Down-regulation of AMPK/PPARδ signalling promotes endoplasmic reticulum stress-induced endothelial dysfunction in adult rat offspring exposed to maternal diabetes

Abstract

Aims

Exposure to maternal diabetes is associated with increased prevalence of hypertension in the offspring. The mechanisms underlying the prenatal programming of hypertension remain unclear. Because endoplasmic reticulum (ER) stress plays a key role in vascular endothelial dysfunction in hypertension, we investigated whether aberrant ER stress causes endothelial dysfunction and high blood pressure in the offspring of dams with diabetes.

Methods and results

Pregnant Sprague-Dawley rats were intraperitoneally injected with streptozotocin (35 mg/kg) or citrate buffer at Day 0 of gestation. Compared with control mother offspring (CMO), the diabetic mother offspring (DMO) had higher blood pressure and impaired endothelium-dependent relaxation in mesenteric arteries, accompanied by decreased AMPK phosphorylation and PPARδ expression, increased ER stress markers, and reactive oxygen species (ROS) levels. The inhibition of ER stress reversed these aberrant changes in DMO. Ex vivo treatment of mesenteric arteries with an AMPK agonist (A769662) or a PPARδ agonist (GW1516) improved the impaired EDR in DMO and reversed the tunicamycin-induced ER stress, ROS production, and EDR impairment in mesenteric arteries from CMO. The effects of A769662 were abolished by co-treatment with GSK0660 (PPARδ antagonist), whereas the effects of GW1516 were unaffected by Compound C (AMPK inhibitor).

Conclusion

These results suggest an abnormal foetal programming of vascular endothelial function in offspring of rats with maternal diabetes that is associated with increased ER stress, which can be ascribed to down-regulation of AMPK/PPARδ signalling cascade.

Graphical Abstract

Graphical Abstract.

1. Introduction

Increasing evidence suggests that cardiovascular diseases (CVDs) are determined not only by a complex interaction between genetic susceptibility, lifestyle, and environment but also by events in utero and early life resulting in resetting of key physiological functions.¹ Epidemiological studies have demonstrated that changes in the intrauterine environment, including malnutrition, placental insufficiency, and hyperglycaemia during specific windows of organogenesis, are important causes of foetal stress, leading to several consequences. These include loss of structure/function, pre-emptive adaptations to an adverse postnatal environment, and finally increased risk of adult diseases, such as hypertension and CVDs.^2,3 This phenomenon, in which the adult phenotype is linked to intrauterine environment, is termed ‘prenatal programming’.⁴

The International Diabetes Federation estimated in 2017 that gestational hyperglycaemia and diabetes occur in 16.2% and 14.0% of pregnancies, respectively,⁵ conferring substantial risk to the offspring. Therefore, it is important to focus on the lifelong consequences of in utero exposure to hyperglycaemia, but the mechanisms involved in the maternal diabetes/hyperglycaemia-programmed hypertension are poorly understood. Emerging evidence from experimental research indicates that endoplasmic reticulum (ER) stress is an important pathophysiological mechanism of hypertension.⁶ The ER plays a fundamental role in protein synthesis, folding, and trafficking, but is also recognized as a primary sensor of cellular stress. Disruption of normal ER homeostasis or function triggers ER stress and leads to the activation of a complex signalling network called the unfolded protein response (UPR), which attempts to re-establish ER homeostasis.⁷ However, long-term UPR activation can disrupt cellular function and lead to chronic disease, including hypertension.⁸ There is also compelling evidence that ER stress is involved in the vascular endothelial dysfunction in hypertension and the inhibition of ER stress increases nitric oxide bioavailability which improves endothelial-dependent vasodilation.⁹

AMPK (5' adenosine monophosphate-activated protein kinase), a sensor of cellular energy, controls systemic energy balance and metabolism; its activation exerts vasoprotective effects by suppressing ER stress, that is associated with atherosclerosis and hypertension.^10,11 PPARδ, a member of the nuclear hormone receptor superfamily, plays a critical role in alleviating ER stress; the activation of PPARδ restores endothelial function in diet-induced obese mice.¹² AMPK has been reported to interact with PPARδ to improve vascular and metabolic functions.¹³ Previous studies have shown that exposure to maternal diabetes in utero causes hypertension in the adult offspring, which could be related to vascular dysfunction, for example, impaired endothelium-dependent relaxation (EDR) in the offspring of diabetic dams, relative to offspring of non-diabetic dams.¹⁴ To date, no study has examined the role of ER stress in the endothelial dysfunction in offspring of diabetic dams. Therefore, we tested the hypothesis that ER stress is involved in vascular endothelial dysfunction and contributes to the hypertension in offspring from diabetic dams, which could be ascribed to the down-regulation of the AMPK/PPARδ signalling cascade.

2. Material and methods

2.1. Animals and treatment

Pregnant Sprague-Dawley rats (250–300 g) were purchased from the Animal Centre of The Third Military Medical University. The rats were intraperitoneally injected with streptozotocin (STZ) (35 mg/kg in 0.4 mM citrate buffer, pH 4.5, Sigma Co., St. Louis, MO) or citrate buffer (1 ml) at Day 0 of gestation, as previously described.¹⁵ The diabetic state was confirmed by measuring the plasma glucose concentration (Accu-Chek Extra Care, Roche Diabetes Care, Paris, France). Only pregnant rats with plasma glucose levels of 15–20 mmol/L were used in the study. For the insulin treatment group (Dia + Ins), dams at Day 0 of gestation were surgically implanted subcutaneously with insulin pellets. Control (Con) and diabetic (Dia) rats underwent sham surgery that did not involve insulin pellet implantation. The diabetic status was confirmed every two days until delivery. All the rats were housed in a temperature- and light-controlled room at 21°C with a 12-h light cycle and had access to food and water ad libitum. The use of animals for these experiments was approved by the Animal Research Ethical Committee of the Third Military Medical University, and these experiments conformed to the NIH guidelines for the care and use of laboratory animals. The euthanasia method stated within the study was used for all animals including the offspring.

