Members of the fibroblast growth factor (FGF) family have been clinically applied to the treatment of ischemic diseases because of their strong angiogenic actions. Although tissue ischemia is predominantly caused by atherosclerosis, the roles of endothelial FGF receptors (FGF-Rs) in atherosclerosis remain obscure. We generated endothelial cell (EC)-targeted constitutively active FGF-R2-overexpressing mice, using the Tie2 promoter (Tie2-FGF-R2-Tg), and crossed them with apolipoprotein E (ApoE)-deficient mice (ApoE-KO) to generate Tie2-FGF-R2-Tg/ApoE-deficient mice (Tie2-FGF-R2-Tg/ApoE-KO). After being fed a Western diet for 8 wk, the Tie2-FGF-R2-Tg/ApoE-KO demonstrated 2.0-fold greater atherosclerotic lesion area on the luminal surfaces of the aortas than the ApoE-KO (P < 0.01). The level of p21Cip1 protein, a cell cycle inhibitor, in the FGF-R2-overexpressing EC was 2.5-fold greater than that in the wild-type (WT) EC at the baseline (P < 0.01). FGF-R2 overexpression in the EC resulted in increased expression of VCAM-1 and ICAM-1, acceleration of apoptosis, and decreased proliferative activity, all of which were normalized by small interfering RNA (siRNA)-mediated knockdown of p21Cip1 (75% reduction in protein level, P < 0.01). Furthermore, the expression of PDGF-B and Egr-1, a PDGF/p21Cip1-inducible transcription factor, in the aortic endothelium of Tie2-FGF-R2-Tg/ApoE-KO was significantly greater than that in ApoE-KO. The proliferation of vascular smooth muscle cells in the aortic media of Tie2-FGF-R2-Tg/ApoE-KO was 2.0-fold higher than that in ApoE-KO (P < 0.01). Thus our study reveals that endothelial FGF-R2 signaling aggravates atherosclerosis by promoting p21Cip1-mediated EC dysfunction and cautions against the use of FGF for therapeutic angiogenesis in the setting of atherosclerosis.
- fibroblast growth factor
- signal transduction
atherosclerosis is a macrophage/vascular smooth muscle cell (VSMC)-mediated inflammatory process, which is initiated by endothelial cell (EC) dysfunction. In physiological conditions, EC produce nitric oxide to exert prosurvival and anti-inflammatory effects and prevent atherosclerosis, while various cardiovascular stresses, such as hypercholesterolemia, cause EC dysfunction, leading to the promotion of synthesis of inflammatory cytokines and adhesion molecules, migration of monocytes and VSMC into the subendothelium, and eventually the development of atherosclerosis (13).
Members of the fibroblast growth factor (FGF) family (FGF1, FGF2, FGF4, and FGF5) have been shown to facilitate angiogenesis by stimulating EC proliferation and migration (27) and maintaining endothelial integrity (18); therefore, they have been investigated as potential treatment for ischemic disease in a variety of animal models and several clinical trials (12, 15). On the other hand, the effects of FGF on atherogenesis have not been clarified. FGF promote the proliferation and migration of VSMC (21). In human atheromatous plaques, FGF-1 and FGF-2 are produced by VSMC and macrophages (7). Atheromatous lesions in apolipoprotein E (ApoE)-deficient mice express many types of FGF receptor (FGF-R), and treatment with FGF-R inhibitor attenuates the proliferation of VSMC and atherosclerosis (20), suggesting that the action of FGF on VSMC promotes atherosclerosis, whereas the role of endothelial FGF-R signaling in atherosclerosis remains obscure.
The biological effects of FGF-2 are mediated by its binding to the high-affinity FGF-R1–4 of the tyrosine kinase family. Activated FGF-Rs phosphorylate various signaling molecules, which culminates in the activation of major signal transduction pathways, such as the mitogen-activated protein kinase (MAPK)-, phospholipase C/intracellular Ca2+ concentration ([Ca2+]i)/protein kinase C (PKC)-, and phosphatidylinositol 3-kinase (PI3-kinase)/Akt- and Src-associated pathways (4), leading to cell proliferation and differentiation.
