Introduction

Diabetes mellitus is a leading cause of morbidity and mortality in western countries due to cardiovascular complications [1]. It has been suggested that hyperglycaemia, insulin resistance, glycation of proteins, oxidative stress and inflammation may be related to atherogenesis in diabetes [2]. The metabolic abnormalities associated with diabetes lead to activation of the renin-angiotensin-aldosterone system (RAAS) with a subsequent increase of angiotensin II (Ang II) and increased AT1-receptor (AT1R) activation [3, 4]. Increased AT1R activation promotes formation of reactive oxygen species (ROS) which are in turn closely linked to the onset and progression of endothelial dysfunction and atherogenesis [5]. Inhibitors of the RAAS system are associated with improvement of insulin sensitivity, reduced rates of new onset of diabetes and decreased ROS formation [6–8]. So far, the causal link between these clinical observations and AT1R inhibition remains unclear. Some angiotensin receptor blockers (ARBs), such as telmisartan are partial agonists of peroxisome proliferator-activated receptors (PPARs) [9–11]. The most abundant isoform, PPARγ, plays an important role in the regulation of adipogenesis and insulin sensitivity [12]. Furthermore, PPARγ activation has been associated with anti-atherosclerotic effects including reduced formation of ROS [13]. Beneficial effects of ARBs may be partially attributed to the activation of PPARγ [9]. In vitro studies investigating the interaction of PPARγ and the AT1R in vascular smooth muscle cells (VSMC) showed that activated PPARγ suppresses AT1R gene expression and vice versa, suggesting that pharmacological blockade or genetic disruption of the AT1R leads to enhanced PPARγ activity thereby mediating anti-atherosclerotic effects in the vascular compartment [14, 15]. However, the relevance of these mechanisms has not been determined in an in vivo model of diabetes. Whether interactions of AT1R and PPARγ play a key role in the pathogenesis of diabetes-induced atherosclerosis remains undetermined.

In the present study we analysed the influence of AT1R-PPARγ interactions on diabetic-induced atherosclerotic lesion formation and endothelial function in an experimental long-term diabetic mouse model. In this well characterized model, injection of the cytotoxin streptozotocin (STZ) results in a reduction in ß-cells and an increase in plasma glucose to diabetic levels [4]. The validity of this model has recently been confirmed as appropriate for the study of diabetes-associated atherosclerosis by the National Institutes of Health (NIH)/Juvenile Diabetes Research Foundation (JDRF)-supported Animal Models of Diabetic Complications Consortium [16]. Our aim was to determine whether pharmacological inhibition or genetic disruption of the AT1R and the PPARγ pathway would interfere with the pathogenesis of diabetic vascular complications.

Methods

Animals and treatment protocols

Female, 6-week-old homozygous apolipoprotein E deficient (ApoE−/−) mice (genetic background: C57BL/6J, Charles River, Sulzfeld, Germany) and AT1A receptor knockout mice (AT1R−/−) with identical genetic background (kindly provided by Dr. Coffmann, University of North Carolina) were used for this study. Thirty-two ApoE−/−-mice and 12 ApoE−/−/AT1R−/−-mice were rendered diabetic by 5 daily intraperitoneal injections of streptozotocin (Sigma-Aldrich, Germany) at a dose of 55mg/kg in citrate buffer or received citrate buffer (0.01 mol/l, pH: 4.5) alone (Figure 1A). All streptozotocin treated animals had blood glucose-levels ≥250 mg/dl 14 days after the induction of diabetes. The same number of ApoE−/−-mice and ApoE−/−/AT1R−/− served as non-diabetic control animals (Figure 1A). In addition, diabetic and non-diabetic ApoE−/−-mice were randomized in 8 groups consisting of 8 animals to receive the AT1R-blocker telmisartan (Sigma-Aldrich, Germany) at a dose of 40 mg/kg body weight per day orally via chow or the selective PPARγ antagonist GW9662 (Sigma-Aldrich) i.p. at a dose of 1mg/kg body weight every second day or telmisartan and GW9962 or vehicle for 18 weeks (Figure 1A). Diabetic and non-diabetic ApoE−/−/AT1R−/−-mice were further randomized in 4 groups consisting of 6 animals to receive either GW9662 or vehicle for 18 weeks (Figure 1A). After induction of diabetes the animals were treated for 18 weeks, had unrestricted access to water and standard mouse chow and were maintained in a room with a 12-hour light/dark cycle and a constant temperature of 22°C. The experimental setting is depicted as flow chart in Figure 1B. After treatment of 18 weeks mice were sacrificed and read-outs were performed (Figure 1B). All animal experiments were performed in accordance with institutional guidelines and the German animal protection law.

