Annexin Peptide Ac2-26 Suppresses TNFα-Induced Inflammatory Responses via Inhibition of Rac1-Dependent NADPH Oxidase in Human Endothelial Cells

The anti-inflammatory peptide annexin-1 binds to formyl peptide receptors (FPR) but little is known about its mechanism of action in the vasculature. Here we investigate the effect of annexin peptide Ac2-26 on NADPH oxidase activity induced by tumour necrosis factor alpha (TNFα) in human endothelial cells. Superoxide release and intracellular reactive oxygen species (ROS) production from NADPH oxidase was measured with lucigenin-enhanced chemiluminescence and 2′,7′-dichlorodihydrofluorescein diacetate, respectively. Expression of NADPH oxidase subunits and intracellular cell adhesion molecule (ICAM-1) and vascular cell adhesion molecule (VCAM-1) were determined by real-time PCR and Western blot analysis. Promoter activity of nuclear factor kappa B (NFκB) was measured by luciferase activity assay. TNFα stimulated NADPH-dependent superoxide release, total ROS formation and expression of ICAM-1and VCAM-1. Pre-treatment with N-terminal peptide of annexin-1 (Ac2-26, 0.5–1.5 µM) reduced all these effects, and the inhibition was blocked by the FPRL-1 antagonist WRW4. Furthermore, TNFα-induced NFκB promoter activity was attenuated by both Ac2-26 and NADPH oxidase inhibitor diphenyliodonium (DPI). Surprisingly, Nox4 gene expression was reduced by TNFα whilst expression of Nox2, p22phox and p67phox remained unchanged. Inhibition of NADPH oxidase activity by either dominant negative Rac1 (N17Rac1) or DPI significantly attenuated TNFα-induced ICAM-1and VCAM-1 expression. Ac2-26 failed to suppress further TNFα-induced expression of ICAM-1 and VCAM-1 in N17Rac1-transfected cells. Thus, Ac2-26 peptide inhibits TNFα-activated, Rac1-dependent NADPH oxidase derived ROS formation, attenuates NFκB pathways and ICAM-1 and VCAM-1 expression in endothelial cells. This suggests that Ac2-26 peptide blocks NADPH oxidase activity and has anti-inflammatory properties in the vasculature which contributes to modulate in reperfusion injury inflammation and vascular disease.


