miR-132-3p and KLF7 as novel regulators of aortic stiffening-associated EndMT in type 2 diabetes mellitus

Pressure myography was performed to directly assess the passive aortic mechanics ex vivo as previously described (PubmedID 26208651). In brief, murine aortae were explanted, placed on specially designed stainless-steel cannulas and secured with silk surgical suture (10-0). The vessel was mounted in the heated chamber of a pressure arteriograph system (Model 110P, Danish Myotechnology, Copenhagen, Denmark) and stretched to in vivo length. Physiological saline solution (PSS) at 37 °C, aerated with 5% CO₂/95% O₂ was used to fill the vessel chamber and for aortic perfusion. After 3 preconditioning cycles, the aortic passive pressure-diameter relationship was determined by an automated protocol. The artery was pressurized from 0 to 180 mmHg in 18 mmHg increments, and the vessel’s outer diameter was simultaneously tracked by continuous computer video analysis (MyoVIEW software, Danish Myotechnology, Copenhagen, Denmark). The resting diameter d0 was quantified under 0 mmHg intraluminal (transmural) pressure and was not significantly different between the experimental groups tested (d0~1000 µm).

Male db/db mice (BKS.Cg-Dock7m+/+Leprdb/J) and their age-matched heterozygous non-diabetic controls (+/db) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). After 20 weeks, mice were exposed to inhalation of a lethal dose of isoflurane (concentration of 5%, Vet One, Meridian, ID, USA) in a closed chamber. Isoflurane was delivered from a vaporizer-system with O₂ as a carrier. Thoracic aortas were harvested and snap-frozen in liquid nitrogen. All animal studies were reviewed and approved by the responsible local animal ethics review boards and procedures were conform to the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. The cryo-sections were fixed with methanol at − 20 °C for 10 min and air-dried before washing three times with PBS. The sections were blocked using SEA BLOCK Blocking Buffer (Thermo Fisher Scientific, Waltham, MA, USA) and incubated at RT for 1.5 h. The primary antibody was added to the sections and incubated at 4 °C overnight. The next day, the slides were three times washed with PBS before adding the secondary antibody and incubating in the dark at RT for 45 min. The slides were washed three times with PBS and incubated with DAPI (1:1000 in PBS) at RT for 5 min. A final washing with PBS was performed before mounting the slides. The following primary antibodies and dilutions (in SEA BLOCK) were used: mouse anti-human CD31 (1:100, M0823, Dako, Carpinteria, CA, USA), rabbit anti-human S100A4 (1:50, A5114, Dako, Carpinteria, CA, USA), rabbit anti-mouse α-SMA (1:100, ab5694, Abcam, Cambridge, MA, USA) and rabbit anti-KLF7 (1:50, ab197690, Abcam, Cambridge, MA, USA). The secondary antibodies Alexa Fluor 647 goat-anti mouse (A21235, Invitrogen, Carlsbad, CA, USA) and Alexa Fluor 568 donkey anti-rabbit (A10042, Invitrogen, Carlsbad, CA, USA) were used in a 1:200 dilution. The stained slides were analysed using the Inverted Zeiss LSM 780 multiphoton laser scanning confocal microscope (Zeiss, Oberkochen, Germany).

Human thoracic aortic samples from patients with diabetes (n = 5) and without diabetes (n = 5) who underwent open aortic surgery were collected during the procedure, snap-frozen and stored at − 80 °C. Groups were matched for age (patients with diabetes: 67.20 ± 2.059 years; patients without diabetes 65.60 ± 2.379 years; p = ns) and smoking status (patients with diabetes 40% smoker; patients without diabetes: 60% smoker; p = ns). Approval for studies on human tissue samples was obtained under informed consent and conducted in accordance with the Declaration of Helsinki.

