From theory to platelets: unraveling the history and complexities of biased signaling Open Access

The origins of biased signaling can be traced back to the late 19th century studies of Rudolf Buchheim and Oswald Schmiedeberg. However, the concept has only been experimentally characterized in the last 40 years. The lifelong goal of Buchheim and Schmiedeberg in ascertaining the relationship between chemical compounds and their biological effects underlies the motivation to study any pharmacological concept.¹ As the fathers of pharmacology,² their curiosities on how the external environment influences biological systems link concepts in biochemistry and physiology. Ultimately, their studies were centered on the end goal of enhancing therapeutics and improving the quality of life. By the late 20th century, this foundational objective would serve as essential building blocks for the emergence of biased signaling in the literature.

The contributions of Jean-Pierre Changeux, Martin Rodbell, Alfred Gilman, and Earl Sutherland shaped the modern definition of biased signaling today. Their contributions to biased signaling theory can be categorized into the following 2 groups: mechanisms in conformational dynamics and signal transduction. In the early 1970s, Changeux introduced the fundamental principles of receptor conformational dynamics.³ He hypothesized that the 3-dimensional structure of proteins was fluid in the absence of a ligand. Changeux et al developed the Monod-Wyman-Changeux model which explained receptor allostery.^1,4 They suggested that ligand-bound proteins have a stable, rigid quaternary structure, whereas free proteins exist as distinct and ever-changing conformations within thermal equilibrium of the ligand-bound conformation.³ This idea was essential for our current understanding of receptor physiology and pharmacology as it was later translated into the functional model of receptor activation of James Black and Paul Leff.⁵ 

Rodbell, Gilman, and Sutherland broadened our understanding of how extracellular stimuli influence intracellular changes. Their work defined the difference between orthosteric ligands, allosteric ligands, and allosteric modulators. Orthosteric ligands bind to the active site of a protein, whereas allosteric ligands bind anywhere on the protein. The term “allosteric modulator” describes the ability of an allosteric ligand to alter receptor activity by influencing the binding or function of the orthosteric ligand. Thus, allosteric modulators affect the sensitivity of a receptor to the orthosteric ligand.⁶ Positive allosteric modulators work with the agonist to stabilize receptor conformations that favor signal transduction by lowering the energy required for signaling activation.⁶ Negative allosteric modulators work against the agonist to increase the energy required to activate signaling.⁶ In this model, the receptor acts as a transmission device that signals to the cell that a specific ligand has arrived. Receptors also behave as ligands themselves, acting on intracellular signaling proteins to amplify the ligand’s effects. This idea is supported by the hypothesis of Sutherland and Rall on primary and secondary messengers mediating cell signaling.⁶ They presented the theory that each receptor acts as a mediator between a primary messenger, such as an agonist, and a secondary messenger, such as an allosteric modulator, through which the cellular signal was converted.^7,8 Rodbell used Sutherland’s observations to identify an allosteric modulator, which he named a “G protein” due to its activation by guanosine triphosphate (GTP) exchange, as the true second messengers of agonist-induced signaling.⁹ Meanwhile, Gilman became the first to isolate a G protein from cell membranes, provided evidence for the existence of multiple G protein isoforms, and concluded that these isoforms could signal through distinct pathways. Thus, he became one of the first researchers to describe a core concept behind biased signaling.¹⁰ 

