|Home | About | Journals | Submit | Contact Us | Français|
The Nuclear Factor (NF)-κB transcription factor family is well-known for their role in eliciting inflammation and promoting cell survival. We discovered that human megakaryocytes and platelets express the majority of NF-κB family members including the regulatory Inhibitor (I)-κB and Inhibitor Kappa Kinase (IKK) molecules.
Investigate the presence and role of NF-κB proteins in megakaryocytes and platelets.
Anucleate platelets exposed to NF-κB inhibitors demonstrated impaired fundamental functions involved in repairing vascular injury and thrombus formation. Specifically, NF-κB inhibition diminished lamellapodia formation, decreased clot retraction times and reduced thrombus stability. Moreover, inhibition of I-κB-α phosphorylation (BAY-11-7082) reverts fully spread platelets back to a spheroid morphology. Addition of recombinant IKK-β or I-κB-α protein to BAY inhibitor-treated platelets partially restore platelet spreading in I-κB-α inhibited platelets, and addition of active IKK-β increased endogenous I-κB-α phosphorylation levels.
These novel findings support a crucial and non-classical role for the NF-κB family in modulating platelet function and reveal that platelets are sensitive to NF-κB inhibitors. As NF-κB inhibitors are being developed as anti-inflammatory and anti-cancer agents, they may have unintended effects on platelets. Based on these data, NF-κB is also identified as a new target to dampen unwanted platelet activation.
The traditional view of platelet biology has dramatically changed as novel activities of these anucleate cells continue to be discovered. Historically, platelets were viewed solely as mediators of thrombosis and hemostasis, but clearly platelet function has expanded to include key roles in inflammation and immunity[1, 2]. Platelets are the product of the megakaryocyte. Remarkably, platelets have unexpected features, including the presence of a substantial and diverse transcriptome, spliceosomal components for mRNA processing, and signal-dependent translation, providing unique mechanisms to rapidly produce new proteins[5, 6]. Recent characterization of the de novo synthesis of platelet mRNAs[7, 8] demonstrates the sophistication of platelet signaling and function, underscoring their role as formidable players in regulating coagulant and inflammatory pathways.
Many novel and unexpected proteins have been identified in platelets including transcription factors. We recently demonstrated that platelets contain the transcription factor peroxisome proliferator-activated receptor gamma (PPARγ) and its heterodimeric partner retinoid X receptor (RXR)[9, 10]. PPARγ ligands attenuate platelet release of pro-inflammatory and pro-coagulant mediators including soluble CD40L (sCD40L) and thromboxane A2 (TXA2)[9, 10], suggesting a new role for platelets in inflammation. Our laboratory further demonstrated that platelet microparticle (PMP)-released PPARγ was capable of transcellular biologic activity. Additionally, it was previously reported that platelets contain some nuclear factor (NF)-κB family members[6, 12-14].
The NF-κB protein family regulates both activation and repression of gene transcription involved in complex signaling pathways, including apoptosis, immune responses and inflammation(and refs within). Five Rel/NF-κB DNA-binding subunits (RelA (p65), RelB (p68), c-Rel, p50 (NF-κB1) and p52 (NF-κB2)) form both heterodimeric and homodimeric complexes and are found in the cytoplasm of most nucleated cells bound to I-κB proteins that maintain these complexes in an inactive state. In response to specific stimuli, differentially formed NF-κB dimers and inhibitory proteins are regulated by an I-κB kinase (IKK) complex via classical or alternative pathways to regulate a multitude of genes. Moreover, Rel/NF-κB activation can be blocked by non-genomic mechanisms such as protein modification or physical association with other proteins. For example, binding of RelA (p65) by PPARγ prevents nuclear translocation and also expedites nuclear export of NF-κB. Although NF-κB regulation has been extensively studied, the prodigious number of physiological processes controlled by these proteins still provides many challenges toward understanding the mechanisms involved in NF-κB signaling pathways.
Based on our prior finding of the transcription factor PPARγ in platelets, we were interested in looking for other transcription factors. Identification of other transcription factors in platelets is important as these proteins may have important non-transcriptional roles. Herein, we present our findings on the presence and activity of NF-κB family members.