Total RNA was extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) and the PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The RNA was treated with Dnase I (Sigma-Aldrich, St. Louis, MO, USA) and the SuperScript II Reverse Transcriptase system (Invitrogen, Carlsbad, CA, USA) was used to synthesize cDNA according to the manufacturer’s protocol. For qRT-PCR, Fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) was used in combination with the StepOne Plus Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The primers used for qRT-PCR are listed in Table 1. The relative expression levels were standardized to GAPDH using the ΔΔCt method.

The micronome of glucose-treated HUVECs and the micronome of aortic vascular smooth muscle cells from db/db mice and db/+controls are available at Gene Expression Omnibus (GEO), NCBI (GSE74296 and GSE74521).

Total microRNA was isolated using a TRIzol-based (Invitrogen, Carlsbad, CA, USA) RNA isolation protocol. Reverse transcription was performed by using the TaqMan microRNA Reverse Transcription kit (Applied Biosystems, Foster Biosystems, Foster City, USA) according to the manufacturer’s instructions. MicroRNA TaqMan assays (Applied Biosystems, Foster City, USA) for hsa-miR-132-3p was used. Amplification took place on a Fast-Real-Time PCR System (Applied Biosystems, Foster City, USA). All fold changes were calculated by the method of ΔΔCt.

A fragment of the human 3′UTR of KLF7 which contains two miR-132-3p binding sites was amplified using Phusion High-Fidelity DNA Polymerase (New England Biolabs, Ipswich, MA, USA) and the primers KLF7-3′UTR-F: 5′ AATTACTAGTACCATCCCTTCAAGACACGT; KLF7-3′UTR-R: 5′ AATTGTTTAAACCCCAGATCTTGAAGGTTGCTG. Both the amplified PCR product and the pMiR-REPORT miRNA expression vector (Invitrogen, Carlsbad, CA, USA) were digested using SpeI and PmeI restriction enzymes (New England Biolabs, Ipswich, MA, USA). Ligation was performed using the LigaFast Rapid DNA Ligation System (Promega, Madison, WI, USA) and transformed into One Shot TOP10 Chemically Competent E. coli (Invitrogen, Carlsbad, CA, USA). Miniprep was performed using the Zyppy Plasmid Miniprep Kit and the sequence of the insert was confirmed using sequencing. The generated pMiR-REPORT-KLF7-3′UTR plasmid was amplified using the HiSpeed Plasmid Midi Kit (Qiagen, Hilden, Germany).

Human embryonic kidney (HEK293) cells were cultured in DMEM medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA, USA) and 1% penicillin–streptomycin (Gibco, Carlsbad, CA, USA). Human umbilical vein endothelial cells (HUVECs; Lonza, Basel, Switzerland) were cultured in Endothelial Cell Growth Medium (EGM-2MV, Lonza, Basel, Switzerland). For glucose treatment, 0.2 × 10⁶ cells were plated and cultured as previously described [29] to induce EndMT. 60 mM d-glucose was supplemented for 4 days and the medium was changed every 2 days. 5.5 mM d-glucose plus 54.5 mM d-mannitol was used as control.

0.5 × 10⁶ HEK293 cells were plated on a 6-well plate and incubated overnight. The next day, the HEK293 cells were transfected by using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA). In short, 2 ug of pMiR-REPORT-KLF7-3′UTR and 1 ug of pGL4.73[hRluc/SV40] (renilla) vector (Promega, Madison, WI, USA) together with 5 nM of hsa-miR-132-3p miRCURY LNA miRNA mimic (YM00472088, Qiagen, Hilden, Germany) or negative control miRCURY LNA miRNA mimic (YM00479902, Qiagen, Hilden, Germany) was added in a tube containing 250 µL Opti-MEM Reduced Serum Medium (Gibco, Carlsbad, CA, USA). Another tube was prepared containing a mastermix of 250 µL Opti-MEM and 5 µL of Lipofectamine 2000 per transfection. The tubes were mixed at RT for 5 min before combining the Lipofectamine 2000 mixture with the DNA mixture. The tubes were mixed again before incubating at RT for 20 min. The transfection complexes were added drop-wise to the cells and incubated overnight. The next day, the medium was changed. The cells were collected 72 h after transfection.