We generated endothelium-targeted transgenic mice in which a constitutively active form of an FGF-R2 mutant is overexpressed, using the Tie2 promoter (Tie2-FGF-R2-Tg) (17). This FGF-R2 mutant was isolated from patients with Crouzon disease, who show cranial deformities (16). We previously reported (17) that FGF-R overexpression in EC reduced myocardial infarct size in association with the enhanced migration of EC/VSMC and subsequent mature vessel formation. We demonstrate here the role of endothelial FGF-R action in atherosclerosis.
MATERIALS AND METHODS
ApoE-deficient (ApoE-KO) mice (19) were purchased from Jackson Laboratory. Lysophosphatidylcholine (LPC) was purchased from Sigma-Aldrich (St. Louis, MO). VEGF was purchased from R&D Systems (Minneapolis, MN).
Generation of Tie2-FGF-R2-Tg and Tie2-FGF-R2-Tg/ApoE-deficient mice.
We previously generated endothelium-targeted transgenic mice that overexpress constitutively activated FGF-R2, using the Tie2 promoter (Tie2-FGF-R2-Tg) (17). Tie2-FGF-R2-Tg mice were crossed with ApoE-KO mice to generate Tie2-FGF-R2/ApoE-KO mice. Tie2-FGF-R2-Tg/ApoE-KO mice showed no major anatomic abnormalities, their survival rate after consuming a Western diet for 8 wk (12 wk old) did not differ from that of ApoE-KO mice (98 ± 2% vs. 97 ± 2%; n = 15), and body weight and mean blood pressure also did not differ between the two genotypes (Tie2-FGF-R2-Tg/ApoE-KO vs. ApoE-KO: 30.1 ± 0.6 vs. 28.2 ± 0.8 g and 94 ± 8 vs. 93 ± 3 mmHg; n = 15 each).
The levels of FGF-R2 and phosphorylated FGF-R2 were examined. Total lysates of the aorta or cultured aortic EC from 4-wk-old mice were subjected to Western blot analysis with anti-FGF-R2 antibody (C-17, Santa Cruz Biotechnology, Santa Cruz, CA) or immunoprecipitation followed by Western blot analysis with antibodies against phosphorylated tyrosine (4G10), human Myc tag, which does not cross-react to mouse Myc (4A6, Upstate Biotechnology), and FGF-R2. The levels of total and phosphorylated FGF-R2 in the aorta and cultured EC of the Tie2-FGF-R2/ApoE-KO mice were significantly higher than those of the ApoE-KO mice [total FGF-R2: 2.0-fold in aorta, 2.1-fold in EC (P < 0.01 each); phosphorylated FGF-R2: 2.6-fold in aorta, 2.7-fold in EC (P < 0.005 each)] (Supplemental Fig. S1A).1
ApoE-KO or Tie2-FGF-R-Tg/ApoE-KO mice (4 wk old, male, C57BL/6 strain) were fed a Western diet including 21% fat and 0.125% cholesterol (Oriental Yeast, Tokyo, Japan), as we described previously (26). Levels of cholesterol, triglyceride, and the major serum lipoprotein classes were measured by anion-exchange high-performance liquid chromatography with perchlorate ion-containing eluent. Their aortas were fixed in 4% paraformaldehyde, frozen, sectioned, and stained with Oil Red O for 20 min at room temperature or immunostained with antibody against monocytes/macrophages (clone MOMA-2, Serotec). All experimental protocols were approved by the Kyoto Prefectural University's Animal Committee.
Primary culture of aortic EC and isolation of mononuclear cells.
EC were isolated from the thoracic aorta with the explant technique and cultured in EBM-2 medium supplemented with 10% FBS and EGM-2 SingleQuots (Clonetics) (17). The cells were used at the second passage.
Mononuclear cells (MNC) were isolated from peripheral blood or bone marrow by Ficoll-Hypaque (Lymphoprep, Nycomed Pharma, Oslo, Norway) gradients.
Terminal deoxyribonucleotidyltransferase-mediated dUTP nick end labeling assay.
