Oxidized low density lipoprotein impairs endothelial progenitor cells by regulation of endothelial nitric oxide synthase

.

Postnatal neovascularization is a crucial compensatory response in chronically ischemic diseases. Increasing evidence indicates that postnatal neovascularization involves the incorporation of endothelial progenitor cells (EPCs) (1,2). Transplantation of EPCs into patients can induce blood flow recovery in ischemic limbs and improve left ventricular function after myocardial infarction (3,4). Therefore, this beneficial property of EPCs is appealing for cell-based therapeutic neovascularization in ischemic tissue. However, in some diseases, such as coronary artery disease and diabetes, the number of circulating EPCs is decreased and their function is impaired (5)(6)(7). Therefore, further understanding of the regulation of EPCs and the mechanism of their regulation may provide new insights into therapeutic neovascularization. However, the exact mechanism leading to a deficiency of EPCs remains unknown. The level of oxidized low density lipoprotein (OxLDL) increases in patients with coronary artery disease or diabetes and serves as an independent predictor for future cardiac events in these patients (8,9). Therefore, OxLDL may be one of the factors that influence the growth and bioactivity of EPCs. Wang and colleagues (10) recently investigated the effects of OxLDL on the number and activity of EPCs, but the underlying mechanism of action still remains largely unknown.
Recent studies have demonstrated that statins and vascular endothelial growth factor (VEGF) rapidly activate serine/threonine kinase Akt in EPCs and induce the mobilization of bone marrow-derived EPCs (11,12). Akt serves as a multifunctional regulator of cell biology (13). Phosphorylated Akt becomes available to phosphorylate its downstream substrates, such as endothelial nitric oxide synthase (eNOS), leading to eNOS activation and nitric oxide (NO) release (14). Furthermore, it has been suggested that eNOS is indispensable for the mobilization of stem/progenitor cells from bone marrow and serves as a significant regulator of hemangioblast activity (15,16). Thus, we hypothesize that the Akt/eNOS pathway is involved in the effects of OxLDL on EPC biology.
OxLDL, through its receptors, exerts its effects on the vessels (17). The classical scavenger receptors are absent or present in very small amounts in endothelial cells (18). In 1997, a lectin-like oxidized low density lipoprotein receptor (LOX-1) was identified in endothelial cells (19). Since then, LOX-1 has been shown to mediate many biological effects of OxLDL in endothelial cells (20). The inhibition of LOX-1 can abrogate the downregulation of phosphorylated Akt and eNOS protein expression (21,22). LOX-1 is expressed in EPCs and related to OxLDL-induced EPC senescence (23). It is of interest to know whether LOX-1 is involved in the effects of OxLDL on the survival and function of EPCs.

Preparation and oxidation of LDL
Blood samples for LDL preparation were taken from healthy volunteers after 12 h of fasting. Written informed consent was obtained from all volunteers in accord with the Institutional Ethics Committee. LDL was separated by density gradient ultracentrifugation as we described previously (24). The purity of LDL was assessed by agarose gel electrophoresis, and the protein content was determined by the modified Lowry method. LDL was dialyzed for 24 h against 0.01 mol/l PBS, pH 7.4, at 48C to remove EDTA and oxidized by exposure to CuSO 4 . EDTA was then added and LDL was dialyzed against PBS to stop oxidization. To assess the extent of LDL oxidation, the thiobarbituric acidreactive substance and electrophoretic mobility on agarose gels were determined. The level of thiobarbituric acid-reactive substance was 2.23 6 1.56 and 24.37 6 8.14 nmol/mg protein in native LDL and OxLDL, respectively. Compared with the native LDL, the OxLDL showed increased electrophoretic mobility on agarose gels of 1.8 6 0.4 times. OxLDL was sterilized by passing it through a 0.22 mm filter.

