Platelets are discoid anucleate cells generated from megakaryocyte (MK) cytoplasm. Hematopoietic stem cells (HSCs) exposed to thrombopoietin (TPO) differentiate into MKs (32, 33). MKs undergo polyploidization through DNA replication without cell division, accumulating from 2n to 64n and even 128n DNA pairs in a multilobe nucleus, a process named endomitosis (34, 35). The formation of an invaginated membrane system (IMS) is another characteristic of MK maturation, and the process is well established (36). Membrane assembly starts at the cell periphery and is positioned precisely between nuclear lobes. The amount of invaginated membrane and extent of nuclear lobulation are correlated, and there is a close association between cleavage furrow formation and inhibited cytokinesis during the formation of the IMS (36). During the process, Golgi complexes and the endoplasmic reticulum (ER) are in close contact with the IMS, suggesting the mechanism by which membranes are formed and lipids are transferred (36).

After maturation via polyploidization and IMS formation, cytoplasmic branches called proplatelets protrude from MKs. Proplatelets are elongated MK protrusions that extend into sinusoidal microvessels in the bone marrow and shed platelets from the tips of the protrusion branches (37). This process was clearly demonstrated in vitro by J. E. Italiano et al. (37). Mature MKs spread and form large pseudopodia on the polarized side opposite the side with polyploid nuclei. The pseudopodia extend and bend dynamically and form new branches into bending sites until the whole cytoplasm transforms into proplatelet tubes, which may undergo anastomosis with each other. Proplatelets contract discontinuously along the long axis to produce areas of swelling into ‘beads’. Subsequently, proplatelet tips adhere and extend to form flat lamellipodia, and during this process, the ends of a proplatelet crawl away from the cell center (37). The process of proplatelet formation and platelet release in vivo ( 38) is slightly different from that in vitro, where sinusoidal vessel walls, blood flow shear stress in microvessels and the microenvironment in bone marrow are absent. With multiphoton intravital microscopy, T. Junt et al. demonstrated that MKs are in close proximity to bone marrow sinusoids and are relatively stationary compared to resident bone marrow cells (39). MKs form protrusions into bone marrow, and these protrusions extend faster in vivo than they do in vitro (3.9 µm/min vs. 0.85 µm/min) (39). Proplatelets extend through the sinusoidal wall into the microvessel lumen in the bone marrow and are cleaved and released into blood circulation. In contrast to an in vitro model, in which all of the MK cytoplasm can be observed to transform into proplatelet nets within 4 hours (37), the proplatelets extended in vivo consist of ~6% of the average volume of an MK, and proplatelets are released approximately every 7 hours (39). This variation indicates that in vivo, proplatelet protrusion and elongation is a gradual process that is effectively controlled. Released proplatelets are easily recognized in the peripheral blood circulation, providing further evidence for the theory that platelets are ultimately formed in peripheral circulatory structures, such as pulmonary arterioles (40, 41), and blood flow shear stress plays critical roles in this process (38).

In the dynamic morphogenesis of proplatelets and final platelet formation, the intracellular cytoskeletal system in MKs, comprising actin, myosin, and microtubules, plays crucial roles ( Figure 1 ).

When actin filament polymerization is inhibited by cytochalasin B (42), the bending dynamics of MK protrusions and branching of proplatelets are significantly decreased (37). Indicating the importance of F-actin in platelet formation, platelet-like swelling areas in a proplatelet consist of a meshwork of densely packed F-actin (37).

In the presence of cytochalasin B, MKs produce proplatelet extension without swollen bead formation (37). This outcome is unsurprising since cytochalasin B reduces actomyosin viscosity and contractile microfilament formation (43), inhibiting segmental contraction and intermediate contractile narrowing in the proplatelets. When myosin-9, also called nonmuscle myosin heavy chain-IIa (NMMHC-IIA), is rendered defective by mutation of the gene that encodes it, MYH9, myosin filament formation is disrupted (44), and contractile function is inhibited (45); therefore, MKs produce giant platelets (46–50). For example, these large platelets are characteristic of May–Hegglin anomaly (47), Fechtner syndrome, Sebastian syndrome (46, 48), and Epstein syndrome (49), which are considered MYH9-related genetic diseases (49, 50).