Control mother offspring (CMO) and diabetic mother offspring (DMO) at the age of 24 weeks were randomized into three groups: (i) untreated group (saline, 1 ml/day, intra-peritoneal injection for 6 weeks); (ii) 4-PBA (an ER stress inhibitor, 1 g/kg/day, intra-peritoneal injection for 6 weeks) group; and (iii) Tudca (an ER stress inhibitor, 150 mg/kg/day, intra-peritoneal injection for 6 weeks) group. Blood pressures were indirectly measured at 24 weeks of age and then every week until 30 weeks of age, using a tail-cuff method (BP-98A; Softron, Tokyo, Japan). Blood pressures were directly measured via a left carotid artery cannulation at 30 weeks of age. Plasma glucose was quantified using a glucose monitor (Accu-chek), and plasma insulin was quantified by radioimmunoassay using a rat insulin kit (Linco Research, St. Charles, MO) at 30 weeks of age. Insulin resistance was evaluated using the homeostasis model assessment (HOMA) index of insulin resistance (HOMA-IR = insulin (mU/L) × glucose (mM)/22.5).¹⁶ Levels of plasma leptin, adiponectin, and adipolin were measured using ELISA (Millipore, Billerica, MA).

2.2. Histomorphometric analysis

Histological analysis was performed in mesenteric arteries and thoracic aortae from the offspring of CMO and DMO at 30 weeks of age. Each segment of mesenteric and thoracic artery was quickly removed, carefully dissected, fixed in 4% paraformaldehyde, embedded in paraffin, and then cut into sections (7 μm). The sections were stained with haematoxylin and eosin (H&E) in order to measure histomorphometric parameters (intima–media thickness and lumen diameter).

2.3. Mesenteric artery preparation and functional study

The rats were euthanized by CO₂ inhalation at 30% volume displacement rate, and the mesenteric arteries were dissected and cleaned of adhering connective tissue in ice-cold and oxygenated Krebs solution containing (mM): 119 NaCl, 4.7 KCl, 2.5 CaCl₂·H₂O, 1.17 MgSO₄·H₂O, 25 NaH₂CO₃, 1.2 KH₂PO₄, 0.027 EDTA, and 11 D-glucose, adjusted to pH 7.35–7.45. Third-order branches of the superior mesenteric artery were cut into several ring segments (∼2 mm in length) for parallel studies on rings obtained from different rats. The rings were mounted on a myograph (DMT, Aarhus, Denmark) for isometric tension recording, using Labchart software (AD instruments, Colorado Springs, CO). All rings were initially stretched to an optimal baseline tension and equilibrated at 37°C for 60 min before the start of each experiment. After testing the constrictor effects of 125 mM high-potassium physiological salt solution (PSS), the vessels were rinsed three times with fresh PSS and allowed to recover to baseline for 15 min. Each ring was then stimulated with increasing concentrations of phenylephrine (Phe, 10⁻⁹–10⁻⁵M; Sigma-Aldrich). Once a sustained tension was reached, either acetylcholine chloride (Ach, 10^-8.5–10^–5M; Sigma-Aldrich) or sodium nitroprusside (SNP, 10^–9–10^–5M; Sigma-Aldrich) was added cumulatively to evoke endothelium-dependent or endothelium-independent relaxations, respectively. To determine the role of eNOS in the EDR studies, the mesenteric arteries were incubated with L-NAME (100 µM) for 30 min, before measuring the effects on PE-induced contraction and Ach-induced relaxation.

2.4. Ex vivo studies of rat mesenteric artery rings

Mesenteric arteries were dissected, and the adjoining connective tissues were carefully cleaned in sterile phosphate-buffered saline (PBS). The arteries were cut into several ring segments (∼2 mm in length) and incubated in Dulbecco’s Modified Eagle’s Media (DMEM, Gibco, Gaithersburg, MD, USA) with 10% foetal bovine serum (Gibco), 100 µg/mL streptomycin, and 100 U/mL penicillin. Tunicamycin (ER stress inducer, 2 µg/mL; Sigma-Aldrich), A769662 (AMPK agonist, 30 µmol/L; R&D Systems, MN, USA), Compound C (AMPK antagonist, 5 µmol/L; Sigma-Aldrich), GW1516 (PPARδ agonist, 100 nmol/L; Alexis Biochemicals, Lausen, Switzerland), GSK0660 (PPARδ antagonist, 500 nmol/L; Sigma-Aldrich), or tempol (10 µmol/L; Sigma-Aldrich) was added into the culture medium that bathed the mesenteric artery rings in an incubator at 37°C for 16 hr. After the incubation, the rings were transferred into fresh Krebs solution for functional studies in myograph.

2.5. ROS measurement

The mesenteric artery rings were frozen and sliced into sections (10 µm) using a Leica CM 100 cryostat before incubation with the fluorescent dye dihydroethidium (DHE, 5 µmol/L; Invitrogen) for 15 min in a light-protected humidified chamber at 37°C. The rings were briefly washed, and then quickly imaged under an Olympus Fluoview FV1000 laser scanning confocal system.

2.6. NADPH oxidase activity assay

Superoxide anions generated by NADPH oxidase were measured in lysates of mesenteric arteries by the lucigenin-enhanced chemiluminescence method. Briefly, NADPH (100 μM, Sigma) and lucigenin (5 μM, Sigma) were added into 1-ml microcentrifuge tubes. Superoxide production was measured every 20 sec for 10 min until enzymatic activity reached a plateau. The values were expressed as relative luminescence units per min per mg of protein. Using this method, the superoxide anion production also represents NADPH oxidase activity.

