Background
The healthy, non-thrombogenic endothelium of the vasculature does not attract nor bind circulating platelets [
1‐
3]. However, following its exposure to proinflammatory cytokines, the non-thrombogenic endothelium becomes activated and converts into a prothrombotic endothelium [
3], resulting in a procoagulant state associated with a predisposition to the adhesion of platelets, atherosclerosis and thrombosis. The adhesion of platelets to the activated endothelium was shown to occur in areas highly prone to atherosclerotic plaque development prior to the detection of lesions, and prior to the infiltration and adhesion of monocytes or leukocytes [
2,
3]. A critical molecule shown to be involved in the process of platelet adhesion to the activated endothelium is the F11R protein, first described by Kornecki et al in 1990 [
4]. F11R is the symbol approved by the Human Gene Nomenclature Committee for the F11 receptor protein (GenBank Accession # 207907; NBC #S56749). In 1995, the amino acid sequences of the N-terminus and internal domains of the platelet F11R molecule were detailed [
5]. A protein termed JAM, described in 1998 [
6] showed correspondingly-identical amino acid sequences to those of the F11R protein, and hence the alias of JAM-A is also provided here. Direct phosphorylation and dimerization of the F11R protein [
5,
7] were shown following the activation of human platelets by physiological agonists. The cloning of the human F11R gene revealed that this molecule is a cell adhesion molecule, member of the Ig superfamily [
8].
Studies of the adhesion of human platelets to cytokine-inflamed endothelial cells (ECs) [
9] determined that homophilic interactions between the F11R molecules expressed constitutively on the platelet surface and the F11R molecules expressed
de-novo on the luminal surface of ECs when stimulated by cytokines, exert over 50% of the adhesive force between these cells. This observation was evidenced by demonstrating the inhibition of the adhesion of platelets to cytokine-inflamed ECs by a recombinant, soluble form of the F11R protein, and by domain-specific F11R peptides with amino acid sequences stretching in the N-terminal region and the 1st Ig fold of the F11R molecule, respectively [
10]. Analysis of the F11R gene identified NF-κB binding sites in the promoter region [
11], indicating that cytokines, during processes of inflammation, can cause up-regulation of the F11R gene. Yet, both the biochemical and genetic evidence thus far only suggests the involvement of F11R in the adhesion of circulating platelets to the cytokine-inflamed endothelium. In this report we demonstrate directly, by utilizing small interfering F11R RNAs (siRNAs), that F11R plays a critical role in the adhesion of platelets to the inflamed endothelium, an important early step in atherogenesis.
Materials and methods
Human endothelial cells and proinflammatory cytokines
Human aortic endothelial cells (HAEC) and human umbilical vein endothelial cells (HUVEC) (frozen vials of 106 cells) were purchased from Cascade Biologics, Inc., Portland, OR, and grown in Medium 200 containing 1% or 2% fetal calf serum (FCS) (Cascade Biologics, Inc., Portland, OR). For the experiments detailed below, both HAEC and HUVEC at 2nd passage, were treated with purified human recombinant TNFα (100 units/ml) (R&D Systems, Inc., Minneapolis, MN) and/or IFNγ (200 units/ml) (Roche Diagnostics, Mannheim, Germany), maintained at 37°C for the indicated periods of time. In a series of dose-response experiments in which the concentrations of TNF-α and IFN-γ were varied, a concentration of 50 pM TNFα is equivalent to100 units/ml TNF-α, and a concentration of 5.8 nM IFNγ is equivalent to 200 units/ml IFNγ.
