KLF2-dependent, Shear Stress-induced Expression of CD59

Complement activation may predispose to vascular injury and atherogenesis. The atheroprotective actions of unidirectional laminar shear stress led us to explore its influence on endothelial cell expression of complement inhibitory proteins CD59 and decay-accelerating factor. Human umbilical vein and aortic endothelial cells were exposed to laminar shear stress (12 dynes/cm2) or disturbed flow (±5 dynes/cm2 at 1Hz) in a parallel plate flow chamber. Laminar shear induced a flow rate-dependent increase in steady-state CD59 mRNA, reaching 4-fold at 12 dynes/cm2. Following 24–48 h of laminar shear stress, cell surface expression of CD59 was up-regulated by 100%, whereas decay-accelerating factor expression was unchanged. The increase in CD59 following laminar shear was functionally significant, reducing C9 deposition and complement-mediated lysis of flow-conditioned endothelial cells by 50%. Although CD59 induction was independent of PI3-K, ERK1/2 and nitric oxide, an RNA interference approach demonstrated dependence upon an ERK5/KLF2 signaling pathway. In contrast to laminar shear stress, disturbed flow failed to induce endothelial cell CD59 protein expression. Likewise, CD59 expression on vascular endothelium was significantly higher in atheroresistant regions of the murine aorta exposed to unidirectional laminar shear stress, when compared with atheroprone areas exposed to disturbed flow. We propose that up-regulation of CD59 via ERK5/KLF2 activation leads to endothelial resistance to complement-mediated injury and protects from atherogenesis in regions of laminar shear stress.

The complement cascade provides an essential defense against bacterial infection and a bridge between innate and adaptive immunity (1). However, by the nature of its cytolytic activity, complement has the potential to inflict injury on bystander host tissues including vascular endothelium. C1q, C3a, C5a, and the C5b-9 membrane attack complex (MAC) 3 have the capacity to exert pro-inflammatory effects on vascular endothelial cells (EC) including induction of cellular adhesion molecules and cytokine secretion, increased leukocyte adhesion, and generation of a pro-thrombotic endothelial surface (2)(3)(4)(5). Mechanisms implicated in complement deposition on the surface of EC include activation of the classical pathway by immune complexes, anti-phospholipid, and anti-endothelial cell antibodies (6) and through recognition of apoptotic cell blebs by the globular head of C1q (7). Alternatively, activation of the lectin pathway may follow exposure to hypoxia reoxygenation (8).
Complement activation on the surface of human EC is regulated by membrane-bound inhibitory proteins: decay-accelerating factor (DAF) (CD55), membrane cofactor protein (MCP, CD46), and CD59. The genes encoding DAF and MCP are clustered on the long arm of chromosome 1, whereas that for CD59 is located on chromosome 11. These proteins use distinct mechanisms for complement regulation. DAF prevents the formation and accelerates the decay of C3 and C5 convertases (9), whereas MCP accelerates the degradation of C3b and C4b by Factor I (10). CD59 inhibits the terminal pathway of complement activation, preventing incorporation of C9 into the MAC (11,12).
Complement activation may be an early prelesional event in atherogenesis, as revealed by colocalization of C5b-9 with lipid deposits in the tunica intima prior to monocyte recruitment (13). Experimental models suggest that vascular endothelial injury is the earliest detectable event in atherogenesis. Apoptosis of EC occurs preferentially at bifurcations and curvatures, where denudation of vascular endothelium enhances the risk of plaque development and local thrombosis. Moreover, aging and exposure to oxidized low density lipoprotein or reactive oxygen species increases EC apoptosis, consistent with a role in the initiation of atherogenesis (14). In addition to local lipidrelated activation of complement, exposure of the underlying * This work was funded by Arthritis Research Campaign Fellowships KO566 extracellular matrix, resulting from apoptosis of EC, may further reinforce complement activation (13,15).
