Circulating Platelets as Mediators of Immunity, Inflammation and Thrombosis

Abstract

Introduction

After erythrocytes, platelets are the most prevalent blood component, and they vary in size from 2–5 µm¹. Once released by their megakaryocytic precursor, platelets enter the bloodstream and circulate for 7–10 days^{2, 3}. Platelets are complex cells that contain three distinct types of granules: α-granules, dense or δ-granules, and lysosomes^{4, 5}. Platelet α-granules contain various proteins, chemokines, cytokines and growth factors that are assembled in platelets by megakaryocytes and are necessary for normal platelet functionality. Platelet δ-granules contain small molecules such as ADP, serotonin, polyphosphates, glutamate, histamine and calcium that are necessary for hemostasis^{4, 6}. Platelet lysosomes contain enzymes such as glycohydrolases and enzymes that degrade glycoproteins, glycolipids and glycosaminoglycans^{4, 7}. Platelets do not have a nucleus but contain prepackaged proteins and various forms of RNA from their precursor cells. Found throughout the vasculature, platelets respond to signals from the endothelium, circulating cells, or other blood components. In this review, we will focus on the complex roles of platelets in thrombosis and immunity facilitated by their various interactions with the endothelium and circulating immune cells.

Established mechanisms for hemostasis and thrombosis

It is well established that in primary hemostasis, platelets adhere to the damaged vessel wall at the site of injury^{2, 3, 8}. This process occurs through multiple signaling cascades and is highly dependent on glycoproteins (GPs) expressed on the platelet surface^{2, 3, 8}. The inhibitory function of the intact endothelium (specifically, the production of prostacyclin and nitric oxide as well as expression of CD39) is decreased upon vessel damage, and extracellular matrix proteins such as collagen are exposed to the circulation^{2, 3}.

Platelet adhesion, or initial binding of platelets to the damaged site, differs depending on the level of shear stress in the circulation. At high shear stress, circulating plasma-derived von Willebrand factor (vWF) rapidly unfolds and deposits on the exposed subendothelial collagen at the injury site. The unfolding of vWF exposes multiple binding sites for GP1b (CD42), a member of the GP1b-V-IX receptor complex^{2, 3, 8}. This process decreases platelet velocity allowing for low shear stress and GPVI-collagen interactions^{2, 3, 8}. Under low shear stress, platelets adhere predominantly to collagen through their GP1a/2a integrins. This phase of adhesion leads to platelet activation as well as a number of distinct platelet physiological and cytoskeletal changes, including a shift from a resting discoid shape to an activated state with numerous pseudopodia (Figure 1)^{2, 3, 8}.

Open in a separate window

Figure 1

Differences in platelet thrombotic vs. immune activation visualized by scanning electron microscopy

Isolated A. human platelets or B. human platelets mixed with leukocytes, were treated with immune (Loxoribine-Loxo, Encephalomyocarditis virus-EMCV) or thrombotic agonists. Panels demonstrate differing levels of activation when platelets are activated via immune vs. thrombotic pathways.

The initial adhesion and activation of platelets is followed by recruitment of additional platelets from the circulation and formation of three-dimensional aggregates through a number of molecular interactions. The recruitment of other platelets, their activation, and platelet aggregation is mediated by platelet generation of thromboxane A2 (TxA2) and release of ADP from their δ-granules. Platelet-platelet interactions during thrombi formation are further mediated and stabilized by the platelet fibrinogen receptor GP2b/3a^{2, 3, 8}.

Depending on the level of the injury, blood can leak into the tissue and expose platelets to tissue factor expressed by cells that comprise the vessel. The presence of tissue factor leads to thrombin generation which mediates firm platelet aggregation. Once formed, the primary hemostatic plug retracts and becomes firmly anchored at the site of injury, preventing re-bleeding. Clot stabilization is described as secondary hemostasis and is characterized by consolidation of the platelet mass through actin/myosin-mediated platelet retraction. At this stage, thrombin, in addition to activating platelets, converts fibrinogen to fibrin leading to the formation of a network of fibrin fibers. In certain settings, normal hemostasis is overwhelmed by pathological factors leading to uncontrolled clot formation and potential vessel occlusion. As previously discussed, uncontrolled clot formation is defined as thrombosis and, in the artery, is primarily mediated by platelets. Arterial thrombosis is a central cause of myocardial infarction (MI) and stroke (embolic or in situ), while venous clot formation may lead to deep vein thrombosis and/or venous thromboembolism (VTE).

In response to blood borne pathogens and consequent tissue damage, platelets are also involved in a coordinated intravascular coagulation response recently termed “immunothrombosis”. During this process, platelets and immune cells form a physical barrier of confinement preventing dissemination of pathogens and potentially leading to activation of the innate and adaptive branches of the immune system^{6, 9}. Interestingly, platelets mediate the crosstalk between the hemostatic and the immune system utilizing similar pathways (Table 1). Consistently, various viral or bacterial infections have been found to contribute to the risk of thrombosis that manifests as arterial thrombosis or VTE and may contribute to atherosclerosis^10–12. Thus, although immunothrombosis may be an efficient way of assisting the immune system, it may significantly contribute to the overall risk of cardiovascular disease.

Contemporary mechanisms for hemostasis and thrombosis

Recent in vivo studies have provided a more complex view of hemostasis and thrombosis. This new model suggests that the hemostatic plug is characterized by a stringent architecture, with a distinct core and outer shell^{13, 14}. Fibrin deposition is distinctly localized at the base of the core in the extravascular space before complete hemostasis is achieved. A platelet activation gradient is also established, with the inner core of the plug composed of tightly packed, activated platelets which are degranulated and P-selectin positive, while the outer shell is composed of less activated loosely packed platelets that do not express P-selectin^{13, 14}. The outer shell, however, is stable and permeable to plasma solutes. Consistent with the activation-gradient of platelet distribution, there is a distinct distribution of platelet agonists throughout the thrombus. The core of the plug contains a high concentration of thrombin and, as the plug becomes more porous, a gradient of ADP and TxA2 develops^{13, 14}.