Quantification of F11R mRNA in HAEC and HUVEC by real-time PCR
HAEC and HUVEC endothelial cells were grown to confluence and treated with cytokines at various times and doses. The treated cells were washed with 1× PBS, lysed, the total RNA extracted utilizing RNeasy Mini Kit (Qiagen, Valencia, CA, USA), and analyzed by real-time PCR on three separate experiments conducted in triplicate. The levels of F11R mRNA were determined by use of an ABI Prism 7000HT Sequence Detection System (ABI; AppliedBiosystem, Foster City, CA). The F11R primers consisted of the forward primer - 740: CCG TCC TTG TAA CCC TGA TT, reverse primer - 818: CTC CTT CAC TTC GGG CAC TA and probe -788: TGG CCT CGG CTA TAG GCA AAC C. The GAPDH forward primer - 620: GGA CTC ATG ACC ACA GTC CA, reverse primer - 738: CCA GTA GAG GCA GGG ATG AT, and the probe - 675: ACG CCA CAG TTT CCC GGA GG. Thermal cycles consisted of: 1 cycle at 48°C for 30 min, 10 min at 95°C and 40 cycles for 15 sec at 95°C, 1 min at 60°C. The probes were dual-labeled with FAM-TAMRA, obtained from ABI. Each mRNA level was expressed as a ratio to GAPDH. The mRNA levels were calculated using a standard curve of RNA isolated from normal human kidney (Stratagene) for the time course and dose curve or QPCR Human Reference total RNA (Stratagene) utilizing the ABI Prism 7000 SDS Software (Applied Biosystems).
Statistical analysis for real-time PCR
The RNAs, derived from ECs grown and treated in tissue culture wells, were isolated individually. Real time PCR procedures were performed in triplicate and averaged for each sample in three separate experiments (n = 9). The data were analyzed by Student's t-test and by mixed linear model analysis using SPSS software. Differences were considered significant at P < 0.05.
Preparation of inhibitors of RNA synthesis, NF-κB and JAK protein kinase
Actinomycin D (Sigma, St. Louis, MO), a known inhibitor of RNA synthesis, was diluted in DMSO to a 500 μg/ml (100X) stock solution. Parthenolide (Sigma), an inhibitor of the nuclear factor kappa B, NF-kB signaling [
12], was diluted in chloroform to a 50 mM (1000X) stock solution. The inhibitor of Janus kinase, JAK protein kinase, the tyrosine kinase inhibitor tyrphostin AG490 [
13], (Sigma) was diluted in ethanol to a 5 mM (100X) stock solution. All stock solutions were diluted in culture media to 1X concentration prior to experimentation. HAEC and HUVEC were grown to confluence and then treated with either actomycin D, parthenolid, or AG490, added in culture media without growth factor supplements for 1 hr at 37°C. Proinflammatory cytokines, TNFα and/or IFNγ were then applied to the media and the ECs were further incubated at 37°C for up to 24 hrs.
Silencing of the F11R gene of HAEC and HUVEC endothelial cells: transfections with small interfering RNAs (siRNAs)
Transfections were performed using Oligofectamine (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Briefly, 9 × 104 HAEC and HUVEC cells were seeded onto 96 well plates in 200 M media supplemented with LSGS without antibiotics, and the transfections of ECs were carried-out with either the stealth F11R siRNA HSS121425 (5'GGGACUUCGGAGUAAGAAGGUGAUUU 3') (300 nM) or the control, non-targeting siRNA No. 2 (Dharmacon). Subsequently, the transfected ECs were incubated in 200 M media containing 1% FBS followed by the application of cytokines TNFα (100 units/ml) and/or IFNγ (200 units/ml) for various periods of time.
Analysis of F11R in HAEC and HUVEC lysates and cell culture media
Monolayers of arterial and venous endothelial cells (90 - 95% confluence) were collected and homogenized in lysis buffer containing 20 mM Tris, 50 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1% sodium deoxycholate, 1% Triton X-100, and 0.1% SDS, pH 7.4 supplemented with protease and phosphatase inhibitors (Sigma-Aldrich) for the preparation of total cell lysate material derived from human arterial and venous endothelial cells. Protein concentration was quantified by the bicinchoninic acid (BCA) assay. Procedures utilizing SDS-polyacrylamide gel electrophoresis (10%, PAGE) followed by immunoblotting were performed as described previously [
14].
Collection and analysis of F11R in the media from cultured endothelial cells
The media derived from the arterial and venous, cytokine-treated and nontreated endothelial cells were collected at the time of cell harvesting and concentrated 200X using the centrifugal filter Centricon YM-10. Identification of the F11R protein within the collected media involved the resolution of all proteins by SDS-PAGE (10%) followed by immunoblotting procedures utilizing anti-F11R antibody, as described previously [
10].
Quantitation of immunoblots
Quantitation of the immunoblots was performed using image J (NIH). Briefly, scanned images of immunoblots were opened in image J, the protein bands were selected using the freeform tool and measured for integrated density. The values were normalized to tubulin levels by dividing the integrated density of the specific band by the integrative density of the tubulin band. ANOVA statistical analysis was performed on the normalized values. All values are the average of three immunoblots ± SEM.