In the past 30 years, the term “biased signaling” has traditionally been interchangeable with “biased agonism,” referring to the ability of an agonist to preferentially activate one signaling pathway over another. This concept is best represented in the context of G protein–coupled receptors (GPCRs).¹¹ Two main signaling pathways are extensively described in GPCR signaling: G protein–mediated signaling and β-arrestin signaling. G protein signaling triggers mobilization of effector proteins, such as phospholipase C-β (PLC-β), protein kinases, and Rho-dependent kinases depending on which isoform is acting on the receptor. G⍺_q activates PLC-β, which increases intracellular calcium flux through phosphoinositide hydrolysis and generation of inositol triphosphate (IP3). Both G⍺q and G⍺_12/13 can increase Rho guanine nucleotide exchange factor (GEF) activity to promote RhoA-GTP formation to activate RhoA. G⍺_i-coupled signaling is characterized by decreasing cyclic adenosine 5′-monophosphate levels, whereas G⍺_s-coupled signaling increases cyclic adenosine 5′-monophosphate.¹² Phosphorylation of the GPCR’s carboxy terminus results in the recruitment of β-arrestin prompting changes in cell shape and receptor trafficking. For a ligand to be considered biased, it must stabilize a single unique receptor conformation that generates a preferential signaling response independently of the cell system analyzed. This idea has been debunked for certain ligands that block some pathways but amplify others. This also implies that these ligands were selectively stabilizing a range of active receptor conformations to preferentially activate one signaling cascade over another. In this framework, biased agonism violates the 2-state model for receptor signaling in which receptors have only one active and inactive state. Instead, they act as passive conduits for the ligand. It also implies that efficacy is not an inherent aspect of the receptor.¹¹ Therefore, GPCRs should not be modeled as rigid “on-off” switches, but they should be thought of as fluid proteins with multiple potential active states and efficacies.

As a more in-depth understanding of biased signaling was cultivated, the term expanded to include biased receptors, system bias, and ligand bias to complement the presence of biased agonists. The idea that receptors could be biased emerged between the late 1960s and 1980s, with one of the first papers providing experimental evidence of biased signaling in serotonergic receptors being published in 1987.¹³ Biased receptors can be generated through mutations in the receptor or alternative splicing mechanisms which affect receptor-ligand binding or second messenger signaling.¹¹ System bias describes an environment in which specific signaling pathways are preferentially activated due to the types of transducers expressed in a cell. For example, a ligand can act as an agonist in one tissue, but this same ligand introduced to another tissue may act as an antagonist.^14,15 

Here, we summarize the current literature on biased signaling mechanisms in the context of platelet receptor signaling. Platelets are essential mediators of diverse biological functions such as hemostasis, host immune responses, and tumor metastasis.¹⁶ These diverse functions are executed through their surface receptors, which mediate signaling (GPCRs and immunoreceptor tyrosine-based activation motif), adhesion (selectins, integrins, and glycoproteins), and pathogen and damage sensing (Toll-like receptor 4 [TLR4]). Each receptor has specific ligands that initiate receptor-dependent intracellular signaling pathways leading to distinct physiological outcomes such as calcium mobilization, cell adhesion, leukocyte recruitment, granule secretion, and platelet aggregation.^17,18 To date, biased signaling in platelets has been described for protease-activated receptors (PAR1 and PAR4) and TLR4.^19-21 These 2 classes of receptors are significant players in platelet responses and attractive therapeutic targets. In this review, we focus on the current hypotheses for biased signaling mechanisms in PAR1, PAR4, and TLR4 through the lens of platelet biology, its potential impact in therapeutics for blood-related disorders, and broadening our understanding of platelet function.

Platelets contain several GPCRs that respond to thrombin (PAR1 and PAR4), adenosine 5′-diphosphate (P2Y₁₂ and P2Y₁), thromboxane (TP), prostaglandins, serotonin (5HT-2A), chemokines (CCR5 and CXCR4), and epinephrine (adrenergic receptor α_2A) with each contributing to platelet function.^18,22 There are 5 primary signaling pathways used by GPCRs, which are as follows: G⍺_q, G⍺_12/13, G_i, G⍺_s, and β-arrestin.²² When GPCRs are activated, a conformation change induces the alpha subunit of the heterotrimeric G protein to separate from the βγcomplex. As a result, there is an exchange of guanosine diphosphate for energy-rich GTP to activate secondary messengers that help elicit the platelet response.²³ 