Cell isolation and culture
Human umbilical cord blood samples were collected in sterile blood packs containing anticoagulant. Written informed consent was obtained from all mothers in accord with the Institutional Ethics Committee. Mononuclear cells (MNCs) were isolated by Histopaque-density gradient centrifugation within 4 h of sample collection. CD34 1 cells were isolated from MNCs with the CD34 Isolation Kit (Miltenyi Biotec) according to the manufacturer's instructions. The viability and purity of CD34 1 cells were evaluated by trypan blue exclusion and flow cytometry. CD34 1 cells were depleted of adherent cells by incubation in fibronectin (Invitrogen)-coated culture flasks for 24 h. This process was repeated three times, with the purpose of removing CD34 1 mature endothelial cells, which shed off the vessel wall and adhered to the culture flasks. The nonadherent CD34 1 cells were reseeded in fibronectin-coated culture flasks and cultured in endothelial cell growth medium-2 (Clonetics) supplemented with 50 ng/ml VEGF (Pepro Tech), 100 ng/ml stem cell growth factor (Pepro Tech), and 20 ng/ml interleukin (kindly provided by Kirin Pharmaceuticals). After 7-10 days of culture, nonadherent cells were removed by washing with PBS, and adherent cells were continuously cultured in endothelial cell growth medium-2. Human umbilical vein endothelial cells (HUVECs) were isolated by in-cubating umbilical vein with 0.25% trypsin for 15 min at 378C and were cultured in M199 with 20% FBS. HUVECs in passages 1 and 2 were used.

EPC characterization
EPCs and HUVECs were subjected to flow cytometric analysis to examine the expression of CD31, VE-cadherin, vWF, CD34, CD133, CD106, CD45, and CD14. Cells were incubated for 30 min at 48C with monoclonal mouse anti-human antibodies against VE-cadherin and vWF (BD PharMingen). Cells were prepenetrated with 0.1% Triton X-100 for 15 min at 378C before being labeled with vWF antibody. Then, cells were incubated with FITCconjugated anti-mouse IgG. For additional analysis, cells were incubated for 30 min at 48C with phycoerythrin-conjugated antihuman CD133 (Miltenyi Biotec) and FITC-conjugated anti-human CD34, CD31, CD106, and CD14 antibodies (BD PharMingen). Isotype-matched antibodies served as controls. Cells were then fixed in 1% polyformaldehyde and quantitatively analyzed by flow cytometry.
To reconfirm EPCs, attached cells were assessed for the uptake of Dil Ac-LDL and immunostaining with CD31 or vWF (Dako Corp.) by confocal microscopy. 49,6-Diamino-phenylindole were used to dye the nuclei of adherent cells.

Treatment of EPCs
EPCs were treated without or with different concentrations of OxLDL (10, 25, and 50 mg/ml) for 48 h. Among these EPCs, two groups of EPCs were pretreated with 10 mg/ml lectin-like oxidized low density lipoprotein receptor monoclonal antibody (LOX-1 MAb; JTX92), a specific blocking antibody to LOX-1, or 100 mmol/l L-arginine (Sigma), a precursor of NO, before exposure to 50 mg/ml OxLDL. To confirm the role of Akt and eNOS in EPC biology, one group of EPCs was treated with 100 mmol/l Nnitro-L-arginine methyl ester (L-NAME; Sigma), an inhibitor of eNOS. Another group of EPCs was treated with 1 mmol/l triciribine, a specific inhibitor of the Akt signaling pathway.

Apoptosis assay
The proapoptotic potential of OxLDL on EPCs was assessed by annexin V/propidium iodide binding assay according to the manufacturer's instructions. To exclude dead cells, only annexin V-positive and propidium iodide-negative cells were counted.
For in situ detection of apoptotic cells, EPCs were fixed with 2% paraformaldehyde. Then, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay was applied according to the manufacturer's instructions (Roche Molecular Biochemicals). EPCs were counterstained with 49,6diamino-phenylindole (Sigma). TUNEL-positive cells were examined in a blinded manner, and the percentage was evaluated.

Adhesion assay
EPCs were labeled with Dil Ac-LDL as described above. After detachment and centrifugation, EPCs were resuspended in adhesion buffer (0.5% BSA in EBM-2), and identical numbers of cells were replated onto fibronection-coated 24-well culture plates, incubated for 30 min at 378C, and then washed three times carefully with adhesion buffer to remove nonadherent cells. The adherent cells were counted by fluorescence microscopy at 2003 magnification by independent blinded investigators. Six independent fields were assessed for each well, and the average number of adherent cells per 2003 field was determined.