Figure 1
figure 1

Experimental setting. (A) Thirty-two ApoE−/−-mice and 12 ApoE−/−/AT1R−/−-mice were rendered diabetic after injections of streptozotocin. The same quantity of ApoE−/− and ApoE−/−/AT1R−/−-mice received vehicle alone and served as non-diabetic controls. Diabetic and non-diabetic ApoE−/−-mice were further randomized in groups of 8 animals to receive telmisartan, GW9662, telmisartan and GW9962 or vehicle for 18 weeks. Diabetic and non-diabetic ApoE−/−/AT1R−/−-mice were further randomized in groups of 6 animals to receive either GW9662 or vehicle for 18 weeks. (B) After treatment of 18 weeks mice were sacrificed and read-outs were performed.

Full size image

Measurements of blood pressure (BP), heart rate, blood glucose and body weight

Systolic blood pressure and heart rate were measured by a computerized tail-cuff system (CODA 6, Kent Scientific) in conscious animals. Blood glucose levels were measured using Accu-Chek®-Sensor, Roche, Mannheim, Germany). Blood samples were collected by tail vein puncture. Body weights were measured weekly and changes in body weight at baseline compared to body weight after 18 weeks of treatment were calculated.

Aortic ring preparations and tension recording

After excision of the descending aorta, the vessel was immersed in chilled, modified Tyrode buffer containing, in mmol/L, NaCl 118.0, CaCl2 2.5, KCl 4.73, MgCl2 1.2, KH2PO4 1.2, NaHCO3 25.0, Na EDTA 0.026, D(+) glucose 5.5, pH 7.4. Three-millimeter rings were mounted in organ baths filled with the above-described buffer (37°C; continuously aerated with 95% O2 and 5% CO2) and were attached to a force transducer, and isometric tension was recorded. The vessel segments were gradually stretched over 60 minutes to a resting tension of 10 mN. The drug concentration was increased when vasoconstriction or vasorelaxation was completed. Drugs were washed out before the next substance was added.

Staining of atherosclerotic lesions and morphometric analysis

Hearts with ascending aortas were embedded in Tissue Tek OCT embedding medium and sectioned on a Leica cryostat (9 μm), starting at the apex and progressing through the aortic valve area into the ascending aorta and the aortic arch and placed on poly-L-lysine (Sigma) coated slides. At least 15 consecutive sections per animal were used for analysis. For detection of atherosclerotic lesions, aortic cryosections were fixed with 3.7% formaldehyde and stained with oil red O working solution. For morphometric analysis, hematoxylin staining was performed according to standard protocols. Stained samples were examined with a Zeiss Axiovert 200 microscope (Carl Zeiss Jena, Germany) and an AxioCam MRc5. Images were acquired with Zeiss AxioVision software Rel. 4.5.0 and processed with Corel Graphic Suite X4. For quantification of atherosclerotic plaque formation in the aortic root, lipid staining area and total area of serial histological sections were measured. Atherosclerosis data are expressed as lipid-staining area in percent of total surface area. The investigator who performed the histological analyses was unaware of the hypothesis of this study and the treatment of the respective animal group.