The 3 main G protein signaling pathways relevant to platelet function are G⍺_q, G⍺_i, and G⍺_12/13. When activated by TP receptor or PARs, G⍺_12/13 triggers GEF to activate Rho family members and induce platelet shape change through pathways that include Rho-activated kinase (p160ROCK) and LIM kinase.^24,25 These activated kinases lead to phosphorylation of myosin light-chain kinase and cofilin to regulate both myosin and actin filament formation resulting in platelet shape change.²⁶ The G_i pathway is activated downstream of P2Y₁₂.²⁷ G⍺_i is associated with inhibition of adenylyl cyclase activity which contributes to platelet aggregation. Activation of the Gβγ_i subunit culminates in phosphorylation of PLC-β, diacylglycerol, and IP3 leading to increased cytosolic Ca²⁺ and platelet degranulation. Gβγ_i signaling also activates phosphatidylinositol 3-kinase, Akt, and Rap1B, resulting in integrin activation and platelet aggregation.^28-30 G⍺_q signaling is initiated by P2Y₁, TP, and PARs activating the PLC-β pathway to initiate phosphoinositide hydrolysis and generation of IP3 to increase cytosolic Ca²⁺.²⁴ Intracellular Ca²⁺ signaling triggers integrin activation and platelet aggregation through the CalDAG-GEF/Rap1/RIAM pathway.³¹ G⍺_q also activates phospholipase A₂ to amplify thromboxane A₂ synthesis in addition to phosphorylation of myosin light-chain kinase for platelet shape change.³² 

β-Arrestins are multifunctional intracellular proteins that interact with several cellular partners, including GPCRs.³³ There are 2 isoforms of β-arrestin, arrestin-1 and arrestin-2. Arrestin-2 is a mediator of platelet aggregation, granule secretion, immune regulation, inflammatory responses, and activation of integrins.³⁴ β-Arrestin is activated through the formation of a GPCR–β-arrestin complex driven by phosphorylation of the carboxy terminus of the activated GPCR.^34,35 β-Arrestin turns off signaling from GPCRs, known as desensitization, by blocking G protein–mediated signaling. β-Arrestin is also responsible for receptor trafficking and internalization which determine the number of surface receptors that can be recycled back to the plasma membrane.³⁶ Resting platelets have a “steady state” of receptors on the surface and within the cell. Once platelets are activated, β-arrestins guide receptors from the cytosol to the membrane, increasing their surface expression.^34,37 Consequently, membrane protein trafficking contributes to platelet-cell interactions and platelet function.³⁸ 

There are 4 PARs (PAR1-4) named in their order of discovery.^39,40 PARs have a distinct activation mechanism in which a protease cleaves the N terminus to generate a tethered ligand. The tethered ligand interacts with the receptor core to induce a conformational change in the receptor to initiate signal transduction through effector proteins. PAR1 and PAR4 are expressed in human platelets. PAR1 was discovered and cloned in 1991 by Vu et al,⁴¹ and PAR4 was discovered in 1998 by Xu et al.⁴² Thrombin is the canonical protease agonist for PAR1 and PAR4 and is a dominant platelet activator of aggregation and thrombogenesis.^43,44 PAR1 and PAR4 signal through the G⍺_q, Gα_12/13, and β-arrestin pathways. Although they were once thought to be functionally redundant, we now know that PAR1 induces rapid transient signaling whereas PAR4 induces sustained signaling, and they cooperate to mediate the full thrombin response.^19,45 

PARs are activated by multiple proteases at canonical and noncanonical sites (Figure 1). The unique protease-mediated activation mechanism of PARs creates distinct ligands depending on the cleavage site. Thus, the physiological responses resulting from PAR activation also depend on the proteolytic cleavage site for the activating protease.⁴⁴ Each of the unique tethered ligands has the potential to act as biased agonists that activate distinct pathways leading to context-dependent signaling (Figure 2).

PAR1 is cleaved by thrombin at the canonical site, Arg41 at subnanomolar concentrations. In addition to the cleavage site, PAR1 contains a hirudin-like binding site that enables it to bind thrombin’s exosite-I with high affinity.⁴⁵ Thrombin-mediated cleavage of PAR1 predominantly results in G⍺_q and G⍺_12/13 pathway activation to facilitate transient calcium mobilization, actin cytoskeletal rearrangement, surface expression of platelet adhesion receptors, and inside-out activation of integrin αIIbβ3 (Figure 2).⁴⁶ Given its key role in platelet activation, PAR1 is a target for developing new antiplatelet agents.