2.7. Measurement of NO metabolites

Endothelium-intact mesenteric rings were incubated in Krebs solution to measure NO metabolites. The arterial rings were stimulated with Ach (10^–6) for 5 min, rapidly removed, dabbed dry with filter paper, and weighed. The incubation solution was assayed for the stable end-products of NO, that is, nitrate (NO₃^–) and nitrite (NO₂^–), using the nitrate reductase method (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

2.8. Primary culture of mesenteric endothelial cells

Microvascular endothelial cells were isolated as described previously.¹⁷ Briefly, the mesenteric arteries were removed and dissected in physiologic saline solution. Incubating the mesenteries with 0.2% collagenase type I and 0.1% bovine serum albumin at 37°C for 1 h detached the endothelial cells, which were collected by centrifugation. The endothelial cells were resuspended and cultured in endothelial cell growth medium (EGM), supplemented with 20 μg/ml bovine pituitary (Lonza, Walkersville, MD) and 20% foetal bovine serum for 48 h. Experiments were performed with endothelial cells at 80–90% of confluence. The identity of endothelial cells was determined by staining with anti-CD31 (endothelial cell marker) antibody and confirmed by immunofluorescence microscopy. The level of TNF-α in the supernatant of primary cultures mesenteric endothelial cells was determined by ELISA (Millipore, Billerica, MA). After treatment with the reagent or vehicle, the cells were harvested for western blotting.

2.9. Confocal immunofluorescence microscopy

To determine the purity of the mesenteric endothelial cells, the expression of an endothelial cell marker protein (platelet endothelial cell adhesion molecule 1, CD31) was verified by immunofluorescence microscopy. Mesenteric arterial endothelial cells in primary culture, grown on coverslips, were fixed with 4% paraformaldehyde for 20 min at room temperature. After washing with PBS, the fixed cells were incubated overnight at 4°C with 3 µg/mL of anti-CD31 antibody (1:200; Millipore). After washing, the coverslips were incubated with Alexa Fluor-488 anti-rabbit secondary antibody for 2 h at 4°C. Coverslips were mounted on SlowFade mounting medium with the nuclear stain DAPI (Electron Microscopy Sciences, Hatfield, PA) and sealed onto glass slides. The samples were imaged using an Olympus AX70 laser scanning confocal microscope (Supplementary material online, Figure S1).

2.10. Western blot analysis

The mesenteric artery endothelial cells in primary culture were homogenized in lysis buffer, and protein samples from these mesenteric artery endothelial cells were separated by SDS–PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with 5% non-fat milk in Tris-buffered saline (TBS) and 0.05% Tween-20 PBS for 90 min at room temperature. Then, the membranes were probed with primary antibodies against phosphorylated AMPKα at Thr¹⁷², phosphorylated eIF2α (p-E74 Like ETS Transcription Factor 2) at Ser52, phosphorylated eNOS (p-endothelial nitric oxide synthase) at Ser¹¹⁷⁷, Thr495, and Tyr657, phosphorylated protein kinase RNA-like ER (p-PERK), phosphorylated inositol-requiring enzyme 1 (p-IRE1) (1:500; Cell Signaling Technology, Danvers, MA, USA), t-eNOS, t-AMPKα, t-elF2α, t-PERK, CHOP (C/EBP homologous protein or DNA Damage Inducible Transcript 3) (1:1000; Cell Signaling Technology), ATF6 (Activating Transcription Factor 6) (1:500; Abcam, Cambridge, UK), PPARδ (1:500, Cayman Chemical, Ann Arbor, MI, USA), and GAPDH (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies at 4°C overnight, followed by washing, and further incubation in horseradish peroxidase-conjugated anti-rabbit secondary antibody (1:10000; Li-Cor Bioscience, Bad Homburg, Germany).

2.11. Statistical analysis

The data are expressed as mean ± SEM. The relaxation was calculated as percentage reduction of the PE-induced contraction. Concentration–response curves were analysed using the Graphpad Prism 5.0 software. Data were analysed by one-way ANOVA, followed by Student–Newman–Keuls test. Differences between specific groups were analysed using a non-parametric Mann–Whitney test comparing two groups. Value of P < 0.05 was considered significant.

3. Results

3.1. Effect of maternal diabetes exposure on blood pressure and general parameters in the rat offspring

The plasma glucose levels were higher, and the plasma insulin levels were lower in STZ-treated dams than control dams. There were no differences in gestation term, mean blood pressure, maternal weight gain, and litter size between the two groups (Supplementary material online, Table S1). Hyperglycaemia in dams was ameliorated via insulin administration (Figure 1A). Blood pressure of conscious adult rat offspring was monitored by the non-invasive tail-cuff method at 24 weeks of age and then every week until 30 weeks of age. The systolic blood pressure was always higher in DMO than in CMO (Figure 1B). The blood pressure, measured directly in the cannulated left carotid artery, was also higher in DMO than in CMO at 30 weeks of age (Figure 1C). Insulin administration to the diabetic dams ameliorated the increased blood pressure of DMO (Figure 1D). As shown in Supplementary material online, Table S2, there were no differences in heart rate and fasting plasma glucose levels between the two groups at 30 weeks of age. However, consistent with our previous report,¹⁸ body weight, plasma insulin, and HOMA-IR were significantly higher in DMO than in CMO. Moreover, compared with CMO, the plasma levels of adiponectin and adipolin were reduced (Supplementary material online, Table S2); the plasma levels of leptin (Supplementary material online, Table S2) and TNF-α levels in the supernatant of primary cultures of mesenteric endothelial cells were higher in DMO (Supplementary material online, Figure S2). To assess the histological structure of the blood vessels, mesenteric arteries and thoracic aortae were stained with H&E. There were no obvious differences in intima–media thickness and luminal diameter of the mesenteric arteries and thoracic aortae between the two groups (Supplementary material online, Figure S3).