The adhesion of platelets to endothelial cells: labeling of human platelets by calcein
Platelet rich plasma (PRP) was prepared from 100 mL of citrated whole blood, by centrifugation at 200 × g for 20 min at 23°C. Calcein (2 μg/mL)(Invitrogen) [
15,
16] was added to the PRP, and the PRP was maintained at 30°C for 1 hr in the absence of light. Platelets were isolated from PRP, washed as detailed previously [
10] and resuspended at final concentrations ranging from 2.5 - 3.5 × 10
8/mL
Assays conducted for measuring the adhesion of platelets to endothelial cells were performed in the dark due to the sensitivity of the calcium probe calcein to light exposure. Initially, HAEC and HUVEC, plated in cell culture wells, were incubated with 1% FBS/BSA in 200 M media for 1 hr at 37°C to block nonspecific binding sites. Aliquots of freshly-prepared, calcein-labeled platelets (3.3 × 10
8/ml) were added to each of the cell-culture wells, and plates were incubated at 37°C for 1 hr. Paraformaldhyde (4%), pH 7.4, was added to each well and incubation continued at 23°C for 15 min. The addition of paraformaldehyde, before washings, did not affect the natural capacity of the platelets to adhere to endothelial cells. The plates were washed 3× with pre-warmed growth factor-free 200 M media. Then aliquots (100 μl) of pre-warmed PBS were added to wells, and wells were read using a Perkin Elmer plate reader Victor 3, 1420 multilabel counter with fluorescein filter, as detailed previously described [
9].
To improve normality of distribution, the dependent variable (number of platelets per endothelial cell) was transformed by dividing by 10, adding 1 and taking the natural log. A mixed linear model was constructed that introduced treatment, cell type and the state of platelet activation (nonactivated vs agonist-activated) (and their mutual interactions) as fixed factors, with plate as a random factor. Since the variance of the dependent variable differed substantially according to plate, treatment and platelet state, variances were estimated separately for each combination of these factors. Due to the unbalanced nature of the study design, Satterthwaite adjustments were applied to numerator degrees of freedom. To offset the issue of multiple testing, Tukey-adjustments were applied to p-values for pair-wise group comparisons. Analysis of model residuals was undertaken to check for model fit and outliers. SAS Release 9.3 (SAS Institute, Cary NC) PROC MIXED software was used. Four outlying observations were excluded from analysis. All of the fixed main effects and their interactions were statistically significant at the 0.001 level, with the exception of the cell type main effect (p = 0.783). Discrepancies of means among the 11 plates were significant (Z = 2.11, p = 0.017). The inter-assay coefficient of variance was 0.7 ± 0.3 (S.E). The intra-assay coefficient of variance for each condition on the same plate was lower [(range from 0.05 to 0.16 ± .02 (S.E.) (Z > 6.00, P < 0.0001)] than the inter-assay coefficient of variance.
Discussion
The results reported here provide direct evidence for the critical role of F11R in the initiation of atherogenesis. This study demonstrates that inhibition by specific siRNA of the
de-novo biosynthesis of F11R, induced in endothelial cells by inflammatory cytokines, significantly inhibits the adhesion of human platelets to inflamed endothelial cells, an adhesion that would lead to production of atherosclerotic plaques in non-denuded blood vessels [
3].
Under physiological conditions, the non-activated, healthy endothelium expresses low levels of F11R- mRNA and the F11R/JAM-A protein resides primarily within the endothelial tight junctions [
6]. Under these conditions, circulating human platelets that constitutively express the F11R protein on their cell surface
4 do NOT adhere to a non-inflamed endothelium [
3]. On the other hand, when endothelial cells are exposed to the proinflammatory cytokines TNFα and/or IFNγ, F11R- mRNA levels rise significantly, followed by increased
de-novo synthesis of the F11R-protein and the insertion of newly-synthesized F11R molecules into the luminal surface of the endothelium [
18]. The present study provides direct evidence for the progression of this chain of events by the use of two blockers of mRNA synthesis: Actinomycin, an overall inhibitor of RNA synthesis, and F11R-siRNA, a specific inhibitor of the synthesis of F11R-mRNA. Both of these inhibitors blocked the enhancement of expression of F11R-mRNA and of the synthesis of the F11R protein in cytokine-stimulated arterial and venous endothelial cells. Most importantly, the critical pathophysiological role of the F11R-protein in the formation of a thrombogenic surface was proven by demonstrating that the inhibition of the expression of F11R-mRNA and thus of the increase in F11R protein in cytokine-exposed endothelial cells prevents the adherence of human platelets to inflamed endothelial cells.
