Introduction

Cardiovascular disease (CVD) is the single largest cause of morbidity and mortality in patients with uremia or reduced kidney function. In patients with end-stage renal disease (ESRD), age-adjusted cardiovascular mortality increases about 30-fold compared with the general population, with sudden death representing the major cardiovascular cause of death [1]. However, the underlying pathophysiological basis of these observations is not clear, and several traditional as well as nontraditional risk factors may contribute to atherosclerotic CVD in patients with ESRD [2]. Excessive oxidative stress is one of these factors that plays a central role and might represent a link between several features observed in ESRD patients that have been associated with cardiovascular morbidity and mortality, namely an increased inflammatory load and the extent of malnutrition [3]. Increased oxidative stress also causally contributes to a reduced endothelial capacity in patients with renal insufficiency [4]. Endothelial dysfunction is thought to play a key role in the development of atherosclerotic CVD and to have independent prognostic implications [5]. Importantly, endothelial capacity is reduced already in patients with early stages of chronic kidney disease, resulting in severe endothelial dysfunction [6].

The type IIA secretory phospholipase A2 (sPLA2) is an acute-phase protein showing increased expression in acute as well as chronic inflammatory states [7]. sPLA2 is expressed in response to proinflammatory stimuli in various cell types including vascular smooth muscle cells (VSMCs) [7, 8] that show constitutive expression of the enzyme that is dramatically upregulated in atherosclerotic areas [9, 10].

In experimental mouse models, sPLA2 has provided a potential link between the dyslipidemia of inflammatory states and increased atherogenesis [11, 12]. Interestingly, plasma sPLA2 levels are significantly increased in human populations with atherosclerotic CVD and normal kidney function [13] and predict cardiovascular events comparable or even superior to C-reactive protein [13].

The aim of this study was (1) to investigate if sPLA2 expression might provide a link to oxidative stress generation in patients with ESRD, a patient population with a proinflammatory state and an excessive increase in CVD mortality, and (2) to explore potential pathophysiological consequences of sPLA2 expression on vascular function in transgenic mice.

Methods

Patient recruitment

The study was approved by the local ethics committee, and informed consent was obtained from all patients. Sixty patients with ESRD who had been undergoing maintenance hemodialysis with a median of 40 (range 26–29) months were enrolled in this study. Of note, taking the average time these patients were on hemodialysis into account, this group might represent a selected group of survivors. Twenty-five subjects with normal renal function were used as a control group. The clinical and biochemical characteristics of patients and control subjects are given in Table 1. Patients were in a stable condition and free from intercurrent illness and infection for at least 3 months. As confirmed by clinical examination, patients were in a good state of health, notably without signs of malnutrition or wasting. All the patients were dialyzed on standard bicarbonate basis for 4–5 h three times weekly using biocompatible polysulfone hemodialysis membranes (F60, Fresenius Medical Care, Bad Homburg, Germany). Dialysis adequacy was estimated using kt/V values [14]. None of the patients had residual renal function (diuresis over 24 h below 100 ml). The venous blood sample was drawn at midweek before dialysis session. Blood was immediately placed on ice and centrifuged at 4,000×g for 5 min at 4°C; plasma was stored at −80°C until analysis.

Table 1 Clinical and biochemical characterization of patients with ESRD and healthy control subjects
Full size table

Quantitation of type IIA sPLA2 protein mass and activity

Quantitation of sPLA2 by ELISA (Cayman Chemical, Ann Arbor, MI, USA) and the sPLA2 activity assay were performed essentially as previously described [15].

Cell culture studies

Human VSMCs were obtained from Promocell (Heidelberg, Germany). Cells were kept according to the manufacturer’s protocol. For serum stimulation experiments, cells were washed twice with phosphate-buffered saline (GIBCO, Karlsruhe, Germany) and then incubated for 24 h in DMEM supplemented with 5% of the respective human sera as indicated.

Experimental animals

C57BL/6 J mice were obtained from Charles River (Sulzfeld, Germany). The human group IIA sPLA2 transgenic mice on the C57BL/6 J genetic background used in this study have been described previously [11]. All animal experiments were approved by the responsible ethics committee of the Landesamt für Gesundheit, Ernährung und technische Sicherheit Berlin.