Apoptotic cells in the frozen aorta sections or cultured EC were identified by terminal deoxyribonucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) with the In Situ Cell Death Detection Kit (Chemicon). Slices of the aorta and the cultured EC were costained with propidium iodide (Sigma) and Hoechst 33342 (Dojin, Kumamoto, Japan), respectively. The percentage of TUNEL-positive cells was calculated in five random fields.
The mice were fed a Western diet for 14 days and injected with 5-bromo-2-deoxyuridine (BrdU) in Ringer solution (25 mg·kg−1·day−1 ip) for the last 5 days. The aorta was removed, and 5-μm paraffin-embedded sections were prepared. Three sections that were at least 50 μm apart were stained with horseradish peroxidase (HRP)-conjugated anti-BrdU antibody (Serotec). In an in vitro study, proliferating cells were detected with the Cell Proliferation Kit (Amersham), according to the manufacturer's protocol. Briefly, subconfluent EC were cultured with DMEM containing 0.5% FBS and VEGF (10 ng/ml) for 48 h with BrdU and were then immunostained with anti-BrdU antibody. The percentage of BrdU-incorporating cells was calculated in five random fields.
Immunostaining of aorta.
Frozen sections of aorta were fixed with 4% paraformaldehyde and incubated with antibodies against platelet-derived growth factor (PDGF) receptor-β, Egr-1, ICAM-1, and VCAM-1 (Santa Cruz Biotechnology) and subsequently with FITC-conjugated secondary antibody.
RNA isolation and real-time PCR.
Total RNA was isolated with an RNA extraction kit (Qiagen, Tokyo, Japan), and cDNAs were generated with the SuperScript III First-Strand Synthesis System (Invitrogen). Real-time PCR was performed in combination with SYBR Green dye (Roche). Molecular expression was analyzed with β-actin as an internal control. An initial denaturation step was performed for 10 min at 95°C, followed by 40 cycles of amplification (55°C for 16 s, 72°C for 16 s, 94°C for 20 s). The following primers were used: VCAM-1: sense 5′-caggctggagattgatctg-3′, antisense 5′-gagagatgtagagttgtagttc-3′; ICAM-1: sense 5′-gtgatgctcaggtatccatc-3′, antisense 5′-gtccactctcgagctcatc-3′; PDGF-B: sense 5′-gtcgagttggaaagctcatctc-3′, antisense 5′-gagatgagctttccaactcgac-3′; β-actin: sense 5′-gtggggcgccccaggcacca-3′, antisense 5′-ctccttattgtcacgcacgatttc-3′.
Western blot analysis.
The aortas and EC were lysed in buffer (in mM: 50 Tris·HCl, 150 NaCl, 2 EDTA, and 8 EGTA, with 1% Triton X-100, 1 μg/ml leupeptin, and 1 μg/ml aprotinin). The lysates were subjected to Western blot analysis with antibodies against p21Cip1, Egr-1, α-tubulin, and lamin A/C. The relative densities of the bands were analyzed by Scion Image. To isolate the nuclear fraction, cultured cells were scraped, suspended in buffer [in mM: 10 Tris (pH 7.4), 10 NaCl, 3 MgCl2, 0.5 DTT, and 0.5 PMSF, with 10% glycerol, 0.25% NP-40], and subjected to centrifugation.
Knockdown of p21Cip1 with small interfering RNA.
Small interfering RNA (siRNA) against mouse p21Cip1 (sc-29428) and nonsilencing RNA (sc-37007) (Santa Cruz Biotechnology) were transfected (final concentration of each 50 nM) into EC (5 × 105 cells, 75–85% confluent) with a Lipofectamine RNAi MAX kit (Invitrogen) according to the manufacturer's instructions. Forty-eight hours after transfection, the efficiency of knockdown was assessed by Western blot analysis.
Migration assay and tube formation assay.
Aortic EC were isolated from wild type (WT) and FGF-R-Tg mice. Assays were performed as we described previously (17).
Data analysis and statistics.
All values are expressed as means ± SE. Repeated-measures ANOVA was used to analyze the time course experiments. Scheffé's test was used for multiple comparisons. For comparisons between two groups, the two-sample t-test was performed. P values < 0.05 (2-tailed) were considered statistically significant.
