|Home | About | Journals | Submit | Contact Us | Français|
Elevations in C-reactive protein (CRP) are associated with increased cardiovascular disease risk and endothelial dysfunction. CRP antagonizes endothelial NO synthase (eNOS) through processes mediated by the IgG receptor Fcγ receptor IIB (FcγRIIB), its immunoreceptor tyrosine-based inhibitory motif (ITIM), and SH2 domain-containing inositol 5’-phosphatase 1 (SHIP-1). In mice CRP actions on eNOS blunt carotid artery reendothelialization.
How CRP activates FcγRIIB in endothelium is not known. We determined the role of Fcγ receptor I (FcγRI) and the basis for coupling of FcγRI to FcγRIIB in endothelium.
In cultured endothelial cells, FcγRI blocking antibodies prevented CRP antagonism of eNOS, and CRP activated Src via FcγRI. CRP-induced increases in FcγRIIB ITIM phosphorylation and SHIP-1 activation were Src-dependent, and Src inhibition prevented eNOS antagonism by CRP. Similar processes mediated eNOS antagonism by aggregated IgG used to mimic immune complex. Carotid artery reendothelialization was evaluated in offspring from crosses of CRP transgenic mice (TG-CRP) with either mice lacking the γ subunit of FcγRI (FcRγ−/−) or FcγRIIB−/− mice. Whereas reendothelialization was impaired in TG-CRP versus wild-type, it was normal in both FcRγ−/−;TG-CRP and FcγRIIB−/−;TG-CRP mice.
CRP antagonism of eNOS is mediated by the coupling of FcγRI to FcγRIIB by Src kinase and resulting activation of SHIP-1, and consistent with this mechanism, both FcγRI and FcγRIIB are required for CRP to blunt endothelial repair in vivo. Similar mechanisms underlie eNOS antagonism by immune complex. FcγRI and FcγRIIB may be novel therapeutic targets for preventing endothelial dysfunction in inflammatory or immune complex-mediated conditions.
C-reactive protein (CRP) is an acute phase reactant and a predictor of increased risk for cardiovascular disorders including myocardial infarction, sudden cardiac death, stroke, peripheral vascular disease, and hypertension.1–3 Studies in cell culture and in mice indicate that CRP antagonizes endothelial nitric oxide synthase (eNOS), resulting in the inhibition of endothelial cell migration, the promotion of endothelial cell-monocyte adhesion, impaired vascular wound repair, impaired vasorelaxation, and hypertension.4–6
CRP and immune complexes bind to cell plasma membranes via IgG Fcγ receptors (FcγR).7–9 We previously demonstrated that not only CRP, but also immune complexes potently antagonize eNOS in cultured endothelial cells.4 FcγR are categorized into activation receptors such as FcγRI, which contain a cytoplasmic immunoreceptor tyrosine-based activation motif (ITAM) located in its γ-chain subunit (FcRγ), and inhibitory receptors such as FcγRIIB, which has an immunoreceptor tyrosine-based inhibitory motif (ITIM) in its cytoplasmic domain.9 We previously showed that CRP inhibition of eNOS requires FcγRIIB, the phosphorylation of its ITIM, and the resulting recruitment and activation of SH2 domain-containing inositol 5’-phosphatase 1(SHIP-1) via phosphorylation of Tyr1020. Activated SHIP-1 attenuates signaling downstream of PI3 kinase, blunting Akt activation and eNOS Ser1179 phosphorylation by Akt, and thereby decreasing eNOS enzyme activation.4, 10 However, how CRP activates FcγRIIB is not known.
To better understand the basis for CRP actions on endothelium, the present study was designed to delineate the mechanisms by which CRP activates FcγRIIB. In immune response cells such as lymphocytes, activating FcγR and inhibitory FcγR are frequently functionally coupled, with the actions of activating FcγR opposing inhibitory FcγR function in some instances, and the actions of activating FcγR being required to initiate inhibitory FcγR function in others.9, 11 Recognizing that FcγRI is coexpressed with FcγRIIB in endothelial cells,4, 12 we raised the hypothesis that FcγRI is required for FcγRIIB-mediated actions of CRP on endothelium. In addition to testing this hypothesis, studies were performed to determine how FcγRI are coupled to FcγRIIB in endothelium. We also investigated whether immune complex-induced eNOS antagonism entails similar FcγRI-FcγRIIB functional linkage. Furthermore, to determine if these processes are operative in vivo, studies of carotid artery reendothelialization were performed in mice derived from crosses of transgenic CRP mice (TG-CRP) with FcRγ chain-null or FcγRIIB-null mice.