Whole blood was obtained under Institutional Review Board approval following informed consent from male and female donors, 21-65 years of age that were NSAID-free for two weeks prior to donation. Blood was collected by venipuncture and platelets were washed and prepared for spreading as described[9, 17]. Platelet purity was determined to be >99%.
Western blot analysis of lysates (5-10 μg/lane) was performed using mouse monoclonal (p50 (E-10), p52 (C-5) and IKKbeta (H-4)), or rabbit polyclonal p65 (C-20), c-Rel, RelB(c-19), IκB-α (C-21) IκB-β, IKK-γ, and Bcl-3) antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) and goat polyclonal GST (GE Healthcare, Piscataway, NJ) followed by goat anti–rabbit, goat anti-mouse (Jackson Immuno Research Lab, West Grove, PA, USA) or donkey anti-goat (Rockland, Gilbertsville, PA) horseradish peroxidase secondary antibody. Platelet activation was performed at 37°C for 30 mins.
Measurements of p50, and RelA (p65) in platelet lysates were obtained using commercially available, highly specific and sensitive TransAM transcription factor assay kits (Active Motif, Carlsbad, CA).
Washed platelets (3×108 cells/mL) were incubated at 37°C/30 minutes with vehicle (DMSO <0.1%), or Bay-11-7082 (1 μM) (Biomol, Plymouth Meeting, PA), or SC-514 (10 μM) (Calbiochem, San Diego, CA). Concentrations were pre-determined by dose-response assays. Pellets were resuspended in Nonidet P-40 (Sigma) lysis buffer containing protease inhibitor cocktail (Sigma). Spread platelets post-treated with inhibitors were washed and incubated with inhibitors up to 30 minutes/37°C. Human recombinant IKK-β (Active Motif) or I-κB-α protein (Santa Cruz) was added to platelets in Tyrode's buffer (Sigma) containing DMSO (<0.1%), as a vehicle for uptake. To examine phosphorylation levels following recombinant IKK-β addition, platelets were lysed in buffer containing 25 mM Tris, 150 mM NaCl, 1% IGEPAL, 1% Sodium Deoxycholate, 0.1% SDS, 1 mM Sodium Fluoride, 10 mM Sodium Pyrophosphate, 1 mM Sodium Orthovanadate and protease inhibitor cocktail (Sigma).
Platelet spreading was performed as described . Cells were stained with phalloidin-Alexa Fluor 488 (Invitrogen, Carlsbad, CA). Platelet spreading was imaged using Nomarski differential interference contrast optics and SPOT RT software (New Hyde Park, NY) with an Olympus BX51 microscope (Melville, NY). Platelets for scanning electron microscopy (SEM) imaged were using a Zeiss 40VP SEM (Thornwood, NY).
Meg-01 megakaryocytic cells were purchased from the American Type Culture Collection (Rockville, MD) and cultured as previously described.
Human PRP was incubated with or without inhibitors (30 minutes/37°C) followed by thrombin (1 U/mL) addition. Clot formation and retraction was recorded using a Kodak Z812 IS digital camera. Clot size was calculated as the ratio of the area of the retracted clot/area of the total PRP volume and recorded as a percent.
Standard clinical laboratory clotting assays were performed to detect deficiencies in certain factors of the extrinsic (PT), or intrinsic (PTT) pathways and fibrinolysis (ECLT). The times reported are referenced to a standardized control.
Glass capillaries were coated with 100 μg/ml collagen (Chrono-Log) and blocked with 0.5% BSA. Perfusion was performed via a motorized multiple syringe pump (New Era Pump Systems) and images were obtained using an IX-81 motorized inverted microscope (Olympus America Inc.) coupled to a CCD camera (Hitachi), and connected to a TV monitor through a DVD recorder (Sony Electronics). IPLab deconvolution software (version 4.04, Scanalytics, Inc.) enabled automated data acquisition in all three dimensions. Whole blood samples were perfused for eight minutes through capillaries at a shear rate of 500 s-1 (flow rate = 368 μL/min), followed by a PBS wash at the same shear rate for approximately 8 minutes during which images were photographed.
Experiments were repeated from a minimum of three individuals except where stated. Results are shown as mean +/- standard error (S.E.). Data were evaluated by one-way or repeated measures analysis of variance (ANOVA). A value <0.05 was considered statistically significant.