Matrix metalloproteinases 1 and 2 (MMP1 and MMP2) initiate biased signaling by cleaving PAR1 at distinct sites from thrombin.⁴⁷ MMP1 cleaves PAR1 at Asp39, creating a novel tethered ligand extended by 2 aa in comparison to thrombin (Figure 1). MMP1 primarily activates G⍺_12/13, which triggers MAPK signaling, leading to platelet shape change. Unlike thrombin, MMP1 induces a small intracellular calcium response and weak platelet aggregation (Figure 2).^48,49 MMP2 is secreted by platelets at the site of platelet plug formation and contributes to sustained activation by cleaving PAR1 at Leu38.⁵⁰ In contrast to other proteases, MMP2 requires the integrin αIIbβ3 as a cofactor to facilitate PAR1 cleavage.²⁰ MMP2-activated PAR1 initiates G⍺_q and G⍺_12/13 signaling leading to downstream MAPK phosphorylation, phosphatidylinositol 3-kinase activation, and calcium mobilization. However, concomitant G⍺_i signaling is required for full platelet aggregation and adenylyl cyclase inhibition.⁵⁰ Thus, the MMP2-PAR1 axis primes platelets to be activated by other stimuli.²⁰ 

PAR1-biased signaling also occurs in endothelial cells. Thrombin cleaves at Arg41 and promotes vascular inflammation through G⍺_q and G⍺_12/13 signaling resulting in disassembly of adherens junctions, rearrangement of the actin cytoskeleton, and disruption of the endothelial barrier.^51,52 Conversely, activated protein C (APC) cleaves PAR1 at Arg46 when it is bound to its cofactor, endothelial protein C receptor. APC exhibits anti-inflammatory and antiapoptotic effects through stabilization of the endothelial barrier by preferentially signaling through β-arrestins.^53,54 In addition, neutrophil elastase and proteinase-3 cleave endothelial PAR1 and signal through G⍺_i to promote MAPK activation.⁵⁵ However, the physiological context in which elastase and proteinase-3 promote PAR1-biased signaling and whether they affect PAR1 signaling in platelets is undetermined.

Thrombin cleaves PAR4 at Arg47, resulting in a unique tethered ligand in comparison to PAR1 (Figure 1). PAR4 lacks the hirudin-like sequence. Therefore, higher thrombin concentrations are required for PAR4 activation. Instead, PAR4 possesses a cluster of anionic residues in its exodomain that slow receptor dissociation from thrombin’s cationic autolysis loop.⁵⁶ PAR4 also has dual prolines that interact with the active site of thrombin to compensate for the absence of the hirudin-like site.^42,45,56 Consequently, PAR4 is a less efficient thrombin substrate than PAR1. However, once activated, it has sustained calcium mobilization in comparison to PAR1.⁵⁷ 

The potential biased signaling from PAR4 has only recently been explored. PAR4 is cleaved by multiple proteases, but all at the canonical Arg47 site.³⁹ In addition to Arg47, a novel PAR4 cleavage site (Ser67/Arg68) for the neutrophil protease cathepsin G was recently identified using mass spectrometry of peptides that mimic the N terminus of PAR4 (Figure 1).⁵⁸ This opens the possibility for context-dependent activation of PAR4 by cathepsin G. Cathepsin G induces calcium mobilization and platelet aggregation, but the mechanism and physiological significance of cathepsin G-mediated PAR4 activation in platelets remain understudied.⁵⁸ Evaluating this mechanism will provide better understanding of the reciprocal activation effect that platelets and neutrophils share with each other.⁵⁹ It will also further identify PAR4 signaling patterns that can be exploited for therapeutic benefit because platelet-neutrophil interactions are significant contributors to the pathogenesis of thrombosis.⁶⁰ 