The susceptibility of branch points and curvatures to atherosclerosis is in large part related to exposure to disturbed blood flow (DF), with low shear reversing or oscillatory flow patterns. In contrast, the arterial tree exposed to unidirectional laminar shear stress (LSS) Ͼ10 dynes/cm 2 tends to be protected (16). This is reflected in the phenotype of EC exposed to LSS, typically characterized by enhanced NO biosynthesis, prolonged cell survival, and an anticoagulant, anti-adhesive cell surface (17)(18)(19). In contrast, endothelium exposed to DF exhibits reduced levels of eNOS, increased apoptosis, generation of reactive oxygen species, permeability to low density lipoprotein, and leukocyte adhesion (reviewed in Ref. 16).
The importance of unidirectional LSS for vascular endothelial cytoprotection, together with the observation that LSS protects against complement-induced EC activation and chemokine synthesis (20), led us to explore whether LSS may be protective against complement-mediated vascular injury through suppression of the C5b-9 membrane attack complex. We show for the first time that CD59, the predominant membrane-bound regulator of the MAC, is preferentially induced by unidirectional atheroprotective LSS, providing enhanced protection against complement activation. This response was independent of NO and dependent upon an ERK5, Kruppel-like factor 2 (KLF2) signaling pathway. In contrast, an atheropronedisturbed flow waveform failed to increase CD59 protein expression. These observations were confirmed in vivo, where CD59 expression was significantly higher on arterial EC in atherosclerosis-resistant areas of the murine aorta. Thus, up-regulation of CD59 may represent an important component of endothelial cytoprotection in regions of LSS, preserving vascular integrity and minimizing both complement activation and atherogenesis.

EXPERIMENTAL PROCEDURES
Materials-Anti-human CD59 mAb (IgG1) Bric 229 was purchased from the International Blood Group Reference Laboratory (Bristol, UK). The anti-murine CD59 mAb (MEL-4) was a gift from BP Morgan (University of Wales School of Medicine). Anti-DAF mAb 1H4 (IgG1) and anti-MCP mAb TRA-2-10 (IgG1) were gifts from D. Lublin and J. Atkinson, respectively (Washington University School of Medicine, St. Louis, MO). UO126 and N G -monomethyl-L-arginine (L-NMMA) were from BIOMOL (Plymouth Meeting, PA), and LY290042 was from Merck Biosciences (Nottingham, UK). Other products were from Sigma-Aldrich.
Endothelial Cell Exposure to Flow-Human umbilical vein ECs (HUVEC) and human aortic ECs (HAEC; purchased from Promocell, Heidelberg, Germany) were cultured as described (21). The use of human EC was approved by Hammersmith Hospitals Research Ethics Committee (reference number 06/Q0406/21). Confluent HUVEC or HAEC cultures were exposed to control static conditions, high shear unidirectional laminar flow (up to 20 dynes/cm 2 ), or disturbed flow with directional changes of flow at 1Hz (Ϯ 5 dynes/cm 2 ) for up to 48 h using a parallel plate flow chamber (Cytodyne, La Jolla, CA) as described previously (22,23). Cell viability was assessed by examination of EC monolayers using phase contrast microscopy, cell counting, and estimation of trypan blue exclusion.
RNAi Design and Transfection-The short interfering RNA (siRNA) duplexes were from Dharmacon Inc. (Lafayette, CO) or Ambion (Austin, TX). siRNA sequences for KLF2 were sense, 5Ј-GCCCUACCACUGCAACUGGUU-3Ј, and antisense, 5Ј-CCAGUUGCAGUGGUAGGGCUU-3Ј); siRNA sequences for ERK5 were sense, 5Ј-GGCUCGGCUUGGUUAAUUCtt-3Ј, and antisense, 5Ј-GAAUAAUCCAAGCCGAGCCtc-3Ј. Corresponding negative control sequences were commercially synthesized. For siRNA delivery, HUVECs were plated at 3 ϫ 10 5 cells/well on fibronectin-coated glass slides in endothelial cell basal medium (Cambrex BioScience, Wokingham, UK) to obtain 50% confluency. siRNA targeting KLF2, ERK5, or scrambled control siRNA (all at 10 -50 nM) was transfected into EC using Oligofectamine-based transfection in endothelial cell basal medium. EC were cultured for 48 h in endothelial cell basal medium and analyzed for target gene expression by qrt-PCR or immunoblotting. The specificity of siRNA targeting was confirmed using a second set of sequences.