Ozaki et al. [
19], were the first to report the changes in the localization of JAM/F11R protein in human umbilical vein endothelial cells that were treated simultaneously with the cytokines TNFα and IFNγ. As this treatment caused a disappearance of JAM from intercellular junctions, but no change in the total level of the protein [
19], the authors concluded that the exposure of endothelial cells to cytokines causes a redistribution of this protein from intercellular junctions to the surface of the plasma membrane of the inflamed endothelium. Our present results demonstrate that such treatment of arterial and venous endothelial cells with the cytokines TNFα and IFNγ induces
de-novo biosynthesis of F11R-mRNA and of the F11R protein. Taken together, all the data indicate that the lack of change in overall levels observed in the redistribution of the F11R/JAM protein in inflamed EC involve the disappearance of F11R/JAM-A molecules of the intercellular junctions that are degraded and/or released to the circulation (as discussed below). These are replaced with newly synthesized molecules of F11R/JAM-A that are inserted into the luminal side of the plasma membrane, that then acquires a thrombogenic surface.
As reported here, the biochemical pathway leading to the upregulation of the F11R gene following exposure of endothelial cells to the cytokine TNFα involves the NF-κB signaling pathway. Parthenolide, an inhibitor of NF-κB, blocked the TNFα-induced expression of the F11R gene - results consistent with our findings of NF-κB binding-sites in the promoter region of the F11R gene [
11]. On the other hand, the upregulation of F11R mRNA by IFNγ was blocked solely by the antagonist AG-490, a JAK tyrosine kinase inhibitor, indicating the involvement of the JAK/STAT signaling pathway in the induction of F11R mRNA and the
de-novo expression of the F11R protein by IFNγ. As the analysis of F11R gene structure indicates the presence of two promoters with regulatory elements consisting of NF-κB, GATA, Inr, ets sequences, TATA, and several GC and CCAAT boxes [
11], thus it is the participation of these regulatory elements that may account for the effects of IFNγ on the induction of F11R mRNA and protein observed here.
An additional important result of the present report is that exposure of endothelial cells to the inflammatory cytokines TNFα and IFNγ results in the release of soluble F11R molecules (sF11R) into the extracellular medium. Thus, the release of F11R appears to be an integral part of the pathological process induced within the vasculature in response to inflammatory cytokines. The important clinical implications of this process were reported previously [
17,
20]. A significant increase in the level of sF11R was found in the serum of patients with coronary artery disease (CAD) associated with high risk of atherosclerosis and heart attack [
17]. Furthermore, in this study the levels of serum-sF11R correlated significantly with the clinical severity of the disease [
17]. In other clinical studies, Salifu et al. [
20] reported of significantly enhanced levels of sF11R in the plasma of renal disease patients prone to atherosclerosis, and Ong et al. [
21] have demonstrated enhanced levels of sF11R in the serum of hypertensive patients. An increase in the level of the cytokine TNFα was also determined in the circulation of CAD patients and hemodialysis patients [
17] and these levels correlated positively with the circulating levels of sF11R. We have proposed that increased levels of sF11R immunoreactivity in plasma or serum can serve as markers for the initiation and progression of atherosclerosis. Similar to the results observed with HAEC and HUVEC, recent studies [
22] have shown that the exposure of cultured primary or immortalized human brain microvascular ECs to proinflammatory cytokines resulted in a decrease of F11R immunostaining at the tight junctions. However, the serum levels of sF11R were NOT altered in patients with multiple sclerosis and ischemic stroke that have demonstrated an inflamed blood-brain barrier. Haarmann et al. [
22], suggest that ECs of the blood-brain barrier are not induced to release sF11R by inflammatory stimuli, and that this resistance serves as a unique protection of the CNS compartment.