Heterodimerization of GPCRs can influence receptor signaling. Specifically for platelets, dimerization with P2Y12 directs PAR4 signaling toward the β-arrestin pathway. Li et al revealed that inhibition of P2Y12 reduces PAR4-mediated Akt activation and recruitment of arrestin-2 to PAR4, which stabilizes the growth of platelet-rich thrombi.⁶¹ Thrombin activation of PAR4 induces direct association with P2Y12 in human platelets.⁶¹ Using site-directed mutagenesis, Khan et al revealed that PAR4-P2Y12 dimerization and PAR4-mediated calcium mobilization are facilitated by a 3-aa sequence, Leu194-Gly195-Leu196, located at the base of TM4 in PAR4.⁶² Overall, these findings revealed that PAR4-P2Y12 heterodimerization biases PAR4 toward arrestin signaling in platelets to stabilize the growth of thrombi.

Because PAR1 and PAR4 are 2 dominant mediators of platelet activation, they have been attractive targets for antiplatelet therapies in the last 20 years. Despite major progress in understanding PAR1 and PAR4 function on platelets, developing pharmacological agents targeting these receptors has been difficult due to the tethered ligand activation mechanism.⁶³ However, progress has been made with small molecules and antibodies that bind to orthosteric or allosteric sites to completely inhibit signaling or bias receptor signaling.^40,63 

The only current US Food and Drug Administration–approved drug that targets PAR signaling is the PAR1 antagonist vorapaxar, which was approved in 2014 for the prevention of thrombotic cardiovascular events.⁶⁴ Vorapaxar is a competitive small-molecule PAR1 inhibitor that interacts at the ligand binding site to orthosterically block PAR1 signaling. Vorapaxar binds to PAR1 with high affinity and has high lipid and protein binding, which prolongs its activity and creates challenges for patient management. Owing to the increased bleeding risk, it comes with a black box warning.⁶⁵ 

Parmodulins were first characterized as reversible, allosteric, small-molecule PAR1 antagonists that bind to the cytoplasmic interface of PAR1 at helix 8 to bias PAR1 signaling to G⍺_12/13 instead of G⍺_q.⁶⁶ Parmodulins do not interfere with the extracellular protease recognition sites on the N terminus or the ligand binding site of the receptor. This unique quality of parmodulins directs PAR1 signaling to a specific pathway without ablating the signaling response altogether. For example, parmodulins have been observed to inhibit prothrombotic signaling without compromising the cytoprotective APC-mediated signaling in endothelial cells.⁶⁷ The compound ML161 (also known as parmodulin 2) inhibits platelet-dense granule release while maintains platelet shape change, further suggesting that parmodulins selectively inhibit G⍺_q signaling, but not G⍺_12/13.⁶⁸ 

Because PAR4 activation results in sustained signaling that leads to stable clot formation, it is an attractive antiplatelet target. Small molecules and inhibitory antibodies have shown promise in preclinical models. In 2017, Bristol Myers Squib reported a series of small molecules that reduced thrombosis in nonhuman primates with less bleeding than US Food and Drug Administration–approved P2Y12 antagonists.⁶⁹ Subsequently, BMS-986141 was evaluated in phase 2 clinical trials for the reduction in recurrent stroke and the prevention of ministrokes, but the study was terminated.⁷⁰ BMS-986120 was evaluated in phase 1 clinical trials for the prevention and treatment of thromboembolic disorders.⁶⁹ BMS-986120 demonstrated selective and reversible antiplatelet effects with high oral availability in humans.⁶⁹ It also reduced ex vivo thrombus formation in conditions representative of deep arterial injury and stenosed coronary artery.⁶⁹ Furthermore, a single dose resulted in complete inhibition of PAR4 activation peptide-stimulated platelet activation within 2 hours, and this inhibition was resolved within 24 hours. Importantly, BMS-986120 did not increase bleeding or adverse events in phase 1 clinical trial.⁶⁹ Smith et al performed computational modeling of imidazothiadiazole compounds such as BMS-986120 in combination with virtual high-throughput screening to identify a novel lead compound.⁷¹ From this, a series of analogs that block thrombin signaling of PAR4 were developed. Notably, these small molecules do not block PAR4 signaling initiated by the PAR4 activation peptide mimetic, AYPGKF.