Introduction
Annexin-1 (also termed lipocortin-1) is the first member of the annexin superfamily which, in humans, consists of at least 12 members each of which has a unique N-terminal sequence [1]. The annexin-1 peptide and its N-terminal derivative Ac2-26 have been shown to have protective effects in both brain and cardiac tissue following ischemia/reperfusion injury [2,3,4]. These effects have been attributed to the anti-inflammatory actions of annexin-1 and Ac2-26.
Annexin-1 is found intracellularly in gelatinase granules of neutrophils [5] and in human serum, particularly in inflammatory conditions such as myocardial infarction [6] and colitis [7]. These findings, together with observations from in vitro and in vivo models of inflammation strongly suggest an anti-inflammatory role for annexin-1. When released, annexin-1 binds to its receptor to mediate cell detachment and inhibits leukocyte transmigration [5,8,9]. Data now suggest that these effects are mediated through the specific interaction of annexin-1 with the formyl peptide receptor (FPR) [10] and FPR like receptor (FPRL-1) [11].
During an inflammatory response, endothelial cells become activated, a process characterized by specific changes in endothelial phenotype including upregulation of cell surface adhesion molecules which enhance leukocyte adhesion and transmigration across the blood vessel wall (reviewed by [12]). There are a number of stimuli for endothelial activation including inflammatory cytokines (e.g. Interleukin-1b, tumour necrosis factor a; TNFa), ischemia-reperfusion and diabetes [13,14,15]. Reactive oxygen species (ROS) are known to be involved in many of the processes involved in endothelial activation including upregulation of adhesion molecules such as intracellular cell adhesion molecule (ICAM-1), vascular cell adhesion molecule (VCAM-1), monocyte chemoattractant protein-1 (MCP-1) and P-selectin [16,17,18]. For example, TNFa-induced intracellular ROS regulates downstream NFkB signaling pathways and the expression of ICAM-1, VCAM-1 and MCP-1 in endothelial cells [16,19,20].
The potential sources of ROS in endothelial cells are multiple including the mitochondrial electron transport chain, NADPH oxidase, xanthine oxidase, cytochrome P450 and uncoupled endothelial NO synthase (eNOS). Of these, NADPH oxidase is the only enzyme which is dedicated to ROS production [15,21]. There are seven isoforms of the catalytic subunit (Nox1 to Nox5, Duox1 and 2) of NADPH oxidase, of these Nox2, Nox4 and Nox5 isoforms are expressed in endothelial cells [22,23,24,25]. The proinflammatory cytokine TNFa has been shown to activate acutely NADPH oxidase assembly via a protein kinase C (PKC) dependent pathway [2,17]. Furthermore, TNFa-mediated inflammatory effects such as increased expression of cell adhesion molecules (ICAM-1 and VCAM-1) have also been attributed to the activation of NADPH oxidase [16,26]. Intervention of TNFainduced Rac1-depedendent NADPH oxidase assembly either by dominant negative Rac1 (competitive inhibitor of Rac1) or cells derived from p47phox knock out animal significantly reduced ICAM-1 and VCAM-1 expression in endothelial cells [2,17,19,27]. These findings suggest TNFa enhanced NADPH oxidase activity and expression of cell adhesion molecules in endothelial cells.
Whilst the anti-inflammatory and anti-migratory effect of annexin-1 on leukocytes is well established, little information is available regarding the biological effect of this peptide on endothelial cells during inflammation. In this study, we investigated the effects of the annexin-1 peptide Ac2-26 on TNFainduced NADPH oxidase-derived ROS production and its role in endothelial cells during inflammation. To the best of our knowledge this is the first study to show that annexin-1 peptide Ac2-26 inhibits NADPH oxidase activity and the inflammatory response in endothelial cells.

Cell Culture
Human dermal microvascular endothelial cells (HMECs) kindly provided by Prof. Thomas J Lawley from Centers for Disease Control and Prevention, Atlanta, GA, USA. Dermal microvascular endothelial cells initially obtained from human foreskin and immortalised with a PBR-322-based plasmid containing the coding region for the simian virus 40A gene product, large Tantigen [28], were cultured in EGM-MV Bulletkit (Lonza) containing 10% fetal bovine serum in endothelial basal medium using standard cell culture techniques.

Measurement of Superoxide
NADPH oxidase activity was assessed by measuring superoxide with lucigenin-enhanced chemiluminescence using a Topcount microplate scintillation counter (PerkinElmer, Waltham, Massa-chusetts, USA) running in single-photon-count mode, as described previously [25]. For HMECs, (10,000 cells/well) were cultured in white OptiPlates (PerkinElmer) for 24 hr. Media was replaced at this time and cells were treated with TNF-a (20 ng/ml) for various times. Inhibitors were added 30 min prior to application of TNFa. Before measurement of superoxide, cells were preincubated with diethyldithiocarbamic acid (3 mM) in Krebs-HEPES buffer (HBSS, in mM: NaCl 98.0, KCl 4.7, NaHCO 3 25.0, MgSO 4 1.2, KH 2 PO 4 1.2, CaCl 2 2.5, D-glucose 11.1 and Hepes-Na 20.0) for 45 min to inactivate endogenous superoxide dismutase. The endothelial NADPH oxidase was stimulated with 100 mM NADPH, and the chemiluminescence was detected with 5 mM lucigenin.