Northern Blotting-RNA was extracted from HUVEC using the RNeasy kit (Qiagen). Total RNA was separated on a 1% agarose/formaldehyde gel, transferred overnight to Hybond-N nylon membranes (Amersham Biosciences), and analyzed by specific hybridization to a radiolabeled cDNA probe for human CD59 (gift from H. Waldmann, University of Oxford, Oxford, UK) as previously described (26). Integrated density values for each band were obtained with an Alpha Innotech ChemiImager 5500 Alpha Innotech (San Leandro, CA), normalized with respect to the 28 S band on ethidium bromide-stained rRNA loading patterns and expressed as percentage change from control.
Flow Cytometry-Flow cytometry was performed as previously described using a Beckman-Coulter flow cytometer (Luton, UK) (26). The results are expressed as the relative fluorescent intensity, representing the mean fluorescent intensity with test mAb divided by the mean fluorescent intensity using an isotype-matched irrelevant mAb.
Complement Deposition Assays-Cell surface C9 deposition and complement-mediated lysis was assessed by flow cytometry as described previously (24). EC were exposed to LSS at 12 dynes/cm 2 for 24 h, harvested, and suspended in veronal-buffered saline containing 0.1% gelatin. The noncomplement fixing, inhibitory CD59 mAb BRIC 229 (27) was used at 20 g/ml. For analysis of C9 deposition, EC were exposed to normal human serum (NHS) or heat-inactivated NHS for up to 3 h at 37°C. C9 binding was detected with mouse anti-human C5b-C9 Technoclone (Vienna, Austria) and fluorescein isothiocyanate rabbit anti-mouse Ig. Complement-mediated cell lysis was quantified by assessing the percentage of cells permeable to propidium iodide, by flow cytometry, following exposure to NHS or heat-inactivated NHS for 3 h.
Confocal Microscopy-Confocal immunostaining and microscopy was by modification of an established method (28). C57BL/6 mice (8 weeks) were killed by CO 2 inhalation, and the vasculature perfused-fixed with 1% paraformaldehyde in phosphate-buffered saline. Thoracic organs were removed en bloc and equilibrated in OCT for 18 h. The aortic arch was dissected, and a transverse slice was cut that encompassed the circumference of the aorta at the origin of the brachiocephalic trunk, the brachiocephalic trunk itself, and the proximal portions of its branches (right common carotid and right subclavian). The block was frozen in isopentane, and serial cryostat sections were cut and incubated with 10% normal goat serum (Dako, Ely, UK), followed by rat anti-mouse CD59 mAb MEL-4 for 30 min. The sections were washed and incubated with 1:200 goat antirat AlexaFluor 568 (Molecular Probes and Invitrogen). The sections used for quantitation were stained only for CD59. The adjacent sections were immunolabeled with anti-CD59 as before and then incubated with biotinylated Griffonia simplicifolia isolectin B4 (Vector Laboratories, Burlingame, CA), followed by streptavidin-AlexaFluor 488 (Invitrogen). Other serial sections were immunolabeled with anti-CD59, followed by AlexaFluor 488, and then incubated with Cy3-labeled mouse monoclonal anti-smooth muscle actin (clone ␣-1A4, Sigma). After the second label, the sections were washed in phosphate-buffered saline and incubated for 10 min with TOPRO-3 (Molecular Probes) and mounted in phosphatebuffered saline/glycerol.