The nonmuscle myosin II molecule is a hexamer composed of two 230 kDa heavy chains, two essential light chains (ELCs) of 17 kDa, and two regulatory light chains (RLCs) of 20 kDa (51). The regulation of nonmuscle myosin II relies on the phosphorylation of serine 19 and threonine 18 on RLCs, which is mediated by different kinases, but in MKs and platelets, the most important players are Rho-GTPases, especially RhoA and CDC42 (52, 53). Tissue-specific gene knockout of RhoA in MKs led to macrothrombocytopenia. RhoA^-/- MKs produce 50% fewer platelets and exhibit a 25% increase in total platelet volume, with the large platelets formed in RhoA^-/- MKs being rounder than those in wild-type MKs (54). RhoA regulates proplatelet formation by inhibiting cytoplasmic protrusion extension (55–57). In vitro, retroviral overexpression of RhoA leads to reduced MK formation, but MKs transfected with retroviruses encoding dominant-negative RhoA produce more proplatelets (56). Inhibition of RhoA and its main downstream effector ROCK leads to reduced phosphorylation of RLCs in NM II and increased proplatelet formation (56). In contrast to the ‘STOP’ signal function of RhoA, CDC42 seems to exert a ‘GO’ signal function, driving MK proplatelet formation. Inhibition of CDC42 expression significantly reduces proplatelet formation, while MKs overexpressing CDC42 produce markedly more proplatelet protrusions (58).

Electron micrographs have clearly revealed the parallel bundles of structural microtubule skeletons in proplatelets, and increasing evidence indicates a fundamental function for microtubules in proplatelet protrusion and elongation (37, 39, 59, 60). Stabilizing microtubules with Taxol significantly decreases the number and length of proplatelet protrusions from MKs and leads to fewer and shorter but thicker protrusions¹⁸. Moreover, microtubules disrupted by nocodazole (39), vincristine sulfate or colchicine show highly suppressed proplatelet formation and elongation (59). A β1-tubulin gene (TUBB1) mutation in humans results in the formation of abnormally large proplatelets and macrothrombocytopenia (60). Mice lacking the transcription factor NF-E2, which inhibits the transcription of β1-tubulin, a main tubulin isoform in MKs, exhibit severe thrombocytopenia. MKs with NF-E2 knocked out undergo polyploidy, and invaginated membranes accumulate but never form proplatelets (61).

At the intracellular and molecular levels, platelet activation involves intracellular calcium flux, negatively charged phospholipid translocation, granule release, and shape change.

Platelet calcium flux is stimulated by agonists mainly through G protein-coupled receptors (GPCRs) or ITAM-linked receptors (ILRs) (62, 63). Receptors for thrombin, ADP and thromboxane A2 (TxA2) are GPCRs and signal through phospholipase (PLC) β. ILRs include GPVI and C-type lectin-like receptor 2 (CLEC-2), which regulate PLCγ isoforms (62). Activation of both pathways generates inositol 1,4,5 trisphosphate (IP3), which binds to the inositol phosphate-sensitive (IPS) receptor on dense tubules to promote Ca²⁺ release and increase the cytoplasmic Ca²⁺ concentration. Subsequently, the increase in Ca²⁺ level activates the cytoskeletal system and regulates various cell processes, such as phospholipid translocation; granule release; cell shape change; and protein trafficking, redistribution and activation (1). Negatively charged phospholipids translocate from the inner leaflet to the outer membrane surface in activated platelets. The negatively charged platelet surface facilitates coagulation by enabling platelet binding to the coagulation enzyme complex, which activates serine proteases and subsequently activates thrombin (64). Activated platelets release α-granules, dense granules, and lysosomes. These components play various roles in regulating physical processes, including hemostasis and coagulation, inflammation, vasoconstriction, and angiogenesis (56, 65–71). Notably, some of these components, such as vWF, fibrinogen, and growth factors, are released into the extracellular microenvironment. Other components, such as integrin α_IIbβ₃, GPVI, components of the GPIb-IX-V complex and P-selectin, are fused or redistributed into the cytoplasmic membrane, where they play critical roles in signal transduction. Activated platelets transform from a regularly round discoid shape to an irregular shape, forming actin-enriched sheets in lamellipodia and numerous extended filopodia that facilitate platelet aggregation and adhesion (1).