Measurement of Intracellular ROS
Total intracellular ROS measurement in endothelial cells was also performed using 29,79-dichlorodihydrofluorescein diacetate (DCFH 2 -DA; Invitrogen, Life Technologies, Victoria, Australia) fluorescence as previously reported [25]. HMECs cultured in 96 well (black) opti-plates (10,000 cells/well) were washed with HBSS prior to loading with DCFH 2 -DA (10 mM). Fluorescence was then measured with excitation and emission wavelengths of 480 nm and 520 nm respectively using a Polarstar microplate reader (BMG Labtech, Germany) at 37uC over a period of 1 hr.

MTS Assay
A tetrazolium-based proliferation assay was used to determine HMEC cell number following superoxide or ROS measurement. Cells were washed with PBS and incubated with 10 ml of CellTiter 96 TM AQ solution (Promega, New South Wales, Australia) and 40 ml of Krebs-HEPES buffer for 1 h at 37uC in a 5% CO 2 incubator. Absorbance at 490 nm was determined using a Polarstar plate reader after incubation. Luminescence count (Count per second; CPS) or fluorescence unit (Relative fluorescence unit; RFU) were normalized with MTS absorbance and value expressed as percentage control.

Transfection of Plasmids
HMECs (100, 000 cells/well) were seeded in six-well plate the day before transfection. Transfection was performed using Lipofectamine2000. Each well contained 500 ng of either green fluorescent protein (GFP) or N17Rac1 (kindly provided by Prof. Richard G Pestell Kimmel Cancer Center, Thomas Jefferson University Philadelphia, PA 19107) mixed with 1 mL of Lipofec-tamine2000 in the presence of 500 ml of Opti-MEM. Cells were transfected for 5 h and then incubated with complete medium for 48 h. Transfection efficiency was determined by counting GFP positive cells vs. total cells. We were able to achieve 70 to 80% GFP positive cells (data not shown). The transfected cells were then treated with TNFa in the absence or presence of Ac2-26 peptide or DPI for 6 h.

Gene Specific Primer PCR
Due to the inconsistency of Nox2 gene expression detection with Applied Biosystems TaqMan Assay on demand, we used gene specific amplification of cDNA for detection of Nox2 gene expression. cDNA was prepared from 200 ng of total RNA. cDNA was generated with the Thermoscript RT-PCR System (Invitrogen, Carlsbad, CA) using gene-specific priming with Nox2-For AGAGGGTTGGAGGTGGAGAATT (Accession No. NM_000397) and GAPDH-For GAAGGTGAAGGTCG-GAGTC (Accession No. NM_002046) at an annealing/extension temperature of 55uC. Human GAPDH was used as a housekeeping gene to normalise all samples. The real-time PCR reactions were performed in a 7300 real-time PCR system (Applied Biosystems) using SYBR Green-based real-time PCR assay with the SYBR Green PCR Master Mix (Applied Biosystems) and in-house designed primers against Nox2 (For AGAGGGTTGGAGGTG-GAGAATT and Rev, GCACAAGGAGCAGGACTAGATGA; Accession No. NM_000397) and GAPDH (For GAAGGT-GAAGGTCGGAGTC and Rev, GAAGATGGTGATGG-GATTTC; Accession No. NM_002046). The specificity of the products was demonstrated by melt curve analysis and gel electrophoresis.

NFkB Activity Assay
The NFkB activities in HMECs were investigated by luciferase activities assay. HMECs (100, 000 cells/well) were seeded in sixwell plates the day before transfection. Transfection was performed using Lipofectamine2000. Each well contained 200 ng of either empty pGL3 (Promega, New South Wales, Australia) or NFkB-driven luciferase (Stratagene, La Jolla, CA) pGL3 vector and 50 ng pRL-SV40 (Promega) mixed with 0.5 ml of Lipofectamine2000 in the presence of 250 ml of Opti-MEM. Cells were transfected for 5 h and then incubated with complete medium for 24 h. After 24 h cells were treated with TNFa in the absence or presence of Ac2-26 peptide or DPI for 24h. NFkBdriven luciferase activity in cells was determined using a Dual-Luciferase reporter assay system (Promega, New South Wales, Australia) and measured on a Polarstar microplate reader. Transfection efficiency was normalized with the Renilla luciferase containing plasmid pRL-SV40 according to manufacturer's instructions.