Using previously-described approaches, primary bovine aortic endothelial cells (BAEC) were harvested, propagated and used within nine passages.10 In selected experiments BAEC were transfected with cDNA encoding hemagglutanin (HA)-tagged human FcγRIIB as described previously.10 In other experiments, primary aortic endothelial cells were harvested from FcγRI+/+ and FcγRI−/− C57BL/6 mice.13 Briefly, mouse thoracic aortae were removed, placed in DMEM with 20% fetal bovine serum (FBS) and 1000 U/ml heparin, washed with DMEM, and filled with DMEM containing 2 mg/ml collagenase type II (Worthington) for 45 min at 37 °C. Endothelial cells were then collected by flushing with 5 ml of DMEM with 20% FBS, pelleted, and plated onto a 35 mm dish coated with collagen type I. After a 2h incubation, cells were washed twice with phosphate-buffered saline and placed in DMEM with 20% FBS, penicillin-streptomycin, 2 mM L-glutamine, 1× non-essential amino acids, 1× sodium pyruvate, 25 mM HEPES, 100 ug/ml heparin, and 100 ug/ml endothelial cell growth supplement (ECGS, Sigma-Aldrich). Mouse primary endothelial cells were used within 3 passages.
eNOS activation was assessed in whole cells by measuring 14C-L-arginine conversion to 14C-L-citrulline during 15 min incubations as previously described.5 eNOS agonists employed were insulin (500nM) and acetylcholine (Ach, 10μM). In select experiments incubations were performed in the presence of human recombinant CRP (5 or 25μg/mL), heat aggregated IgG (100μg/mL), PP2 (10μM, Calbiochem, Gibbstown, NJ), piceatannol (10μM, Calbiochem), anti-FcγRI antibodies (10μg/mL, anti-CD64, clones 10.1 & H-250, Santa Cruz Biotechnology, Inc., Santa Cruz, CA.), or anti-FcRIII antibody (10μg/mL, anti-CD16, clone 3G8, Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Immunoblots were performed to assess Src kinase phosphorylation, FcγRIIB phosphorylation, and SHIP-1 phosphorylation. Antibodies used were anti–phospho-Src (Tyr416) (Cell Signaling Technology), anti-c-Src (Santa Cruz Biotechnology), anti-phospho-FcγRIIB (Tyr292) (Abcam), anti-FcγRIIB (Fitzgerald), anti-phospho-SHIP-1 (Tyr1020) (Cell Signaling Technology), and anti-SHIP-1 (Santa Cruz Biotechnology). Results shown were confirmed in 3 or more independent experiments.
The care and use of all study animals was approved by the Institutional Animal Care and Use Committee at the University of Texas Southwestern Medical Center. Mice null for the FcR γ-chain (FcRγ)−/− mice (C57BL/6, Taconic Farms, Hudson, NY) were crossed with CRP transgenic mice (TG-CRP, C57BL/6) expressing a transgene consisting of coding sequence for rabbit CRP linked to promoter and regulatory elements for the cytosolic form of rat PEPCK.5 The progeny were crossed and experiments were performed in littermates in the resulting four groups of offspring: FcRγ+/+, FcRγ+/+;TG-CRP, FcRγ−/−, and FcRγ−/−;TG-CRP. Using an identical strategy with mating between TG-CRP and FcγRIIB−/− mice, littermates in the following four groups were also generated: FcγRIIB+/+, FcγRIIB+/+;TG-CRP, FcγRIIB−/−, and FcγRIIB−/−;TG-CRP. Carotid artery reendothelialization following perivascular electric injury was studyed in male mice at 11–13 weeks of age as previously described.5, 14 Ninety-six hours post-injury, the animals were injected with Evans blue dye, arteries were harvested, and the area of denudation which incorporated the dye was quantified by image analysis (FluorChem SP, AlphaEase FC software, Alpha Innotech) performed by an investigator blinded to group assignment. In a subset of mice serum was obtained on the day of carotid artery denudation, and serum CRP levels were determined by ELISA.5, 15 CRP levels (mean±SEM, n=5–6) in non-transgenic mice were <1μg/mL. In FcRγ+/+;TG-CRP and FcRγ−/−;TG-CRP littermates they were 39±7 and 61±4 μg/mL, respectively. In FcγRIIB+/+;TG-CRP and FcγRIIB−/−;TG-CRP littermates they were 29±6 and 30±4 μg/mL, respectively.
