EPAS1 Attenuates Atherosclerosis Initiation at Disturbed Flow Sites Through Endothelial Fatty Acid Uptake

Abstract

Background: Atherosclerotic plaques form unevenly due to disturbed blood flow, causing localized endothelial cell (EC) dysfunction. Obesity exacerbates this process, but the underlying molecular mechanisms are unclear. The transcription factor EPAS1 (HIF2A) has regulatory roles in endothelium, but its involvement in atherosclerosis remains unexplored. This study investigates the potential interplay between EPAS1, obesity, and atherosclerosis.

Methods: Responses to shear stress were analyzed using cultured porcine aortic EC exposed to flow in vitro coupled with metabolic and molecular analyses and by en face immunostaining of murine aortic EC exposed to disturbed flow in vivo. Obesity and dyslipidemia were induced in mice via exposure to a high-fat diet or through Leptin gene deletion. The role of Epas1 in atherosclerosis was evaluated by inducible endothelial Epas1 deletion, followed by hypercholesterolemia induction (adeno-associated virus-PCSK9 [proprotein convertase subtilisin/kexin type 9]; high-fat diet).

Results: En face staining revealed EPAS1 enrichment at sites of disturbed blood flow that are prone to atherosclerosis initiation. Obese mice exhibited substantial reduction in endothelial EPAS1 expression. Sulforaphane, a compound with known atheroprotective effects, restored EPAS1 expression and concurrently reduced plasma triglyceride levels in obese mice. Consistently, triglyceride derivatives (free fatty acids) suppressed EPAS1 in cultured EC by upregulating the negative regulator PHD2. Clinical observations revealed that reduced serum EPAS1 correlated with increased endothelial PHD2 and PHD3 in obese individuals. Functionally, endothelial EPAS1 deletion increased lesion formation in hypercholesterolemic mice, indicating an atheroprotective function. Mechanistic insights revealed that EPAS1 protects arteries by maintaining endothelial proliferation by positively regulating the expression of the fatty acid-handling molecules CD36 (cluster of differentiation 36) and LIPG (endothelial type lipase G) to increase fatty acid beta-oxidation.

Conclusions: Endothelial EPAS1 attenuates atherosclerosis at sites of disturbed flow by maintaining EC proliferation via fatty acid uptake and metabolism. This endothelial repair pathway is inhibited in obesity, suggesting a novel triglyceride-PHD2 modulation pathway suppressing EPAS1 expression. These findings have implications for therapeutic strategies addressing vascular dysfunction in obesity.

Keywords: atherosclerosis; diet, high-fat; endothelial cells; obesity; plaque, atherosclerotic.

Figure 1.

EPAS1 is enriched at an atheroprone site exposed to low and oscillatory shear stress. A, PAEC were seeded on μ-slides and cultured under high shear stress (HSS), low shear stress (LSS), or low oscillatory shear stress (LOSS) for 72 hours using the Ibidi system. Protein levels of EPAS1 were quantified by immunoblotting. Representative images and mean values normalized to the level of PDHX (pyruvate dehydrogenase complex component X) (n=4) are shown. Differences between means were analyzed using a Kruskal-Wallis test. B, Time-averaged wall shear stress (WSS) (TAWSS) and oscillatory shear index (OSI) were mapped onto the geometry of the murine aortic arch. Representative images are shown and regions exposed to LOSS (inner) or HSS (outer) are marked. C, Aortic arches were isolated from C57BL/6J mice aged 6 to 8 weeks, and en face immunostaining was performed using anti-EPAS1 antibodies (red). Endothelium was costained (EC; green) and nuclei detected using DAPI (4′,6-diamidino-2-phenylindole; blue). Scale bar: 20 µm. EPAS1 levels were quantified at LOSS and HSS regions (n=7). Each data point represents an animal. Differences between means were analyzed using a paired t test.

Figure 2.

High-fat feeding and obesity reduced EPAS1 levels at atheroprone aortic endothelium. A, C57BL/6N mice aged 5 weeks were exposed to high fat diet (HFD) (n=6) or to standard chow (n=11) for 25 weeks. B, Lep^ob/ob mice (n=6) and littermate controls (wild-type, n=5) aged 22 weeks were analyzed. A and B), Aortic endothelial cells were stained en face using anti-EPAS1 antibodies (red). Endothelium was costained (CD31; green) and nuclei detected using DAPI (4′,6-diamidino-2-phenylindole; blue). Scale bar: 50 µm. EPAS1 fluorescence levels were quantified at low oscillatory shear stress (LOSS) and high shear stress (HSS) regions, and mean±SDs are presented. Each data point represents an animal. Differences between means were analyzed using a 2-way ANOVA with Sìdak’s multiple comparisons test in A, while B was analyzed using an aligned ranked transform ANOVA test. MFI indicates mean fluorescence intensity.

Figure 3.

Hyperglycemia had no statistically significant effect on EPAS1. Mice (C57BL/6J) aged 20 weeks were treated with streptozotocin (STZ) or vehicle control (initial) and analyzed 2 weeks later (final), n=5 animals per group. Glycemia (A) and plasma triglycerides (B) were measured. C, Endothelial cells were stained en face using anti-EPAS1 antibodies (red). Endothelium was costained (CD31; green) and nuclei detected using DAPI (4′,6-diamidino-2-phenylindole; blue). Scale bar: 50 µm. EPAS1 fluorescence levels were quantified at low oscillatory shear stress (LOSS) and high shear stress (HSS) regions, and mean±SDs are presented. Each data point represents an animal. Differences between means were analyzed using an aligned ranked transform ANOVA test (B and C). MFI indicates mean fluorescence intensity.

Figure 4.