Measurement of vascular reactive oxygen species

ROS release in intact aortic segments was determined by L-012 chemiluminescence, as previously described [5]. Chemiluminescence was assessed over 15 minutes in a scintillation counter (Lumat LB 9501, Berthold) at 1-minute intervals. The vessel segments were then dried, and dry weight was determined. ROS release is expressed as relative chemiluminescence per milligram of aortic tissue.

Statistical analysis

Data are presented as mean ± SEM. Statistical analysis was performed using the ANOVA test followed by the Neuman–Keuls post hoc analysis. P < 0.05 indicates statistical significance.

Results

Blood pressure, heart rate and metabolic parameters

Total cholesterol levels, fasting blood glucose, change in body weight, blood pressure and heart rate were measured in all groups (Table 1). Mice were treated as described in the method section and depicted in Figure 1A and Figure 1B. Blood pressure (mmHg) and heart rates (bpm) were measured in all groups by tail-cuff measurements. After 18 weeks systolic blood pressure was reduced in ApoE−/−/AT1−/−-mice compared to ApoE−/−-mice. Heart rates were significant higher in diabetic animals. Measurements of fasting blood glucose levels showed pathological glucose levels in diabetic animals compared to non-diabetic animals. GW9662 treatment in diabetic ApoE−/− and diabetic ApoE−/−/AT1−/−-mice resulted in the highest elevation of fasting blood glucose levels. Cotreatment with telmisartan and GW9662 markedly attenuated this effect in diabetic ApoE−/−-mice. Diabetic ApoE−/−-mice treated with telmisartan and diabetic ApoE−/−/AT1−/−-mice showed the lowest fasting blood glucose levels compared to the other diabetic animals. At baseline (d0) body weight was identical in all groups (data not shown). After 18 weeks of standard chow diet, all groups of diabetic ApoE−/−-mice had significant more loss of body weight than the corresponding groups of diabetic ApoE−/−/AT1R−/−-mice. Administration of GW9662 led to a significant decrease in body weight in diabetic ApoE−/− and diabetic ApoE−/−/AT1R−/−-mice compared to vehicle treated diabetic ApoE−/− and diabetic ApoE−/−/AT1R−/−-mice. Interestingly, GW9662 treated diabetic ApoE−/−-mice lost significant more body weight then GW9662 treated diabetic ApoE−/−/AT1R−/−-mice. In contrast, all groups of non-diabetic ApoE−/−-mice and non-diabetic ApoE−/−/AT1R−/−-mice had a uniform increment in body weight after 18 weeks. Total cholesterol levels were higher in diabetic ApoE−/−-mice and diabetic ApoE−/−/AT1−/−-mice compared to non-diabetic groups. Highest total cholesterol levels were detected in GW9662 treated diabetic ApoE−/−-mice, indicating poor glucose metabolism and increased lipolysis in GW9662 treated diabetic ApoE−/−-mice. Co-treatment with telmisartan reduced this effect significantly. GW9662 had no effect in diabetic ApoE−/−/AT1−/−-mice. All parameters are shown in Table 1.

Table 1 Blood pressure, heart rate and metabolic parameters
Full size table

Vascular function

Vascular function was assessed in isolated aortic ring preparations. In contrast to diabetic ApoE−/−/AT1R−/−-mice, endothelium dependent vasodilatation was significantly impaired in diabetic ApoE−/−-mice indicating that AT1R-deficiency attenuates endothelial dysfunction in diabetic animals. Endothelium-dependent vasodilatation was significantly impaired in GW9662 treated diabetic ApoE−/−-mice, whereas treatment with telmisartan led to a significant improvement of endothelium-dependent vasodilatation. Cotreatment with GW9662 abolished the beneficial effect of telmisartan on endothelial function (Figure 2A). In non-diabetic ApoE−/−-mice treated with vehicle or telmisartan or telmisartan and GW9662 endothelial function was not significantly affected (Figure 2B). Endothelium independent vasorelaxation induced by nitroglycerin was similar in all groups (data not shown). In addition, vasoconstriction induced by phenylephrine or KCL was similar in all groups (data not shown).