The porous outer shell of the thrombus may allow for recruitment of leukocytes necessary for injury repair or pathogen elimination. At the site of injury in veins (low shear stress), in addition to platelet activation and thrombus formation, thrombin plays a central role in leukocyte recruitment to the hemostatic plug. This process is mediated predominantly by activated platelets and not endothelial cells¹⁵. Activation/cleavage of the platelet thrombin receptor PAR4 promotes leukocyte recruitment and migration¹⁵ and platelet P-selectin enables leukocyte interaction with the thrombus^{15, 16}. Thrombin also mediates fibrin generation which limits leukocyte migration in the clot by forming a physical barrier¹⁵. Finally, platelet GP1b can bind thrombin depleting its effect and limiting leukocyte trafficking into the thrombus¹⁵.

Under lower shear stress conditions either in veins or in arteries (ischemia reperfusion), platelet adhesive interactions can also lead to formation of a CXCL7 chemotactic gradient within the thrombus body. This gradient guides intravascular migration of leukocytes through the thrombi to the sites of injury via their CXCR1/2 receptors¹⁶. This suggests that platelets, by mediating hemostasis, can also facilitate leukocyte recruitment to the clot enabling close contact of the immune system with potential pathogens that may have penetrated the skin barrier. Further studies are necessary to elucidate how platelet distribution changes in vivo in pathogen-associated venous thrombi.

Platelets and endothelial cells

The interaction between platelets and endothelial cells has been previously and extensively reviewed^{17, 18}. Platelet rolling over intact endothelium is observed predominantly in veins and increases with endothelial activation in response to inflammatory stimulus or infection. Under low shear stress in veins, platelet-endothelial interactions are mediated by platelet-GP1bα and endothelial-vWF¹⁸. Under higher shear stress in small venules, endothelial P-selectin and platelet PSGL1 or GP1bα mediate platelet rolling¹⁸. In the absence of additional signals, platelets disengage and return to the circulation. When the endothelium is under stress, firm adhesion of platelets will occur through endothelial ICAM1 (or through endothelial αVβ3) and platelet GP2b/3a in a fibrinogen-dependent manner. Endothelial stress is manifested during inflammation or infection by an increase in endothelial surface expression of P-selectin, E-selectin, VCAM and ICAM-1¹⁹. Endothelial cells also regulate platelet activation and thrombus size by secreting thrombo-regulators such as nitric oxide (NO) and prostacyclins, and removing ADP/ATP through CD39-ectonucleotidase. NO can also be secreted by platelets and can inhibit additional platelet recruitment, reduce P-selectin expression and promote disaggregation²⁰.

In addition to integrin/selectin surface expression, platelets can also affect the endothelium by increasing their surface expression of CD154 (CD40L) or by secreting cytokines such as IL-1β. Seconds after thrombin stimulation of human platelets in vitro, CD154 surface expression occurs. Platelet CD154, in turn, can lead to expression of endothelial E-selectin, VCAM1 and ICAM1, as well as secretion of MCP1 and IL-8 from endothelial cells²¹. In addition to surface protein expression, platelets can synthesize proteins from stored RNA templates^{22, 23}. One of these proteins, the cytokine IL-1β, is synthesized in response to inflammatory stimuli²² (Figure 2). Released IL-1β can increase endothelial permeability, as well as recruitment and attachment of leukocytes to the endothelium^{24, 25}. This leads to platelet alteration of the inflammatory cell transmigration to damaged or infected tissue. Interestingly, IL-1β increases the binding potential of platelets to collagen and fibrinogen and, in the presence of collagen, promotes aggregation²⁶.

Open in a separate window

Figure 2

Platelet-mediated interactions with vascular or circulating cells

Platelets interact with endothelial and immune cells in the circulation, orchestrating a response to microbes, inflammatory stimuli and vessel damage. Through their TLRs (or inflammatory signals) platelets can change their surface expression and release their granule content thereby engaging different immune cells. Platelets form HAGs and initiate innate immune responses in the presence of TLR agonists and viruses such as EMCV, CVB, dengue, flu, HIV. Platelets can interact with DC through their P-selectin, activate them to become Ag-presenting through their CD154. By releasing α- or δ-granule content which leads to IgG (IgG1, IgG2, IgG3) production and control of T-cell function, platelets engage the adaptive immune response. Similarly, platelets are able to activate the endothelium, make it more permeable and mediate leukocyte trafficking to the inflamed endothelium. Proteins in bold represent changes of expression on the platelet surface. Continuous lines represent direct binding; dotted lines represent interaction through secretion. Abbreviations: HAG-heterotypic aggregates; Ag-antigen; 5-HT-serotonin; PF4-platelet factor 4; CMV- cytomegalovirus; EMCV-encephalomyocarditis virus; CVB-coxsackievirus B;

Receptors in platelet-immune cell interactions

To understand the role of platelets in thrombosis in diverse clinical settings, their role in immunity should be addressed. Under certain infectious conditions or inflammatory stimuli, platelets directly interact with circulating leukocytes by changing surface expression of P-selectin or CD40⁵. Platelet P-selectin and CD40 are recognized by leukocyte PSGL1 and CD154, respectively, leading to formation of platelet-leukocyte aggregates. Platelets form these aggregates predominantly with circulating monocytes or neutrophils, thereby contributing to the innate immune response⁵. Platelets also contribute to adaptive immunity by interacting with and activating dendritic cells (DCs) through the CD40-CD154 axis. Platelet-mediated activation of DCs causes the DCs to present antigen to T-cells⁵. In addition, platelet activation can also lead to the release of δ-granule content and secretion of molecules such as serotonin and RANTES (CCL5) that are also known to mediate T-cell activation and differentiation^27–29. Notably, platelets contain the mRNA transcripts for all Toll-like receptors (TLR1-TLR10)^{4, 30, 31}. TLRs are pathogen-associated molecular pattern recognition receptors that are key regulators in the initiation of the innate immune response to foreign organisms.