Human platelets, primary megakaryocytic (human cord blood-derived CD34+ cells), and Meg-01 cells, a cell line extensively used as a model of human megakaryocytes, were analyzed by Western blot using antibodies specific for NF-κB family members. For each cell type, bands were found corresponding to classic Rel/NF-κB and I-κB subunits (Figure 1). The protein levels of some family members were reduced following agonist activation, as was observed in platelets for RelA. In contrast, RelA, p50, p100 and p52 increased following activation of primary human megakaryocytic cells with a phorbol ester, PMA. RelB was present in Meg-01 cells and induced in primary human megakaryocytic cells, but only weakly present in platelets. Agonist stimulated platelets contained higher levels of I-κB-α and Bcl-3, but PMA activation did not substantially change levels of these proteins in primary human megakaryocytic cells. Platelets were previously shown to upregulate Bcl-3 protein upon agonist stimulation. In primary human megakaryocytic cells, the levels of inhibitor kinases, IKK-β and IKK-γ (NEMO) were altered by PMA activation, whereas there was no change in platelets in response to agonist stimulation. The doublet bands evidenced in the IKK-γ immunoblot (Figure 1, IKK panel) likely represent a phosphorylated form of IKK-γ.
To detect whether platelet NF-κB can form complexes and bind to its DNA consensus a TransAM NF-κB activity assay was used to measure protein binding to plate-bound NF-κB-DNA oligonucleotides. The subsequent detection of bound proteins with NF-κB specific antibodies demonstrated DNA-binding activity for RelA (p65) and p50 (Figure 2) and was confirmed by electromobility shift assay (data not shown).
To ascertain whether NF-κB plays a physiological role in platelets, widely-used small molecule inhibitors of the NF-κB pathway were tested to determine their potential effects on platelet spreading. The first, SC-514, is a selective and reversible inhibitor of IKK-β that decreases levels of I-κB-α phosphorylation/degradation and the second, BAY-11-7082 (BAY), irreversibly inhibits I-κB-α phosphorylation to reduce NF-κB activity . As shown in Figure 3A, untreated column, platelets spread fully on immobilized fibrinogen. In contrast (Figure 3A, BAY column) lamellipodia formation was severely hindered in BAY inhibitor-treated platelets, although these platelets can form filipodia (yellow arrows). Platelets treated with SC-514, exhibited an intermediate morphology (Figure 3A; SC-514 column). Some platelets spread similarly to untreated platelets (white arrow), others have morphology similar to BAY inhibitor-treated platelets (yellow arrow) and a third group form branched lamellipodia (orange arrow). NF-κB inhibitor alteration of platelet spreading on collagen is similar to fibrinogen spreading (data not shown).
To further explore inhibitor effects on the platelet cytoskeleton, spread platelets were treated with Triton X-100 to expose the detergent-resistant cytoskeleton. Figure 3B (untreated column) shows normal actin filament formation as compared to the more disorganized distribution in SC-514 treated platelets (Figure 3B, SC-514 column) and the loss of cytoskeletal reorganization in BAY inhibitor-treated platelets (Figure 3B, BAY column). Phalloidin staining for detection of actin filaments (Figure 3C) further demonstrates the cytoarchitectural changes in the inhibitor-dosed platelets.
Quantification of our data further supports the conclusion that lamellae formation is abrogated in BAY-11-7082-treated platelets and reduced following SC-514 treatment (Figure 3D). The effects of SC-514 were even better appreciated when platelet surface area was examined (Figure 3E), demonstrating a significant reduction in the surface area covered compared to untreated platelets. A time course determined platelet spreading did not recover in inhibitor-treated platelets after 90 minutes (not shown).
To better understand the involvement of NF-κB proteins in platelet shape change, fully spread platelets were treated with increasing doses of BAY inhibitor. Figure 4A demonstrates that following post-treatment with BAY inhibitor, large circumferential lamellipodia were retracted and platelets returned to a convoluted spheroid shape. This effect was dose-dependent as increasing concentrations of BAY produced greater numbers of rounded platelets. Quantification of this process is shown in Figure 4B.