Data and Statistics
Data are presented as mean 6 standard error of the mean (SEM). The mean data were analyzed with one-way analysis of variance (one-way ANOVA) followed by Newman-Keuls post hoc or t-test (for multiple comparisons). A P value of less than 0.05 was regarded as statistically significant.

Effects of TNFa on Superoxide Generation, ICAM-1 and VCAM-1 Expression in HMECs
Treatment of HMECs with TNFa (20 ng/ml) stimulated NADPH-dependent lucigenin enhanced chemiluminescence ( Figure 1A). At 6 h, chemiluminescence was increased to 11464.7% of control (n = 6, P,0.05, one way ANOVA) and increased to 123.265.8% of control (n = 5, P,0.01, one way ANOVA) after 24 h. TNFa is a well known pro-inflammatory cytokine and we [29] and others [13,23], have previously shown that TNFa induces gene expression of adhesion molecule in endothelial cells. Similar to its stimulatory effect on NADPHdependent superoxide anion generation, TNFa induced both gene and protein expression of ICAM-1 ( Figure 1B and C) and VCAM-1 ( Figure 1D and E) expression at 6 h and 24 h.

The Stimulatory Effect of TNFa is NADPH Oxidase Dependent
To investigate which enzyme sources are involved in the TNFastimulated increase in superoxide production, cells were treated with pharmacological inhibitors of these enzymes 30 min prior to superoxide measurement (Figure 2A). Whilst a number of the inhibitors altered basal superoxide generation (i.e. superoxide generated in the absence of TNFa), inhibition of xanthine oxidase (allopurinol, 100 mM), cyclooxygenase (indomethacin, 3 mM), endothelial nitric oxide synthase (L-NAME, 100 mM) as well as mitochondrial sources (rotenone, 1 mM) failed to inhibit the TNFa-stimulated increase in superoxide generation. Only inhibition of NADPH oxidase with DPI (1 mM) abrogated the TNFa stimulated increase in superoxide generation, suggesting that NADPH oxidase is a major source of TNFa-stimulated superoxide release in these cells. We next explored whether interfering with the assembly of a functional NADPH oxidase affects TNFa mediated ICAM-1 and VCAM-1expression by transfecting HMECs cells with a plasmid carrying dominant negative Rac1 (N17Rac1). Endothelial cells transfected with N17Rac1 plasmid showed a marked increase in N17Rac1 protein compared to cells transfected with the control GFP plasmid ( Figure S3 D), indicating an efficient transfection of endothelial cells with N17Rac1. Interestingly, N17Rac1 transfected endothelial cells showed a marked reduction in the TNFa-induced upregulation of ICAM-1 ( Figure 2B and D) and VCAM-1 ( Figure 2C and D). Similarly, DPI (1 mM) significantly decreased the TNFa-induced upregulation of both ICAM-1( Figure 2E and G) and VCAM-1 ( Figure 2F and G), confirming that NADPH oxidase-derived ROS are involved in TNFa-stimulated expression of cell adhesion molecules.