Platelets express functional TLR4^32–34, TLR2^35–37 and TLR9^{38, 39}, and these TLRs can initiate thrombotic responses of varying intensities (Table 2). Platelet-TLR4 and -TLR2 can also mediate interactions with neutrophils^{33, 36}. Additionally, platelets express functional TLR7⁴⁰ and TLR3^{41, 42}; however, these TLRs do not induce prothrombotic responses (aggregation mediated by thrombin⁴⁰), but instead alter platelet-leukocyte interactions. Stimulation of TLR7 leads to the platelet surface expression of P-selectin and CD154, suggesting select α-granule release⁴⁰. Stimulation of TLR3 does not lead to P-selectin or CD154 surface expression but, at high agonist concentration (possibly equivalent to late stages of infection), potentiates arachidonic acid-, ADP- or collagen-mediated aggregation^{41, 42}. Interestingly, analysis of platelets from 1864 people shows that women have significantly higher expression of all TLRs as compared to men³⁰. TLR expression in platelets from women associated with an increase of plasma P-selectin while, in men, TLR expression associated with an increase in IL-6, TNFRII and soluble ICAM-1³⁰.

Importantly, platelet immune activation differs from hemostatic, thrombin-mediated platelet responses. Scanning electron micrographs of human platelets alone and in the presence of leukocytes, illustrate a distinct morphological shape change dependent on their function (Figure 1)⁴⁰. Platelet hemostatic function demonstrated by stimulation with thrombin causes complete disappearance of the platelet’s discoid shape and conversion into long intertwined “spaghetti-like” pseudopodia (Figure 1). Platelet immune function demonstrated by stimulation with TLR7 agonist or TLR7-activating virus leads to smaller platelet groups in which single platelets can be identified with fewer pseudopodia connecting to other platelets or to leukocytes (Figure 1). This suggests that platelets undergo distinct forms of activation depending on the stimulatory signal and their functional role. Different levels of activation appear necessary for the proper formation of either the hemostatic plug or recruitment of leukocytes (see “Contemporary mechanisms for hemostasis and thrombosis”).

Platelets and neutrophils

The platelet interaction with neutrophils is central to initiating the immune response (Figure 2). In humans, neutrophils are the most abundant blood leukocyte and, together with monocytes, are the major initiators of the innate immune response. Platelets form heterotypic aggregates with neutrophils as a function of TLR7, TLR2, and TLR4 activation in the circulation^{33, 36, 40}. Platelet-neutrophil interactions are mediated by the platelet P-selectin/neutrophil PSGL1 axis. Neutrophils can also phagocytose activated platelets in vivo with internalization mediated by phosphatidylserine platelet surface expression⁴³. Interestingly, as a result of TLR7 stimulation, platelet components (as measured by CD41) are internalized by neutrophils⁴⁰. It is currently unclear whether platelet microparticles⁴⁴ or the entire platelet is internalized as a function of TLR7-stimulation⁴⁰.

Platelet-neutrophil interactions are also important for neutrophil extracellular trap (NET) formation, a process termed netosis whereby pathogens are captured and neutralized^45–47. Staphylococcus aureus alpha toxin for instance mediates secretion of β-defensin 1 from platelets which can lead to NET formation⁴⁸. In the setting of lipopolysaccharide (LPS) activation of TLR4, platelets contribute to netosis by reducing the time required for NET formation³³ (Table 2). Although various TLRs are implicated in organ netosis, the role of platelet TLRs, with the exception of TLR4, has not been elucidated (Table 2).

Platelets and monocytes

Platelets can interact with monocytes as a consequence of inflammatory stimuli or infection (Figure 2). Platelet-monocyte aggregates may be a more sensitive method for assessing platelet activation as compared to P-selectin expression⁴⁹. Circulating platelet-monocyte aggregates are elevated in patients with acute myocardial infarction^{50, 51} and, in older patients, platelet-monocyte aggregates are associated with vascular thromboembolism⁵². Murine studies have shed light on the complex interaction between platelets and monocytes. During chronic inflammation, simultaneous activation of platelets and neutrophils leads to secretion of RANTES and human neutrophil peptide 1 (HNP1), respectively⁵³. These proteins form heterodimers responsible for monocyte/macrophage adhesion and recruitment in a model of myocardial infarction⁵³. Thrombin-activated human platelets also express the chemokine receptor CX3CR1 leading to the formation of platelet-monocyte complexes through the CX3CL1 ligand on monocytes⁵⁴.

