) (Lunz et al., 2019), co-immunoprecipitates with FLNA in thapsigargin-treated human platelets. Following activation-induced Ca²⁺ release from intracellular stores, ER-localized STIM1 interacts with Orai1, a Ca²⁺ release-activated channel on the plasma membrane (Galan et al., 2011). This interaction promotes store-operated calcium entry (SOCE) and the Ca²⁺-dependent platelet functions. Further, the authors reported that following siRNA knockdown of FLNA in platelets, the STIM1-Orai1 interaction (and the Ca²⁺) rise was accentuated. The authors conclude that FLNA regulates store-operated calcium entry (SOCE) process by restraining rises in [Ca²⁺]_i thus avoiding Ca²⁺ overloading (Lopez et al., 2018). The finding of increased [Ca²⁺]_i in the FLNA-knockdown platelets appears to stand in contrast with data obtained from FLNA-deficient mouse platelets, in which actin assembly is diminished (Falet et al., 2010). It is possible that FLNA’s modulation of [Ca²⁺]_i fluxes is independent of its role in modulating shape change. Additional research is required to fully dissect the role(s) of FLNA in this complex biological system.

As noted in Table 1, FLNA interacts with multiple signaling molecules that modulate actin assembly. Syk is a tyrosine kinase that is a critical element of the GPVI-driven signaling pathway (Manne et al., 2015). Stimulation of ITAM- and hemITAM–containing receptors GPVI or CLEC2 lead to their phosphorylation and recruitment and activation of the tyrosine kinase Syk (Poole et al., 1997; Suzuki-Inoue et al., 2006). Syk then promotes activation of downstream signaling leading to increases in [Ca²⁺]_i and inside-out activation of αIIbβ3 integrin (Ozaki et al., 2005). Notably, Syk was shown to be essential for lamellipodial formation and platelet spreading (Hughan et al., 2007). Moreover, FLNA Ig repeat 5 binds directly to Syk and regulates the ITAM-Syk signaling pathway by promoting Syk recruitment to the plasma membrane (Falet et al., 2010). Another kinase, protein kinase C (PKC), is activated by rises in [Ca²⁺]_i and has been shown to regulate actin assembly in platelets (Harper and Poole, 2007; Harper and Poole, 2010). Interactions between FLNA and PKC have been documented in fibroblasts (Glogauer et al., 1998) and in HeLa cells (Tigges et al., 2003), and evidence also indicates that FLNA and PKC share a common function in regulating cell spreading (Kim et al., 2010) although this was not directly tested in platelets. Nevertheless, these data collectively support the notion that FLNA likely integrates multiple intracellular kinase pathways that regulate the actin cytoskeleton.

In addition to the proteins described above, there are multiple FLNA interacting proteins whose interactions with FLNA have not been specifically documented in platelets to date. Of particular relevance to actin cytoskeletal dynamics are the Rho GTPases. The formation of the cell membrane extensions characteristic of activated platelets (e.g. filopodia and lamellipodia) is catalyzed by the Rho GTPases Cdc42 and Rac1 (Nobes and Hall, 1995). Since Cdc42 and Rac1 reportedly bind FLNA Ig repeat 24 (Ohta et al., 1999), it is reasonable to speculate that FLNA directly modulates their activity since actin assembly is curtailed in FLNA-deficient platelets relative to controls (Falet et al., 2010). Conversely, another FLNA binding partner is FilGAP, which inactivates Rac1 (Ohta et al., 2006). FLNA binds FilGAP at the Ig repeat 23 and regulates FilGAP activity in cells (Ohta et al., 2006), possibly by approximating FilGAP and Rac1. These data suggest that FLNA may serve to constrain actin assembly by facilitating the inactivation of Rac1. Further work is clearly required to elucidate the precise determinants through which FLNA regulates actin dynamics in platelets. BOX 2

• Normal actin assembly in platelets is disrupted in the absence of FLNA.