Activated platelets fulfill their function in hemostasis, thrombosis and inflammation via adhesion (to collagen or other types of cells) and aggregation (to other platelets). In injured vessels, collagen in the subendothelial matrix is exposed and binds to two prominent receptors on platelets, GPVI and GPIa/IIa (integrin α₂β₁). Endothelial cells undergoing injury or inflammatory responses release von Willebrand factor (vWF) from Weibel-Palade bodies (WPBs), which bind to the platelet GPIb/V/IX complex (1, 72). These binding events initiate the activation of a signaling cascade, leading to Ca²⁺ release and the subsequent activation processes in platelets and ultimately to GPIIb/IIIa (integrin α_IIbβ₃) activation (1, 73). Activated platelets redistribute more GPIIb/IIIa onto the cell surface by granule secretion, and more importantly, these receptors are activated, showing high affinity for their ligand fibrinogen (74). One fibrinogen binds two molecules of GPIIb/IIIa, promoting stable platelet aggregates (75).

From the information presented thus far, it is clear that platelet activation is a profoundly dynamic and orchestrated process. Therefore, it is not surprising that the cytoskeletal system exhibits fundamental functions in this process (1, 76–78). Actin polymerization is not only critical for filopodia formation and extension (1) but also important for α-granule secretion (76). In activated platelets, peripheral microtubule coils expand due to the increased Ca²⁺ concentration and fold into the cell center. This change promotes platelet transformation from a discoid to a spherical shape (78).

The platelet cytoskeleton is regulated by the Rho GTPase family members RhoA (54, 79–82), Rac1 and Cdc42, which collectively mediate platelet activation. Specifically, RhoA is essential for platelet shape change, α-granule and dense granule secretion, integrin α_IIbβ₃ activation, integrin-mediated clot retraction, and stable thrombus formation (54). Rac1 deficiency blocks granule secretion, lamellipodia formation and platelet aggregation (80, 81). Cdc42 and actin polymerization are critical for integrin α₂β₁ (also known as GPIa/IIa) activation, which induces the tight attachment of platelets to collagen (82).

In the blood, platelet activation needs to be quick and effective to limit blood loss and restore blood vessel integrity. Simultaneously, platelet activation requires tight control to limit the scale of thrombosis and to maintain vascular patency and blood supply to tissues. Therefore, platelet functions must be both positively and negatively regulated to maintain a balance of effective reactions and controlled scale.

Positive drivers of platelet activation include thrombin, adenosine diphosphate (ADP), TxA2, and epinephrine. All of these agonists cooperate with one or more GPCRs and subsequently elevate the Ca²⁺ concentration and facilitate various platelet activation processes (83, 84).

The negative regulators of platelet activation maintain platelet quiescence in the blood circulation and control the degree of thrombosis. These inhibitory factors are generated in endothelial cells (nitric oxide, NO and prostacyclin PGI₂), on the platelet surface (platelet endothelial cell adhesion molecule-1, PECAM-1), or in platelets (85). Knowledge obtained to date on these regulators has been effectively summarized elsewhere (85, 86).

Another emerging group of platelet functional regulators comprises neuronal guidance proteins (NGPs). Our understanding of these factors is still in its infancy, but increasing evidence indicates their promising prospects.