Platelets and eosinophils

Platelets are described to affect eosinophil function as well. Eosinophils are granulocytes that compose about 2% of all leukocytes and it has been proposed that they may contribute to the innate immune response to helminth parasite (worms) infection. Pathogenically, eosinophil function has been attributed to the profound damage they cause in an allergic asthma. Direct interaction between platelets and eosinophils in the circulation during helminth infection in vivo has not been described. However, co-incubation experiments of platelets and eosinophils (for 48h) show that platelets prolong eosinophil life by secreting granulocyte-macrophage colony-stimulating factor (GM-CSF)⁵⁵. This suggests that platelets may enable eosinophils in fighting helminth infection. However, by delaying eosinophil apoptosis platelets may also contribute to asthma-mediated tissue damage. A murine model of allergic inflammation has suggested that platelets through their P-selectin recruit eosinophils to the lungs⁵⁶ and consequently may propagate tissue damage. Allergen sensitized mice had an increase of platelet-eosinophil aggregates in the circulation and in the lung tissue and aggregate interactions were mediated through platelet P-selectin and eosinophil PSGL1⁵⁶.

Platelets, B-cells and T-cells

When interacting with lymphocytes, platelets are more selective and bind preferentially to larger, activated cells⁵⁷. In the presence of LPS or ADP, aggregation between platelets and B-cells is almost unaffected. However, when platelets are co-incubated with B-cells for three days in vitro, B-cells increase their production of IgG1, IgG2 and IgG3, suggesting that platelet content can contribute to B-cell function and alter adaptive immunity⁵⁸.

T-cell activation more robustly increases platelet aggregation via both T-cytolytic and T-helper cells⁵⁷ (Figure 2). This process is mediated by platelet GP2b/3a, CD154, and lymphocyte CD11b⁵⁷. CD154 is expressed under various stimulants (such as TLR7 activation⁴⁰) and is capable of enhancing T-cell immunity to viral challenge. Evidence suggests that, by interacting with lymphocytes, platelets may facilitate the recruitment of these cells to an injured vessel at a site of inflammation or infection⁵⁹, and this is a central step in T-cell trafficking. Activated platelets may mediate T-cell functions further by releasing a number of factors, such as platelet factor 4 (PF4, CXCL4)-RANTES or serotonin^{27, 28, 60}. The release of these factors impacts T-cell function in various ways. PF4, for instance, is necessary for the limitation of Th17 expansion and differentiation⁶¹. Serotonin, the largest peripheral circulation reservoir of which is stored in platelet δ-granules, can push naïve T-cells to activate and proliferate^{27, 28}.

Platelets and dendritic cells

As previously mentioned, platelets interact with and influence another important cell in adaptive immunity, the DC. Platelets, through the CD40-CD154 axis, can induce DCs to present antigen to T-cells⁵. Additionally, platelet-DC interactions can be mediated by the P-selectin/PSGL1 axis, followed by firm adhesion through platelet junctional adhesion molecule C (JAM-C) with MAC-1 on DCs⁶² (Figure 2). Interestingly, this recruitment and activation of DCs is followed by phagocytosis of platelets by DCs⁶². The mechanism by which DCs change after platelet phagocytosis is still largely unknown. It is possible also that serotonin originating from phagocytosed platelets may be utilized by DCs and released as needed for T-cell proliferation and differentiation.

Platelets and natural killer cells

The interaction between platelets and natural killer (NK) cells is the least understood interaction within the circulation (Figure 2). In murine studies, tumors coated with platelets become impermeable to NK cell destruction⁶³ with fibrinogen involvement⁶⁴. It has been proposed that circulating metastatic cells attract platelets and influence them to release their granule content. As a result, human NK cells lose their cytotoxicity and their ability to produce IFNγ, possibly through the downregulation of NKG2DL mediated by platelet-TGFβ release⁶⁵. However, the baseline communication between platelets and NK cells remains unknown as does the effect on this interaction during infection or autoimmune disease.

Platelet-leukocyte interactions in atherogenesis

The role of platelet-leukocyte interactions during inflammation has been determined to be contributory in murine models of atherosclerosis. It is now established that activated platelets can increase endothelial atherogenic potential and lead to elevated platelet-leukocyte aggregate formations in the circulation. Platelets may further contribute to atherogenesis through the delivery of proinflammatory factors to leukocytes and endothelial cells⁶⁶. The platelet-derived factors implicated in this setting are surface P-selectin and the platelet-derived chemokines RANTES and PF4⁶⁶. PF4/RANTES heterodimer disruption reduces atherosclerosis in ApoE-deficient mice by attenuating monocyte recruitment⁶⁰. Murine models lacking P-selectin show that this adhesion receptor is a key mediator of leukocyte recruitment into lesions and promotes their advancement in ApoE-deficient mice⁶⁷. In this case, both platelet and endothelial P-selectin contribute to lesion development; however, endothelial P-selectin plays a crucial role in atherosclerotic lesion growth⁶⁸. Similarly, murine models lacking CD40 in ApoE-deficient mice demonstrate that platelet CD40 promotes atherosclerosis by activation and recruitment of leukocytes to endothelial cells⁶⁹. When platelets lacking CD40 are injected in the circulation of mice, expression of VCAM1, PECAM, VE-Cadherin and P-selectin are decreased on endothelial cells. This decrease in adhesion molecule expression is accompanied by less advanced atherosclerotic plaques with reduced levels of macrophages and neutrophils⁶⁹. Increased expression of platelet CD154, in turn, contributes to accelerated atherosclerosis by increasing platelet-platelet interactions and inflammatory response through mediation of platelet-endothelium and platelet-leukocyte interactions⁷⁰. Additionally, platelet CD154 suppresses Treg cell recruitment leading to accelerated atherosclerosis⁷⁰.