To evaluate the potential role of FcγRI in endothelium in the actions of CRP, studies of eNOS activation by insulin were performed in BAEC (Fig. 1). Under control conditions, CRP (25ug/ml) caused an 82.7% decrease of eNOS activation by insulin, mirroring our previous findings.4 In contrast, in cells treated with either of two FcγRI blocking antibodies, 10.1 or H-250, CRP had no effect on eNOS activation. However, blocking antibody to FcγRIII did not alter CRP inhibition of eNOS activity. Thus, CRP antagonism of eNOS activation is FcγRI-dependent.
Knowing from previous work that CRP antagonism of eNOS is mediated by FcγRIIB4 and now having also implicated FcγRI, we determined how FcγRI are coupled to FcγRIIB and the ensuing signaling events prompted by CRP in endothelial cells. Since Src kinase or spleen tyrosine kinase (Syk) frequently links activating FcγR function to inhibiting FcγR function in immune response cells,9 we tested the requirement for these kinases in CRP inhibition of eNOS activity. In preliminary experiments we determined that eNOS activation by acetylcholine (Ach) in BAEC is Src- and Syk-independent (data not shown), making it feasible to employ Ach as an agonist in studies of Src or Syk in CRP antagonism of eNOS. Ach increased eNOS activity and this was fully inhibited by CRP as previously observed4 (Fig. 2A). However, CRP did not inhibit Ach-stimulated eNOS activity in the presence of the Src inhibitor PP2. In parallel experiments, the antagonism of Ach-induced eNOS activation by CRP was unaffected by the Syk inhibitor piceatannol (Fig. 2B), which effectively prevented Syk activation by heat-aggregated IgG in Raji cells (data not shown).
To further interrogate the potential role of Src in CRP action in endothelial cells, we evaluated whether CRP activates Src by immunoblot analysis of Src phosphorylation at Tyr416. Using HDL as a positive control for Src stimulation in endothelial cells,14, 16 we observed that within 10 min of CRP treatment (25 ug/ml) Src phosphorylation was increased (Fig. 3A). Heat-inactivated CRP (25μg/mL) did not stimulate Src phosphorylation (data not shown), indicating that the response requires the intact protein. Dose-response studies further revealed that Src activation occurs at CRP concentrations as low as 2μg/mL (Fig. 3B), which is well within the range for modest chronic elevations in CRP that are associated with increased cardiovascular disease risk in humans,17, 18 and which mirrors the dose-response for eNOS antagonism by CRP.4
The role of FcγRI in CRP activation of Src was then interrogated in studies employing primary mouse endothelial cells isolated from aortas of FcγRI+/+ and FcγRI−/− mice. Whereas HDL caused comparable activation of Src in FcγRI+/+ and FcγRI−/− endothelial cells, CRP activation of Src was observed in FcγRI+/+ endothelial cells but absent in FcγRI−/− endothelial cells (Fig. 3C).
To test the coupling of CRP-induced, FcγRI-mediated Src activation to FcγRIIB, BAEC were transfected with cDNA encoding human FcγRIIB and 48h later the phosphorylation of FcγRIIB Tyr292 in the ITIM domain in response to CRP was evaluated by immunoblot analysis. This approach was used because the abundance of endogenous FcγRIIB in BAEC was insufficient to detect changes in receptor phosphorylation (data not shown). CRP caused a robust increase in transfected FcγRIIB phosphorylation, and this was fully prevented by Src inhibition with PP2 (Fig. 4A). The requirement for Src in the activation of SHIP-1 was tested in non-transfected BAEC. CRP activated SHIP-1 phosphorylation, and this was prevented by Src inhibition with PP2 (Fig. 4B). These cumulative findings indicate that through FcγRI, CRP activates Src, leading to FcγRIIB ITIM phosphorylation and SHIP-1 recruitment and activation.