Neuronal guidance proteins (NGPs), also named axon guidance proteins, are various proteins that guide neurons in nervous system development both during embryogenesis and in neonates (87, 88). NGPs are composed of (but not limited to) netrin and its receptor called deleted in colorectal cancer (DCC); slit and its receptor Robo; Ephs and their receptors called ephrins; RGM and its receptor neogenin; Wnt and the receptor Frizzled; and protocadherins (Pcdhs) and semaphorins and their receptors called plexins (23, 87, 88). Many NGPs are evolutionarily conserved (89). NGPs engage in overlapping attraction or repulsion signaling with billions of neurons to induce the assembly or collapse of growth cones (23), suppress or promote the growth of axons and dendrites (24), modulate synaptic contacts (25) and prune axons (26) to refine neuronal circuits. During this complicated process, some NGPs provide long-distance chemoattractive or chemorepulsive signals to guide neuron axons; for example, netrin-DCC and Slit-Robo navigate neurons during spinal cord development (87, 90). Some NGPs serve as surrounding repulsive signals to inhibit inappropriate synaptic contacts, such as that between semaphorins and plexins (87, 91), and Pcdhs generate self-avoidance signaling that inhibits neurons from inducing nonfunctional synapse formation (87, 92). Although various and redundant signaling proteins are involved, the effectors for all NGPs are consistent and unique and originate in the cytoskeletal system. By binding to their receptors, NGPs regulate the action of small GTPases and subsequently influence cytoskeleton rearrangement and neuron motility (87, 88, 93). Given the fundamental function of the cytoskeletal system in platelet formation and activation, it is not surprising that NGPs play critical roles in regulating platelet activity.

Increasing evidence has proven that NGPs are involved in many physiological and pathological processes, such as immune reactions (29), tumor invasion and metastasis (94, 95), and tissue repair and regeneration (96, 97). In this review, we provide up-to-date knowledge about their regulatory functions in platelet formation and activation. All NGPs involved in platelet formation and activation are shown in ( Table 1 ).

Semaphorin 3A (Sema3A) is a secreted homodimer that functions through the receptor complex comprising neuropilin-1 and PlexinA1 (106). Sema3A binds to neuropilin-1, while PlexinA1 mediates intracellular signaling (107, 108). In the nervous system, Sema3A inhibits sympathetic neuron migration and modulates sympathetic neuron arrest and aggregation in the proper position (109). The genes for Sema3A, neuropilin-1 and PlexinA1 (PLXNA1) are all orthologous between humans and mice, and sequence alignment analysis with the Constraint-based Multiple Alignment Tool (COBALT) from the National Center for Biotechnology Information (NCBI) has shown a 100% match for all three protein-encoding genes (110). The functions of Sema3A in the immune system have been established: i) Sema3A inhibits T-lymphocyte activation, proliferation, and cytokine production (111–113); ii) Sema3A stimulates dendritic cell (DC) activation and plays a chemorepellent role in DC migration; and iii) Sema3A regulates monocyte and macrophage migration and polarization.

Western blotting and RT−PCR have demonstrated that the Sema3A receptors neuropilin-1 and PlexinA1 are abundantly expressed on human platelets (98). Sema3A inhibits GPIIb/IIIa activation on human platelets and subsequent platelet aggregation (98). Sema3A has also been proven to downregulate α-granule and dense granule secretion of human platelets (98). Moreover, it suppresses human platelet adhesion and spreading on fibrinogen-coated and uncoated surfaces, indicating that this inhibitory function is either GPIIb/IIIa-dependent or GPIIb/IIIa-independent (98). All of these inhibitory functions can be explained by the significant inhibition mediated by Sema3A on Rac-1 activation in stimulated human platelets, which decreases cofilin phosphorylation and inhibits actin polymerization (98). In addition, the inhibitory function of Sema3A does not depend on either intracellular Ca²⁺ concentrations or changes in cAMP or cGMP levels in activated platelets (98).