To date, there have been no antagonists described that block specific pathways downstream of PAR4. However, Thibeault et al synthesized a library of peptides derived from the PAR4 parent activation peptide (AYPGKF-NH₂) that show promise.¹⁹ By sequentially mutating each amino acid in the parent peptide sequence, they identified the key residues important for PAR4 activation and potential biased agonists for that preferentially blocked G protein–coupled or β-arrestin–dependent signaling pathways. Notably, most peptides did not fully abrogate one signaling pathway in favor of another. Rather, they observed a more robust response from one pathway, highlighting foundational biased signaling principles hypothesized by Rodbell.¹¹ Compounds such as these hold potential for pathway-specific modulation of PAR4 signaling.

TLRs are part of the innate immune system that recognizes molecular patterns in response to damage or foreign pathogens.⁷² TLRs are transmembrane, leucine-rich repeat proteins expressed on platelets, neutrophils, T lymphocytes, B lymphocytes, dendritic cells, and natural killer cells.⁷³ Their primary function is to mediate the innate immune response by stimulating the synthesis of proinflammatory cytokines and chemokines.^73,74 There are 10 human TLRs and 13 mouse TLRs, 4 of which are expressed on platelets (TLR1, TLR2, TLR4, and TLR6).⁷³ 

TLR4 is the most studied on platelets where it recognizes the lipopolysaccharide (LPS) component of the cell envelope of gram-negative bacteria. LPS stimulation of platelets promotes platelet-neutrophil aggregates that stimulate the formation of neutrophil extracellular traps (NETs) and subsequent platelet activation.^75,76 LPS treatment of platelets results in a concomitant increase of soluble CD40 ligand (sCD40L) in plasma, and increases in sCD40L levels are suggested to directly involve TLR4. However, it is unclear whether LPS induces platelet-neutrophil aggregates directly or indirectly through TLR4-mediated increased levels of sCD40L.⁷⁵ Direct platelet activation through TLR4 is mediated through both cyclooxygenase and thromboxane A₂ pathways.⁷⁷ However, there are conflicting reports on whether LPS can stimulate expression of platelet-surface activation markers such as P-selectin.⁷⁸ 

TLR4 on platelets has pleomorphic effects depending on the physiological context. For example, it is essential for platelet activation and aggregation during hemorrhagic shock to prevent abnormal clot formation.⁷⁹ However, platelet TLR4 activation also leads to inflammatory injury in the lung and liver.⁷⁹ In addition, the genetic deletion of TLR4 on mouse platelets weakens platelet activation and decreases pulmonary vascular hypertrophy, thus improving pulmonary hypertension outcome.⁸⁰ In a murine cancer model, platelet TLR4 helps promote distant metastasis through forming platelet-tumor cell aggregates in an integrin-dependent manner.⁸¹ In severe sepsis, high levels of S100A8/A9 bind to platelet TLR4, leading to a positive feedback loop involving platelet pyroptosis and resulting in excessive inflammatory cytokine release.⁸² 

Unlike PARs, TLR4-biased signaling is dictated by its localization within the cell rather than interaction with different external agonists. TLR4 activation begins with the formation of a heterotetrametric complex with TLR4, myeloid differentiation factor 2 (MD-2), CD14, and LPS. LPS stimulates signaling through the TLR4 MyD88-dependent or -independent pathway depending on the LPS chemotype (Figure 3).⁸³ 

The proinflammatory MyD88-dependent signaling pathway is dependent on the TLR4 complex remaining at the plasma membrane (Figure 3).⁸⁴ In this pathway, several cytosolic proteins form a large multimeric complex called the Myddosome. Once activated in nucleated cells, a component of the Myddosome dissociates to ultimately allow NF-κB translocation to the nucleus and enhance transcription of tumor necrosis factor-α and interleukin-1β.⁸⁵ In platelets, NF-κB is hypothesized to stimulate platelet granule secretion, aggregation, and cytokine expression.⁸⁶ Platelets do not have a nucleus nor genomic DNA, but they do contain messenger RNA transcripts that can be differentially processed after stimulation by LPS or thrombin.⁸⁷ The affected transcripts include proinflammatory cytokines such as interleukin-1β and platelet agonist producers such as cyclooxygenase-2, though the mechanism remains unclear.⁸⁸ 