Figure 2
figure 2

Vascular function. After 18 weeks aortic segments of diabetic (A) and non-diabetic (B) ApoE−/− and ApoE−/−/AT1R−/−-mice were isolated and their functional performance was assessed in organ chamber experiments. Endothelium-dependent vasodilation induced by carbachol is shown. Diabetic ApoE−/−-mice displayed severe impairment of endothelial function compared to ApoE−/−/AT1R−/−-mice and telmisartan-treated ApoE−/−-mice. Treatment of diabetic ApoE−/− and ApoE−/−/AT1R−/−-mice with GW9662 antagonized the protective vascular effects of AT1R deficiency or AT1 antagonism. *P < 0.05 vs. diabetic ApoE−/−, #P < 0.05 vs. diabetic ApoE−/−/AT1R−/− and ‡P < 0.05 vs diabetic ApoE−/− + Telmisartan + GW9662, n = 6-8 per group.

Full size image

Vascular oxidative stress

Vascular release of ROS radicals was measured by L012-chemiluminescence assays in intact aortic segments. Figure 3A and Figure 3B illustrate that vascular ROS release was significantly higher in diabetic animals than in non-diabetic animals. Diabetic ApoE−/− -mice had significantly higher ROS levels than diabetic ApoE−/−/AT1R−/−-mice. AT1R-deficiency in diabetic ApoE−/−-mice and telmisartan treatment in diabetic ApoE−/−-mice significantly decreased vascular ROS release. Co-administration of GW9662 abolished this effect, whereas treatment with GW9662 alone induced the highest ROS release in diabetic ApoE−/−-mice (Figure 3A). In non-diabetic animals AT1R-deficiency and treatment with telmisartan reduced vascular ROS release in a comparable fashion (Figure 3B).

Figure 3
figure 3

Oxidative stress. Vascular ROS formation in isolated aortic segments of diabetic (A) and non-diabetic (B) ApoE−/− and ApoE−/−/AT1R−/−-mice were assessed by L-012 chemiluminescence. Diabetic ApoE−/−-mice showed enhanced oxidative stress compared to ApoE−/−/AT1R−/−-mice and telmisartan-treated ApoE−/−-mice. Treatment of diabetic ApoE−/− and ApoE−/−/AT1R−/−-mice with GW9662 raised the levels of oxidative stress to the range found in diabetic ApoE−/−-mice. *P < 0.05 vs. diabetic and non-diabetic ApoE−/−, #P < 0.05 vs. diabetic and non-diabetic ApoE−/−/AT1R−/−, ‡P < 0.05 vs diabetic ApoE−/− + GW9662, n = 6-8 per group.

