Cell culture

Human aortic endothelial cells from healthy individuals (HAEC) and from patients with diabetes (D-HAEC) were purchased from Lonza (Basel, Switzerland). The cells were grown in fibronectin-coated 75 cm2 flasks with endothelial cell basal growth media-2 (EBM2; Lonza) supplemented with EGM-2 microvascular endothelial SingleQuotsTM kit (Lonza) and 10% fetal calf serum (FCS; Gibco; Waltham, MA, USA). By using trypsin–EDTA solution (Sigma-Aldrich; Saint Louis, MI, USA) for 3 min, cells were detached and reseeded in fibronectin-coated cell culture dishes. HAEC were cultured in EGM-2 containing 2% FCS and exposed for 20 h either to normal (5 mmol/l) or high glucose (25 mmol/l) concentrations.

To test the effect of pharmacological inhibition of methyltransferase EZH2 or NADPH oxidase, HAEC and/or D-HAEC were treated with EZH2 methyltransferase inhibitor GSK126 (0.5–15 µmol/l; GlaxoSmithKline; Brentford, UK) or NADPH oxidase inhibitor apocynin (100 µmol/l; Sigma-Aldrich; St. Louis, MI, USA), respectively. Treatments with corresponding vehicles (DMSO and ethanol, respectively) were used as control experiments. Since GSK126 (5 µmol/l) did not affect cell viability (Additional file 1: Fig. S1), we used this concentration throughout all the experiments.

Cell viability assay

Cell viability of HAEC treated with GSK126 was assessed by a commercially available MTT assay kit (Millipore; Burlington, MA, USA). Briefly, cells were exposed for 20 h with increasing concentrations of GSK126 0.5–15 µmol/l. MTT solution (100 µl) was added followed by incubation at 37 ºC for 4 h. Then, isopropanol with 0.04 N HCl was added for colour development. Absorbance was measured on a VersaMax microplate reader (Molecular Devices; San José, CA, USA) at a test and reference wavelength of 570 nm and 630 nm, respectively.

EZH2 silencing, and UTX and JMJD3 overexpression

HAEC were transfected with targeted double-stranded siRNA oligonucleotides against methyltransferase EZH2 or its negative control (Thermo Fisher Scientific; Waltham, MA, USA) for silencing studies. HAEC were also transfected with a plasmid vector coding for UTX (pCMV-HA-UTX; ID: 24168), JMJD3 (pCMV-HA-JMJD3; ID: 24167), or their corresponding empty vector (pCMV-HA; ID: 32530) (Addgene; Watertown, MA, USA). Transfection was performed using Lipofectamine 2000 Transfection Reagent (Invitrogen; Waltham, MA, USA) following the manufacturer’s instructions. Briefly, the cells were washed with D-PBS (Thermo Fisher Scientific) and submerged into OPTI-MEM medium (Thermo Fisher Scientific) containing 100 nmol/l of the corresponding siRNA or 3 µg of corresponding plasmid vector for 24 h. Thereafter, the transfection medium was removed and a fresh culture medium was added. Finally, after 2 or 3 days of transfection for overexpression or silencing studies, respectively, cells were collected for RT-qPCR, immunoblotting, and ESR spectroscopy analysis. All constructs were verified by DNA sequencing, and transfection efficiency was checked by RT-qPCR and immunoblotting.