The luciferase assay was performed using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA). In short, the cells were lysed with 1:5 diluted passive lysis buffer by shaking at RT for 30 min. Centrifugation at 13,000 rpm for 10 min was performed to collect the supernatant (containing the cell lysate). The cell lysate was added in triplicates to a 96 well plate before adding Luciferase Assay Reagent II (for Firefly luciferase activity) and Sto&Glo Reagent (for Renilla luciferase activity) to detect the luminescence with the help of a luminescence plate reader.

0.2 × 10⁶ HUVECs were plated on a 6-well plate and incubated overnight. The next day, the HUVECs were transfected by using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA). In short, for miR-132-3p overexpression, 10 pmol of hsa-miR-132-3p miRCURY LNA miRNA mimic (YM00472088, Qiagen, Hilden, Germany) or Silencer Negative Control No.1 siRNA (AM4635, Ambion, Austin, TX, USA) was added in a tube containing 250 µL Opti-MEM (Gibco, Carlsbad, CA, USA). For KLF7 KO, 50 pmol of KLF7 Silencer Select siRNA (Ambion, Austin, TX, USA) or Silencer Negative Control No.1 siRNA (AM4635, Ambion, Austin, TX, USA) was added in a tube containing 250 µL Opti-MEM (Gibco, Carlsbad, CA, USA). Another tube was prepared containing a mix of 250 µL Opti-MEM and 5 µL of Lipofectamine 2000 per transfection. The tubes were incubated at RT for 5 min. before combining the Lipofectamine 2000 mixture with the siRNA/miRNA mixture. The Lipofectamine 2000 mix was combined with the siRNA/miRNA mix and incubated at RT for 20 min. The transfection complexes were added drop-wise to the cells and incubated for 4 h. After 4 h the medium was changed. The next day d-glucose and d-Mannitol treatment was started. The cells were collected 72 h after transfection.

Total RNA was extracted using TRIzol Reagent (Ambion, Austin, TX, USA) according to the manufacturer’s protocol. The RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used to synthesize cDNA according to the manufacturer’s protocol. For qRT-PCR, Fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) was used in combination with the Viia7 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The primers used for qRT-PCR are listed in Table 1. The relative expression levels were standardized to GAPDH using the ΔΔCt method.

Total microRNA was isolated using a TRIzol Reagent (Ambion, Austin, TX, USA) according to the manufacturer’s protocol. Reverse transcription was performed by using the TaqMan microRNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA) and microRNA-specific stemloop primers (Table 2). For miR-qRT-PCR, analysis was performed on a mixture containing 5 ng cDNA equivalent, 0.1 µM sense primer, 0.1 µM antisense primer (Table 3) and Fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA). Analysis was performed on the Viia7 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The relative expression levels were standardized to U6 using the ΔΔCt method.

Cells were two times washed with PBS before being fixed with 2% PFA at RT for 30 min. After another two washes with PBS, cells were permeabilized with 0.5% Triton in PBS for 15 min. Cells were washed again 3× in PBS and blocked with blocking buffer (5% donkey or goat serum in PBS) for 10 min. Cells were incubated with primary antibody (1:100 in blocking buffer) overnight at 4 °C. The next day, cells were washed with PBS, PBS + 0.05% Tween-20 and again with PBS before incubation with secondary antibody (1:500) in DAPI (1:5000) and 2% Human Serum for 60 min at RT in the dark. Cells were washed with PBS, PBS + 0.05% Tween-20 and PBS and stored in PBS at 4 °C until analysis. The following primary antibodies were used: mouse anti-human CD31 (M0823, Dako, Santa Clara, CA, USA), rabbit anti-human SM22α (ab14106, Abcam, Cambridge, MA, USA). Alexa Fluor 647 donkey anti-mouse (A31571, Invitrogen, Carlsbad, CA, USA) and Alexa Fluor 568 goat anti-rabbit (A21069, Invitrogen, Carlsbad, CA, USA) were used as secondary antibodies. The cells were analyzed using the EVOS FL cell imaging system (Life Technologies, Carlsbad, CA, USA).