The inhibitory effect of Sema3A on platelet activation is consistent with some clinical research (114). Guo Q et al. showed that in certain autoimmune diseases, such as systemic lupus erythematosus (SLE), the serum concentration of Sema3A is significantly lower than that in serum from healthy individuals. In SLE patients with thrombocytopenia, the Sema3A concentration in serum is even lower than that in uncomplicated SLE cases and is highly correlated with the platelet count (114).

Semaphorin 4D (Sema4D, also named CD100) is a transmembrane protein expressed on various cell types, including platelets, neutrophils, T cells, B cells, monocytes and dendritic cells (DCs), in the immune system and can be found in the lungs, brain, kidneys, heart, and spleen (105, 124–127). It was discovered on human T lymphocytes and named CD100 (128) in 1992, and its discovery was the primary evidence for Semaphorin expression in the immune system (125). On the cell membrane, Sema4D forms homodimers with monomers linked by disulfide bridges (129). The extracellular region of Sema4D is cleaved, releasing a soluble form, and this exodomain cleavage is mediated by the metalloprotease ADAM17 (130). Sema4D binds three receptors, PlexinB1, PlexinB2 and CD72 (105, 131), and its binding affinity for these receptors seems to vary depending on the cell type with the expressed receptors and the cell condition (131). The SEMA4D, PLXNB1 and PLXNB2 genes are orthologous between humans and mice and are conserved in humans, rhesus monkeys, mice, rats, chickens and zebrafish (110).

Research has proven that Sema4D and its receptors play critical roles in the immune system. For example, Ponnat I. et al. proved that Sema4D binds to PlexinB1 on monocytes and DCs to influence the immune cell migration process (27). A Sema4D-knockout mouse model of foreign antigen-induced crescentic glomerulonephritis has been found to recruit fewer macrophages to the glomeruli and exhibit fewer activated T and B cells in lymph nodes than wild-type mice (132). Nishide M. et al. (105) illustrated that the soluble Sema4D concentration was increased in patients presenting with antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) and that this increase was accompanied by decreased expression of Sema4D on the neutrophil surface. Soluble Sema4D facilitated endothelial cell inflammation, while PlexinB2 on endothelial cells bound to membrane Sema4D on neutrophils and inhibited neutrophil extracellular trap (NET) formation.

The role of Sema4D in regulating platelet activation has been demonstrated mainly by the research group of Lawrence F. Brass (100, 133, 134), who proved with western blots and flow cytometry that human platelets expressed Sema4D and that the expression increased 2-fold after PMA stimulation, after which total cleavage occurred that was mediated through ADAM17 action (100). This cleavage seemed to depend on and follow platelet aggregation since blocking the binding of fibrinogen to integrin α_IIbβ₃ inhibited Sema4D cleavage (100). Platelets also express the receptors of Sema4D with CD72 in human platelets and with PlexinB1 in both human and mouse platelets. Immunoblotting confirmed that the expression of CD72 on human platelets was significantly upregulated by PMA stimulation (100). Platelets from Sema4D-knockout mice showed impaired aggregation in vitro, while coagulation and thrombus formation after vascular injury were inhibited in vivo ( 100). Brass et al. concluded that platelet membrane Sema4D promoted platelet activation and aggregation by binding CD72 or PlexinB1 on adjacent platelets (100).

Subsequent research from the same group revealed the mechanism for impaired collagen-induced platelet aggregation in Sema4D-knockout mice. They found that Sema4D was crucial for splenic tyrosine kinase (Syk) activation in collagen-stimulated platelets (133). Knocking out Sema4D in mice suppressed the activation of Syk, which subsequently caused lower levels of Ca²⁺ to be released after collagen-induced platelet activation (133). The important role played by Syk in regulating the cytoskeletal system has been shown by other studies (135–137). In thrombin-stimulated platelets, Syk is relocated to the actin filament network and promotes actin polymerization (136). Syk also regulates microtubules by binding and phosphorylating β-tubulin and α-tubulin (135, 137).