Chromatin immunoprecipitation coupled with quantitative PCR (ChIP-qPCR) assay

Chromatin immunoprecipitation (ChIP) assay was performed in HAEC exposed to either normal or high concentrations of glucose in the presence and in the absence of GSK126 using the Magna ChIP assay kit (Millipore) as described before [24]. HAEC were fixed in 1% of paraformaldehyde (PFA) solution (Santa Cruz Biotechnology; Dallas, TX, USA). Cross-linking was quenched by the addition of 125 mM of glycine (Sigma) for 10 min. After quenching, cells were scraped in D-PBS and centrifuged. Cells were then lysed in SDS lysis buffer and sonicated to obtain chromatin fragments of 200–500 bp using a water bath sonicator (Diagenode; Denville, NJ, USA) 30 s “on”/ 30 s “off” for 15 min. Immunoprecipitation of soluble chromatin was carried out by using specific antibodies against H3K27me3 (Millipore) and JunD (Abcam; Cambridge, UK). Mouse IgG antibody (Abcam) was used as a negative control. Antibody-bound chromatin fractions were precipitated with dynabeads coated with protein A (Invitrogen). Purified DNA sequences were detected by using real-time qPCR (RT-qPCR) and specific primers against SOD1, SOD2, JunD, and NOX4 promoters (Additional file 1: Table S1). Primers for the promoter region (TSS ± 1000 bp) of each gene were designed using the Oligo Perfect Primer Designer software (Thermo Fisher Scientific). ChIP-qPCR quantifications were performed using the comparative cycle threshold method and reported as the percentage of the antibody-bound chromatin against the total input DNA.

RNA extraction and real-time (RT-qPCR)

As previously described [25], total RNA was extracted from HAEC and D-HAEC of all experimental groups using TRI Reagent® (Sigma) and Direct-zol™ RNA miniprep kit (Zymo Research; Irvine CA, USA), following the manufacturer’s instructions. Quantification of recovered RNA was assessed using a NanoDrop2000 spectrophotometer (Thermo Fisher Scientific). RNA was retrotranscribed to cDNA using the high-capacity cDNA conversion kit (Applied Biosystems; Foster City, CA, USA). Transcript levels were quantified by RT-qPCR with the FastStart Universal SYBR Green Master Mix (Roche; Basel, Switzerland) on an ABI 7900HT PCR cycler (Applied Biosystems). The cDNA was amplified with a thermal profile at the following conditions: hot-start activation at 95 °C for 10 min, followed by 40 cycles of denaturation (95 °C for 10 s), then annealing/extension (60 °C for 30 s), and finally, a dissociation cycle (melting curve; 60 to 95 °C, increasing 0.5 °C / 30 s) to verify that only one product was amplified. The primer concentrations were 10 pmol. Expressions of genes of interest were normalized by the expression of the housekeeping gene β-Actin and the relative quantification was calculated using the ΔCT formula. All samples for each experiment were run on the same plate in triplicate and the average values were calculated. Primers used in RT-qPCR are enlisted in Additional file 1: Table S2.

Immunoblotting

Total proteins from HAEC in all experimental groups were extracted using native lysis buffer (Abcam) and cell debris was removed by centrifugation for 10 min at 12,000 × g at 4 °C. Immunoblotting was performed as previously reported [25]. Briefly, 30 µg of protein were subject to SDS-PAGE gel electrophoresis followed by transfer onto Immonobilon-P filter papers (Millipore) using the Trans-Blot Turbo Transfer System (Bio-Rad, Hercules, CA, USA). Membranes were blocked with 5% skimmed milk or 5% BSA in 0.1% Tween 20 (Sigma) in PBS for 1 h at room temperature (RT) and incubated overnight at 4 °C with specific primary antibodies against H3K27me3 (Millipore), H3 (Abcam), EZH2 (Abcam), UTX (Abcam), JMJD3 (Invitrogen), SOD1 (Abcam), SOD2 (Upstate Biotechnology), NOX4 (Santa Cruz Biotechnology) (1:1000), and GAPDH (Sigma) (1:10,000). Anti-rabbit and anti-mouse secondary antibodies were purchased from Santa Cruz Biotechnology and used at 1:10,000. The immunoreactive bands were detected by an enhanced chemiluminescence system (Millipore, USA). The intensity of the bands was quantified by densitometry using Image J v2.9.0 (NIH, Bethesda, MD, USA) and normalized by H3 (for H3K27me3) or GAPDH (for SOD1, SOD2, and NOX4) signals.