In a first step, we characterized the structural stiffness of the aortic wall in db/db mice and their +/db counterpart controls using ex vivo pressure myography. We found that the aortic segments taken from db/db mice exhibited significantly increased “material” stiffness of the explanted tissue (compared to +/db controls, Additional file 1: Figure S1).

To assess whether EndMT occurs during aortic stiffening, we performed confocal microscopy of co-immunofluorescent staining of the endothelial marker CD31 in combination with the mesenchymal markers α-SMA and S100A4 respectively in aortic sections of db/db mice (a murine model of T2D). We observed a robust co-localization of CD31 with both α-SMA and S100A4 in aortas of db/db mice which was almost absent in control (db/+) mice (Fig. 1A, B). The observed co-localization of CD31 with both α-SMA and S100A4 was significantly more pronounced in aortas of db/db mice when compared to aortas of control mice (for α-SMA 10 versus 2.8, p < 0.05, Fig. 1C; for S100A4 10.6 versus 1.6, p < 0.05, Fig. 1D). We then examined the mRNA expression of the EndMT transcription factors Snail, Slug and Twist in aortic tissue of db/db mice to further test the presence of EndMT. Indeed, Snail, Slug and Twist were significantly upregulated in aortic tissue from db/db mice when compared to control mice (Snail 13.5-fold, p < 0.01; Slug 9.4-fold, p < 0.001; Twist 6.1-fold, p < 0.001, Fig. 1E). To investigate if EndMT is also associated with aortic stiffening in patients with T2D, we also studied aortic tissue from patients with T2D and control subjects, and we found that SNAIL, SLUG and TWIST are significantly upregulated in T2D patients (SNAIL 2.6-fold, p < 0.01; SLUG 2.5-fold, p < 0.05; TWIST 3.4-fold, p < 0.001, Fig. 1F). Altogether, this demonstrates that EndMT can be observed in the context of T2D.

To identify the underlying mechanism by which EndMT is regulated in the context of T2D, we overlapped the micronome of aortas from db/db mice with the micronome of high glucose-treated human umbilical vein endothelial cells (HUVECs) (available at Gene Expression Omnibus (GEO), NCBI, GSE74296 and GSE74521). Three microRNAs (miRs) were significantly downregulated in both datasets: miR-9, miR-30 and miR-132-3p (Fig. 2A). Since miR-132-3p was the highest downregulated in glucose-treated HUVECs and the second highest downregulated in aortas of db/db mice, we decided to further examine the expression of this miR in aortic tissue of T2D patients. Indeed, we observed a significant downregulation of miR-132-3p in aortic sections of T2D patients when compared to control subjects (2.4-fold, p < 0.05, Fig. 2B). Since miR-132-3p is universally downregulated in high glucose-treated HUVECs and in aortic tissue of db/db mice and diabetic patients, we identified miR-132-3p as possible regulator of EndMT-triggered aortic stiffening in T2D.