Full size image

Atherosclerotic lesion formation

Development of atherosclerotic lesions was quantified in diabetic and non-diabetic animals using oil red O-staining and macroscopic analysis of the aortic root after 18 weeks. Figure 4A-F and Figure 4H-M shows representative aortic root preparations of the different groups of animals. In contrast to vehicle treated non-diabetic ApoE−/−-mice (Figure 4H), vehicle treated diabetic ApoE−/−-mice (Figure 4A) displayed more atherosclerosis in the aortic root. In age-matched diabetic and non-diabetic ApoE−/−/AT1R−/−-mice, atherosclerotic lesions were significantly diminished (Figure 4B and 4I). Concurrent with the significantly improved endothelial function in telmisartan-treated diabetic ApoE−/−-mice (Figure 4C), a significant reduction in atherosclerotic lesion formation was observed compared to vehicle treated diabetic animals. Application of GW9662 in diabetic ApoE−/−-mice showed pronounced atherosclerotic lesion formation (Figure 4E). Co-administration of GW9662 and telmisartan attenuated this effect (Figure 4D). In addition, application of GW9662 significantly increased atherosclerotic lesion formation in diabetic ApoE−/−/AT1R−/−-mice (Figure 4F) compared to vehicle treated diabetic ApoE−/−/AT1R−/−-mice (Figure 4I). PPARγ inhibition in GW9662 treated diabetic ApoE−/−-mice in the highest extent of atherosclerotic lesion formation (Figure 4. In non-diabetic mice (Figure 4H-N) atherosclerotic lesion formation was significantly less pronounced compared to corresponding diabetic groups (Figure 4A-G). AT1R-deficiency lowered atherosclerotic plaque burden in non-diabetic ApoE−/−/AT1R−/−-mice compared to non-diabetic ApoE−/−-mice, indicating that AT1R-deficiency results in decreased atherogenesis not only in diabetic but also in non-diabetic conditions. Subgroup analysis in non-diabetic mice indicated that treatment with telmisartan led to reduced atherosclerotic lesion formation in ApoE−/−-mice (Figure 4J), whereas GW9662 (Figure 4L) application showed a tendency towards more atherogenesis. Importantly, as in diabetic animals, AT1R-deficiency mediated significant atheroprotective effects in non-diabetic animals. Vehicle treated ApoE−/−/AT1R−/−-mice (Figure 4I) had significantly reduced plaque burden compared to ApoE−/− controls (Figure 4H). GW9662 treated ApoE−/−/AT1R−/−-mice (Figure 4M) had significant less atherosclerotic plaques than ApoE−/−-mice treated with GW9662 (Figure 4E). Quantitative analysis of atherosclerotic lesion formation in diabetic and non-diabetic animals is shown in Figure 4G, Figure 4N and in Additional file 1: Table S1.

Figure 4
figure 4

Atherosclerotic lesion formation. Diabetic and non-diabetic ApoE−/−-mice were treated 18 weeks with telmisartan, GW9662, telmisartan and GW9662 or vehicle, whereas diabetic and non-diabetic ApoE−/−/AT1R−/−-mice treated with GW9662 or vehicle. Representative histological cross-sections of the aortic root were stained with oil red O to analyse atherosclerotic plaque development, Figure 4A-F (diabetic animals) and Figure 4H-M (non-diabetic animals). Quantitative analysis of atherosclerotic lesion formation indicated as plaque area in % of total area is depicted in Figure 4G (diabetic animals) and Figure 4N (non-diabetic animals). Diabetic ApoE−/−-mice displayed increased atherosclerotic lesion formation. AT1R-deficiency and telmisartan treatment in ApoE−/−-mice resulted in a significantly reduced area of atherosclerotic lesions, whereas GW9662 antagonized the atheroprotective effects of AT1R deficiency or AT1 antagonism. *P < 0.05 vs. diabetic and non-diabetic ApoE−/−, #P < 0.05 vs. diabetic ApoE−/−/AT1R−/−, ‡P < 0.05 vs diabetic and non diabetic ApoE−/− + GW9662, n = 6-8 per group.

Full size image

Discussion

The present study demonstrates the major role of the AT1R in diabetes-induced atherogenesis. Here we show that AT1R blockade using a pharmacological approach and AT1R deficiency by gene deletion attenuates atherosclerosis and improves endothelial function in diabetic and non-diabetic ApoE−/−-mice. In addition, we show for the first time that the antiatherosclerotic effects of AT1R-inhibition are in part mediated via PPARγ in vivo. PPARγ-inhibition by GW9662 treatment resulted in enhanced ROS generation and atherosclerotic lesion formation in comparison to untreated diabetic ApoE−/−/AT1R−/−-mice.