Effects of Ac2-26 on Superoxide Generation and ICAM-1 and VCAM-1 Expression
We next examined the effect of the annexin peptide Ac2-26 on both basal and TNFa stimulated superoxide release, ICAM-1and VCAM-1 gene expression since Ac2-26 has previously been shown to block the response to TNFa. At concentrations below 1.5 mM basal superoxide release was unchanged ( Figure S1A). Pretreat-ment of cells with Ac2-26 30 min prior to TNFa stimulation (20 ng/ml) dose-dependently reduced superoxide generation ( Figure 3A). Ac2-26 (0.5 mM) significantly reduced TNFa-stimulated superoxide release from 12366% to 10363% of control (n = 5, P,0.05 vs TNFa alone). For further studies an Ac2-26 concentration of 0.5 mM was used since this reduced TNFastimulated superoxide release without affecting basal release. To validate the inhibitory effect of Ac2-26 on ROS generation, we also assayed total ROS production using DCFH 2 -DA fluorescence and found that Ac2-26 completely abolished the TNFa-induced intracellular ROS generation in endothelial cells ( Figure 3B). Moreover, TNFa-induced gene expression of ICAM-1 ( Figure 3C) and VCAM-1 ( Figure 3D) was also suppressed by Ac2-26, demonstrating its anti-inflammatory activity upon TNFa stimulation without exerting basal effects.

The Anti-inflammatory Effect of Ac2-26 is FPRL-1 Receptor Specific
To confirm that the inhibitory effect of Ac2-26 is specific, we used the FPRL-1 antagonist WRW4. HMECs were treated with WRW4 (10 mM) for 30 minutes prior to Ac2-26 and TNFa treatment. WRW4 did not alter basal or TNFa-stimulated superoxide production ( Figure 3E) but abrogated the Ac2-26 inhibition of TNFa-induced ROS production ( Figure 3E). This pattern was also observed for basal and TNFa-induced ICAM-1 gene upregulation ( Figure S2). These results indicate that Ac2-26 acts via FPRL-1 receptor to block the TNFa responses.

TNFa Induced ICAM-1 and VCAM-1 Expression Requires Rac-1 Dependent NADPH Oxidase Activation
Having shown that Ac2-26 actions are NADPH oxidase dependent, we examined whether this effect was due to alterations in NADPH oxidase subunit expression. We measured the gene expression of Nox2 and Nox4 catalytic subunits following 24 h TNFa stimulation in HMECs. Surprisingly TNFa (2-50 ng/ml) suppressed Nox4 gene expression ( Figure 4A), suggesting Nox4 was not involved in TNFamediated superoxide release. Co-treatment of Ac2-26 did not modify the inhibitory effect of TNFa on Nox4 gene expression. Neither did TNFa alone nor in combination with Ac2-26 affect the gene expression of Nox2 (Figures 4C and 4D).Similarly, TNFa alone or in combination with Ac2-26 did not affect gene expression of associated subunits: p22phox, p67phox ( Figures  S3A and S3B) and protein expression of Rac1 (Figures S3C and  S3D). This suggests that TNFa does not regulate the gene expression of Nox2 or other important subunits in endothelial cells.
Since we were unable to detect any change in gene expression (Nox2, p22phox and p67phox), we speculated that Ac2-26 might interfere with the assembly of NADPH oxidase thereby suppressing TNFa mediated ICAM-1 and VCAM-1 gene expression. HMECs were therefore transfected with plasmid carrying a dominant negative Rac1 (N17Rac1) before addition of either TNFa or Ac2-26. In cells transfected with a control GFP plasmid, TNFa stimulated expression of ICAM-1( Figure 4E and G) and VCAM-1 ( Figure 4F and G) was  Figure 4E and G) and VCAM-1 ( Figure 4F and G) was no longer observed in N17Rac1 transfected cells. These findings suggest that TNFa activates Rac1 dependent NADPH oxidase activity to increase expression of ICAM-1 and VCAM-1, and this enzyme assembly process is suppressed by Ac2-26. Furthermore, we found that pretreatment of Ac2-26 for 0.5 h also reduced the PMA stimulated superoxide generation in phagocytic cells (DMSO differentiated HL-60 cells [29] (Figure S4). Overall our findings suggest that Ac2-26 prevents TNFa mediated NADPH oxidase activation to suppress both ICAM-1 and VCAM-1 expression.