Arterial relaxation studies

The direct effects of acetylcholine (ACH) on arterial relaxation were evaluated in 2-mm rings of thoracic aorta from 3-month-old male sPLA2 transgenic mice and wild-type controls as previously described [16]. Selected studies were performed in rings treated with indomethacin, NS398, SC-560, and tiron.

mRNA expression analyses by real-time quantitative PCR

Total RNA from cells and mouse aortas was isolated using TRIZOL. Expression analysis by real-time quantitative reverse-transcription PCR was performed as previously published [17]. mRNA expression levels presented were calculated relative to the average of the housekeeping gene cyclophilin.

Detection of intracellular O −2 in aortic tissue

The oxidative fluorescent dye hydroethidine was used to evaluate in situ the production of superoxide [18].

Quantification of isoprostanes and 2,3-dinor-6-keto-PGF1α

Plasma and aortic levels of the isoprostane 8,12-iso-iPF2α-VI were measured by gas chromatography–mass spectrometry as previously described [19]. To assess cyclooxygenase (COX)2 activity in mice, urine was collected for 24 h from groups of animals, and 2,3-dinor-6-keto-PGF1α levels were quantified by stable dilution isotope gas chromatography/mass spectrometry [20]. COX2 activity in humans was measured according to a protocol of Patrignani and co-workers [21].

Statistical analysis

Statistical analyses were performed using the statistical package for social sciences (SPSS Inc., Chicago, IL, USA). Data are presented as means ± SEM. Statistical analysis was performed using the Mann–Whitney U test to compare different groups; Spearman's rank correlation coefficient was used to assess possible associations between different parameters, and the dose–response curves were compared using Friedman’s test. All P values presented are two-tailed. Statistical significance for all comparisons was assigned at P < 0.05.

Results

Increased oxidative stress in ESRD patients is correlated with elevated plasma sPLA2 levels

First, we assessed plasma levels of type IIA sPLA2 in patients with ESRD and explored a potential link to oxidative stress. Compared with controls, ESRD patients (n = 60) undergoing hemodialysis displayed increased in vivo oxidative stress as assessed by plasma levels of the isoprostane 8,12-iso-iPF2α (0.81 ± 0.10 vs. 2.73 ± 0.44 ng/mL, p < 0.001, Fig. 1a). Patients with ESRD had on average 6.3-fold higher levels of sPLA2 compared with healthy controls (n = 25; ESRD 1,156 ± 65 versus controls 184 ± 5 ng/dL, p < 0.001, Fig. 1b). Plasma sPLA2 activity was 3.4-fold elevated in patients with ESRD (ESRD 141 ± 16 vs controls 42 ± 2U/mL, p < 0.001, data not shown) and correlated well with plasma sPLA2 concentration measurements (r = 0.50, p < 0.001, Fig. 1c), indicating that sPLA2 in patients with ESRD is catalytically active. Plasma sPLA2 levels in ESRD patients were neither correlated with the time that these patients had been undergoing maintenance hemodialysis (r = −0.19, ns) nor with hemoglobin (r = −0.10, ns) or serum albumin levels (r = −0.12, ns). There was a trend towards a correlation between plasma sPLA2 and CRP levels (r = 0.25, p = 0.062). Neither treatment with statins nor ACE inhibitors or angiotensin II type 1 receptor antagonists was associated with a significant change in plasma sPLA2 levels. Body weight in ESRD was higher than in the control group; however, no influence of body weight on sPLA2 levels was observed. Higher weight in ESRD patients might indicate a status of relative good health with minor relevance of the malnutrition/inflammation complex on the observed results. Importantly, plasma isoprostanes as marker of oxidative stress were positively correlated with plasma sPLA2 levels in ESRD (r = 0.61, p < 0.001, Fig. 1d).