Sections were examined with a Zeiss LSM 510 Meta inverted confocal microscope (Thornwood, NY). Scan and photomultiplier settings were set to optimize signal/noise ratio for each emission wavelength. Using these photomultiplier settings, there was no detectable cross-over between channels. However, to eliminate any possibility of data skew by signal contamination, quantitation was performed on CD59-only stained sections. Processing was with Zeiss LSM Image Browser and quantitation by export of the images into Image J. The regions of interest were selected, and the histogram function was used to calculate the distribution of pixel intensities on the red channel (corresponding to AlexaFluor 568).
Animals-C57BL/6 mice were purchased from Harlan Olac (Bicester, Oxford, UK) and housed under controlled climactic conditions in microisolator cages with autoclaved bedding. Irradiated food and drinking water were readily available. All of the animals were housed and studied according to UK Home Office guidelines. Sentinel mice were housed alongside test animals and regularly screened for a standard panel of murine pathogens.
Statistics-The data are expressed as the means of individual experiments Ϯ S.E. The data were grouped according to treatment and analyzed using GraphPad Prism software (San Diego, CA), and the analysis of variance with Bonferroni correction or an unpaired Students t test. Kolmogorov-Smirnov analysis was used to compare immunofluorescence intensity in different regions of the murine aorta using an Excel plug-in Stat-Plus Professional. The differences were considered significant at p values of Ͻ0.05.

CD59 Expression Is Induced by Unidirectional LSS-To
assess the influence of shear stress on the surface expression of the complement inhibitory proteins MCP, DAF, and CD59, HUVEC in a parallel plate flow chamber were exposed to unidirectional LSS (12 dynes/cm 2 ) for up to 48 h and analyzed by flow cytometry. No change in the cell surface expression of either MCP or DAF was seen (Fig. 1). In contrast, a significant increase in CD59 expression of up to 2.5-fold was seen following exposure to unidirectional LSS for 24 h, and this was sustained at 48 h (Fig. 1).
Northern analysis detected alternatively spliced variants of CD59 (as described previously (24)) and demonstrated that induction by LSS was associated with an increase in mRNA, first detectable 2 h after the onset of flow and continuing to rise over 16 h. This pattern was confirmed by qrt-PCR, which revealed a sustained increase of up to 6-fold, 24 -48 h postinitiation of LSS (Fig. 2). To determine whether LSS-induced up-regulation was dependent upon the magnitude of shear force, HUVEC were exposed to increasing LSS for 24 h, and changes in CD59 mRNA were analyzed by qrt-PCR. CD59 mRNA levels rose progressively as LSS increased to 12 dynes/ cm 2 (Fig. 2C).
Experiments performed on HAEC to represent EC derived from a vascular bed affected by atherosclerosis also showed a significant up-regulation of CD59 mRNA and cell surface pro- tein in response to LSS (Fig. 3). The extracellular matrix upon which EC are cultured may also influence their responsiveness to shear stress (29). To address this, HUVEC were cultured on fibronectin, gelatin, and collagen type I and exposed to LSS for 24 h. However, the nature of the underlying matrix did not alter the ability of LSS to induce CD59, with equivalent induction seen under all conditions (data not shown).
LSS-induced CD59 Is Cytoprotective-To address the functional significance of CD59 up-regulation, the effect of LSS on cell surface deposition of C9 was measured. HUVEC were cul-tured under static conditions or in the presence of LSS (12 dynes/cm 2 ) for 24 h prior to exposure to 20% NHS and flow cytometric analysis, using an antibody against a neo-epitope on C9 that is revealed upon C5b-9 complex formation. In addition, complement-mediated EC lysis was quantified using propidium iodide uptake (24). Exposure to NHS resulted in measurable C9 deposition on the surface of HUVEC cultured under static conditions. A significant reduction in C9 deposition was seen on EC exposed to LSS (p Ͻ 0.05), with levels comparable with that seen on EC exposed to serum that had been heatinactivated to prevent complement activation (Fig. 4A). Likewise, LSS conferred protection against complement-mediated lysis when compared with EC cultured under static conditions (p Ͻ 0.05) (Fig. 4B). The inhibitory, noncomplement fixing CD59 mAb BRIC 229 (27) was used to confirm the role of CD59 in protection against lysis. BRIC 229 reversed the protective effect of LSS, and its presence resulted in lysis of Ͼ75% of EC (Fig. 4B).