Ac2-26 Antagonised TNFa Induced NF-kB Promoter Activity
To further elucidate the effect of Ac2-26 on TNFa downstream signaling process, we transfected endothelial cells with human nuclear factor kappa B(NF-kB) promoter cloned in pGL3 luciferase reporter vector (pGK3/NF-kB) and tested the effects of TNFa alone and in combination with Ac2-26. As expected, TNFa significantly increased the NF-kB promoter activity ( Figure 5) and this was reduced to the same extent by Ac2-26 and DPI.

Discussion
Annexin 1 and its derived N-terminus peptide Ac2-26 are well known to exert anti-inflammatory activities in inflammatory cells such as neutrophils and macrophages [30,31,32]. Information about their actions in other cells that are important to inflammation particularly endothelial cells is limited [33]. The present study addressed this aspect in human endothelial cells, focusing on the effect of Ac2-26 on ROS generation, ICAM-1and VCAM-1 induction following stimulation with the proinflammatory cytokine TNFa. This is the first study to illustrate that Ac2-26 inhibits the activation of Rac1-depedent NADPH oxidase to suppress stimulated superoxide generation in endothelial cells, leading to downregulation of both ICAM-1 and VCAM-1 expression.
We and others have previously shown that TNFa-mediated responses such as superoxide generation and adhesion molecule expression are NADPH oxidase-dependent in human endothelial cells [2,16,17,29]. We confirmed this by demonstrating that TNFa mediated superoxide release was sensitive to DPI and the induced ICAM-1and VCAM-1 gene upregulation were blocked by inhibition of NADPH oxidase with the use of dominant negative Rac1 (N17Rac1) and DPI. Several studies have shown that Rac1 is required for the assembly of an active NADPH oxidase in human endothelial cells [16,17]. Chen et al. [16,19] recently reported that Rac1 and an NADPH oxidase-dependent pathway was responsible for the TNFa-mediated responses including increases in ICAM-1, VCAM-1 and MCP-1 expression. Furthermore, Li et al. showed that TNFa binds to its adaptor proteins, TNF receptor associated factors (TRAF4), to promote the membrane association of p47phox in HMECs [17]. Similarly endothelial cells isolated from either p47phox or Nox2 knockout animals also showed a reduction in TNFa-induced ICAM-1 gene expression [2,17]. Likewise, subcutaneous infusion of a specific Nox2 inhibitor peptide gp91ds-tat reduced angiotensin II-induced ICAM-1 protein expression in the rat aortic endothelium [34]. Clearly, TNFa activates the assembly of a Nox2-based NADPH oxidase to promote superoxide generation leading to an increase in ICAM-1 gene expression.
TNFa has also been shown to modify the gene expression of catalytic subunits such as Nox2 in several cell types. We have previously demonstrated that TNFa increases Nox2 gene expression in HEK293 cells [35]. Transgenic mice with an overexpression of TNFa in endothelial cells also exhibited an increase in Nox2 mRNA expression [36]. Similarly, Nox2 mRNA in monocytic and microglial cell lines was upregulated following lipopolysaccharide/interferon gamma treatment via the NF-kB signaling pathway [37]. In contrast, here we could not find any effect of TNFa on the gene expression of Nox2 and the associated subunits p22phox and p67phox. Nox4 is another important Nox subtype found in human endothelial cells and its superoxide generating capacity is independent of Rac1. Surprisingly we found TNFa suppressed Nox4 mRNA in the present study whilst others showed that TNFa elevated Nox4 gene expression in porcine cerebral endothelial cells [38]. The difference in TNFa mediated gene alteration between these cells [38] may relate to the specific subtype of protein kinase C second messenger that is activated upon TNFa stimulation, and remains to be clarified. For example, Frey et al [39] previously showed that protein kinase Cf regulates TNFa induced NADPH oxidase activation and this