To determine whether the effects of BAY-11-7082 were dependent on NF-κB function, the selectivity of this inhibitor was measured. GST-tagged recombinant human (rh)IKK-β or rhI-κB-α was added to platelets pre-treated with BAY inhibitor and spread on a fibrinogen. After thorough washing, adherent platelets were imaged, or removed and lysed. A Western blot revealed bands representing the exogenously added rhIKK-β or rhI-κB-α in lysates treated with recombinant protein only (Figure 5A). An anti-GST antibody identified the same bands in the samples containing the recombinant proteins. All samples contained the endogenous IKK-β and I-κB-α proteins at their expected molecular weights (data not shown, the GST tag adds 27kD to the recombinant proteins). As shown by light microscopy for rhI-κB-α and graphically for IKK-β, addition of these recombinant proteins partially restored lamellipodia formation in BAY-treated platelets (Figure 5B & C). Untreated or vehicle treated platelets were not significantly changed following addition of recombinant protein (data not shown). Further, active rhIKK-β was transferred to platelets to investigate the phosphorylation status of the IKK-β substrate, I-κB-α. An antibody that recognizes only the phosphorylated form of I-κB-α indicated that the level of phosphorylation was significantly increased only in platelets that received rhIKK-β. Moreover, fluorescently labeled rhIKK-β was internalized in BAY-treated platelets (data not shown).
To investigate whether SC-514 or BAY inhibitors dampen platelet aggregation, PRP was stimulated with ADP or thrombin, with or without inhibitor pre-treatment and aggregation measured over time. Interestingly, aggregation was unaffected by either SC-514 or BAY inhibitors over a wide range of inhibitor concentrations (1-200 μM) (data not shown).
Clot retraction is a platelet-dependent process that results in aggregate contraction essential for thrombus consolidation and resistance to shear stress during wound healing. PRP was pre-treated with increasing doses of BAY and SC-514 inhibitors followed by activation with thrombin to observe whether these inhibitors interfered with normal clotting. Significant retraction in untreated or vehicle-treated PRP occurred quickly (Figure 6A). The lower doses of BAY inhibitor (<25 μM) were slightly slower and retracted to ≤20% by 45 minutes (data not shown). In contrast, retraction time was even slower for BAY50 and significantly protracted at higher doses of BAY inhibitor (Figure 6A). Clot retraction in SC-514 treated platelets was not as dramatic as observed in the BAY-treated samples, although it too was significantly delayed compared to controls (data not shown).
To test whether inhibitor effects were platelet-specific or whether other aspects of the clotting cascade were also affected by BAY-11-7082, we performed pro-thrombin time (PT), partial thromboplastin time (PTT) and euglobulin clot lysis time (ECLT) assays on human platelets. PT and PTT are measures of clotting efficacy of either the extrinsic pathway of coagulation, initiated by the release of tissue factor (PT), or the intrinsic pathway and factors common to both pathways (PTT). ECLT measures fibrin dissolution following clot formation. PRP was left untreated, vehicle-treated or treated with BAY (50-200 μM). The test results for PT, PTT and ECLT (data not shown) were all within normal values indicating that the effects of BAY are platelet-specific.
To further investigate platelet adhesion and thrombus formation in a more physiologically relevant setting, a flow adhesion assay was used to study platelet function under shear conditions. As expected, untreated (data not shown) and vehicle-treated platelets formed stable, compact thrombi that remained invariant throughout the buffer wash (Figure 6B, VEH). In contrast, BAY-treated whole blood formed loose aggregates and the thrombi that did form were highly unstable. Figure 6B tracks the release of one embolus in BAY-treated platelets to illustrate the dynamic process that occurred. In all BAY-treated samples, emboli and cells were continually flowing through the buffer. This effect was greatest at the highest concentration of BAY inhibitor and was reduced as inhibitor dosage decreased. By the end of the buffer wash, inhibitor-treated samples contained large areas that were reduced to a cell monolayer compared to controls (data not shown).