Endothelial FGF receptor signaling accelerated atherosclerosis.
ApoE-KO and Tie2-FGF-R2-Tg/ApoE-KO mice were fed a Western diet for 8 wk. Levels of total cholesterol, LDL-C, HDL-C, and triglyceride in serum at 8 wk after Western diet did not differ between the genotypes (ApoE-KO vs. Tie2-FGF-R-Tg/ApoE-KO: 1,231 ± 132 vs. 1,197 ± 124 mg/dl, 260.4 ± 18.2 vs. 266.3 ± 12.6 mg/dl, 11.9 ± 0.8 vs. 13.8 ± 1.5 mg/dl, and 21.6 ± 3.1 vs. 22.8 ± 2 mg/dl, respectively; n = 15 each), whereas the atheromatous plaque area on the luminal surfaces of aorta and in the aortic sinuses of Tie2-FGF-R2-Tg/ApoE-KO mice were 2.0-fold (P < 0.01; Fig. 1A) and 1.6-fold (P < 0.01; Fig. 1B) greater than those of ApoE-KO mice. A subanalysis revealed that the atheromatous lesions of the ApoE-KO mice were mainly located in the aortic arch (Fig. 1A), whereas in the Tie2-FGF-R2-Tg/ApoE-KO mice, in addition to the aortic arch the atheromatous area was observed in the abdominal aorta (Fig. 1A). The area of MOMA-2-positive monocyte/macrophage infiltration in the aortic sinus of the Tie2-FGF-R2-Tg/ApoE-KO mice was also 79% greater than that in the ApoE-KO mice (P < 0.01; Fig. 1C). Meanwhile, no obvious atherosclerosis developed in the aortas of the FGF-R2-Tg or WT mice given the Western diet for 8 wk (data not shown).
Endothelial FGF-R2 signaling promotes synthesis of adhesion molecules.
The expression of adhesion molecules in the endothelium causes monocyte migration into the subendothelium, leading to the initiation of atherosclerosis (13). We studied whether endothelial FGF-R2 signaling affects the expression of adhesion molecules. Although VCAM-1 and ICAM-1 mRNA levels were increased in both genotypes at day 14, the increases in the Tie2-FGF-R2-Tg/ApoE-KO mice were significantly higher than those of the ApoE-KO mice [VCAM-1, 2.0-fold (P < 0.01); ICAM-1, 1.4-fold (P < 0.05)] (Fig. 2A), and both were found to colocalize with the CD31-positive endothelium by immunostaining (Fig. 2B). By day 14, serum LDL had already increased to its peak level in both genotypes (ApoE-KO vs. Tie2-FGF-R2-Tg/ApoE-KO: 256 ± 13 vs. 272 ± 18 mg/dl).
Consumption of the Western diet increased the serum concentration of oxidized low-density lipoprotein (OxLDL), which plays a crucial role in the development of atherosclerosis. LPC is an oxidized phospholipid and has the most potent proatherogenic activity of all OxLDL (10, 14). We therefore stimulated EC with LPC (30 μg/ml, 24 h) in vitro. VCAM-1 and ICAM-1 mRNA levels in FGF-R-overexpressing EC were 1.7-fold (P < 0.01) and 1.3-fold (P < 0.05) higher than those of WT EC, respectively (Fig. 2C).