Mechanosensitive Signaling Pathways Regulating Expression of CD59-Activation of signaling pathways such as those regulated by PI3-K/Akt and ERK1/2 and enhanced eNOS-dependent NO biosynthesis may contribute to the vasculoprotective effects of LSS (30). Thus, we initially adopted a pharmacological approach to explore the role of these pathways in the induction of CD59 by LSS. HUVEC were pretreated with PI3-K antagonist LY290042, or MEK-1 inhibitor UO126, at concentrations we have shown to inhibit phosphorylation of Akt and ERK1/2, respectively (31,32). Inclusion of LY290042 had no significant inhibitory effect on the induction of CD59 mRNA in response to LSS (Fig. 5A). Although UO126 reduced expression, this did not reach significance (Fig. 5B). Likewise, the presence of nitricoxide synthase inhibitor L-NMMA, at concentrations capable of inhibiting the induction of CD59 by atorvastatin under hypoxic conditions (24), did not inhibit LSS-induced CD59 protein (Fig. 5C) or mRNA expression (not shown).
In contrast, inclusion of the protein synthesis inhibitor cycloheximide abrogated the LSS-mediated increase in CD59 mRNA (Fig. 5D), suggesting dependence upon de novo synthesis of an inducible intermediary protein in response to LSS. Furthermore, it has been reported that in addition to MEK-1, FIGURE 2. CD59 induction is dependent upon the magnitude of shear force. A and B, HUVEC were exposed to LSS for up to 48 h and CD59 mRNA quantified by Northern blotting (A, with fold change calculated by densitometric quantification of three separate experiments) and qrt-PCR (B). C, HUVEC were exposed to varying LSS (0 -12 dynes/cm 2 ) for 24 h, and CD59 mRNA was quantified by qrt-PCR. The data are expressed as the means Ϯ S.E. from two to five experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. UO126 may inhibit ERK5 (33); hence the decrease in flow-induced CD59 mRNA observed in the presence of UO126 (Fig.  5B) may in fact reflect a reduction in ERK5 activation. These data, combined with (i) the identification of ERK5 (BMK-1) as a shear-inducible cytoprotective member of the MAPK family (34), (ii) identification of the LSS-inducible transcription factor KLF2 as a downstream target of ERK5 (35,36), and (iii) microarray data suggesting overexpression of KLF2 may increase CD59 mRNA in EC (36), led us to investigate the role of ERK5/KLF2 further. The KLFs are a subclass of the zinc finger transcription factors, within which KLF2 has emerged as an important factor in the maintenance of endothelial homeo-stasis (37,38). We adopted an siRNA approach, which reduced the expression of ERK5 transcripts in EC by 80% (supplemental Fig. S1A). Further analysis demonstrated that the induction of both CD59 mRNA (supplemental Fig. S2A) and surface protein by LSS was significantly attenuated in HUVEC treated with ERK5 siRNA, when compared with scrambled siRNA controls (p Ͻ 0.01) (Fig. 6A).
LSS significantly induced expression of KLF2 mRNA, a response that was reversed by the presence of specific siRNA oligonucleotides targeting KLF2 (supplemental Fig. S1B). In addition, knockdown of ERK5 significantly inhibited induction of KLF2 by LSS (supplemental Fig.  S1C). Analysis of CD59 mRNA in HUVEC pretreated with KLF2-specific siRNA demonstrated that LSS induction of CD59 was inhibited, an effect that was not seen with scrambled control siRNA (supplemental Fig. S2B). This was further confirmed by flow cytometric analysis, which showed that cell surface expression of CD59 on EC treated with KLF2 siRNA and exposed to LSS was equivalent to that on static cells (Fig. 6B).