Immunofluorescence studies

EZH2 and H3K27me3 were immunostained in HAEC and D-HAEC using specific antibodies. Briefly, cells grown on glass coverslips were fixed in 4% PFA for 10 min, permeabilizated with 0.1% Triton™ X-100 (Sigma) in D-PBS for 10 min, blocked with 1% BSA (Sigma) for 1 h at RT, and incubated with 1:100 primary anti-EZH2 (Thermo Fisher Scientific) and anti-H3K27me3 (Millipore) antibodies overnight at 4 ºC. After washing with PBS, 1:200 Alexa Fluor 594- and 488-conjugated antibodies (Abcam) were used as secondary antibodies for 2 h in darkness at RT, and nuclei were stained with Hoechst (2 µg/ml in PBS; Sigma) for 5 min. Aortic rings from WT and db/db mice were directly processed for immunohistochemistry or placed into 24-well plates containing EGM-2 culture media with GSK126 (5 µmol/l) or vehicle for 20 h. Then, they were fixed for 24 h in 4% PFA, hydrated in graded ethanol, embedded in paraffin and mounted on coated glass slides (SuperFrost Plus; Thermo Fisher Scientific). For antigen retrieval, slides were subjected to high-pressure boiling in citrate buffer (pH 6.0). After blocking with goat serum (Abcam), aortic cross-sections were incubated overnight (4 °C) with either of the following primary antibodies: rabbit polyclonal anti-H3K27me3 (Millipore), mouse monoclonal anti-EZH2 (Invitrogen), mouse monoclonals anti-SOD1 and anti-SOD2 (Santa Cruz Biotechnology) (1:100). A rabbit monoclonal antibody against the endothelial marker CD31 (Abcam; 1:100) was also used. Alexa Fluor 488 and 647-conjugated secondary antibodies (Abcam) were used (1:200) and nuclei were stained with Hoechst (2 µg/ml in PBS) for 20 min. Coverslips were mounted on slides with a fluorescence mounting medium (Agilent; Santa Clara, CA, USA).

Confocal microscopy

Cells and tissue samples were visualized with a Leica TCS SP8 confocal laser scanning (Wetzlar, Germany) and a Nikon Eclipse Ti2 (Tokyo, Japan) microscope, respectively. Cells were examined with a 63X oil lens fitted with immersion oil (Leica; Wetzlar, Germany). 6–12 cells were randomly selected for each experimental condition and, depending on the cell depth, 5–8 stacks per channel were collected and projected in a single image. Aortic sections were visualized with 20X objective and 10 stacks per channel were collected. After the acquisition, image analysis was performed using ImageJ software.

Measurement of superoxide anion

Superoxide anion (O2−) generation in HAEC and D-HAEC was assessed by electron spin resonance (ESR) spectroscopy analysis using the spin trap 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethyl-pyrrolidine (CMH) as described elsewhere [26]. Briefly, 6-well plated cells were washed with D-PBS and resuspended in 600 μL of Krebs-HEPES buffer supplemented with the following composition (µmol/l): deferoxamine methanesulfonate salt (DF; 25), diethyldithiocarbamic acid sodium salt (DETC; 5), and CMH (200) (Noxygen; Elzach, Germany). After 30 min incubation at 37 °C, cell suspensions were snap-frozen in liquid nitrogen and stored at − 80 °C. ESR spectra were recorded using a NOX-E.5-ESR spectrometer (Bruker; Billerica, MA, USA). Signals were quantified by measuring the total amplitude after correction of baseline and subtraction of background [26].

NF-κB p65 binding activity assay

NF-κB p65 binding activity in HAEC and D-HAEC was measured by the TransAM™ NF-κB p65 activation protein assay kit (Active Motif; Carlsbad, CA, USA), following the manufacturer’s instructions. Briefly, 40 μg of whole cell lysate was added into a 96-well plate immobilized with consensus sequences for NF-κB p65 subunit (GGGACTTTCC) for 1 h at RT. Wells were washed with washing buffer following incubation with anti-NF-κB p65 antibody (Active Motif) for 1 h at RT. Horseradish peroxidase-conjugated secondary antibody was then added and the plate was incubated for an additional hour at RT. NF-κB p65 DNA binding was assessed by spectrophotometer at 450 nm on a VersaMax microplate reader (Ocean Springs, MI, USA).