Tie2 is expressed not only in endothelial cells but also in bone marrow (BM)-derived inflammatory cells (2), suggesting that FGF-R2 under Tie2 promoter is abundantly expressed in inflammatory cells, leading to atherosclerosis. We therefore examined the expression level of FGF-R2 in the inflammatory cells from the peripheral blood (PB) and the BM and in the atheromatous plaques in Tie2-FGF-R2-Tg mice. Tie2-(Myc-tagged)-FGF-R2 was detected in both PB- and BM-derived MNC (Supplemental Fig. S1B, left); however, these levels at day 14 after diet consumption represented just ∼25% and ∼32% of those in the cultured EC (P < 0.005 and P < 0.01, respectively) (Supplemental Fig. S1B, right), whereas the levels of total (Tie2-overexpressed plus endogenous) and phosphorylated FGF-R2 at day 14 did not increase in PB MNC in Tie2-FGF-R2-Tg mice and slightly (∼1.5-fold, P < 0.05 each) increased in BM MNC (Supplemental Fig. S1B, left). Furthermore, frozen sections of the atheromatous aorta at day 56 were coimmunostained with antibodies against Myc tag and CD31 (Supplemental Fig. S1C). The results showed that Tie2-FGF-R2 was expressed in the CD31-positive endothelium but not in the inflammatory cells in atheromatous plaques (Supplemental Fig. S1C). At day 14, no expression of CD11b mRNA, a myeloid cell marker, was observed in the aorta (Supplemental Fig. S1D), and MOMA-2-positive macrophages were also barely detectable by immunostaining (data not shown). Thus Tie2-FGF-R2 in circulating inflammatory cells did not play a major role in development of atherosclerosis in Tie2-FGF-R2-Tg/ApoE-KO mice.
We studied whether FGF-R-mediated angiogenic activity is altered in hypercholesterolemia-induced atheromatous lesion. LPC is a major oxidized phospholipid whose concentration is increased in high OxLDL levels and has strong inflammatory activities. We stimulated EC with LPC to study its effect on FGF-R-mediated angiogenic activities, such as EC migration and tube formation. VEGF-induced migration (Supplemental Fig. S2A) and tube formation (Supplemental Fig. S2B) in FGF-R2-overexpressing EC were 1.8-fold and 2.3-fold (P < 0.01 each) higher than those of WT EC, whereas treatment with LPC (10 μg/ml) markedly inhibited both activities. It has been consistently reported that LPC attenuates FGF-2-mediated EC migration through inhibition of ERK (22). Thus FGF-R-mediated angiogenic activities are attenuated in patients with hypercholesterolemia and in the atheromatous lesion.
p21Cip1-mediated growth inhibition, apoptosis, and synthesis of adhesion molecules in FGF-R-overexpressing EC.
p21Cip1, a cell cycle-dependent kinase inhibitor, has been shown to promote the expression of adhesion molecules in the aorta (11). Indeed, we found that the p21Cip1 protein level in the Tie2-FGF-R2 EC at the baseline was 2.5-fold higher than that in the WT EC (P < 0.01) (Fig. 3A), and, interestingly, siRNA-mediated knockdown of p21Cip1 in FGF-R2-overexpressing EC [81% reduction in mRNA levels and 71% reduction in protein levels of p21Cip1 compared with the control (scrambled) RNA-treated group; P < 0.005 each, as shown in Fig. 2D] attenuated LPC-induced synthesis of VCAM-1 and ICAM-1 mRNA and protein [VCAM-1 75% and 63% reduction (P < 0.005 each) and ICAM-1 40% and 45% reduction (P < 0.01 each), compared with control RNA-treated groups] (Fig. 2E).
The effect of an increased p21Cip1 level on VEGF-induced cell growth was evaluated by BrdU uptake. The number of BrdU-incorporating WT EC was increased to 11.2-fold higher than the baseline at 48 h after stimulation with VEGF (10 ng/ml), whereas in the FGF-R2-overexpressing cells, this increase was much lower (3.5-fold vs. baseline) (Fig. 3B), and knockdown of p21Cip1 increased the number of BrdU-positive cells to a level similar to that observed in the WT cells (P < 0.01 vs. control RNA-treated group) (Fig. 3B).
Furthermore, after consumption of a Western diet for 14 days, the number of TUNEL-positive cells in the endothelia of the thoracic aortas of the Tie2-FGF-R2-Tg/ApoE-KO mice was 88% greater than that in the ApoE-KO mice (P < 0.01; Fig. 3C). An in vitro study showed that LPC (30 μg/ml, 12 h)-induced apoptosis in FGF-R2-overexpressing EC was 75% higher than that in WT EC, and knockdown of p21Cip1 reduced LPC-mediated apoptosis to the level seen in WT EC (P < 0.01; Fig. 3D). Thus, although FGF-R2 is known to promote the cell growth and survival of EC, sustained activation of FGF-R2 had inhibitory effects.