Figure 1.

Effect of ER stress inhibition, and glucose-lowering intervention in diabetic dams on blood pressure in offspring. (A) Plasma glucose levels of control and diabetic mothers during pregnancy. Con, control group; Dia, diabetes group; Dia + Ins group, diabetes + insulin group. ^*P < 0.05 vs. others (n = 6/group), one-way ANOVA, and subsequent Student–Newman–Keuls test. (B) Non-invasive blood pressure measurement by tail-cuff plethysmography in offspring rats from 24 to 30 weeks of age. Data are expressed as the means ± S.E.M (n = 6/group). ^*P < 0.05 vs. DMO + 4-PBA and DMO + Tudca, ^#P < 0.05 vs. others, one-way ANOVA and subsequent Student–Newman–Keuls test. (C) Invasive blood pressure measurement through left carotid artery cannulation in 30-week-old offspring rats. Data are expressed as the mean ± S.E.M (n = 6/group). ^**P < 0.01, one-way ANOVA, and subsequent Student–Newman–Keuls test. (D) Non-invasive blood pressure measurement by tail-cuff plethysmography in rat offspring from 4 to 24 weeks of age. Con, control group; Dia, diabetes group; Dia + Ins group, diabetes + insulin group. ^*P < 0.05 vs. others (n = 6/group), one-way ANOVA, and subsequent Student–Newman–Keuls test.

3.2. Effect of maternal diabetes exposure on vascular function in the rat offspring

To determine if maternal diabetes had an effect on vascular reactivity in the rat offspring, mesenteric arteries were treated with cumulative concentrations of PE (10⁻⁹–10⁻⁵mol/L) and KCl (125 mM). There was a significantly increased contractile response in mesenteric arteries from DMO compared with CMO (Figure 2A and  Supplementary material online, Figure S4A). Moreover, treatment with L-NAME augmented the PE-induced contraction in arteries from CMO but not DMO (Figure 2B), suggesting that the arteries of DMO may have a decreased basal production of NO. To assess the EDR of mesenteric arteries, the vessels were treated with Ach. The EDR was attenuated in DMO compared with CMO (Figure 2C). Pretreatment with L-NAME (100 µM) inhibited the Ach-induced relaxation in both CMO and DMO such that the Ach-induced relaxation was no longer different between them (Figure 2D). The nitric oxide donor, SNP (1 nM-10 μM), was used to determine the effect of endothelium-independent relaxation in these arteries. There was no difference in the SNP-induced relaxation in the arteries from CMO and DMO (Figure 2E). To determine the effect of glucose-lowering intervention in diabetic dams on the vascular endothelial function in DMO, diabetic dams at Day 0 of gestation were surgically implanted with insulin pellets subcutaneously. We found that insulin administration to the diabetic dams normalized the Ach-induced vasorelaxation in DMO (Figure 2F). To determine if endothelial dysfunction can occur before the development of hypertension, the EDR of mesenteric arteries was studied in 8-week-old DMO which were normotensive. We found that the EDR was attenuated in DMO compared with CMO (Supplementary material online, Figure S4B).

Figure 2.

Effect of ER stress inhibition, and glucose-lowering intervention in diabetic dams on the function of mesenteric arteries in offspring. Vascular functional evaluation included PE-induced contraction (A and B), Ach-induced relaxation (C and D), and SNP-induced relaxation (E) in mesenteric arteries from CMO and DMO. DMO at 24 weeks of age were treated with vehicle or ER stress inhibitors, 4-PBA (1 g/kg/day), or Tudca (150 mg/kg/day) for 6 weeks. (F) Ach-induced relaxation in mesenteric arteries in rat offspring at 24 weeks of age. Con, control group; Dia, diabetes group; Dia + Ins group, diabetes + insulin group. Data are expressed as the mean ± S.E.M (n = 6/group). ^*P < 0.05 vs. others, one-way ANOVA, and subsequent Student–Newman–Keuls test.

3.3. Increased ER stress is linked to endothelial dysfunction in adult offspring exposed to maternal diabetes

Disruption of normal ER homeostasis is accompanied by vascular endothelial dysfunction and increased blood pressure.⁹ To determine the role of ER stress in the vascular endothelial dysfunction in adult offspring exposed to maternal diabetes, the CMO and DMO were treated with ER stress inhibitors sodium 4-phenylbutyrate (4-PBA) and tauroursodeoxycholic acid (Tudca) for 6 weeks. Compared with control DMO, 4-PBA and Tudca treatments attenuated the PE-induced vasoconstriction (Figure 2A) and improved the Ach-induced vasorelaxation (Figure 2C). L-NAME treatment augmented PE-induced vasoconstriction (Supplementary material online, Figure S5A) and impaired Ach-induced vasorelaxation (Supplementary material online, Figure S5B) in DMO with 4-PBA and Tudca treatments. By contrast, in CMO, 4-PBA and Tudca treatments had no effect on PE-induced vasoconstriction (Supplementary material online, Figure S6A) and Ach-induced vasorelaxation (Supplementary material online, Figure S6B). The ability of L-NAME to augment the PE-induced vasoconstriction (Supplementary material online, Figure S6C) or impair the Ach-induced vasorelaxation (Supplementary material online, Figure S6D) was also not affected by 4-PBA or Tudca in CMO. These in vitro studies agreed with the ability of 4-PBA and Tudca treatments to decrease the blood pressure in DMO (Figure 1B and C) but had no effect on the blood pressure in CMO (Supplementary material online, Figure S7A and B). The high blood pressure in DMO was due mainly to increased endothelium-dependent vasoconstriction because the SNP-induced endothelium-independent relaxation in mesenteric arteries was not affected by ER stress inhibitors in the CMO or DMO (Supplementary material online, Figure S8 and Figure 2E).