Monocyte adhesion assay

Monocyte adhesion to HAEC exposed to the different experimental conditions was tested in the presence and in the absence of TNF-α (5 mmol/l, 24 h). Briefly, human monocytes THP-1 were kindly donated by Dr. Magdalena Paolino (Karolinska Institute; Stockholm, Sweden) and grown in RPMI-1640 culture medium (Biowest; Nuaillé, France) supplemented with 10% FBS, 2 mol/l L-glutamine, 4,5 g/l glucose, and 1% penicillin/streptomycin, following the supplier’s recommendations. THP-1 cells were stained with 2 μmol/l of CalceinAM fluorescent dye (Abcam) in serum-free media at 37 °C for 30 min, and washed with 10 mL of PBS. Stained cells were then resuspended in EBM2:RPMI-1640 medium (1:1), supplemented with 10% FBS, and co-cultured with HAEC under rotating conditions for 1 h at 37 °C. After incubation, non-adhering cells were removed with PBS and monolayers were fixed with 1% PFA. Co-cultures were observed under confocal microscopy and the number of adherent THP-1 cells was determined with ImageJ software.

Experimental animal models

Animal care and protocols were approved by the regional ethical committee and conformed to the Guide for Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication No. 85–23, revised 1996). Wild-type (WT) C57BL/6 male mice (6–8 weeks old) housed in the animal experimental core facility of Karolinska Institutet (Stockholm, Sweden) were sacrificed at the age of 12–18 weeks for isometric tension studies. Immunohistochemistry studies were performed in aortas from db/db and WT male mice (15–20 weeks old; Charles River Laboratories; Sulzfeld, Germany). All animals were kept in a 12:12-h light–dark cycle with free access to standard chow and water. Mice were anaesthetized with sodium pentobarbital (50 mg/kg i.p.) followed by thoracotomy. The entire aorta from the heart to the iliac bifurcation was excised and placed immediately in cold modified Krebs–Ringer bicarbonate solution (pH 7.4) of the following composition (mmol/l): NaCl (118.6), KCl (4.7), CaCl2 (2.5), KH2PO4 (1.2), MgSO4 (1.2), NaHCO3 (25.1), glucose (11.1), and calcium EDTA (0.026) (Sigma). The aortas were cleaned from adhering fat and connective tissues under a dissection microscope and subsequently cut transversely into 2 mm rings.

Isometric tension studies

Aortic rings from C57BL/6 mice were placed into 24-well plates containing EGM-2 culture media with either normal (5 mmol/l) or high glucose concentration (25 mmol/l) in the absence and in the presence of GSK126 (5 µmol/l) for 20 h. Then, the aortic rings were mounted in an isometric force transducer (multi-Myograph 610 M; Danish Myo Technology, Hinnerup, Denmark), suspended in an organ chamber filled with 6 mL Krebs–Ringer bicarbonate solution at 37 °C, and bubbled with 95% O2 and 5% CO2. The internal diameter was set at a tension equivalent to 0.9 times the estimated diameter at 100 mmHg. Resting tension was gradually increased to 2 mN. Following a 30 min equilibration period, aortic rings were exposed to potassium chloride twice (50 mmol/l, Sigma). After pre-contraction with phenylephrine (10−6 mol/l; Sigma), endothelium-dependent and independent relaxations were assessed by acetylcholine (Ach; 10−9–10−4 mol/l; Sigma) and nitroprusside (SNP; 10−10–10−5 mol/l, Sigma), respectively. Incubation of aortas with a high concentration of mannitol (25 mmol/l) did not affect vascular responses to Ach and SNP (Additional file 1: Fig. S2). Isometric forces for the different data points were quantified with LabChart™ software, and relaxations were expressed as a percentage of the pre-contraction plateau tension. Several aortic rings from the same animal were studied in parallel.

Statistical analysis

All data are presented as mean ± standard error of the mean (SEM), and statistical analysis was performed using GraphPad Prism Software (version 9.0.1). The normality distribution of the samples was assessed by the Shapiro–Wilk normality test. Unpaired t for parametric and Mann–Whitney tests for non-parametric data were used to determine the significance between two groups. One-way ANOVA with Tukey’s multiple comparisons and Kruskal–Wallis with Dunn’s multiple comparison tests were used for comparison among several groups of parametric and non-parametric data, respectively. Two-way ANOVA was used for the myograph studies. A p-value less than 0.05 was considered statistically significant.