Moreover, the function of Sema4D in dyslipidemia-induced atherosclerosis has been described (134). Platelet activation plays important roles in promoting atherosclerosis in dyslipidemia (134, 138). Both native and oxidized low-density lipoprotein (LDL) lead to platelet hypersensitivity to agonists and increased aggressive adhesion, granule secretion and aggregation, which increases the risk of athero-occlusion and death via cardiovascular disease (138). In mice with dyslipidemia, platelet accumulation in the injured endothelium is 3-fold greater than that in the endothelia of healthy mice (138). Sema4D knockout inhibits collagen-induced platelet accumulation and contact in vitro, and it leads to decreased platelet accumulation in the acutely injured endothelium in mice with normal lipid levels and those with dyslipidemia (134).

Cleavage of the Sema4D exodomain (130) seems to involve the same mechanism as that underlying the shedding of GPIbα (139), GPVI (140), and PECAM-1 from platelets (141). In resting platelets, calmodulin binds the Sema4D cytoplasmic domain Arg762-Lys779, whereas inhibition or deletion of calmodulin causes Sema4D cleavage without triggering platelet activation (with the P-selectin level as the measured marker) or ADAM17 reaction (130).

Although platelets are anucleate cells and do not carry genomic DNA, increasing evidence has proven that platelets can respond to stimuli at the protein translational level because they contain abundant messenger RNA (mRNA); microRNA (miRNA), which functionally regulates mRNA transcription; and necessary organelles, such as the rough endoplasmic reticulum and ribosomes (142–145). Platelets express 32% of all human genes at the mRNA level (146, 147) and can synthesize various proteins, including the major membrane glycoproteins GPIb, GPIIb, and GPIIIa and granule proteins such as vWF and fibrinogen (148). In platelets, mRNA translation is regulated by miRNAs, which represent the majority of all small RNAs (~80%) (142).

The function of PlexinB2 in platelet formation and activation has been highlighted by research performed with miRNA-126-3p (101). In human platelets, the mRNA of PLXNB2 (the gene encoding the protein PlexinB2) has been confirmed to be a target of miR-126-3p. After MKs were transfected with miR-126-3p, the expression of PLXNB2 mRNA and protein was significantly downregulated (101). Compared to the mock cells, miR-126-3p-transfected human MKs expressed 30% more CD62P in thrombin-stimulated PLS (platelet-like structures) and exhibited 156 ± 14.9% greater adhesion to the fibrinogen-coated chamber (101). In line with this finding, silencing PlexinB2 in human MKs enhanced platelet adhesion to the fibrinogen-coated channel. These results indicate the inhibitory function of PlexinB2 in platelet activation.

However, as PlexinB2 is the main receptor of Sema4D, the inhibitory function of PlexinB2 appears to slightly contradict the facilitative effect of Sema4D on platelet responses.

Ephrins and Eph receptors belong to the receptor tyrosine kinase (RTK) superfamily, and ligand binding induces tyrosine phosphorylation of their cytoplasmic region (149). Eph receptors are composed of two classes, EphA (EphA1-EphA10) and EphB (EphB1-EphB6), which are distinguished and named according to extracellular domain sequence (149). The ligands of Eph receptors, ephrinA1-A5 and ephrinB1-B3, are membrane-binding proteins and are anchored to the cytoplasmic membrane via their GPI domain (ephrinA) and transmembrane region (ephrinB) (149). Membrane-bound ephrin binding induces Eph receptor phosphorylation, but soluble ephrin binding to Eph receptors does not trigger receptor c-terminal phosphorylation (149). During nervous system development, the Eph receptor density gradient on retinal ganglion cells is similar to the density gradient of ephrin expressed on subcortical neurons, which helps maintain proper neuronal axon projection via a ‘topographic mapping’ function (87). EPH and EPH-related receptors are evolutionarily conserved; for instance, the EPHA4 and EPHB1 genes are orthologous between humans and mice (110). These proteins have been proven to be important after an inflammatory response (150); for example, EphA2 and ephrin-A1 regulate endothelial permeability by increasing the Src kinase level and upregulating Rho-GTP expression, which subsequently leads to the opening of adherens junctions (151, 152). Additionally, Eph/ephrin plays roles in angiogenesis (153) and the response to spinal cord injury (154).