ER stress is the result of complex intracellular signalling cascades that include three ER stress transduction pathways: PERK, ATF6, and IRE1. Our present study showed that ER stress markers, including the phosphorylations of PERK, IRE1, and elF2α, and the expressions of CHOP and ATF-6 were increased in DMO compared with CMO, which were normalized by treatment with ER stress inhibitors (Figure 3A– E). We also evaluated NO production by measuring the nitrate/nitrite level and the activity of eNOS by measuring the phosphorylation of eNOS at Ser¹¹⁷⁷, Thr⁴⁹⁵, and Tyr⁶⁵⁷. Compared with CMO, NO production was decreased in mesenteric tissue from DMO (Supplementary material online, Figure S9A). The phosphorylation of eNOS at Ser¹¹⁷⁷ was decreased (Figure 3F), and the phosphorylations of eNOS at Thr⁴⁹⁵ and Tyr⁶⁵⁷ were increased (Supplementary material online, Figure S9B–D) in primary cultures of mesenteric endothelial cells of DMO which were restored to CMO levels after ER stress inhibition. By contrast, in CMO, 4-PBA and Tudca had no effect on the expression of ER stress markers and the phosphorylation of eNOS at Ser¹¹⁷⁷, Thr⁴⁹⁵, and Tyr⁶⁵⁷ in primary cultures of mesenteric endothelial cells (Figure 3A–F and  Supplementary material online, Figure S9B–D). Total eNOS expression was similar in all groups (Figure 3F and  Supplementary material online, Figure S9B). Insulin administration to the diabetic dams significantly reduced ER stress (CHOP) and increased eNOS at Ser¹¹⁷⁷ phosphorylation in DMO, relative to DMO of insulin-untreated diabetic dams (Figure 3G and H).

Figure 3.

Effect of ER stress inhibition, and glucose-lowering intervention in diabetic dams on ER stress markers and eNOS phosphorylation in primary cultures of mesenteric endothelial cells from offspring. The expressions of phosphorylated (p) ER stress markers, such as p-PERK (A), p-IRE1 (B), CHOP (C), p-eIF2α (D), and ATF6 (E), and p-eNOS at Ser¹¹⁷⁷ (F) in mesenteric endothelial cells in primary culture were analysed by western blotting. Mesenteric endothelial cells were obtained from 30-week-old CMO and DMO. The CMO and DMO at 24 weeks of age were treated for 6 weeks with vehicle, or ER stress inhibitors, 4-PBA (1 g/kg/day), or Tudca (150 mg/kg/day). CHOP (ER stress markers) (G), and phosphorylation and expression of eNOS at Ser¹¹⁷⁷ (H) in mesenteric endothelial cells in primary culture were analysed by western blotting. Con, control group; Dia, diabetes group; Dia + Ins group, diabetes + insulin group. Data are expressed as the means ± S.E.M (n = 6/group). ^**P < 0.01, one-way ANOVA, and subsequent Student–Newman–Keuls test.

3.4. Role of AMPK/PPARδ signalling on the disruption of ER homeostasis and endothelial dysfunction in DMO

At present, little is known about the role of AMPK/PPARδ signalling on the disruption of ER homeostasis and endothelial dysfunction in mesenteric arteries of DMO. Our present study showed that in primary cultures of mesenteric endothelial cells from DMO, AMPKα phosphorylation (p-AMPK) at Thr¹⁷² and PPARδ expression were decreased (Figure 4A and B), whereas ER stress markers such as the expressions of CHOP and ATF6, and the phosphorylations of PERK, IRE1, and elF2α, were increased in DMO compared with CMO (Figure 3A– E). To determine the effect of glucose-lowering intervention in diabetic dams on AMPK/PPARδ signalling in offspring, diabetic dams at Day 0 of gestation were surgically implanted with insulin pellets subcutaneously. We found that insulin administration to the diabetic dams significantly increased AMPK phosphorylation and PPARδ expression in primary cultures of mesenteric endothelial cells from insulin-treated DMO, relative to the insulin-untreated DMO (Figure 4C and D).

Figure 4.

Effect of ER stress inhibition, and glucose-lowering intervention in diabetic dams on the phosphorylation of AMPKα and expression of PPARδ in primary cultures of mesenteric endothelial cells from offspring. The phosphorylation of AMPKα (A) and expression of PPARδ (B) in mesenteric endothelial cells in primary culture were analysed by western blotting. The CMO and DMO at 24 weeks of age were treated for 6 weeks with vehicle or ER stress inhibitors, 4-PBA (1 g/kg/day), or Tudca (150 mg/kg/day). The protein phosphorylation of AMPKα (C) and expression of PPARδ (D) in mesenteric endothelial cells in primary culture were analysed by western blotting. Con, control group; Dia, diabetes group; Dia + Ins group, diabetes + insulin group. Data are expressed as the mean ± S.E.M (n = 6/group). ^**P < 0.01, one-way ANOVA, and subsequent Student–Newman–Keuls test.

We wondered whether the inhibition of ER stress was AMPK/PPARδ signalling-dependent. Incubation with an AMPK agonist (A769662, 30 µmol/L) markedly improved the Ach-induced vasorelaxation in mesenteric arteries of DMO, which was inhibited by an AMPK antagonist (Compound C, 5 µmol/L) or a PPARδ antagonist (GSK0660, 500 nmol/L) (Figure 5A). A PPARδ agonist (GW1516, 100 nmol/L) also reversed the impairment of EDR in DMO but was unaffected by Compound C (Figure 5B), indicating that PPARδ is downstream to AMPK.

Figure 5.