PPARγ belongs to the family of peroxisome proliferator-activated receptors, which control the expression of various genes attributed with potentially vasoprotective effects such as improved glucose metabolism, lipid homeostasis and reduction of oxidative stress [17, 18]. Interestingly, PPARγ affects the expression of the AT1 receptor in vascular cells in vitro. Several authors investigated this issue in vascular smooth muscle cells and reported a downregulation of AT1 receptor expression on the transcriptional level following stimulation of PPARγ [14]. Tham et al. analysed the effects of AT1 receptor activation in respect to PPARγ expression in vitro and in vivo and reported a downregulation of PPARγ expression following stimulation with Ang II [19]. This mutual interaction is stressed by clinical observations in patients with metabolic syndrome defined by the coincidence of arterial hypertension and impaired glucose homeostasis [20]. Hypertension and the development of diabetes mellitus might interfere much closer than presently anticipated and the suggested interlocking regulation of PPARy and the AT1 receptor might be the key factor in this respect. On the other hand this interaction might explain why the use of Ang II blockers is associated with improved glucose metabolism. PPARγ activation by PPARγ-agonists induces vascular protection through the improvement of lipid metabolism, anti-inflammation and anti-proliferation [10, 18, 21]. However, the role of the RAAS in mediating vascular protective effects of PPARγ-agonists is presently not fully understood, particular not in metabolic condition of diabetes [22].

One characteristic of the PPARs is that their activation can occur through a broad spectrum of ligands even with rather low affinity [23]. This implies that particular care must be taken when assessing the PPARγ-dependence of AT1R-signaling-pathway. Signalling pathway connected with PPARγ activation have been investigated in a variety of recent publications [24]. Focussing on vascular inflammation Ji and coworkers [18] found that PPARγ agonist treatment of vascular cells in vitro and in vivo significantly reduced proinflammatory effects of Ang II. The modulatory effects of PPARγ were related to diminished activation of the proinflammatory toll-like receptor 4 (TLR4). TLR4 in turn has been attributed with consecutive activation of the IP10/PKC/NF-k B pathway. In contrast to Sugawara et al. PPARγ activation did not affect AT1 receptor expression but significantly reduced AT1 receptor dependent ERK1/2 regulation [14]. Studies published by other authors support the relevant role of the AngII/TLR4-axis in this respect [25].

Our results emphasize that AT1R-PPARγ-interactions are in part responsible for atheroprotective effects of AT1R-deficiency. Application of the specific PPARγ-antagonist GW9662 in diabetic ApoE−/−/AT1R−/−-mice resulted in enhanced ROS generation and atherogenesis in comparison to untreated diabetic ApoE−/−/AT1R−/−-mice. However, the here demonstrated ability of GW9662 to inhibit PPARγ does not rule out occurrence of a mechanism independent of PPARγ-inhibition or counteracting AT1R-action by altering its signaling cascade olbeit in the literature, GW9662 has not been associated with pleiotropic pharmacological effects independent of PPARγ-inhibition or associated with AT1R-interference.

Diabetes mellitus is associated with a profound risk of developing atherosclerosis and its complications such as myocardial infarction [26], stroke [27] and peripheral vascular disease [28]. In patients with diabetes, atherosclerotic lesion progression is accelerated if compared to the non-diabetic population [29]. ARBs have recently been shown to prevent the onset of diabetes in hypertensive patients and to reduce cardiovascular and renal disease progression in diabetic patients with hypertension [30–36]. Whether a specific AT1 Blocker shows higher levels of PPARγ agonism and whether this effect results in a clinical benefit remains an unsolved question. Our data showed that telmisartan improved oxidative stress, endothelial function and atherosclerosis while GW9662 treatment showed increased ROS levels with deleterious effects on endothelial function and atherosclerosis. These effects are much more pronounced in diabetic animals compared to non-diabetic animals. This diabetes-dependent aggravation might be partly related to the diabetes-specific over-activation of the RAAS.

In conclusion, AT1R-deficiency or pharmacological inhibition of the AT1R and activation of PPARγ with telmisartan in diabetic individuals have beneficial effects on oxidative stress, endothelial function and atherosclerotic plaque development especially in metabolic and RAAS abnormalities associated with diabetes indicating to a relevant interaction of PPARγ and the RAAS in vivo. These findings suggest a potential utility of AT1R inhibitors with partial PPARγ agonistic activity in the prevention and treatment of diabetic macrovascular complications.