Endothelial FGF-R signaling promoted PDGF synthesis and induced VSMC proliferation in aorta.
EC dysfunction causes the proliferation and migration of VSMC into the subendothelium (13). We evaluated VSMC growth by BrdU uptake. BrdU-positive VSMC were barely detectable in the aortas of ApoE-KO mice at the baseline, while in Tie2-FGF-R2-Tg/ApoE-KO mice 3.2% of VSMC were BrdU positive (P < 0.01). At day 14, there were more BrdU-positive VSMC in the aortas of Tie2-FGF-R2-Tg/ApoE-KO mice than in those of ApoE-KO mice (10.4% vs. 5.4%, P < 0.01; Fig. 4A).
PDGF has been shown to promote the growth of VSMC. Indeed, the PDGF-B mRNA level in the aortas of Tie2-FGF-R2-Tg/ApoE-KO mice was 2.1-fold higher at baseline and 1.8-fold higher at day 14 than those of ApoE-KO mice (P < 0.05 each) (Fig. 4B). Also, our in vitro study showed that the PDGF-B mRNA level in FGF-R2-overexpressing EC was increased by 88% at baseline and by 82% at 6 h after LPC stimulation compared with those of WT EC (P < 0.05 each) (Fig. 4B). The PDGF-R-β level in the aortic media of Tie2-FGF-R2-Tg/ApoE-KO mice was markedly higher than in that of ApoE-KO mice (Fig. 4C).
Endothelial FGF-R signaling increased synthesis and nuclear translocation of Egr-1, a PDGF/p21Cip1-inducible transcription factor.
Egr-1 is a PDGF/p21Cip1-inducible transcription factor (5, 3). We evaluated the expression level of Egr-1 in aorta by immunostaining. The Egr-1-positive area was increased in the aortic endothelium at day 14 in both genotypes; however, its level in Tie2-FGF-R2-Tg/ApoE-KO mice was 2.5-fold higher than in ApoE-KO mice (P < 0.01, n = 8; Fig. 4D), whereas Egr-1 was barely detectable in either genotype at baseline (data not shown). The Egr-1 protein level in FGF-R2-overexpressing EC in vitro was increased to 2.0-fold at 1 h after LPC stimulation, whereas it did not change in WT EC (P < 0.01) (Supplemental Fig. S3A). Furthermore, the LPC-induced nuclear translocation of Egr-1 in FGF-R2-overexpressing EC was found to be markedly higher than that observed in WT EC by immunostaining (15% vs. 37%, P < 0.01; Fig. 4E) and Western blotting (1.9-fold, P < 0.01; Supplemental Fig. S3B).
Because of its strong angiogenic effect, FGF-2 is used to treat ischemic diseases. Considering that tissue ischemia is mostly caused by atherosclerosis-induced vascular stenosis, the effect of endothelial FGF-R signaling on atherosclerosis needs to be clarified. The initial event in atherogenesis is EC dysfunction, which induces adhesion molecules and inflammatory cytokines/chemokines to promote the migration of inflammatory cells into the subendothelium. In this study, we clarified that endothelial FGF-R signaling promotes hypercholesterolemia-induced atherosclerosis. Further analysis revealed that endothelial FGF-R signaling increased the synthesis of p21Cip1 and PDGF. Subsequently, p21Cip1 facilitates the transcription of adhesion molecules in EC, whereas PDGF induces VSMC proliferation, leading to the promotion of atherosclerosis. Furthermore, Heeschen et al. (6) reported that plaque vascularization is associated with growth of atheromatous lesion in the aorta; thereby angiogenic action of FGF-R itself may promote atherogenesis. Thus endothelial FGF-R signaling aggravates atherosclerosis, and we therefore caution against the use of FGF to induce therapeutic angiogenesis in the setting of atherosclerosis.
The mechanism by which p21Cip1 promotes the transcription of inflammatory molecules has not been fully elucidated. Direct binding of p21Cip1 to transcriptional factors, such as E2F1, STAT3, and Myc, inhibits their transcriptional activities (1), whereas it has been reported that p21Cip1 binding to the transcriptional repression domain of p300 (transcription coactivator), CRD1, derepresses p300 activity (25), leading to p300-mediated activation of various transcription factors.