Effect of AMPK/PPARδ signalling pathway on the endothelial function of mesenteric arteries from DMO and CMO incubated with tunicamycin. (A and B) Ach-induced relaxation in mesenteric arteries from DMO and CMO incubated with A769662, AMPK agonist (30 µmol/L), or GW1516, PPARδ agonist (100 nmol/L) alone or in the presence of Compound C, AMPK inhibitor (5 µmol/L), GSK0660, PPARδ antagonist (500 nmol/L) for 16 h. ^*P < 0.05 vs. others (n = 6/group). (C and D) Ach-induced relaxation in mesenteric arteries from CMO incubated with tunicamycin (Tuni, 2 µg/mL), A769662 (30 µmol/L), GSK0660 (500 nmol/L), GW1516 (100 nmol/L), or compound C (5 µmol/L) for 16 h. Data are expressed as the means ± S.E.M (n = 6/group). ^*P < 0.05 vs. others, one-way ANOVA, and subsequent Student–Newman–Keuls test.

A positive role of ER stress in endothelial dysfunction was also demonstrated by impaired EDR, induced by ex vivo exposure of mesenteric arteries to tunicamycin (ER stress inducer, 2 µg/mL) in mesenteric arteries of CMO (Figure 5C and D). Co-incubation with the AMPK agonist A769662 (30 µmol/L) or the PPARδ agonist GW1516 (100 nmol/L) restored close to normal the tunicamycin-induced impairment of EDR (Figure 5C and D). Tunicamycin had no effect on AMPK phosphorylation or PPARδ expression (Figure 6A– C). In the presence of tunicamycin, the AMPK agonist A769662 increased the levels of AMPK phosphorylation and PPARδ expression while the PPARδ agonist GW1516 had no effect on AMPK phosphorylation but increased PPARδ expression (Figure 6A– C). These results suggest that PPARδ is downstream of AMPK.

Figure 6.

Effect of AMPK/PPARδ signalling pathway on ER stress and eNOS at Ser¹¹⁷⁷ phosphorylation in primary cultures of mesenteric endothelial cells from 30-week-old CMO, incubated with tunicamycin. A and D, Representative bands. B, C, E, F, G, and H, Densitometry of western blots. Phosphorylation and expression of AMPKα (B), expression of PPARδ/GAPDH (C), the levels of ER stress markers p-PERK, p-IRE1, and ATF6 (E–G), and phosphorylation at Ser¹¹⁷⁷ and expression of eNOS (H) in mesenteric endothelial cells in primary culture from CMO, incubated with tunicamycin (Tuni, 2 µg/mL), A769662 (30 µmol/L), GSK0660 (500 nmol/L), GW1516 (100 nmol/L), or Compound C (5 µmol/L) for 16 h. Data are expressed as the means ± S.E.M (n = 6/group). ^**P < 0.01, one-way ANOVA, and subsequent Student–Newman–Keuls test.

Tunicamycin increased the levels of ER markers, including ATF6 expression, and PERK and IRE phosphorylations, in primary cultures of mesenteric endothelial cells from CMO, effects that were inhibited by the AMPK agonist A769662 and PPARδ agonist GW1516 (Figure 6D– G). The PPARδ antagonist GSK0660 blocked the ability of the AMPK agonist A769662 to inhibit the tunicamycin-induced increase in the expression or phosphorylation of ER markers; the AMPK antagonist Compound C had no additional effect on the PPARδ agonist GW1516-mediated inhibition of tunicamycin-induced increase in these ER stress markers (Figure 6D– G). The tunicamycin-induced decrease in the phosphorylation of eNOS at Ser¹¹⁷⁷ was also normalized by the AMPK agonist A769662 and the PPARδ agonist GW1516 (Figure 6H). However, the PPARδ antagonist GSK0660 prevented the ability of the AMPK agonist A769662 to increase the phosphorylation of eNOS at Ser¹¹⁷⁷, and the AMPK antagonist, Compound C did not prevent the ability of PPARδ agonist GW1516 to increase in the phosphorylation eNOS at Ser¹¹⁷⁷, in agreement with the ex vivo studies, indicating that PPARδ is downstream to AMPK.

3.5. AMPK/PPARδ signalling activation ameliorates ER-stress-associated oxidative stress in rat mesenteric arteries

ER stress has been reported to initiate a burst of oxidative stress in the ER lumen by targeting the mitochondria which triggers the increase in reactive oxygen species (ROS) production.¹⁹ The mesenteric artery of DMO, relative to CMO, had increased ROS and NADPH oxidase activity (Figure 7A and B). Treatment with ER stress inhibitors, 4-PBA and Tudca, for 6 weeks normalized the increased oxidative stress in DMO but had no effect in CMO (Figure 7A and B). Exposure of mesenteric arteries of CMO to tunicamycin (2 µg/mL, 16 h) markedly increased ROS production (Supplementary material online, Figure S10). The increase in ROS elevation was reversed by co-treatment with AMPK agonist A769662 (30 µmol/L) or PPARδ agonist GW1516 (100 nmol/L). The PPARδ antagonist GSK0660 (500 nmol/L) prevented the ability of the AMPK agonist A769662 to decrease ROS generation. However, the ability of the PPARδ agonist GW1516 to inhibit the ROS production induced by tunicamycin was not affected by Compound C (Supplementary material online, Figure S10), indicating that PPARδ is downstream of AMPK. To confirm that oxidative stress promotes endothelial dysfunction in DMO, we treated the mesenteric arteries of DMO with the antioxidant tempol. The results showed that antioxidant treatment with tempol ex vivo rescued the impaired Ach-induced relaxation (endothelium-dependent) of mesenteric arteries from DMO; there was no effect in CMO (Figure 7C).

Figure 7.