We previously demonstrated that similar to the pentraxins CRP and SAP, heat-aggregated IgG used to mimick immune complex causes eNOS antagonism.4 Although it has been shown that common Fc receptor-mediated processes underlie phagocytosis and cytokine secretion by leukocytes in response to pentraxins and immune complex,19 it is unknown whether CRP and immune complex induce similar FcγR-mediated events in endothelial cells. Similar to CRP (Fig. 2A), heat-aggregated IgG employed to mimic immune complex completely inhibited Ach-stimulated eNOS activity, and the suppression was fully reversed by PP2 antagonism of Src (Fig. 5A). We further tested the effect of aggregated IgG on the phosphorylation of the ITIM of FcγRIIB in endothelial cells, and found that similar to CRP (Fig. 4A), aggregated IgG stimulated FcγRIIB phosphorylation, and the phosphorylation was prevented by PP2 (Fig. 5B). Aggregated IgG also stimulated SHIP-1 phosphorylation in endothelial cells, and the activation of SHIP-1 was completely inhibited in the presence of PP2 (Fig. 5C). Thus, mirroring the events prompted by CRP, in endothelial cells aggregated IgG activates Src, leading to FcγRIIB activation and SHIP-1 activation.
To determine the role of FcγRI in CRP actions on endothelium in vivo, mice null for the FcRγ-chain (FcRγ−/−), which contains the ITAM domain of FcγRI, were crossed with TG-CRP mice and the resulting progeny were crossed. Carotid artery reendothelialization was then evaluated in the four following groups of littermates: FcRγ+/+, FcRγ+/+;TG-CRP, FcRγ−/−, and FcRγ−/−;TG-CRP. Compared with FcRγ+/+ controls, reendothelialization was impaired in FcRγ+/+;TG-CRP mice (Fig. 6A), as indicated by the larger area of remaining denudation that incorporated Evans blue dye in the latter group. Cumulative studies revealed that the area of remaining denudation was 4.0-fold greater in FcRγ+/+;TG-CRP versus FcRγ+/+ mice (Fig. 6B). In contrast, despite having elevated CRP, FcRγ−/−;TG-CRP had normal reendothelialization (Fig. 6A, B).
To determine if partnership between FcγRI and FcγRIIB is required for CRP actions on endothelium in vivo, additional reendothelialization experiments were performed in the following four groups of littermates: FcγRIIB+/+, FcγRIIB+/+;TG-CRP, FcγRIIB−/−, and FcγRIIB−/−;TG-CRP. Compared with FcγRIIB+/+ controls, reendothelialization was impaired in FcγRIIB+/+;TG-CRP mice (Fig. 7A,B), with the area of remaining denudation increased 3.4-fold in the latter group. In contrast, although CRP was similarly elevated in FcγRIIB−/−;TG-CRP mice, they had normal reendothelialization (Fig. 7A, B). These collective observations, which mirror the findings for Fc receptor coupling in the actions of CRP on cultured endothelial cells, indicate that both the γ–chain of FcγRI and FcγRIIB are required for CRP antagonism of reendothelialization in vivo.
CRP levels are strongly correlated with increased risk for cardiovascular disease and with endothelial dysfunction related to decreased NO bioavailability.1, 3, 20–22 We previously demonstrated that CRP potently antagonizes eNOS activation by diverse agonists at levels of CRP that have been associated with increased risk of cardiovascular disease,23 and that the inhibitory Fc receptor FcγRIIB is critically involved in this process.4 This represents a potentially major mechanism by which vascular NO bioavailability may be diminished under numerous disease conditions. Here we show that the activating Fc Receptor, FcγRI, and functional coupling of FcγRI to FcγRIIB are required for CRP inhibition of eNOS in endothelial cells, thereby identifying another key cell surface receptor that participates in the negative regulation of NO production by the endothelium. We further show that consistent with these observations in cell culture, both FcγRI and FcγRIIB are required for CRP to blunt endothelial repair in vivo, which is related to a diminution in bioavailable NO.5 Elucidation of these processes mediated by the FcγRI-FcγRIIB tandem provides further evidence that CRP is likely a causal factor in certain forms of cardiovascular disease, and not merely a marker of chronic, low-grade inflammation. In addition, since CRP impairs insulin signaling to eNOS in endothelial cells,10 which normally promotes blood flow that augments glucose disposal in skeletal muscle,24 these newly-identified mechanisms are also potentially relevant to the insulin resistance that accompanies chronic inflammatory conditions including obesity.