Fig. 1
figure 1

ESRD patients display increased oxidative stress correlated with elevated plasma levels of sPLA2. a Plasma levels of the isoprostane 8,12-iso-iPF2-VI. b plasma levels of sPLA2 in patients with ESRD (n = 60) and controls (n = 25). Data are given as means ± SEM. c Correlation between plasma sPLA2 levels and enzymatic activity. d Correlation between plasma sPLA2 levels and plasma isoprostanes. Asterisk, significantly different from control values (at least p < 0.05)

Full size image

sPLA2 expression increases vascular oxidative stress in transgenic mice

Next, we investigated whether increased sPLA2 expression and oxidative stress generation within the vessel wall are causatively linked. We investigated these effects in sPLA2 transgenic mice which have no signs of renal failure. The plasma levels of human sPLA2 in these mice are 95 ± 7 µg/dL. Aortic levels of the isoprostane 8,12-iso-iPF2α-VI are elevated in sPLA2 transgenic mice compared with wild-type C57BL/6 controls not expressing sPLA2 (40.8 ± 11.4 vs. 11.9 ± 1.1 pg/mg, p < 0.05, Fig. 2a), suggesting that sPLA2 expression results in increased lipid peroxidation and in vivo oxidative stress within the vascular wall.

Fig. 2
figure 2

sPLA2 transgenic animals have increased ROS production and in vivo oxidative stress. a Aortic content of the isoprostane 8,12-iso-iPF2-VI in sPLA2 transgenic mice and controls (n = 5). Data are given as means ± SEM. Asterisk, significantly different from controls (p < 0.05). b In situ formation of superoxide radicals in aortas from sPLA2 transgenic mice (sPLA2 tg) and controls (C57BL/6) using the oxidative dye hydroethidine detected by fluorescent microscopy as detailed in “Methods”

Full size image

As a further measure of vascular reactive oxygen species (ROS) generation, EtBr fluorescence was determined by confocal microscopy in aortas from sPLA2 transgenic mice and controls. In contrast to controls, tissue sections of aortas from sPLA2 transgenic mice showed a marked basal increase in EtBr fluorescence (Fig. 2b), indicating significantly increased production of ROS within the media of the vessel wall of sPLA2 transgenic mice. In the presence of the free oxygen radical scavenger tiron (10 mmol/L), EtBr fluorescence in aortas from sPLA2 transgenic mice was reduced to levels observed in controls.

Increased vascular oxidative stress in sPLA2 transgenic mice mediates severe endothelial dysfunction

Next, we studied the potential pathophysiological effects of increased sPLA2-mediated oxidative stress generation by assessing vascular function in the sPLA2 transgenic mice. There is an established link between increased oxidative stress and reduced endothelial function [4]. To examine the effects of ACH (for endothelium-dependent vasorelaxation) and sodium nitroprusside (SNP; for endothelium-independent vasorelaxation) on vascular tone in sPLA2 transgenic mice and controls, both substances were added to aortic rings precontracted with phenylephrine (PE 1 µmol/L; maximal vasoconstriction 8.1 ± 0.8 mN, n = 16). Maximal vasoconstriction achieved in response to 1 µmol/L PE was not significantly different in aortas from sPLA2 transgenic mice and controls. ACH and SNP induced dose-dependent vasodilation in either sPLA2 transgenic mice or controls (Fig. 3a, b). Maximal ACH-induced vasodilation was significantly lower in sPLA2 transgenic mice compared with controls, indicating severe endothelial dysfunction in sPLA2 transgenic mice (maximal vasodilation in response to 10 µmol/L ACH in sPLA2 transgenic mice 36 ± 8% and in controls 80 ± 2% of PE-induced vasoconstriction, p < 0.05; n = 8; Fig. 3a). EC50 of ACH did not differ between both groups of experimental animals (EC50 [log mol/L]: sPLA2 transgenic −6.2 ± 0.5 and wild type −6.3 ± 0.1, Fig. 3a). Maximal SNP-induced vasodilation in sPLA2 transgenic mice and controls was not significantly different, indicating normal endothelial-independent vascular function in sPLA2 transgenic mice (Fig. 3b).