Potential mechanisms by which inflammation may lead to the formation of F11R detected in the plasma or serum of cardiovascular patients may involve the shedding of endothelial cell membrane-microparticles, as-well-as the release of soluble fragments of F11R by the action of circulating extracellular proteases. The occurrence of both these types of events have been previously reported. In early studies reported in 1986, we have demonstrated that exposure of human platelets to granulocytic elastase (released during inflammation) results in the release of soluble fragments of the platelet fibrinogen receptor, α
2β
3 integrin, and consequently in the direct binding of fibrinogen and the aggregation of platelets by fibrinogen [
23]. Evidence for the potential involvement of the disintegrin- metalloproteases in the proteolytic cleavage of JAM-A was provided by Koenen et al. [
24], who detected a soluble form of the F11R/JAM molecule with molecular mass of 33kDa in the conditioned media of inflamed HUVEC
in culture, as well as
in-vivo in cytokine-treated mice [
24]. The generation of endothelial-membrane microparticles has been reported by Combes et al. [
25] and by VanWijka et al. [
26]. Thus, the shedding of F11R-containing microparticles from platelets and endothelial cell membranes, and the action of proteases degrading the protein in intercellular junctions of EC that disappear during inflammatory processes, and/or on the surface of the plasma membrane of platelets, may all represent alternate mechanisms operating during inflammatory processes that are responsible for the appearance of soluble and microparticle-bound F11R molecules in the plasma and serum of patients with cardiovascular diseases.
We previously have shown that significant levels of the F11R mRNA and protein are expressed in vessels of CAD patients exhibiting clinical symptoms of coronary artery disease associated with atherosclerotic plaques [
18]. The increased expression of F11R at sites of atherosclerotic lesions was shown by others to be highest in unstable atherosclerotic plaques [
27], thereby demonstrating the involvement of F11R in both atherogenesis and atherothrombosis.
We have previously identified three different types of cells present in the atherosclerotic plaque express high levels of F11R. These are platelets, endothelial cells and smooth muscle cells [
4,
28]. Accordingly, the pathophysiological functioning of the F11R protein was examined for each cell type, and demonstrated to involve platelet-endothelial cell adhesive interactions, platelet aggregation, and the migration and proliferation of cytokine-stimulated smooth muscle cells. Stellos et al. [
29] reported a role for the F11R in the repair of the injured, inflamed endothelium, by showing that JAM-A/F11R molecules expressed on endothelial progenitor cells are required for the re-endothelialization of the vasculature, yet another critical role for F11R. Our previous studies utilized two F11R peptide-antagonists to determine that F11R provides well over 50% of the adhesive force operating between platelets and inflamed EC [
9]. The involvement of JAM-A in neointima formation following wire-injury of carotid arteries was reported by Zernecke et al. [
30]. Interactions between activated platelets, through their release of the chemokine RANTES, and its deposition onto endothelial cells were shown to be dependent on JAM-A [
30]. The results of the present study obtained with an experimental approach that specifically silences the F11R gene, provide direct evidence for the critical role of F11R in the adhesion of platelets to the endothelium under inflammatory conditions, which is an early, initial stage of plaque formation in atherogenesis. Accordingly, we propose that specific antagonists of the pathological actions of F11R represent a new target for the development of novel drugs for the prevention and treatment of atherosclerosis, heart attacks, stroke, and other cardiovascular disorders triggered by inflammatory processes.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BMA: Participated in design of studies, carried out all experiments and was involved in the drafting of the manuscript. These studies constitute a partial requirement for the attainment of her PhD in the Department of Medicine and Cell Biology/Anatomy.
JDM: Has made significant contributions to the conception and interpretation of the data.
MOS: Has made significant contributions to this work, has participated in analysis and interpretation of data, has performed the statistical analysis and was involved in drafting of the manuscript.
EK: Has been involved in experimental design, data analysis, the writing of the manuscript. Was critically important for the intellectual content of this work, and has given final approval of the version to be published.
YHE: Has been involved in experimental design, data analysis, the writing of the manuscript, and critically important for intellectual content of this work.
AB: Has made significant contributions to the conception, design and supervision of all experiments, performed data analysis and interpretation of data, supervised and coordinated all studies, and drafted the manuscript. All of the authors have read and approved the manuscript.