We found that activated endothelial FGF-R2 signaling induces growth inhibition and apoptosis through the overexpression of p21Cip1 (Fig. 3, B and D). However, contrary to our findings, FGFs have been shown to promote EC growth and survival (27). Tyrosine kinases, such as FGF-R, activate signal transduction for cell growth and survival, whereas since unlimited stimulation of cell growth gives rise to a risk of cancer, the activity of tyrosine kinase signaling is strictly regulated by several inhibitory systems. The induction of p21Cip1 counteracts the overactivation of FGF-R2 signaling. Constitutively activated FGF-R2 mutants in humans or mice have been consistently shown to promote Stat-1-mediated synthesis of p21Cip1, which causes growth inhibition in osteoblasts and chondrocytes, leading to skeletal deformities (23, 26).
A Western diet elevates the serum concentration of OxLDL, which have potent proatherogenic properties. Of the known lipid components of OxLDL, LPC is thought to be the main active component that causes OxLDL-mediated EC damage (9). During LDL oxidation, as much as 40% of the phosphatidylcholine included in an LDL molecule is converted to LPC. OxLDL with higher LPC contents cause greater impairment of endothelium-dependent aortic ring relaxation than LDL with lower LPC contents (14). We found that LPC-mediated synthesis of adhesion molecules and apoptosis is enhanced by the overexpression of FGF-R2 (Figs. 2C and 3D). LPC binds to G protein-coupled receptors containing G2A with high affinity and GPR4 with low affinity, which causes an increase in [Ca2+]i and PKC activation (9). Although cross talk of FGF-R with LPC receptor signaling remains obscure, given that FGF-R activates [Ca2+]i/PKC to promote cell migration (4), [Ca2+]i/PKC may be involved in cross talk during FGF-R- and LPC receptor-mediated intracellular signaling.
FGF-R overexpression in EC increased PDGF synthesis in the aorta (Fig. 4B), whereas VSMC growth was upregulated in the aortas of FGF-R-Tg/ApoE-KO mice fed a Western diet (Fig. 4A). In light of a report showing that hypercholesterolemia promoted PDGF synthesis in the endothelium and that blocking the PDGF/PDGF-R system attenuated atherosclerosis (8, 10, 24), enhanced release of PDGF in EC is likely to promote the hyperproliferation of VSMC in the aorta of FGF-R-overexpressing mice.
FGF-R-overexpressing EC showed increased expression of Egr-1 (Fig. 4, D and E). Egr-1, a zinc finger transcription factor, is poorly expressed in the normal artery wall but is induced in VSMC and EC in the fibrous caps of human obstructive carotid atherosclerosis lesions. Egr-1 works as a master regulator of the expression of the PDGF family, the FGF family, and endothelial adhesion molecules and is implicated in a myriad of cardiovascular conditions, such as lesion development, hypertrophy, ischemia, and angiogenesis (5). Egr-1-deficient mice showed attenuated PDGF synthesis in the aorta and decreased atherosclerosis (5), suggesting that Egr-1 plays a crucial role in the synthesis of PDGF associated with the development of atheromatous lesions in Tie2-FGF-R2-Tg mice. Furthermore, Egr-1 binds to the p21Cip1 promoter and directly activates its transcription (3), suggesting that Egr-1 synthesis is an upstream event of p21Cip1 induction in activated FGF-R signaling.
In conclusion, endothelium-targeted constitutively activated FGF-R signaling facilitated atherosclerosis. Endothelial FGF-R signaling promoted p21Cip1-mediated apoptosis, growth inhibition, and the synthesis of adhesion molecules in EC and also enhanced PDGF production and VSMC proliferation, thereby facilitating atherosclerosis. The FGF family has been extensively used for revascularization therapy in ischemic disease. Considering that ischemic disease involves atherosclerosis-based vascular damage, the possibility of the FGF family aggravating atherosclerosis should be taken into account and further studies of this matter are needed.
No conflicts of interest, financial or otherwise, are declared by the author(s).
↵1 Supplemental Material for this article is available online at the Journal website.
- Copyright © 2011 the American Physiological Society