involves phosphorylation of the associated subunits p47phox or p67phox [40]. On the other hand, Xu et al. [41] found that PKCe downregulates Nox4 and PKCa upregulates Nox4, suggesting an interrelated regulatory mechanism between protein kinase C subtypes and Nox4 gene regulation following TNFa stimulation. Nevertheless, our findings suggest that TNFa-induces Nox2 based NADPH oxidase activity rather than the expression of either Nox2 or Nox4 gene expression to stimulate ROS generation in HMECs. This is well supported by a recent study which showed that TNFa treatment stimulated total ROS production in HEK cells overexpressing Nox2 but not Nox4 [42]. The other potential candidate for superoxide generation in human endothelial cells is Nox5 [24] and its activation relies solely on calcium [43]. Furthermore Montezano et al. [44] demonstrated that Nox5 activation was independent of Rac following angiotensin II stimulation in HMECs. Therefore Nox5 is unlikely to contribute to TNFa induced superoxide release in the present study.
TNFa-induced NADPH oxidase derived ROS have been shown to regulate downstream signaling by mitogen-activated protein (MAP) kinases including p38 MAPK, extracellular signalregulated kinase (Erk1/2) and c-Jun terminal kinases (JNK) in endothelial cells [17,20]. Li et al. demonstrated that the binding of TRAF4 to p47phox promoted the phosphorylation of p38 MAPK and Erk1/2 but not JNK [17]. On the other hand, Lin et al. reported that inhibition of JNK or p38 MAPK but not of Erk suppressed NADPH oxidase dependent-TNFa mediated ICAM-1 and VCAM-1 gene upregulation [20]. Although these findings are inconsistent, they both support a role for NADPH oxidase-derived ROS in the regulation of MAPK phosphorylation. It has been established that MAPK phosphorylation activates the transcription factor NF-kB to promote cell adhesion molecule gene expression in endothelial cells [45]. Consistent with the findings of Yin et al. [45], we demonstrated that TNFa enhanced the promoter activity of NF-kB. Min et al. recently demonstrated that TNF-related activation-induced cytokine (TRANCE) or receptor activator of NF-kB ligand increases ICAM-1 protein and NF-kB DNA binding in human endothelial cells and both responses are abrogated by inhibition of protein kinase C, phospholipase and antioxidant [46]. Phospholipase and protein kinase C have previously been shown to regulate NADPH oxidase activation [14], so TNFa activation of the Rac1 pathway could involve phospholipase and this aspect warrants further investigation.
Endothelial cells play an important role in inflammation since they serve as the primary transendothelial migration points for blood borne leucocytes to reach inflamed tissues [47]. Accumulating evidence has implicated a role for NADPH oxidase-derived ROS in endothelial activation during the inflammatory response [2,12,16,17,48]. Limited studies conducted by us [4] and others [33] have shown a protective action of annexin peptide Ac2-26 in cell types other than inflammatory cells such as cardiac myocytes and endothelial cells. Using human endothelial cells, we confirmed the inhibitory effect of Ac2-26 on superoxide generation and ICAM-1 gene upregulation was indeed FPRL-1 receptor specific. We further demonstrated a novel inhibitory mechanism exerted by Ac2-26 on superoxide generation by showing Ac2-26 prevents TNFa mediated assembly of Rac1-dependent NADPH oxidase, leading to a downregulation of NF-kB-induced ICAM-1 and VCAM-1expression. How Ac2-26 interferes with the activation of NADPH oxidase is not fully understood. Annexin-1 or Ac2-26 peptide has previously been shown to inhibit phospholipase A2 [49,50] and phosopholipases are involved in the activation of NADPH oxidase in several cell types [14,51]. We therefore propose that Ac2-26 may target phospholipase and this then interferes with NADPH oxidase assembly to reduce superoxide generation.