Evidence is provided herein for a new pathway of platelet regulation by NF-κB. Our studies demonstrate that primary human megakaryocytic cells, the Meg-01 cell line, and platelets contain nearly all known NF-κB family members including c-Rel, p105/p50, p100/p52, I-κ-Bs and IKKs. Moreover, expression levels and modification of specific NF-κB proteins was dependent on cell type and sensitive to agonist stimulation. For example, RelA (p65) and RelB were highly expressed in unstimulated Meg-01 cells. In contrast, normal platelets expressed low levels of RelB and human primary megakaryocytic cells required agonist stimulation to produce detectable amounts. We also observed activation-specific changes in the levels of some proteins (e.g. RelA and p50) and phosphorylated protein (IKK-γ) in both megakaryocytes and platelets. mRNA transcripts for several NF-κB family members have been reported in human platelets, including I-κB-α[3, 22, 23]. These observations support the concept that NF-κB proteins are important in platelet function.
Platelets are known to contain other transcription factors such as PPARγ and its binding partner, RXR[9, 10, 24]. Both proteins regulate platelet function by nongenomic mechanisms, including RXR interaction with Gq and PPARγ ligand dampening of platelet activation[9, 10]. Additionally, PPARγ and RXR are released in platelet microparticles and elicit a PPARγ ligand-dependent transcellular response in a monocytic cell line. Further studies are in progress to determine whether NF-κB subunits, like PPARγ, are released during platelet activation.
A role for NF-κB in platelet physiology is also supported by data from the Bloodomics Consortium that identified a transcript in platelets encoding a novel gene COMMD7 (copper metabolism gene MURR1 domain containing 7) that correlated with both platelet function and coronary artery disease risk. COMMD7 is a member of the COMM domain protein family involved in the negative regulation of NF-κB activity (Goodall and Ouwehand, unpublished observations, 2009). A Western blot probed with an anti-COMMD7 antibody revealed the presence of the CommD7 protein in platelet lysates (data not shown).
In further support of NF-κB function in platelets, we demonstrate that inhibitors of NF-κB have profound effects on platelet signaling pathways involved in shape change and spreading, modulation of clot retraction, and thrombus stability under flow conditions. The Rho GTPase, Rac1, which is ubiquitously expressed in all mammalian cells, stimulates actin polymerization and lamellae formation in platelets. Here, we show that NF-κB inhibition attenuates lamellae formation, an essential function during vascular wound healing. Intriguingly, the Rho GTPase family was previously reported to be specifically involved in regulating NF-κB activity in nucleated cells, including a role in cytoskeletal organization of critical cellular functions[25, 26]. A significant finding herein was that BAY treatment of fully spread platelets reversed lamellipodia formation, reinstating platelets to a spherical morphology. These new data support a dynamic role for NF-κB in the regulation of platelet cytoskeletal architecture.
Our data also show that NF-κB inhibition significantly hampered clot retraction, although platelets were still able to aggregate as measured by aggregometry at low levels of shear. Our aggregation data are in contrast to the findings of Malaver et. al, who showed that BAY-11-7082 inhibited platelet aggregation. These differences may be attributed to the concentration of the platelet agonist used in PRP (0.8 U/mL thrombin vs 0.05 U/mL thrombin, Malaver et. al). We found by our investigative methods that thrombin addition stimulated clot formation in PRP, but platelets failed to form a tightly condensed clot in a dose-dependent manner. Furthermore, flow adhesion assays indicated that NF-κB-inhibited platelets failed to form stable thrombi on a collagen matrix at physiological levels of shear stress. Although platelets were recruited to the matrix, the aggregates that did form were less compact compared to untreated or vehicle-treated platelets, and emboli were copiously released. It was previously demonstrated that inhibition of signaling events that support platelet activation following integrin activation and platelet aggregation are necessary for clot retraction and for promoting thrombus growth and stability. Our data support a role for NF-κB inhibitors interfering with outside-in signaling via the fibrinogen-binding integrin, αIIbβ3. Initial fibrinogen binding mediates aggregation (inside-out signaling), which subsequently triggers outside-in responses that stabilize aggregates and initiates platelet spreading, vesicle secretion and clot retraction. These observations support a central role for platelet NF-κB during vascular injury, and suggest that dysregulation of NF-κB function may contribute to vascular disease processes.