Migration assay
The migration of EPCs was performed in a Transwell Chamber (Corning Costar). Briefly, EPCs were gently detached, harvested, and resuspended in chemotaxis buffer (EBM-2, 0.5% BSA). One hundred microliters of chemotaxis buffer containing 1 3 10 5 cells was added to the upper compartment, and 600 ml of chemotaxis buffer with 100 ng/ml VEGF was added to the lower compartment. After incubation at 378C for 4 h, the filters were removed and the cells in the lower compartment were counted using flow cytometry, with appropriate gating, for 20 s at a high flow rate. The migratory rate was expressed as the percentage of input cells migrating into the lower chamber (average events for 20 s/average input cell events for 20 s 3 100%). All groups of experiments were performed in triplicate.

Tube-formation assay
An ECM gel (Sigma) was thawed at 48C overnight and placed on a 96-well culture plate at 378C for 1 h to allow solidification. EPCs treated without or with OxLDL were harvested and replated (10,000 cells/well) on the top of the solidified ECM gel in EBM-2 medium supplemented with 0.5% BSA and VEGF (100 ng/ml). Cells were incubated at 378C for 12 h. Tube formation was defined as a structure exhibiting a length four times its width. The networks of tubes were photographed from six randomly chosen fields with a microscope. The total length of the tube structures in each photograph was measured using Adobe Photoshop software (25).

Determination of NO generation
NO is unstable, but it has the stable end products nitrite and nitrate. Hence, the best index of NO generation is the sum of nitrite and nitrate. The culture medium of EPCs was harvested and stored at 2808C until assay. The level of nitrite/nitrate was measured as described previously (26). Briefly, nitrate was converted to nitrite with nitrate reductase, and total nitrite was measured with the Griess reagent. The absorbance was determined at 540 nm with a spectrophotometer.

Semiquantitative RT-PCR
Total RNA was extracted from cultured EPCs using the RNA extraction kit (Qiagen). To generate cDNA, the DNase-treated RNA (100 ng) was reverse-transcribed with oligo(dT) and Moloney murine leukemia virus reverse transcriptase at 378C for 1 h for PCR amplification to estimate the expression of LOX-1 and eNOS. Two specific primers were used to identify and amplify LOX-1 (sense primer, 59-TTACTCTCCATGGTGGTGCC-39; antisense primer, 59-AGCTTCTTCTGCTTGTTGCC-39). The PCR product was 188 bp. For PCR, 32 cycles were used at 948C for 40 s, 558C for 1 min, and 728C for 1 min. Two specific primers were used to identify and amplify eNOS (sense primer, 59-ATGAAGCACCTGGAGAATGAG-39; antisense primer, 59-TCGGAGCCATACAGGATTG-39). The PCR product was 299 bp. For PCR, 30 cycles were used at 948C for 40 s, 508C for 1 min, and 728C for 1 min. The RT-PCR-amplified samples were visualized on 1% agarose gels using ethidium bromide. Each mRNA band was normalized with b-actin mRNA as an internal reference. The band of interest was analyzed by scan analysis software and expressed as the ratio to the b-actin mRNA band.

Western blot analysis
EPCs lysates were separated on SDS-PAGE gels and transferred to polyvinylidene difluoride membranes (Millipore). The membranes were blocked by incubation in Tris-buffered saline (10 mmol/l Tris, pH 7.5, and 100 mmol/l NaCl) containing 0.1% (v/v) Tween 20 and 5% (v/v) nonfat dry milk for 2 h, followed by a 2 h incubation at room temperature with rabbit polyclonal antiphospho-Akt-Ser473 antibody (Cell Signaling Technology), mouse monoclonal anti-eNOS antibodies (BD Transduction Laboratories), or goat anti-LOX-1 polyclonal antibody (Biovision). The membranes were washed extensively in Trisbuffered saline containing 0.1% (v/v) Tween 20 before incubation for 1 h with a horseradish peroxidase-conjugated secondary antibody (Amersham Life Sciences). The membranes were then washed and developed using ECL substrate (Amersham Life Sciences). The membranes were reprobed with an anti-actin antibody (Santa Cruz Biotechnology) and quantified by densitometry.