Platelets and metastatic cells in the circulation

Platelets contribute to the increased thrombotic risk of cancer patients. High platelet count is associated with decreased survival in various malignancies and thrombocytopenia, in turn, is associated with decreased metastatic potential of tumors. Entry of metastatic cancer cells into the blood stream leads to platelet activation. Metastatic cells express PSGL1 and secrete high levels of tissue factor⁷¹. Platelet-cancer cell interactions and cancer-induced thrombi formation in the vasculature may inhibit recognition and elimination by the immune system. Additionally, the thrombi formation can lead to tethering of the cancer cell to the endothelium, promoting survival and invasive potential. More recent studies in mice have proposed that, as a consequence of interaction with tumor cells, platelet signals recruit granulocytes to the tumor, thereby, guiding the formation of an early prometastatic microenvironment. Depletion of platelets by cell-specific depleting antibodies inhibits granulocyte recruitment to metastatic cells in lungs and depletion of granulocytes leads to fewer metastases and reduced metastatic progression⁷².

Platelet-derived TGF-β is an important factor contributing to metastasis. In vitro studies have shown that thrombin-mediated platelet-derived TGF-β can alter NK cell antitumor reactivity⁶⁵. In vivo studies eliminating TGF-β specifically from platelets have proposed that this cytokine, when originating from platelets, is solely responsible for inducing an invasive epithelial mesenchymal transition to metastasis⁷³. These studies utilized the PF4Cre murine model which, at the time, was believed to eliminate TGF-β only from the megakaryocyte/platelet lineage⁷³. We have since learned that a subset of epithelial cells can express PF4⁷⁴ in addition to activated CD4+ and CD8+ T-cells⁶¹. These studies suggest that other tissue-resident cells or epithelial cells themselves can contribute to the invasive potential of cancer cells mediated by TGF-β. The role of the platelet-mediated increase in metastatic potential through TGF-β is further complicated because thrombin stimulation also leads to the release of PF4 from platelet α-granules⁶¹. PF4, in turn, inhibits TGF-β signaling in a SMAD2/3 dependent manner⁶¹. Regardless of TGF-β signaling, metastatic cancer cells entering the blood stream hijack the hemostatic function of platelets by releasing tissue factor and, consequently, generate thrombin and initiate thrombus formation and neutrophil recruitment. The formed thrombi can attach to the endothelium and promote viability and metastatic potential of the cancer cell.

Platelets in bacterial infections

During bacterial infections, platelets actively mediate the host response through interactions with circulating leukocytes. In addition to PF4 and RANTES, platelet α-granules contain various antimicrobial compounds such as platelet connective tissue activating peptide 3 (CTAP-3), platelet basic protein, thymosin β-4 (Tβ-4), and fibrinopeptide (A and B)^{75, 76}. These platelet-derived antimicrobial compounds are known to target various bacterial organisms (Figure 3)^{4, 77, 78}. Our understanding of how platelets mediate these interactions in the circulation is still developing.

Open in a separate window

Figure 3

Platelet and circulating cell interactions during infection initiate the innate or adaptive immune response

Platelets achieve cell-to-cell communication during bacterial or viral infection either by direct interaction with WBCs through surface expression of platelet proteins or through indirect protein release from their α- or δ-granules. EMCV-activated platelets interact with neutrophils in a TLR7-dependent manner. CVB-activated platelets bind to neutrophils in a PS-dependent manner. Dengue and influenza increase microparticle release; dengue-mediated microparticles contain IL-1β. HCMV-activated platelets interact with neutrophils, monocytes, B-cells, T-cells and DCs, suggesting activation of innate and adaptive immune responses. Vaccinia-bound platelets have reduced aggregation potential in the presence of ADP, collagen, or thrombin. The specific pathways by which platelets respond to Herpes simplex virus (HSV)1 or HSV2 are currently unknown. During bacterial infection, platelet interactions with complement C3 opsonized bacteria through GP1b (CD42) lead to slowing of bacterial clearance. DCs recognize the platelet-bacterial complexes, thereby, inducing adaptive immunity. These platelet-bacterial interactions are true for Gram-positive or Gram-negative bacteria. Abbreviations: WBC-white blood cells; 5HT-serotononin; VEGF-vascular endothelial growth factor; PS-phosphatidylserine; C3-complement C3; EMCV-encephalomyocarditis virus; hCMV-human cytomegalovirus; CVB-coxsackievirus B; HIV-human immunodeficiency virus;

Platelets can interact with various strains of bacteria including the Staphylococci family, Neisseria gonorrhea, Porphyromonas gingivalis, and Helicobacter pylori^{77, 79, 80}. Platelets adhere and aggregate around bacteria, typically utilizing their GP2b/3a, FcγRIIa, and IgG receptors in conjunction with fibrinogen or fibronectin^{81, 82}. After platelets interact with bacteria, they release antimicrobial compounds contained in their granules. S. Aureus alpha toxin mediates release of β-defensin from platelets and, consequently, β-defensin induces NET formation⁴⁸. Released PF4 can bind to Gram-negative bacteria, leading to the exposure of PF4 heparin-like epitopes. Exposed PF4-epitopes, in turn, increase antibody binding to the bacterial surface leading to the opsonization and possibly phagocytosis of bacteria by neutrophils^{83, 84}. This mechanism is particularly important since some Gram-negative bacteria, such as Yersinia pestis, are poorly recognized by the mammalian TLR4^{85, 86}. In vivo studies using another Gram-negative bacteria, Porphyromonas gingivalis, show that, during infection, platelets interact with neutrophils forming heterotypic aggregates in a TLR2-dependent manner and TLR2 can also promote platelet aggregation³⁶. This study provides evidence that, by recognizing bacterial components, platelets can activate platelet thrombotic and/or inflammatory pathways.