Effect of maternal diabetes on ROS production, NADPH oxidase activity, and oxidative stress-induced endothelial dysfunction in mesenteric arteries from offspring. CMO and DMO at 24 weeks of age were treated for 6 weeks with vehicle, or ER stress inhibitors, 4-PBA (1 g/kg/day), or Tudca (150 mg/kg/day). (A) Representative images and quantified DHE fluorescence in rat mesenteric arteries. (B) NADPH oxidase activity was measured by lucigenin chemiluminescence in lysates of rat mesenteric arteries. Data are expressed as the mean ± S.E.M (n = 6/group). ^**P < 0.01, one-way ANOVA, and subsequent Student–Newman–Keuls test. (C) Ach-induced relaxation in mesenteric arteries from CMO and DMO was detected. The mesenteric arteries from 30-week-old CMO and DMO were incubated with tempol (10 µmol/L) for 16 h. ^*P < 0.05 vs. others (n = 6/group), one-way ANOVA, and subsequent Student–Newman–Keuls test.

4. Discussion

Many animal and epidemiological studies suggest that in addition to genetic and environmental factors, an abnormal intrauterine environment during a critical period of foetal development produces lifelong adverse consequences in the offspring, for example, increased risk for hypertension in adult life.²⁰ The present study demonstrated that exposure to maternal diabetes resulted in foetal-programmed vascular dysfunction and hypertension in the offspring through increased ER stress, via downregulation of the AMPK/PPARδ signalling pathway.

In utero exposure to maternal diabetes is associated with the development of CVD, including hypertension, in the offspring.²¹ An epigenetic downregulation of Sirt1 gene²² and blunted AMPK response²³ are associated with susceptibility to heart ischemia/reperfusion injury in offspring. Several mechanisms related to hypertension have been proposed to cause hypertension in offspring, such as dysregulation of the sympathetic nervous system²⁴ and/or the renin-angiotensin system,²⁵ increased oxidative stress²⁶ and cyclooxygenase activity,²⁷ and abnormal vascular structure.³ We have reported that programming of adult hypertension in DMO was related to increased renal oxidative stress and impaired renal dopamine D₁ receptor function.¹⁸ In the current study, we extended those observations in the kidney to the vasculature. In utero exposure to maternal diabetes has been reported to induce structural remodelling of conductance and resistance vessels.³ However, consistent with another report,²⁸ our present study showed no difference in intima–media thickness or luminal diameter of the mesenteric arteries or thoracic aorta between CMO and DMO. The reason for this discrepancy is not clear but is not age-related; Dib et al.³ found the abnormal vascular structure as early as 3 months of age while Vessieres et al.²⁸ and we studied rats at 3-8 months of age. However, in the hypertensive state, the reorganization of the vessel wall (increased connections between vascular smooth muscle cells and extracellular matrix) could contribute to the maintenance of a normal level of wall stiffness despite the increase in the pressure wall stress.

Impairment of vascular function has been observed in offspring of dams fed high-fat diet,²⁹ low-nutrition diet,³⁰ as well as dams with inflammation³¹ or diabetes,³² among others. Those studies have provided evidence that the vascular dysfunction in the offspring is mainly related to impaired endothelial function caused by abnormalities in the NO pathway, resulting in vasoconstriction. Endothelial dysfunction is a likely cause of the functional changes in the microvasculature that are important in the pathogenesis of hypertension. The reduction in NO-mediated vasodilation is important in the initiation of hypertension, but NO deficiency in the kidney also impairs the ability to excrete a sodium load, perpetuating the hypertension.³³ However, the underlying mechanism of the vascular endothelial dysfunction and programming of hypertension in offspring of mothers with diabetes is still not clear.

Emerging evidence suggests that ER stress plays an important role in mediating the development and progression of hypertension.⁶ ER stress is regulated by three ER membrane-associated proteins that are engaged in complex signalling pathways: cleavage of ATF6 and activation of IRE-1 and PERK. These proteins are upregulated in the aortae of spontaneously hypertensive rats and obese mice, and contribute to vascular endothelial dysfunction.^9,12 The chemical chaperone 4-PBA and endogenously produced bile acids, such as Tudca, are known to modulate ER function, stabilize protein conformation, improve ER folding capacity, and facilitate mutant protein trafficking.³⁴ In the present study, we found that pharmacological inhibition of ER stress normalized the impaired EDR and decreased the blood pressure of DMO, indicating that ER stress is involved in the vascular endothelial dysfunction in DMO. To determine a causal link between ER stress and vascular endothelial dysfunction, ER stress was induced by tunicamycin in mesenteric arteries of CMO. We found that in CMO, tunicamycin directly impaired endothelial function that was related to ER stress, proved by the increase in the expression of ER markers, p-ERK, p-IRE1, and ATF-6 and reduction in eNOS activity in mesenteric arteries. More importantly, alleviation of ER stress in DMO in vivo with 4-PBA or Tudca rescued the endothelial dysfunction and decreased the elevated pressure. These results demonstrate the relationship between hypertension, vascular endothelial dysfunction, and ER stress.

AMPK, which is important in the regulation of cellular energy, is also an important regulator of vascular homeostasis. AMPK suppresses vascular smooth muscle cell proliferation and migration and exerts a direct vasorelaxing effect in isolated aortic rings.^35,36 Extensive studies have shown that AMPK exerts its therapeutic effects by suppressing ER stress, promoting eNOS activity, and improving vascular endothelial functions in diabetes mellitus and hypertension.^37,38 Maternal hyperglycaemia has also been found to decrease mRNA expression of hepatic AMPK pathway genes, which are associated with metabolic disorders and insulin resistance in adult offspring.³⁹ Indeed, we found that AMPKα phosphorylation was down-regulated in mesenteric arteries from DMO and that AMPK activation was able to improve vascular endothelial function and rescue ER stress-induced endothelial dysfunction in DMO.