Fig. 3
figure 3

sPLA2 transgenic mice display severe endothelial dysfunction due to increased aortic ROS formation. Aortic rings from sPLA2 transgenic mice and controls were precontracted with PE, and the direct relaxation response to ACH (a) and SNP (b) was evaluated (n = 8 experiments). Data are given as means ± SEM. Asterisk, significantly different from wild type (p < 0.05). c, d Thoracic aortic rings from sPLA2 transgenic mice (c) and controls (d) were precontracted with PE, and direct relaxation responses to ACH in the absence or presence of tiron (10 mmol/L) were evaluated (n = 4 experiments each). Asterisk, significantly different from sPLA2 values without tiron added

Full size image

We then investigated the effect of the ROS scavenger tiron (10 mmol/L) on endothelium-dependent vasodilation in sPLA2 transgenic animals and controls. In the presence of tiron, the endothelium-dependent vasodilation induced by increasing doses of ACH was significantly improved (maximal vasodilation in response to 100 µmol/L ACH in sPLA2 transgenic mice 36 ± 8% of PE-induced vasoconstriction, sPLA2 transgenic + tiron (10 mmol/L) 87 ± 5% (p < 0.05) of PE-induced vasoconstriction, Fig. 3c). In wild-type controls, the endothelium-dependent vasodilation induced by ACH was not significantly affected by the presence of tiron (10 mmol/L; Fig. 3d). The maximal vasodilation induced by SNP (100 µmol/L) either in sPLA2 transgenic mice or controls was not affected by tiron (data not shown).

Increased vascular oxidative stress in sPLA2 transgenic mice is dependent on COX2 expression

As a potential mechanism for the increased generation of ROS and oxidative stress within the vascular wall of sPLA2 transgenic mice, we evaluated a possible functional coupling between sPLA2 and cyclooxygenases. It has been previously shown in nonvascular cells that sPLA2 can induce and upregulate COX2 expression [22, 23]. No significant difference was found regarding the expression of COX1 in aortas of wild-type controls and sPLA2 transgenic mice (1.00 ± 0.09 versus 1.21 ± 0.11, respectively, ns, data not shown). However, expression of COX2 was increased by 3.3-fold in aortas from sPLA2 transgenic mice compared with controls (3.33 ± 0.19 vs. 1.00 ± 0.05, respectively, p < 0.001, Fig. 4a). PGE2 is a major cyclooxygenase product and has been shown to be involved in the progression of vascular inflammatory disease [24]. Aortic levels of the major cyclooxygenase product PGE2 were increased by 2.3-fold in sPLA2 transgenic mice compared with controls, further stressing increased cyclooxygenase activity within the vascular wall in response to sPLA2 overexpression (102 ± 21 vs. 44 ± 9 pg/μg protein, respectively, p < 0.05, data not shown). As further experimental support for increased vascular COX2 activity, urinary excretion of 6-keto-PGF1α was significantly higher in sPLA2 transgenic mice compared with wild-type littermates (wild type 1,258 ± 216 vs. sPLA2 2,033 ± 212 pg/mg creatinine, p < 0.05, Fig. 4b).

Fig. 4
figure 4

sPLA2 overexpression results in increased vascular COX2 expression and COX2-dependent ROS formation. a COX2 mRNA expression in aortas from sPLA2 transgenic mice and controls determined by quantitative real-time PCR. b Urinary 2,3-dinor-6-keto-PGF1α levels as marker for vascular COX2 activity measured in urine collected for 24 h from sPLA2 transgenic mice and controls. Data are given as means ± SEM. Asterisk, significantly different from controls (at least p < 0.05). c In situ detection of superoxide radicals in aortas from sPLA2 transgenic mice with and without addition of the COX1/2 inhibitor indomethacin (10 µmol/L for 60 min) using the oxidative dye hydroethidine. d, e Thoracic aortic rings from sPLA2 transgenic mice (d) and wild-type controls (e) were precontracted with PE, and direct relaxation responses to ACH in the absence or presence of the COX1/2 inhibitor indomethacin (10 µmol/L), the specific COX1 inhibitor SC-560 (10 µmol/L), or the specific COX2 inhibitor NS-398 (20 µmol/L) were evaluated (n = 4 experiments each). Data are given as means ± SEM. Asterisk, significantly different from sPLA2 values without added agents (at least p < 0.05)

Full size image

Subsequently, we addressed the effects of altered COX activity within the vascular wall of sPLA2 transgenic mice on aortic ROS production as assessed by EtBr fluorescence. Aortic ROS formation in sPLA2-overexpressing vessels was dramatically reduced to the levels seen in controls by the addition of the unselective COX1/2 inhibitor indomethacin (10 µmol/L; Fig. 4c).