Studies using Lysteria monocytogenes have recently shed light on the role of platelets in inducing adaptive immunity during bacterial infection. Platelet binding to intravascular bacteria leads to slowing of the cytotoxic clearance of the pathogen, keeping bacteria available for induction of an adaptive immune response by splenic CD8+ T-cells⁸⁷. This platelet-mediated process is dependent on the opsonization of bacteria by complement C3, subsequent platelet binding by GP1b, and the capture of the bacteria-platelet aggregates by cells of the immune system⁸⁷. Platelet-bacteria interactions in the presence of C3 lead to different kinetics of clearance between free (fast clearance) and platelet-bound (slow clearance) bacteria. This observation is true for Gram-positive strains such as L. monocytogenes, Bacillus subtilis, Enterococcus faecalis and Staphylococcus epidermidis, and for Gram-negative strains such as Escherichia coli, Salmonella typhimurium and Klebsiella pneumoniae⁸⁷. This study suggests a sophisticated role for platelets in balancing clearance of bacteria and induction of adaptive immunity.

Platelets in viral infections

Platelets interact with various types of viruses. Viral infections are often associated with thrombocytopenia and in some cases thrombosis. Viruses can be simply classified as those with a DNA or RNA genome. RNA viruses are further subdivided into double-stranded and single-stranded viruses. DNA viruses from the herpes viral family such as herpes simplex virus type 1 (HSV1), cytomegalovirus (CMV) and vaccinia, have been found associated with platelets, but it is unclear if they can be internalized by platelets^88–90. RNA viruses such as HIV, hepatitis C virus (HCV), dengue, influenza, coxsackievirus B (CVB) and encephalomyocarditis virus (EMCV) are much smaller in size and can be internalized by platelets^{40, 91–95}. Here, we have focused on the interaction of virally-activated platelets with circulating cells.

In general, platelet function during viral infection can be beneficial and/or detrimental to the host depending on the length of time post infection. Human CMV (hCMV) binds to platelet-TLR2, resulting in release of CD154, IL-1β, and VEGF (Figure 3). hCMV-activated platelets do not show increased aggregation or adhesive potential but rapidly form platelet-neutrophil heterotypic aggregates. In addition to their interactions with neutrophils, hCMV-platelets also exhibit increased interaction with monocytes, B-cells, T-cells and DCs, suggesting a multifactorial interaction with the innate and adaptive immune systems and induction of proangiogenic responses⁸⁸. In vitro studies utilizing vaccinia virus show that ADP-, collagen-, or thrombin-mediated aggregation in isolated platelets is inhibited by the virus but there is an increase in serotonin release⁹⁰ (Figure 3). In vivo studies, however, show that infection with vaccinia leads to fatal intravascular coagulation at 24h post infection⁹⁶. In vitro studies using HSV1-infected human umbilical vein endothelial cells (HUVEC) show platelet binding due to increased thrombin generation from the damaged cells⁹⁷. These observations suggest that other circulating or vascular cells can contribute to the overall prothrombotic balance.

RNA virus infections exhibit more diverse platelet responses. Many RNA viruses cause various degrees of thrombocytopenia at different stages of infection. In the setting of murine EMCV infection, platelets interact with neutrophils to form heterotypic aggregates as early as 1h post infection and this leads to a sudden drop in platelet count. This drop also occurs at 24h post infection. Elimination of platelets before EMCV infection leads to reduced survival similar to mice deficient in TLR7⁴⁰.

In humans, infection with dengue, an RNA virus, is characterized by profound hemorrhagic fever and thrombocytopenia. In a rhesus macaque model, dengue viral infection leads to platelet-monocyte interactions at 24h post infection while platelet-neutrophil interactions peak at 3d post infection⁹⁸. In this study, blood contained monocytes that had engulfed platelets containing dengue virus⁹⁸. Increased permeability of the endothelium during dengue infection is also facilitated by platelets. Dengue infection mediates assembly of the NLPR3 inflammasome and activation of Caspase-1 in platelets. Caspase-1, in turn, alters secretion of IL-1β-rich microparticles⁹⁹. Platelet-monocyte interactions are also increased in people infected with HIV1 or influenza^{100, 101} and are influenced by inflammatory stimuli in a COX-dependent manner¹⁰².

During infection with CVB, platelets form aggregates with the neutrophil population and directly improve the outcome of CVB-induced myocarditis⁹⁵. CVB is internalized by platelets causing phosphatidylserine exposure on the platelet surface without affecting apoptosis⁹⁵. The virus cannot replicate in platelets and platelet depletion causes an increased viral load in heart tissue as well as increased myocarditis⁹⁵.

Influenza infection illustrates the potential for dysregulation of the hemostatic-inflammatory balance. In patients infected with influenza H1N1 virus, there is an elevation of platelet-monocyte aggregates and a two-fold increase in platelet aggregation potential as measured by PAC1 incorporation¹⁰³ (Figure 3). As the infection progresses, coagulation potential increases, as evidenced by the elevation of macrovesicle tissue factor activity¹⁰⁴. In addition, in non-survivor influenza-infected patients, there is a trend toward lowered platelet and increased white blood cell counts¹⁰⁴. H1N1 activates human platelets independently through both FcγRIIA signaling and through thrombin generation¹⁰⁵, while H1N1–platelet interaction increases the expression of the active form of the fibrinogen receptor GP2b/3a, arachidonic acid metabolism, and microparticle release¹⁰⁵. Studies utilizing a lethal dose of the H1N1/PR8 strain in C57BL/6J mice suggest that, in critical non-survivor settings, platelet degranulation and activation contribute to lung inflammation and reduced survival¹⁰⁶. In summary, platelets contribute to the overall procoagulant and inflammatory states during influenza infection, but it is unclear if they are necessary for overall immunity and vascular health.