CD59 Expression Is Differentially Regulated by LSS and DF-Atheroprotected and atheroprone regions of the aorta are exposed to unidirectional LSS and a low velocity reversing flow pattern, respectively. To compare the effect of these different flow patterns on CD59 expression, EC were exposed to LSS (12 dynes/ cm 2 ), or an oscillatory flow pattern (Ϯ5 dynes/cm 2 at 1 Hz) to model DF. Changes in endothelial cell morphology and CD59 expression were compared with cells cultured under static conditions. Preliminary experiments performed to validate the model demonstrated characteristic morphological changes in response to LSS (Fig. 7A) and induction of intercellular adhesion molecule-1 by LSS and vascular cell adhesion molecule-1 by DF, respectively (22) (supplemental Fig. S3). Furthermore, whereas LSS induced KLF2 mRNA expression, no such response was seen in EC exposed to DF (supplemental Fig. S3C). In subsequent experiments, a 4.3-fold increase in CD59 mRNA above static levels was seen following 24 h of LSS. In contrast, CD59 mRNA induction was reduced to 2-fold in EC exposed to DF, significantly lower than LSS (p Ͻ 0.05) (not shown). Likewise, FIGURE 4. LSS enhances EC resistance to complement-mediated lysis. HUVEC were exposed to LSS (12 dynes/ cm 2 ) (gray bars) or cultured under static conditions for 24 h (black bars). EC were then left untreated (UT) or exposed to 20% NHS or heat-inactivated serum (HIHS) for up to 3 h. A, C9 deposition was measured by flow cytometry using an antibody against a neo-epitope on C9 that is revealed upon C5b-9 complex formation. B, percentage EC lysis was calculated as the number of propidium iodide positive cells expressed as a percentage of total cells. CD59 activity was inhibited by pretreatment with mAb Bric 229. The data are expressed as the mean relative fluorescent intensities Ϯ S.E. from three experiments. *, p Ͻ 0.05. LSS induced a significant increase in EC surface CD59 expression when compared with static cultured EC (p Ͻ 0.01), whereas no change was seen in response to DF (Fig. 7B).
CD59 Expression Is Differentially Expressed in the Murine Aorta-A murine model was used to explore further the in vitro observation of differential regulation of CD59 by unidirectional LSS and DF. Regions of murine aorta exposed to LSS and DF can be determined by geometry, with straight portions exposed to LSS and the inner curve of the aortic arch and the aortobrachiocephalic junction exposed to DF (39). Murine aortae were perfused-fixed, sectioned longitudinally, and analyzed by immunohistochemistry for CD59 expression. As seen in Fig. 8 (A-C), CD59 expression was greater on vascular endothelium in areas of the aorta and brachiocephalic artery (not shown) predicted to be exposed to atheroprotective LSS, when compared with the inner curve of the aorta and the aortobrachiocephalic branch point, exposed to DF. To confirm the presence of intact endothelium at the DF sites, sections were stained with G. simplicifolia isolectin B4 (Fig.  8C). Similar analysis demonstrated intact EC at LSS sites (not shown). Quantification of endothelial CD59 expression at multiple sites within LSS and DF exposed areas of the aorta confirmed a significant difference (Fig. 8D). These results demonstrate differential expression of CD59, the most potent regulator of the MAC, with specific reduction on endothelium in atheroprone sites of the aorta.