As a next step, the pathophysiological consequences of increased COX-dependent ROS formation within aortas of sPLA2 transgenic mice were investigated. By using the unselective COX1/2 inhibitor indomethacin (10 μmol/L), the severely impaired endothelial function of sPLA2 transgenic mice could be completely restored to the level observed in controls (Fig. 4d). In the presence of the selective COX1 inhibitor SC-560 (1 µmol/L), endothelium-dependent vasodilation in sPLA2 transgenic mice did not change (Fig. 4d), while in the presence of the selective COX2 inhibitor NS-398 (20 µmol/L) a significant increase (p < 0.05) in endothelium-dependent vasodilation could be observed in sPLA2 transgenic mice (Fig. 4d). In wild-type controls, the endothelium-dependent vasodilation induced by ACH was not significantly affected by the presence of either indomethacin (10 µmol/L), NS-398 (20 µmol/L), or SC-560 (1 µmol/L; Fig. 4e). These data demonstrate that the severe endothelial dysfunction in sPLA2 transgenic mice is dependent on increased vascular COX2 activity.

Increased COX2-dependent PGE2 formation in ESRD patients is correlated with plasma sPLA2 levels

To investigate whether COX2 activity is also increased in ESRD patients and whether this might be linked to sPLA2 levels, first, COX2-dependent production of PGE2 within blood as a marker of COX2 activity was measured. PGE2 generation was significantly increased in ESRD patients compared with controls (42.05 ± 7.01 vs. 19.57 ± 5.59 ng/mL in controls, p < 0.005, Fig. 5a). Interestingly, COX2 activity was correlated with plasma sPLA2 levels in ESRD patients (r = 0.61, p < 0.05, Fig. 5b). Next, we investigated the potential of serum from healthy controls with normal and ESRD patients with high levels of sPLA2 (ESRD 1,462 ± 78 ng/dL (n = 7) versus controls 120 ± 16 ng/dL, p < 0.001) to induce COX1 and COX2 expression in human aortic VSMCs: no significant difference was found regarding the expression of COX1 using serum from healthy controls compared with serum from ESRD patients (1.00 ± 0.09 versus 1.14 ± 0.14, respectively, ns, Fig. 5c). However, expression of COX2 was increased by 5.4-fold in hVSMC in the presence of ESRD serum compared with control serum (5.44 ± 1.01 versus 1.00 ± 0.05, respectively, p < 0.001, Fig. 5d). These data indicate that similar pathophysiological mechanisms might be operative in ESRD patients as delineated in the human sPLA2 transgenic mouse.

Fig. 5
figure 5

ESRD patients display increased COX2 activity correlated with plasma sPLA2 levels. a COX2-dependent PGE2 production. b Correlation between plasma sPLA2 levels and COX2-dependent PGE2 formation in ESRD (n = 8) patients. mRNA expression of COX1 (c) and COX2 (d) in human aortic VSMCs following incubation with serum from patients with ESRD and controls (n = 8). Data are given as means ± SEM. Asterisk, significantly different from control values (at least p < 0.05)

Full size image

Discussion

In the present study, we demonstrate that circulating levels of functionally active sPLA2 are dramatically increased in ESRD patients closely correlated with plasma isoprostanes, a highly sensitive marker of in vivo oxidative stress. In addition, using human sPLA2 transgenic mice, we provide for the first time evidence that increased sPLA2 expression negatively impacts on endothelial function via a COX2-mediated increase in oxidative stress. Furthermore, our data indicate that in ESRD patients similar pathophysiological mechanisms might be operative.

sPLA2 is a cardiovascular risk factor with intrinsic proatherosclerotic biological activity. Transgenic overexpression of sPLA2 in mice results in increased atherosclerosis formation already on a chow diet [12]. sPLA2 decreases plasma high-density lipoprotein (HDL) cholesterol levels by increasing the HDL catabolic rate [11, 25]; sPLA2 induces the formation of aggregated/fused low-density lipoprotein (LDL) particles [26], and sPLA2-modified HDL loses its protective properties against LDL oxidation, which has been attributed to a loss of paraoxonase activity [27]. Macrophage expression of sPLA2 significantly enhances LDL oxidation even in the absence of HDL particles [17] and is sufficient to increase atherosclerotic lesion formation [17, 28]. Our present study extends these observations by showing that sPLA2 expression also contributes to vascular oxidative stress generation, resulting in severe endothelial dysfunction.