Statistical analysis
All data are expressed as means 6 SD. Differences between group means were determined by one-way ANOVA followed by a Newman-Keuls test. P , 0.05 was considered statistically significant.

Characteristics of EPCs developed from cord blood CD34 1 cells
After two rounds of immunomagnetic sorting, the purity of enriched CD34 1 cells was .90%. When CD34 1 cells were cultured on fibronectin, a fraction of cells adhered to fibronectin, whereas others still suspended in culture medium. A number of cell clusters appeared, and nu- merous spindle-shaped cells sprouted from the edges of the cell clusters after 4-7 days in culture. Cell clusters composed of round cells at the center and spindle-shaped cells at the periphery might resemble the structure of blood islands (Fig. 1A). Cord-like structures, commonly formed by endothelial cells, were observed after 10 days in culture (Fig. 1B). After 2 weeks of culture, spindle-shaped adherent cells reached confluence (Fig. 1C). At this time, flow cytometric analysis disclosed that EPCs expressed endothelial markers, including CD31, vWF, VE-cadherin, and CD34 but not CD106, CD45, and CD14. Stem cell marker CD133 was still expressed (Fig. 1D). However, HUVECs only express the marker of endothelial cells (Fig. 1E).

Regulation of Akt phosphorylation by OxLDL
To test our hypothesis that OxLDL influences eNOS activity through the Akt/eNOS pathway, immunoblotting was performed with a specific phosphorylated Akt antibody at the serine 473 phosphorylation site. OxLDL induced a dose-dependent decrease of Akt phosphorylation, with a 56 6 4% reduction at 50 mg/ml. Triciribine also decreased Akt phosphorylation, with a 64 6 3% reduction. However, pretreatment with LOX-1 MAb inhibited the OxLDLinduced decrease of Akt phosphorylation (Fig. 6A, B).

Regulation of eNOS and LOX-1 protein expression by OxLDL
To determine whether OxLDL influences eNOS protein expression, eNOS protein expression was assessed by immunoblotting. Incubation of EPCs with OxLDL dosedependently reduced eNOS protein expression, with a 58 6 7% reduction at 50 mg/ml. Pretreatment with LOX-1 MAb partially restored the downregulation of eNOS protein expression by OxLDL, with a reduction of only 7 6 4% (Fig. 7A, B).