DISCUSSION
The data presented herein suggest that LSS enhances vascular endothelial resistance to complement-mediated injury through induction of CD59 expression. Vascular wall injury contributes to the initiation of atherogenesis, a complex multifactorial inflammatory disease process propagated by both local and systemic factors. Activation of the classical or alternative complement pathways (40 -42) may be involved in the pathogenesis of atherosclerosis from the prelesional stage (13), through early arterial wall lesions (43) and intermediate and advanced plaques (44), via generation of C3a, C5a, and the C5b-9 MAC. LSS is an essential component of vascular endothelial homeostasis, contributing to resistance against apoptosis and the maintenance of an anti-proliferative, anti-oxidant, anti-thrombotic, anti-adhesive endothelial barrier (reviewed in Ref. 45). Although LSS-dependent induction of eNOS and NO biosynthesis is an important regulator of many of these mechanisms (17), the cytoprotective effects of LSS remain to be fully elucidated.
CD59 is a glycosylphosphatidylinositol-anchored, 18 -25-kDa molecule, belonging to the Ly-6 superfamily of cell surface proteins (11,12). Through its ability to bind to the ␣-␥-subunit of C8, CD59 prevents the incorporation of C9 into C5b-9 (46). Thus, CD59, which is constitutively expressed on the vascular endothelial surface, is the predominant membrane-bound regulator of the MAC. In light of the role of complement activation in atherogenesis and of LSS in atheroprotection, we explored the effect of shear stress on endothelial cell CD59 expression. We have shown that 24 -48 h of atheroprotective LSS enhances CD59 protein expression in HUVEC and HAEC and that this response is attenuated in EC exposed to an oscillatory flow pattern (Ϯ5 dynes/cm 2 at 1 Hz), modeling disturbed flow characteristic of atheroprone regions of the vasculature. Comparison  of the level of CD59 expression at sites of the murine aorta exposed to LSS and DF supported the relevance of these observations to the in vivo situation. Moreover, the enhanced expression of CD59 in response to LSS resulted in increased endothelial resistance to C5b-9 deposition and complement-mediated injury.
Urbich et al. (20) demonstrated an increase in EC clusterin in response to LSS, and as we show for CD59, this response was independent of enhanced NO synthesis. Of note, they did not detect a change in CD59 protein expression, which may reflect differences in experimental conditions including the time courses used. Urbich analyzed CD59 protein expression following 18 h of LSS, whereas we first saw a significant increase in protein expression at 24 h. Clusterin, a multifunctional cytoprotective molecule, acts to inhibit complement-mediated cytolysis by binding to C5b-6 and preventing formation of the MAC. Thus, increased expression of both soluble (clusterin) and membrane-bound (CD59) inhibitors of the MAC may contribute significantly to the atheroprotective actions of LSS through reduction of complement-mediated EC activation and injury. Of these, CD59 is considered to be the most potent endothelial regulator of the MAC. In contrast, DF, in addition to reducing protection against complement, may in fact increase complement activation by enhancing alternative pathway activity through induction of endothelial properdin synthesis. EC synthesis of properdin, a stabilizer of C3 and C5 convertases, is a significant contributor to plasma levels and is markedly enhanced following exposure of EC to DF (47).
KLF2 is one of 17 KLFs, a subclass of the zinc finger transcription factors. KLF2 is flow-inducible and differentially expressed in areas of the aorta exposed to LSS and DF (48). KLF2 activity has emerged as an important regulator of endothelial cytoprotective genes including eNOS, thrombomodulin, and heme oxygenase-1, which exert anti-inflammatory, anti-thrombotic, and anti-oxidant effects (25, 36 -38). An ERK5/myocyte enhancing factor 2 pathway regulating KLF2 transcription has been identified (35,36). Thus, KLF2 represents an important transcriptional effector in the cytoprotective actions of LSS (38). We have now added induction of CD59 and protection against complement-mediated injury to the atheroprotective profile of shear stress, a response that mirrors the effect of LSS and DF on KLF2 expression. Using siRNA we have demonstrated that CD59 induction by LSS is dependent-upon ERK5 and KLF2 activity. In support of a role for KLF2, microarray analysis revealed a 2.65-fold increase in CD59 mRNA following adenoviral overexpression of KLF2 in HUVEC for 24 h (36). In contrast, lentiviral overexpression of KLF2 for 7 days failed to detect a change in CD59 mRNA while showing a 1.98-fold induction of DAF (38). Thus, further studies are required to delineate the regulation and outcomes of KLF2 pathway activity, including its relationship to upstream mechanotransducers and signaling mediators and with downstream target gene promoters. Analysis of the 5-kb region upstream of the CD59 gene transcription start site revealed potential KLF2 5Ј-CACCC-3Ј binding sites. Detailed examination of the CD59 promoter and definition of the precise relationship with KLF2 are beyond the scope of the current manuscript and will be addressed in future studies.