Interestingly, endogenous annexin has been detected in endothelial cells following an inflammatory insult [52], suggesting that spontaneous local production of annexin may well be a natural cell defense mechanism for the endothelial cells. However, elevating the production of annexin might not always be favorable. Williams et al. recently showed that cleaved C-terminal annexin peptide increased the clustering of ICAM-1 proteins around neutrophils, facilitating their transmigration across endothelial cells, whilst Ac2-26 (annexin derived peptide with N-terminal) has the opposite effect [53]. Furthermore, we demonstrated that Ac2-26 inhibited Rac1-dependent NADPH oxidase-derived superoxide generation in both phagocytes and endothelial cells. However, Ac2-26 attenuated (inhibits ,25%) phagocytic NADPH oxidase (mainly Nox2) at 5 mM, whereas 0.5 to 1.5 mM concentration of Ac2-26 completely abolished NADPH oxidase activity in endothelial cells. This finding is consistent with a previous study by Karlsson et al. that 20 to 100 mM of Ac2-26 is required to block neurophil NADPH oxidase activity induced by formylated peptide N-Formyl-Met-Leu-Phe (fMLP) [54]. This may be due to the higher expression level of NADPH oxidase subunits and activities in phagocytic as compared to endothelial cells. Indeed, earlier studies have shown that mRNA copy number of Nox2 is many fold higher in phagocytic cells than endothelial cells [23]. Thus, sub-micromolar levels of Ac2-26 are sufficient to block endothelial NADPH oxidase activity.
We and others have previously shown that both phagocytic and vascular NADPH oxidases are important in inflammatory processes such as extravasation and transmigration of inflammatory cells and activation of endothelial cells, including increases in cell adhesion molecule expression and endothelial dysfunction, which are important sources for the initiation and progression of reperfusion injury and artery diseases such as atherosclerosis [2,48,55,56]. Therefore, exogenous application of Ac2-26 could have therapeutic potential stemming from suppression of ROS generation derived from both phagocytic and vascular NADPH oxidases.
In conclusion, our findings suggest that Rac1-dependent superoxide generation is essential for TNFa-mediated upregulation of ICAM-1 and VCAM-1 in endothelial cells. This is the first study to demonstrate that Ac2-26 inhibits superoxide generation in endothelial cells by interfering with the assembly of NADPH oxidase. Figure S1 The effects of annexin-1 peptide Ac2-26 on superoxide generation. Ac2-26 did not affect superoxide generation detected by lucigenin-enhanced chemiluminescence. (TIF) Figure S2 The effect of annexin peptide Ac2-26 via FRL-1on ICAM-1 gene expression. FPRL-1 antagonist WRW4 inhibited the effect of Ac2-26 (0.5 mM) by restoring the TNFa stimulated mRNA expression of ICAM-1 in the presence of Ac2-26. Cells were pretreated 30 min with WRW4(10 mM) then incubated with Ac2-26 alone (0.5 mM) or TNFa (20 ng/ml)+Ac2-26 for 6h. TNFa was added 30 min after Ac2-26. mRNA expression was normalized to control with TNFa stimulation. Data are mean 6 SEM, n = 3 to 5. * P,0.05 vs control without TNFa stimulation; { P,0.05 vs control with TNFa. (TIF) Figure S3 The effects of annexin-1 peptide Ac2-26 on Nox subunit and Rac1 expression in HMECs. TNFa (2-50 ng/ml), alone (A) and in combination with Ac2-26 (B) did not affect mRNA expression of p22phox and p67phox. (C) TNFa treatment for 6 or 24 h did not alter Rac1 protein expression. (D) TNFa and in combination with either Ac2-26 or DPI did not  Figure S4 The effects of annexin-1 peptide Ac2-26 on superoxide generation in DMSO differentiated HL-60 cells. PMA stimulated the superoxide generation detected by lucigenin-enhanced chemiluminescence in DMSO differentiated HL-60 cells and this is reduced by pretreatment with Ac2-26. Data are mean 6 SEM, n = 8 to 9. * P,0.05 vs control (Ctrl) without PMA, { P,0.05 vs control with PMA. (TIF)