Sulforaphane rescues EPAS1 in obese mice. C57BL/6N mice aged 5 weeks were exposed to high fat diet (HFD) for 25 weeks (pretreatment). They were then treated with sulforaphane (SFN; daily IP injections 5 mg/kg for 3 days) or with vehicle for 3 days, with both groups receiving a HFD for that period, n=6 animals per group. Plasma triglycerides (A) and glycemia (B) were measured pretreatment and in SFN-treated and vehicle-treated groups. C, Aortic endothelial cells (EC) were stained en face using anti-EPAS1 antibodies (red), and fluorescence was quantified at low oscillatory shear stress (LOSS) and high shear stress (HSS) regions in SFN-treated and control groups. Endothelium was costained (EC; green) and nuclei detected using DAPI (4′,6-diamidino-2-phenylindole; blue). Each data point represents an animal. Differences between means were analyzed using a 2-way ANOVA using a Sìdak’s multiple comparison test (A and B) or an aligned ranked transform ANOVA test (C).

Figure 5.

Free fatty acids suppress EPAS1 via the destabilizing enzyme PHD2. A, Porcine aortic endothelial cells were exposed to low oscillatory shear stress (LOSS) for 72 hours using the orbital system in the presence or absence of OA (0.25 mmol/L) or PA (0.25 mmol/L). Some cultures were treated with sulforaphane (SFN; 10 μM) or NAC (n-acetyl cysteine) (1 mM) either alone or together with OA or PA for 24 hours. Protein levels of EPAS1, PHD2, and NRF2 were quantified by immunoblotting. Representative images and mean values normalized to the level of α-tubulin (n=3) are shown. Differences between means were analyzed using an aligned ranked transform ANOVA test. B, C57BL/6J mice aged 10 weeks were exposed to high fat diet (HFD) or control chow diet for 10 weeks, n=6 per group. Aortic arch endothelial cells (EC) were stained en face using anti-PHD2 antibodies (red), and fluorescence was quantified at the LOSS region. Endothelium was costained (EC; green) and nuclei detected using DAPI (4′,6-diamidino-2-phenylindole; blue). Scale bar: 20 µm. Each data point represents an animal. Differences between means were analyzed using a nonparametric t test. MFI indicates mean fluorescence intensity.

Figure 6.

Alteration of the prolyl hydroxylase domain (PHD)-EPAS1 pathway in clinical obesity. A through C, Obese (body mass index >38; n=15) and nonobese subjects (n=15) were analyzed. Serum levels of EPAS1 (A), free fatty acids (FFAs) were measured (n=10) obese and (n=8) nonobese controls (B), and total antioxidant capacity (TAC) was measured in obese and nonobese subjects, n=15 each group. Mean values are shown with SEs. D through F, Endothelial cells (EC) were isolated from adipose samples from obese (n=12) and nonobese (n=13) subjects. Levels of PHD2 (D), PHD3 (E), and EPAS1 (F) transcripts were quantified by quantitative real-time polymerase chain reaction. Each data point represents a human subject. Differences between means were analyzed using a Mann-Whitney U test (A, C, and F) or a parametric t test (B, D, and E).

Figure 7.

Endothelial Epas1 protects against atherosclerosis. A, Timeline of Epas1 deletion in a model of hypercholesterolemia. Epas1^EC-KO mice aged 6 weeks and Epas1^EC-WT mice received 5 intraperitoneal injections of tamoxifen and 1 injection of PCSK9-adeno-associated virus at specified time points. After 8 weeks fed with a high-fat diet, the mice were culled and plaque area quantified. B, Plasma cholesterol levels were measured in Epas1^EC-KO mice and Epas1^EC-WT controls, n=7 per group. C, Quantification of plaque burden in the aorta was determined by calculating the percentage of aortic surface area covered by plaque for Epas1^EC-KO mice and Epas1^EC-WT controls, n=7 per group. D, Quantification of plaque burden in the aortic roots of Epas1^EC-KO mice (n=6) and Epas1^EC-WT controls (n=5). Scale bar: 200 µm. Each data point represents 1 mouse, and mean±SEMs are shown. Differences between means were analyzed using an unpaired t test (A through C) or a Mann-Whitney U test (D).

Figure 8.

Endothelial EPAS1 controls fatty acid metabolism via LIPG (endothelial type lipase G) and CD36 (cluster of differentiation 36) expression. A through C, PAEC were treated with small hairpin RNA (shRNA) targeting EPAS1 (n=3) or with scrambled control (n=4) and exposed to low oscillatory shear stress (LOSS) for 72 hours using the orbital system. A, Some cultures were exposed to exogenous OA (0.25 mmol/L). Basal oxygen consumption rates (OCR) were measured. Each dot represents a technical replicate from 3 PAEC donors: control shRNA (n=20), EPAS1 shRNA KD (n=24), and EPAS1 KD treated with OA (n=8). Average values are shown ±SEs. Expression of CD36 (B) or LIPG (C) was quantified by quantitative real-time polymerase chain reaction, and mean±SEs are shown. D and E, Epas1^EC-KO mice and littermate controls lacking Cre (Epas1^EC-WT) were injected with tamoxifen aged 6 weeks and analyzed 2 weeks later, n=6 each group. En face staining of LOSS and high shear stress (HSS) regions of the aortic arch using anti-CD36 (D) or anti-LIPG (E) antibodies. Endothelium was costained (endothelial cell; green) and nuclei detected (DAPI [4′,6-diamidino-2-phenylindole]; blue). Scale bar: 20 µm. Each data point represents an animal. Differences between means were analyzed using a 1-way ANOVA (A), a Mann-Whitney U test (B and C), or 2-way ANOVA (D and E).