Plasma sPLA2 levels are elevated in patient populations with atherosclerotic CVD, [13, 29] and are predictive for the risk of future coronary events in patients with normal kidney function [13, 30]. Notably, plasma sPLA2 levels determined in these patient populations were substantially lower than the levels we found in ESRD patients. In all previous studies, there was a considerable overlap in plasma sPLA2 levels between the groups of patients and controls. In the present study, however, there was a unique threshold value of circulating sPLA2, clearly identifying the patients with ESRD.

Endothelial dysfunction is thought to represent an important independent clinical predictor for atherosclerosis development as well as cardiovascular events [5, 31]. This condition is already present in early stages of chronic kidney disease [6, 32] and highly pronounced in patients with ESRD [33, 34]. It is also well documented that patients with ESRD display increased ROS production and oxidative stress [35, 36] reflected by increased plasma isoprostane levels [37] that were in our present study correlated with plasma sPLA2 levels. Increased oxidative stress substantially contributes to endothelial dysfunction since superoxide anions (O −2 ) inactivate nitric oxide (NO), a principal endothelium-derived relaxing factor [34, 38]. Consistently, overexpression of sPLA2 in mice results in reduced endothelial capacity by increasing the production of ROS within the vessel wall. This increase occurs specifically within the media, the area of the vessel wall that is rich in VSMCs, one of the major cellular sources of sPLA2 expression [8].

sPLA2 is an extracellular phospholipase that cleaves phospholipids at the sn-2 position. The sPLA2 reaction mainly yields arachidonic acid, which is then available as precursor for the production of prostaglandins [39]. Interestingly, aortas of sPLA2 transgenic mice show increased expression of COX2 mRNA, an increased content of the major COX product PGE2, and increased urinary levels of 2,3-dinor-6-keto-PGF1α, an eicosanoid indicative of increased vascular COX2 activity. On the other hand, the ESRD patients investigated also had increased COX2-dependent production of prostaglandins correlated with plasma sPLA2 levels, indicating that similar molecular mechanisms might be operative in human patients. To test this hypothesis, administration of a COX2-specific inhibitor to ESRD patients would be desirable. However, for safety reasons, selective COX2 inhibitors are currently not indicated in this group of patients, as there were no studies approving these medications in ESRD. Although in ESRD patients administration of a COX2 inhibitor to establish causality is not possible due to ethical considerations, pharmacological inhibition of COX2 in sPLA2 transgenic mice resulted in a significant decrease of ROS production and a normalization of the endothelial dysfunction. While the COX product PGE2 has been demonstrated to be capable of inducing vasoconstriction [40], there is evidence that in our system the role of PGE2 might be minor. Not only COX inhibition but also ROS scavengers improved endothelial dysfunction in sPLA2 transgenic mice pointing to COX as a source of ROS rather than of PGE2.

How the increase in COX2 expression is mediated at the transcriptional level is presently not clear. Studies performed in different contexts noted increased COX2 expression when synoviocytes [23] and mast cells [22] were incubated with purified sPLA2 enzyme. In addition, positive feedback regulation of COX2 by PGE2 has been reported [41], and it is therefore tempting to speculate that this might also contribute to linking sPLA2 and COX2 expression in our experimental model. However, at the current stage, it cannot be excluded that other factors or uremic toxins contribute to the induction of COX2 in response to serum from ESRD patients.

In summary, we provide the first evidence for a dramatic increase in plasma sPLA2 levels as well as enzymatic activity in patients with ESRD tightly correlated with oxidative stress. Human sPLA2 expression in transgenic mice resulted in a COX2-dependent increase in ROS generation within the vascular wall and reduced endothelial capacity. Therefore, pharmacological inhibition of sPLA2 in patients with ESRD might represent a novel therapeutic strategy to decrease the oxidative burden in this high-risk patient group, conceivably translating into a reduction of cardiovascular morbidity and mortality.