We further demonstrate that Src kinase is critically involved in the coupling of FcγRI to FcγRIIB in endothelial cells. In immune response cells in which Fc receptors serve their classical functions, Syk is the kinase that primarily links the activating phosphorylation of the ITAM within the γ subunit of FcγRI to the activating phosphorylation of the ITIM in FcγRIIB.25, 26 Interestingly, whereas we now show that in response to CRP, Src activation leads to eNOS antagonism, the stimulation of eNOS by numerous agonists including HDL, estrogen and shear stress also entails Src activation, which preceeds PI3 kinase and Akt activation that promotes eNOS Ser1177 phosphorylation.16, 27, 28 These disparate roles of Src may be explained by the participation of different Src family members. Alternatively, since Src activation entails poorly understood mechanisms that involve plasma membrane receptor clustering and Src interacts with a number of receptors,29 these disparate functions for Src may be explained by different responses by Src to FcγRI-FcγRIIB clustering versus eNOS activating receptor clustering, or differences in the mode of physical linkage between Src and Fc receptors versus eNOS activating receptors.30 Having demonstrated an inhibitory role for Src in the regulation of eNOS for the first time, additional studies are now warranted to distinguish between these possibilities.
Recognizing that the classical ligand for Fc receptors is IgG, we previously demonstrated that heat-aggregated IgG mimicking immune complex causes eNOS antagonism to a degree that is comparable to that attained by CRP.4 We now show that similar to CRP, the actions of aggregated IgG on eNOS are mediated by Src activation causing FcγRIIB ITIM phosphorylation and resulting activation of SHIP-1. Recognizing how eNOS contributes to insulin-mediated glucose disposal,24 these mechanisms may participate in the pathogenesis of not only the cardiovascular disease, but also the insulin resistance that frequently complicate immune complex-mediated diseases such as rheumatoid arthritis or systemic lupus erythematosis.31, 32
From a therapeutic perspective, the identification of FcγRI as the initiator of CRP action in endothelial cells provides a new cell surface target for potential interventions. Studies in humans and in mice indicate that attenuated function of FcγRIIB likely contributes to the pathogenesis of systemic lupus erythematosis.33 Therefore, seeking to normalize CRP-related endothelial dysfunction by decreasing FcγRIIB activity, which can be accomplished by selective antibody-mediated blockade,34 runs the risk of promoting autoimmune disease. Alternatively, strategies that selectively prevent FcγRI activation can now be considered to blunt CRP-induced endothelial dysfunction without shifting the balance of activating and inhibitory Fc receptor activity towards excessive immune function. Now having greater evidence of an important role for SHIP-1 in eNOS antagonism by both CRP and immune complex, the phosphatase may also warrant consideration as a therapeutic target in endothelial cells.
Although there is clear evidence of a causal role for CRP in hypertension and endothelial dysfunction in mice,6, 35, 36 CRP has not yet been implicated as a contributing factor in the pathogenesis of such disorders in humans. However, now knowing that FcγRI and FcγRIIB are both important participants in CRP signaling in endothelium, it becomes apparent that there are multiple modifiers of CRP actions of relevance to vascular health. There are numerous polymorphisms in Fc receptors and some have known impact on receptor function, and there is also great variability in Fc receptor gene copy number.37 Thus, as we continue to identify the participating signaling molecules, genetic and molecular modifiers of CRP action should be considered in our assessment of the role of CRP in cardiovascular disease pathogenesis and our use of CRP as a risk factor in patient populations.
CRP levels are strongly correlated with increased risk for cardiovascular disease and with endothelial dysfunction. Previous studies both in cell culture and in mouse models have shown that CRP inhibits eNOS activation, and that the inhibitory Fc receptor FcγRIIB mediates this process. Nevertheless, how CRP activates FcγRIIB in endothelium is not known. This study shows that the activating Fc receptor FcγRI mediates CRP actions by coupling to FcγRIIB through the activation of Src kinase, and that both receptors are required for CRP to blunt endothelial repair in vivo. We demonstrate that similar mechanisms underlie eNOS antagonism by immune complexes. Three aspects of our work are novel and significant. First, we identify FcγRI as a new cell surface receptor that participates in the negative regulation of endothelial function. Second, we demonstrate for the first time an inhibitory role for Src in eNOS regulation. Lastly, our results raise the novel possibility that the FcγRI-FcγRIIB tandem in the endothelium plays an important role in the cardiovascular complications of immune complex-mediated diseases such as rheumatoid arthritis and systemic lupus erythematosis.
The authors thank Christopher Longoria for his technical assistance.
Sources of Funding
The work was supported by NIH grant HL075473 (to P.W.S.), by the Children’s Medical Center Foundation (to P.W.S.), by the Crystal Charity Ball Center for Pediatric Critical Care Research (to P.W.S.) and the Lowe Foundation (to P.W.S.), and by the Physician-Scientist Training Program at UT Southwestern (to N.C.S.).
In August 2011, the average time from submission to first decision for all original research papers submitted to Circulation Research was 16 days.