Platelets in Plasmodium parasite (malaria) infection

There are six plasmodium species that can infect humans and lead to development of malaria. The role of platelets in Plasmodium parasite infection changes as the infection progresses. Studies using non-inflammatory malaria with P. falciparum or P. chabaudi show that platelets kill the parasite inside of the red blood cells (RBCs) by secreting PF4¹⁰⁷. PF4, in turn, interacts with the Duffy antigen on RBCs through an unknown mechanism and mediates survival of infection¹⁰⁸.

Although these studies suggest a protective role during early infection, a model of cerebral malaria using P. berghei ANKA (PbA) iRBCs clone 2.34 concludes that platelets play multifactorial roles depending on the pathogenic stage of infection¹⁰⁹. Platelets in this case were activated throughout the first 3 days of infection as measured by an increase in circulating PF4. This early activation of platelets is protective possibly due to a platelet-dependent reduction in parasite burden and leukocyte recruitment through an increase of the IL-1β pool¹⁰⁹. As infection progresses, however, platelet activation, possibly mediated by the overall immune response and tissue factor release, leads to microvascular thrombosis as found in the postmortem brains of patients with cerebral malaria^4,110.

Platelets and RNA transfer

The ability of platelets to communicate with other vascular cells through receptor signaling, direct protein interactions, or released granule content has long been understood. More recently, other methods of platelet vascular communication have been described including RNA transfer. Bidirectional RNA transmission is illustrated by transfer of platelet-derived transcripts to leukocytes or endothelial cells and by the uptake of leukocyte- or vascular-derived transcripts into platelets¹¹¹ (Figure 4). Co-culture of platelet-like particles (PLPs) with either a HUVEC cell line or a human monocytic cell line (THP-1) demonstrates functional transfer of platelet-derived mRNA transcripts¹¹¹. Endogenous analysis of THP-1 cells post PLP incubation reveals an increased expression of α, γ, and ε globin genes as a consequence of RNA transfer¹¹¹. Confirmation of this phenomenon in vivo utilizing mouse infusion experiments demonstrates that platelets carrying TLR2 mRNA can transfer their transcripts to mice with TLR2-deficient blood mononuclear leukocytes¹¹¹.

Open in a separate window

Figure 4

Platelet vascular and circulating cellular communication by non-protein-dependent mechanisms

Platelets and platelet-like particles (PLPs) have been shown to communicate with vascular cells in several protein-independent manners, specifically through mitochondrial and RNA transfer. RNA can be transferred from platelets to endothelial cells, monocytes, macrophages, and cancer cells. Inversely, RNA can be taken up by platelets from endothelial cells, monocytes, and cancer cells. Mitochondrial transfer from platelets to neutrophils can also occur. Microparticles mediate a significant number of these communication pathways.

Analysis of human platelets post co-incubation with HUVECs and THP-1 cells reveals that platelets can also take up transcripts from surrounding cells. There are distinct changes in expression of platelet transcripts including those associated with platelet-vascular adhesion and apoptosis¹¹². Platelet RNA uptake has been confirmed by the presence of TLR2 in TLR2-negative platelets post infusion of TLR2-positive leukocytes (peripheral blood mononuclear cells, PBMCs) into TLR2 knockout mice¹¹². Interestingly, platelet RNA uptake can occur in the presence or absence of thrombin stimulation¹¹².

Platelet microparticles, miRNA and vascular inflammation

In addition to the hemostatic and inflammatory responses described, platelet activation also leads to microparticle formation. This is a particularly important observation since platelets are a major source of microparticles in the circulation⁴. Proteins that mediate platelet function related to hemostasis or immunity are also found in platelet microparticles. This includes adhesion molecules such as P-selectin, chemokines such as RANTES, and cytokines such as IL-1β, among others^{99, 113}. Platelet microparticles can promote monocyte recruitment to the inflamed endothelium of the atherosclerotic plaque¹¹³. Increased circulating platelet microparticles have been associated with atherosclerosis development in diabetic patients. Infections such as influenza or dengue also lead to increased platelet microparticle release⁹⁹. During vascular injury, platelet microparticles can increase the regenerative potential of endothelial progenitor cells by enhancing their recruitment, differentiation and release of proangiogenic factors¹¹⁴.

In addition to proteins, platelet microparticles contain various forms of RNA including micro RNA (miRNA), which are small non-coding RNAs that function in post-transcriptional regulation of gene expression. Platelets may affect surrounding cells by transferring miRNA^115,116. The platelet miRNAs, miR-223 or miR-320b, are present in platelet-derived microparticles post thrombin stimulation and can transfer into endothelial cells^{117, 115} (Figure 4). Functionally, it has been shown that miR-320b transfer into endothelial cells can decrease ICAM1 expression¹¹⁵. Similarly, miR-126-3p, highly expressed in platelets, can be taken up and enriched in macrophages, resulting in an increased phagocytic phenotype¹¹⁶. Platelet microparticle-derived miR-223 transfer to cancer cells may play a role in cancer progression. This miRNA transfer leads to downregulation of the tumor suppressor EPB4L13 and a concurrent increase in tumor cell invasion¹¹⁸. Alternatively, transfer of platelet-derived miR-24 to solid tumor cells results in disrupted mitochondrial function, leading to tumor cell death and proliferation decline¹¹⁹. These studies support the hypothesis that platelet RNA can affect function of vascular or circulating cells through RNA transfer.

In support of the relevance of miRNA transfer, is the observation of decreased miRNA expression only in platelets at the site of injury and not in the peripheral circulation¹¹⁵. While the differential expression of miRNAs in platelet-derived microparticles is interesting, it is crucial to establish how these microparticle miRNAs affect the function of recipient cells.