DISCUSSION
Recent studies suggest that circulating EPCs serve as a biological marker for vascular function (27) and that reduced numbers of circulating EPCs predict future cardiovascular events (28). In this study, we found that treatment of EPCs with OxLDL resulted in significant apoptosis by annexin V/propidium iodide binding assay and in situ TUNEL assay. Because the level of OxLDL increases in patients with coronary artery disease or diabetes (8,9), our finding may explain why the number of circulating EPCs is decreased in these patients. Not only does OxLDL inhibit the differentiation of stem cell to EPCs, it also induces the apoptosis of differentiated EPCs. Imanishi et al. (29) reported that OxLDL at concentrations of <10 mg/ml inhibited EPC differentiation but did not induce a significant apoptosis of EPCs, which was in accordance with our results. Only at a dose of .10 mg/ml did OxLDL significantly induce EPC apoptosis. However, pretreatment of EPCs with L-arginine, a precursor of NO, markedly decreased the number of apoptotic cells in response to OxLDL. Furthermore, when NO production was inhibited by L-NAME, a known inhibitor of eNOS, EPC apoptosis was increased significantly. These results suggest that EPC survival depends on NO production and that the proapoptotic effect of OxLDL may be mediated in part by interfering with NO generation.
NO is a significant regulator of neovascularization. Mice deficient in eNOS display impaired ischemia-induced angiogenesis and reduced EPC mobilization (15,30). Estrogen was discovered to increase EPC number, which was attributable to its ability to induce NO (31). Our study provided new evidence for this point and demonstrated that in the presence of either OxLDL or the eNOS antagonist L-NAME, the adhesive, migratory, and tube-formation activities of EPCs were reduced. However, this reduction could be abrogated with L-arginine coculture. We further showed that OxLDL dose-dependently decreased Akt phosphorylation and eNOS protein as well as mRNA expression in EPCs. Phosphorylated Akt is available to phosphorylate eNOS, leading to eNOS activation (14). So the level of Akt phosphorylation partially reflects eNOS activity (32). This study showed that triciribine, a specific inhibitor of Akt, inhibited Akt phosphorylation and the function of EPCs. These data suggest that OxLDL not only inhibited eNOS activity through the Akt/eNOS pathway but also down- regulated eNOS expression, leading to a significant decrease in NO generation, which may promote EPC apoptosis and overall poor function.
In agreement with our study, a recent study demonstrated that in the presence of L-NAME, the tube formation of EPCs was reduced and that this reduction was abrogated with diethylenetriamine-NO, a donator of NO, coculture (33). Another study further demonstrated that asymmetric dimethylarginine, an endogenous inhibitor of eNOS, inhibited the mobilization, differentiation, and function of EPCs (34). However, one article showed that the capacity of circulating EPCs to migrate was not affected by eNOS inhibition with L-NAME (35). It is difficult to explain these discordant findings, because the culture conditions, including culture duration, culture medium, and concentrations of FBS and growth factors, were distinct and the characteristics of EPCs strongly depended on the culture conditions (36,37). Moreover, in our study, CD34 1 cells were used as the source of EPCs, which have been suggested to have higher NO-producing capacity than MNCderived EPCs (38).
We also revealed that the reduction of eNOS expression and activity by OxLDL was mediated by LOX-1. It has been shown that OxLDL exerts its biological effects via the activation of its receptors. LOX-1, found predominantly on endothelial cells, has a different biochemical structure from other receptors. Several investigators have demonstrated that LOX-1 and eNOS can regulate each other. For example, the inhibition of LOX-1 upregulates eNOS expression, and eNOS deficiency leads to the upregulation of LOX-1 expression (21,39). Almost all previous studies about LOX-1 were performed with mature endothelial cells. Imanishi et al. (23) showed that LOX-1 was expressed by EPCs and that its mRNA expression was upregulated by OxLDL in a concentration-dependent manner. We demonstrated that the protein expression was also upregulated by OxLDL in a concentration-dependent manner. Furthermore, the detrimental effects of OxLDL on EPC survival and activity, the downregulation of eNOS expression and activity, and the upregulation of LOX-1 expression were all inhibited by pretreatment with specific blocking antibody to LOX-1. Hence, the detrimental effects of OxLDL on EPCs and eNOS were mediated by LOX-1.
Wang and colleagues (10) have described the effects of OxLDL on EPCs derived from MNCs but not from purified CD34 1 cells; in addition, they did not investigate the mechanism underlying the effects of OxLDL on EPCs, which was the emphasis of our study. In our study, CD34 1 cells were used as the source of EPCs, rather than MNCs. This is because MNC-derived EPCs expressed monocyte/ macrophage markers but only a few expressed the endothelium-specific marker CD144 or the stem/progenitor marker CD133 (40). However, CD34 1 cell-derived EPCs displayed endothelial features while keeping progenitor markers.
In summary, OxLDL inhibits EPC survival and function, which occurs in part by attenuating EPC eNOS expression and activity. Accumulating evidence indicates that the impairment of EPCs may contribute to the progression of atherosclerosis (41). Therefore, our study may bring new evidence for the noxious effects of OxLDL in the pathogenesis of atherosclerosis. Moreover, this may be of clinical relevance for patients with cardiovascular disease and those who undergo neovascularization procedures. Gene and cytokine therapy with VEGF in such patients may not produce a maximal angiogenic response because eNOS, the downstream signaling of VEGF, is inhibited. Therefore, it is conceivable that supplemental therapy with downstream targets of VEGF, such as NO, may be more effective at inducing neovascularization. Because the adverse effects of OxLDL could be restored by LOX-1 MAb, the antagonist of LOX-1 may be a promising agent to promote neovascularization in patients with cardiovascular disease.