Further investigation revealed that blood treated with NF-κB inhibitors coagulated normally, as evidenced by normal prothrombin, partial thromboplastin and euglobulin clot lysis times, indicating that inhibitor effects were platelet specific. Interestingly, Ono et al. recently demonstrated a fibrin-independent platelet contractile mechanism essential for the initial stages of hemostasis and in promoting thrombus contraction . It will be interesting to discover whether NF-κB plays a role in this regulation.
Of particular importance to our hypothesis was the finding that inhibition of IKK-β activity was attenuated by addition of active recombinant IKK-β protein. Our data show that the effects on platelet function following BAY inhibitor treatment were overcome by addition of active recombinant IKK-β protein. With that in mind, we also investigated a specific target of IKK-β phosphorylation, I-κB-α, and discovered that rhIKK-β addition increased I-κB-α phosphorylation levels. Moreover, addition of exogenous rhI-κB-α protein restored platelet spreading, supporting that BAY-11-7082 is targeting NF-κB driven mechanisms.
We very recently initiated studies to comprehensively determine where NF-κB proteins reside (e.g. granules, cytoskeleton, adhesion contacts, mitochondria) and to identify the types of NF-κB protein interactions in both untreated and BAY inhibitor-treated platelets. Our immunofluorescent pilot data indicated that both IKK-β and I-κB-α are centrally located in the unspread platelet. Interestingly, in spreading platelets, IKK-β remains centralized, while I-κB-α begins to appear throughout the lamellipodial region (data not shown).
Although the underlying mechanism(s) of NF-κB function in platelets require(s) further analysis, our overall findings provide a strong rationale that NF-κB proteins have an important role in human platelet physiology. The expanding versatility of platelet regulation leads us to envision possible mechanisms. Firstly, our data and others indicate that RelA and p50 retain the ability to bind DNA. Although platelets lack nuclear DNA, it is plausible that a novel mechanism of mitochondrial gene regulation may function in platelets. It has been reported that the NF-κB subunits RelA and p50, as well as I-κB-α play an important role in regulating mitochondrial mRNAs in other cell types[29, 30]. Secondly, it is conceivable that NF-κB proteins regulate post-transcriptional gene expression in platelets by directly binding RNA in a signal-dependent manner. In fact, we have demonstrated a novel function for glucocorticoid receptor whereby this transcription factor binds directly to monocyte chemoattractant protein (MCP) -1 mRNA, decreasing its stability in arterial smooth muscle cells. Our final point is that platelet Rel/NF-κB/I-κB proteins could elicit effects through direct protein/protein interactions. For example, PPARγ exerts some of its anti-inflammatory effects by directly binding nuclear-localized NF-κB complexes to facilitate nuclear exit and sequestration of NF-κB in the cytoplasm. Also consistent with our data is the fact that Bcl-3, an I-κB-α family member, specifically interacts with a Src-family tyrosine kinase, Fyn, and is involved in cytoskeletal regulation and inhibition of clot retraction in platelets. Interestingly, Fyn is an upstream regulator of the Rho family of GTPases and associates with αIIbβ3 to regulate platelet adhesion and spreading[34, 35].
Our data support the idea that transcription factors, such as NF-κB have important roles in platelet physiology. It is well-documented that dysregulation of NF-κB function contributes to the development of many human diseases, including chronic inflammatory diseases and cancer. Approaches to specifically inhibit the NF-κB pathway are under active development as therapeutic interventions. Thus, it is important to now recognize that platelets and megakaryocytes contain nearly all NF-κB family members, and that these cells could be affected by NF-κB inhibitory drugs. This new understanding of a NF-κB-platelet connection reveals a novel target, namely NF-κB, for dampening platelet activation. The impact of influencing NF-κB function in platelets and megakaryocytes could lead to a host of new therapeutics for platelets.
We are grateful for the following assistance: Kelly F. Gettings and the Blood Bank services; Jamie J. O'Brien for preparing human primary megakaryocytic cells; and Brian H. Smith, Aggregometry. We also thank the following individuals for their excellent technical services: Karen L. Bentley and Ian M. Spinelli, University of Rochester EM Core.
Grant support: This work was supported by RC1HL100051, HL095467, T32-DE07165, DE011390, HL078603, HL086367, ES01247, an EPA Center Grant (R827354) and NS054578.