Complement activation is tightly controlled so as to avoid host injury. Although lysis of nucleated cells is rare, sublytic C5b-9 may induce tissue factor expression and release of soluble factors from EC and vascular smooth muscle cells including platelet-derived growth factor, interleukin-1, interleukin-6, and MCP-1 (49 -51). The resultant proliferation of EC and vascular smooth muscle cells combined with induction of cellular adhesion molecules, monocyte chemotaxis, EC apoptosis, and thrombosis may contribute to atherogenesis. Hence, maintenance of cell surface CD59 to limit C5b-9 deposition is essential. We propose that in areas of the vasculature where this is insufficient, such as sites exposed to DF, a threshold is exceeded whereby levels of C5b-9 deposition are reached that induce EC injury with pro-inflammatory, proatherogenic sequelae. A concept supported by our recent report of increased atherosclerosis in CD59/low density lipoprotein receptor-deficient mice. 4 Moreover, our observation that the induction of cell surface CD59 expression by LSS is significantly reduced by the immunosuppressive drug cyclosporine A (data not shown) suggests this may be a contributory factor in cyclosporine A-mediated vasculopathy.
A further indication of the importance of CD59 in vasculoprotection comes from the study of diabetes mellitus. EC protection mediated by CD59 may be compromised in diabetes mellitus by two mechanisms: glycation of CD59 leading to loss of function (52) and hyperglycemic-induced shedding of cell surface protein (53). Dysfunctional glycated CD59 is detectable in the urine of patients with diabetes FIGURE 8. Analysis of CD59 expression in the murine aorta. Murine aortae were analyzed by immunohistochemistry and laser-scanning confocal microscopy. Endothelial CD59 expression (red) indicated by arrows in regions of the aorta exposed to laminar flow (A) and low shear disturbed flow (B) at the inner arch of the aorta. C, inner arch of the aorta stained with G. simplicifolia (EC marker, red), CD59 (bright green), elastin (darker green), and Topro (nuclear dye, blue). D, image analysis quantification demonstrating maximal CD59 expression on EC in areas of the aorta exposed to LSS and reduced CD59 on EC at the inner curve of the aorta (predicted disturbed flow; n ϭ 3 mice). *, p Ͻ 0.001 (Kolmogorov-Smirnov analysis). Scale bars, 50 m. mellitus, and colocalizes with the increased MAC deposited on the vascular endothelium of target tissues (54). The consequent release of growth factors and the pro-inflammatory, pro-thrombotic sequelae in the vasculature may be a significant factor in the accelerated atherogenesis associated with diabetes mellitus.
In conclusion, our data reveal induction of the complementinhibitory protein CD59 to be a novel cytoprotective outcome of LSS-induced activation of the ERK5/KLF2 signaling pathway. The demonstration of differential regulation of CD59 by LSS and DF suggests that CD59 expression may be a contributory factor in the protection afforded by LSS against atherogenesis. Through the inhibition of the terminal MAC, CD59 has the potential to exert anti-inflammatory and vasculoprotective effects. Moreover, data demonstrating that HMG-CoA reductase antagonists activate KLF2 (25,55,56) and enhance CD59 expression (24) suggest that modulation of KLF2 represents an important component of the vasculoprotective profile of statins and emphasizes the therapeutic potential of KLF2-related signaling pathways.