PPARδ promotes cardiovascular function by induction of angiogenesis, inhibition of atherosclerosis, and protection of endothelial function in diabetic mice and diet-induced obese mice.^40,41 In this study, we found that PPARδ was down-regulated in mesenteric arteries from DMO. This result agrees with previous studies showing that maternal diabetes causes the down-regulation of PPARδ expression in the heart⁴² and liver⁴³ of the offspring. We found that PPARδ mediated the ability of A769662 (AMPK activator) to attenuate ER stress and the subsequent endothelial dysfunction in mesenteric arteries from DMO. Thus, our studies provide solid evidence that supports AMPK/PPARδ as upstream of ER stress. Although ER stress inhibitors restored p-AMPK and PPARδ levels in DMO, we speculated that the effect could be a secondary to another mechanism. Previous studies showed that the ROS scavenger thioredoxin 1 cleaves disulphides in proteins, prevents AMPK oxidation and maintains its activity,⁴⁴ suggesting that the down-regulation of ROS may be the link by which ER stress inhibitors restore p-AMPK and PPARδ levels in DMO. The precise mechanism on how AMPK and PPARδ negatively regulate ER stress remains incompletely understood, but, PPARδ activation can directly inhibit the transcription of ATF4, which is a transcriptional effector of the PERK branch of the ER stress/UPR pathway.⁴⁵ Further investigation is needed to uncover what PPARδ responsive genes are related to protein degradation and regulation of ER stress in the vasculature.

Increased oxidative stress plays an important role in the vascular endothelial dysfunction observed in hypertension.⁴⁶ ER stress initiates a burst of oxidative stress in the ER lumen and by targeting the mitochondria, triggers the elevation of cytosolic Ca²⁺ and ROS production.¹⁹ Moreover, ER stress also causes eNOS uncoupling related to ROS generation and increased NADPH oxidase activity in endothelial cells.⁴⁷ Our present study showed that maternal diabetes exposure led to increased ROS production in mesenteric arteries of the offspring, and ER stress inhibition abolished superoxide generation and normalized NADPH oxidase activity and phosphorylation of eNOS in DMO. Inhibition of tunicamycin-induced ER stress in mesenteric arteries from CMO also abolished the tunicamycin-induced superoxide generation. These studies support a causal link between ER stress and its downstream activation of oxidative stress. Indeed, the superoxide scavenger, tempol, restored the vasorelaxant effect of Ach in DMO without affecting the vasorelaxant effect of Ach in CMO. The mechanisms by which oxidative stress impairs NO bioavailability has been reported including a reduction in endothelial NO synthase, an uncoupling of eNOS enzymatic activity, and oxidation of the NO targets.⁴⁸ In addition, superoxide (O₂⁻)-mediated scavenging of NO generates peroxynitrite and decreases NO bioavailability.⁴⁹

As reported in previous studies, the mechanism of Ach-mediated vasodilation is complex. In addition to NO, other factors, such as prostacyclin and endothelium-derived hyperpolarizing factor (EDHF), are also involved.⁵⁰ EDHF hyperpolarizes vascular smooth muscle cells by opening K⁺ channels and stimulating Na⁺ pump activity that are crucial in regulating vascular tone.⁵¹ In the present study, we showed that inhibition of NO production by L-NAME could not completely block the Ach-mediated relaxation in DMO, indicating that the EDHF pathway could be involved in the regulation of the vascular tone and could have partially compensated for the diminished NO production in mesenteric arteries from DMO.

Limitation

We used STZ to induce type I diabetes in the dams and found that the offspring of diabetic dams develop hypertension. It should be noted that most patients seen in the clinic have type II diabetes. However, there is also an association between maternal diabetes and early onset CVD in offspring of mothers with either type 1 diabetes or type 2 diabetes. Intrauterine exposure to mothers with either type 1 diabetes (n = 22,055) or type 2 diabetes (n = 6,537) leads to an excess risk of early onset CVD, particularly heart failure and hypertension, in offspring (up to 40 years of age).⁵² Whether or not a phenomenon occurs in type II diabetic animal models, similar to that found in type 1 diabetic animal models, is not known and needs to be determined in the future.

The lifelong adverse consequences of abnormal intrauterine environment during pregnancy²⁰ may involve epigenetics. For example, CpG hypermethylation of the sterol regulatory element binding transcription factor 2 promoter in the foetal liver and brain is associated with disorders of neuronal and metabolic functions in the offspring.⁵³ In the present study, we found that impaired AMPK/PPARδ signalling cascade with subsequent enhancement of ER stress and ROS production, and a decrease in eNOS activity, play an important role in the vascular endothelial dysfunction in the offspring of diabetic dams. However, how hyperglycaemia in the maternal dam affects AMPK/PPARδ signalling cascade in the offspring is not completely clear. It is possible that epigenetic changes could be the key link between the dam and the offspring, which needs to be determined in the future.

In conclusion, the present study shows that foetal exposure to maternal diabetes programs mesenteric endothelial dysfunction in the offspring through down-regulation of AMPK/PPARδ signalling cascade with a subsequent enhancement of ER stress and ROS production, and a decrease in eNOS activity. The AMPK/PPARδ signalling cascade could be a potentially effective target in the treatment of foetal-programmed hypertension.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Author’s contributions

H.L. and C.L. developed the hypothesis and designed the experiments; C.F. and C.Y. contributed to western blot; C.C., X.L., and P.A.J. performed the data analyses and revision of the manuscript; J.W., X.G., and Z.X. performed the data acquisition; and H.L., C.H., and C.Z. drafted the manuscript. All authors revised the manuscript and approved the final version.

 

Translational Perspective.

Increasing evidence indicates that foetal programming is an important factor that contributes to the development of cardiovascular disease including hypertension and atherosclerosis later in life. This study explains the roles of AMPK/PPARδ/ER stress signalling cascade and endothelial dysfunction in maternal diabetes-programed adult hypertension in the offspring and provides a potential target for the prevention and therapy of this disease.