Human platelets have been proven to express the Eph kinases EphA4 and EphB1 and the ligand ephrinB1 by Lawrence F. Brass with western blots and fluorescence staining (102). With actin visualization by rhodamine-phalloidin, they proved that clustering of both EphA4 and ephrinB1 promoted human platelet adhesion and spreading on a fibrinogen-coated surface. In addition, α-granule secretion and P-selectin expression on human platelets were also induced by clustering of both EphA4 and ephrinB1, as demonstrated by flow cytometry. All these responses indicated cytoskeletal reorganization, although the platelet cytosolic Ca²⁺ concentration was not increased by EphA4 and ephrinB1 clustering (102). The interaction of EphA4 and ephrinB1 activated Rap1B, a member of the Ras superfamily, in human platelets (102, 103). Moreover, blocking the Eph/ephrin interaction inhibited human platelet aggregation, suggesting that the Eph/ephrin interaction plays a critical role in stabilizing platelet plugs (102, 103). In subsequent research, this research group demonstrated that blocking the Eph/ephrin interaction significantly inhibited platelet clot retraction, which is fundamental for thrombus stability (155). The basic mechanism is based on the Eph/ephrin interaction promoting integrin β₃ binding to myosin, which provides the force needed for platelet clot retraction (155).

Insects and vertebrates employ a symmetric bilateral nervous system in which the two sides are mirror images that are closely connected with contralateral commissural axons that cross the midline structure. During the development of this symmetric nervous system, commissural axons are guided across the midline by long- and short-range attractive and repulsive signals (156). The long-range signals are emitted by chemoattractants called netrins, and the short-range signals are emitted by the contact-mediated repellent Slits and Roundabout (Robo) (156). Slits and Robo are evolutionarily conserved (89, 156). For example, the Slit2 gene is conserved in humans, rats, mice, zebrafish and C. elegans, and an analysis with multiple sequence alignment (MSA) showed 100% matches between human and mouse Robo 1 (110). Robo receptors are immunoglobulin proteins that inhibit axons from crossing the midline. When axon growth cones express high Robo levels, they never cross the midline; in contrast, midline-crossing axons express high Robo receptor levels only after crossing the midline, not before. Robo gene mutation leads to axon crossing and recrossing multiple times. Subsequent research has identified slits as ligands of Robo receptors, and in slit-mutant embryos, axon growth cones cross the midline but do not migrate further (157).

Evidence has proven that slits and Robo are important for the development of non-nervous system organs, such as the lungs, kidneys and heart (158–161). An increasing number of studies have shown that slits and Robo play fundamental roles in inflammation. Slit2 has been proven to inhibit lymphocyte and neutrophil migration due to chemotaxis (162, 163), decrease leukocyte adhesion (164, 165), and suppress leukocyte transendothelial migration (164, 165).

The Robo-1 receptor has been shown to be distributed on the surfaces of both human and murine MKs and platelets by Patel S. with western blots, flow cytometry and immunofluorescence microscopy (104). Through the leucine-rich regions in its N-terminus, Slit2 directly binds Robo-1 (104). By binding to Robo-1, slit2 inhibits human platelet spreading on a fibrinogen-coated surface by inhibiting the formation of lamellar sheets between filopodia, platelet adhesion to immobilized collagen under fluid shear stress and nonstress conditions, and platelet granule secretion (104). Subsequently, slit2 inhibits thrombus formation in injured vasculature and prolongs the bleeding time in a murine tail bleeding model (104). These inhibitory effects are realized by the effect of slit2 on Akt activation in human platelets. Immunoblotting has proven that slit2 inhibits Akt phosphorylation in human platelets but exerts no effect on Rac1, Cdc42, extracellular signal-regulated kinase (ERK), or p38 mitogen-activated protein kinase (MAPK) activation (104).

The regulatory function of NGPs in platelet activation is summarized in Figure 2 .