MyD88-independent signaling is only triggered once the ligand-bound TLR4/MD-2/CD14 complex is internalized. In nucleated cells, MyD88-independent signaling is initiated by endocytosis of the TLR4–MD-2 heterotetrametric complex in the presence of dynamin and membrane-bound CD14.⁸⁹ Once internalized, TLR4 signaling is mediated through effector proteins such as TRIF (toll interleukin-1 receptor domain-containing adapter-inducing interferon beta) and TRIF-related adapter molecule.⁹⁰ This ultimately leads to phosphorylation of interferon regulatory factor 3 (IRF3), which translocates to the nucleus to stimulate the transcription of type 1 interferons.^91,92 In platelets, MyD88-independent signaling is hypothesized to stimulate platelet granule secretion, aggregation, or cytokine production.^93,94 Platelets lack membrane-bound CD14, a necessary component for internalization of TLR4 and LPS. Thus, the requirements for TLR4–MD-2 internalization in platelets are unknown. It is speculated that high levels of soluble CD14 in the plasma compensate for its absence on the membrane to internalize the complex.⁹⁵ This may act to prevent “priming” platelet activation and NET formation during minor bacterial infections where low concentrations of LPS would be present while preserving platelet response during severe inflammation.⁹³ Thus, activation of the TLR4 MyD88-independent signaling by soluble CD14 could prevent unnecessary endothelial damage by preventing platelet-mediated NETosis during acute infection.⁹³ 

Beyond platelets, TLR4 exhibits biased signaling mechanisms in macrophages. Macrophage TLR4 demonstrates a MyD88-independent pathway bias in response to E6020 over monophosphoryl lipid A, 2 synthetic LPS lipid A analogs used as vaccine adjuvants. Despite the signaling bias, both adjuvants demonstrated comparable effects on increasing antibody response.⁹⁶ These findings suggest that a biased signal may not correlate with differing outcomes in immune response. In human macrophages, TLR4 helps differentiate between different bacteria to activate biased signaling pathways.⁹⁷ Pseudomonas aeruginosa signals through TLR4 to p65 using both MyD88-dependent and -independent signaling. Similarly, Escherichia coli uses both pathways to induce IRF3 phosphorylation with or without the endosome. Salmonella enterica only induces p65 and IRF3 through the endosome. Resulting macrophage activation from different bacteria involved different transcription factors such as p65 or IRF3 and different dependencies on endosomes.⁷⁹ Further studies are needed to observe whether there is a similar difference in platelet TLR4 bias signaling.

Emerging evidence demonstrates that PAR1, PAR4, and TLR4 are capable of biased signaling. This raises the potential for novel therapeutics that direct signaling toward specific pathways. The potential to selectively block one arm but spare another may have therapeutic benefits by providing safer alternatives. In some cases, selective activation of specific pathways may have protective physiological effects by facilitating adhesion of platelets to immune or endothelial cells. Furthermore, platelet life span is regulated through calcium-dependent apoptosis.⁹⁸ Biased agonists have the potential to extend or reduce the half-life of platelets in circulation. These all need to be formally tested. We should continue our push to understand the structural basis for activation of each receptor and how these are influenced by other proteins in the platelet membrane. Novel screening approaches that have been used by Smith et al and Thibeault et al who identify new compounds will provide new avenues to explore for both PAR and TLR signaling.^19,71 

The authors thank Xu Han for assistance in creating the figures.

M.T.N. reports research funding from National Institutes of Health, National Heart, Lung, and Blood Institute grant HL098217. N.C.K. reports funding from the American Society of Hematology Hematology Inclusion Pathway Graduate Student Award.

Contribution: N.C.K., G.H.H., G.J.H., and M.T.N. wrote and critically reviewed the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Marvin T. Nieman, Department of Pharmacology, Case Western Reserve University, 2109 Adelbert Rd W309B, Cleveland, OH 44106-4965; email: marvin.nieman@case.edu.