Platelets and mitochondrial transfer

As both protein and RNA communication between platelets and the vasculature can occur through microparticle production, there is the possibility that other platelet contents may transfer. Platelets contain mitochondria (averaging four mitochondria/platelet) that are both functional and have been shown to impact platelet activation and subsequent thrombosis^{44, 120, 121}. Platelet-derived mitochondria may play a role in communication as platelets release functional mitochondria when stimulated with numerous agonists via both platelet microparticles and as extracellular free organelles⁴⁴. These mitochondria may impact overall vascular health and level of inflammation⁴⁴. Extracellular mitochondria have been implicated as damage-associated molecular patterns (DAMPs)^{44, 122–125}. Additionally, mitochondria can act as a substrate for sPLA2-IIA, a phospholipase known to be induced during inflammatory settings⁴⁴. In its interaction with extracellular mitochondria, sPLA2-IIA leads to mitochondrial dysfunction and subsequent mitochondrial content release (arachidonic acid, lysophospholipids and mitochondrial DNA), all of which reinforce a proinflammatory environment. sPLA2-IIA has also been implicated in the internalization of platelet microparticles by neutrophils⁴⁴ (Figure 4). Released mitochondria interact and are internalized by neutrophils, where mitochondria-neutrophil interactions appear to impact neutrophil functionality (increased rolling and interaction with vascular walls, increased induction of NETs), morphology (development of pseudopodia), and substance release (leukotrienes)⁴⁴. Finally, extracellular mitochondria expression is linked to a number of proinflammatory settings. Mouse studies show that sPLA2-IIA-dependent release of mitochondrial factors leads to increased proinflammatory cytokine release and decreased body temperature⁴⁴. Human studies reveal increased levels of extracellular mitochondria in the synovial fluid of multiple inflammatory autoimmune pathologies such as rheumatoid arthritis and osteoarthritis⁴⁴, and these levels correlate with the known elevation in platelet microparticles^{44, 126–128}. These observations suggest that extracellularly released platelet mitochondria may function as a potent communicator of vascular inflammation.

Antiplatelet therapies and complex platelet function

In addition to continual use of established platelet inhibitors, there has been great progress in the development of antiplatelet/antithrombotic therapies for the treatment of various thrombotic pathologies. Cyclooxygenase (COX) in platelets, responsible for TxA2 generation and three dimensional clot formation, has been targeted in the regulation of platelet activity. Platelets contain predominantly COX1 while endothelial cells can contain COX1 and COX2. Aspirin is an irreversible inhibitor of COX activity, and as platelets do not regenerate COX protein, this drug is an effective inhibitor of platelet aggregation potential. Aspirin has provided beneficial effective secondary prevention of arterial thromboembolic events.³⁵ Inhibition of the prothrombotic activity of platelets by aspirin is also considered to be beneficial in infectious diseases such as influenza. Relevant to the role of platelets in cancer, it is suggested that aspirin may be beneficial in preventing the platelet microthrombi formation around cancer cells in the circulation that may aid metastasis. A recent meta-analysis study concludes that long term aspirin use is associated with reduced risk of overall cancer¹²⁹.

A range of P2Y12 inhibitors are also extensively used in patients with cardiovascular syndromes. In vitro blocking of P2Y12 by Cangrelor reduces platelet deposition on the injured endothelium and delays vessel sealing, thereby leading to prolonged escape of plasma-born molecules into the microvaculature¹³⁰. The results observed with inhibition of GP2b/3a signaling are more complex. While still, on occasion, used intravenously to inhibit platelet function in the setting of acute unstable cardiac ischemia¹⁴, oral formulations were associated with increased morbidity and mortality¹³¹. Still unknown is whether inhibiting P2Y12 or GP2b/3a will alter the immune responding functions of the platelet.

Interestingly, the thrombin inhibitor, hirudin, which binds and inhibits thrombin and prevents fibrin generation, does not block the sealing of vessel injury¹³⁰. This observation suggests that the core and the shell of clots can form without fibrin but the stability and retraction may be affected. Finally, TLR agonists/antagonists may be useful in promoting platelet immune function and survival of infection possibly depending on the specific receptor. This is consistent with the observation that elimination of platelets before infection with the TLR7-activating virus, EMCV, leads to decreased survival of mice⁴⁰. As therapeutics designed to target platelets progress, the type of injury, mechanism/type of infection, and timing of infection must be considered.

Conclusion

Platelets exhibit complex interactions with circulating cells and the vessel wall during a broad range of inflammatory processes that manifest in many types of disease. However, there is overlap between the platelets’ hemostatic and inflammatory roles. For example, platelets may engage leukocytes and the immune system by utilizing proteins involved in hemostatic plug formation. More recent data suggest that these interactions are regulated by processes beyond traditionally-defined receptors, e.g. RNA or mitochondrial transfer to other vascular or circulating cells. These complex and sometimes contradictory interactions appear to directly affect disease. During certain stages of infection, platelets initiate activation of both innate and adaptive immunity which is beneficial to the host. However, uncontrolled endothelial damage and inflammation as a result of viral infection progression can lead to adverse prothrombotic responses and increased cardiovascular risk. Thus, platelets appear to have broad and multifaceted functions firmly grounded in their vascular cell interactions. Maintaining these complicated and distinct states of activation is imperative to regulating vascular homeostasis and health. As our understanding of the role of platelets in the vasculature and circulation grows, we have the potential to design better platelet-targeted treatments for patients who present with a wide range of vascular, immune, and oncological diseases.

Acknowledgments

Nonstandard Abbreviations and Acronyms

Footnotes

References