To further investigate the role of miR-132-3p in regulating EndMT, we identified possible targets of miR-132-3p in humans through the online prediction tool TargetScan. This revealed two well-known EndMT regulators: SMAD2, which facilitates TGF-β signaling, and ZEB2, a transcription factor regulating epithelial to mesenchymal transition (EMT) (Fig. 2C). SMAD2 mRNA expression was unaltered in aortas of T2D patients when compared to control subjects whereas ZEB2 was significantly downregulated (SMAD2 not significant; ZEB2, 4.1-fold, p < 0.05, Fig. 2D). This suggests that these two predicted targets of miR-132-3p do not play a role in this context. We next examined the mRNA expression of those predicted miR-132-3p targets with the most putative binding sites for miR-132-3p in their 3′UTR: KLF7, PTEN, DNMT3a and ZBTB20 (Fig. 2C). Out of these four predicted targets, only KLF7 is significantly upregulated in aortas of diabetes patients whereas DNMT3a, ZBTB20 and PTEN remain unaltered when compared to control subjects (KLF7 2.2-fold, p < 0.01; DNMT3/ZBTB20/PTEN not significant, Fig. 2E). This suggests that miR-132-3p targets KLF7 during aortic stiffening in diabetic patients.

To assess whether KLF7 plays a role during EndMT-triggered aortic stiffness, we performed co-immunofluorescent staining of CD31 in combination with KLF7 in aortic sections of db/db mice. With confocal microscopy, we observed a robust co-localization of CD31 with KLF7 in aortas of db/db mice which was almost completely absent in control mice (Fig. 3A). The observed co-localization of CD31 with KLF7 was significantly more in aortas of db/db mice when compared to aortas of control mice (4.2-fold, p < 0.05, Fig. 3B). Finally, we examined the mRNA expression of KLF7 in aortic tissue of db/db mice. We also observed a significant upregulation of KLF7 in aortas of db/db mice when compared to control mice (4.9-fold, p < 0.001, Fig. 3C). This suggests that KLF7 plays a role in EndMT-triggered aortic stiffness in T2D.

To confirm that miR-132-3p targets KLF7, we cloned a fragment of the human 3′UTR of KLF7 in a pMiR-REPORT luciferase vector and transfected HEK293 cells with both the pMiR-REPORT-KLF7-3′UTR and a miR-132-3p mimic or negative control (Fig. 4A). We show that co-transfection with the miR-132-3p mimic results in a significant reduction of the luciferase activity of the KLF7 3′UTR when compared to the negative control, demonstrating that miR-132-3p indeed targets KLF7 (1.6-fold, p < 0.01, Fig. 4B).

To further explore the role of miR-132-3p in high glucose-induced EndMT, we overexpressed miR-132-3p in high glucose-treated HUVECs. Co-immunostaining of CD31 in combination with SM22α showed that high glucose treatment results in a decrease in the expression of CD31 and cobblestone morphology and an increase in SM22α expression accompanied with a spindle-shaped morphology, all indicative of EndMT (Fig. 4C). Moreover, qRT-PCR analysis revealed upregulation of KLF7, SNAIL, SLUG and TWIST upon high glucose treatment, and downregulation of miR-132-3p upon high glucose treatment, confirming the presence of EndMT, the downregulation of miR-132-3p and the upregulation of KLF7 in high glucose-induced EndMT (miR-132-3p 1.9-fold, p < 0.01; KLF7 1.6-fold, p < 0.01; SNAIL 1.75-fold, p = 0.06; SLUG 1.71-fold, p = 0.06; TWIST 1.75-fold, p < 0.01, Fig. 4D–H). After having confirmed the overexpression of miR-132-3p (in mannitol conditions 5083-fold, p < 0.0005; in high glucose conditions 4279-fold, p < 0.005, Fig. 4D), we demonstrated that overexpression of miR-132-3p in combination with high glucose treatment ameliorates high glucose-induced EndMT as shown by an increase in CD31 expression and of cobblestone morphology, and a decrease in SM22α expression and of spindle-shaped morphology (Fig. 4C). Moreover, overexpression of miR-132-3p decreases the expression of KLF7, SNAIL, SLUG, and TWIST in both high glucose (KLF7 1.82-fold, p < 0.0005; SNAIL 2.03-fold, p < 0.01; SLUG 2.03-fold, p < 0.0001; TWIST 1.95-fold, p < 0.001, Fig. 4E–H) and mannitol conditions (KLF7 1.56-fold, p < 0.05; SNAIL 1.81-fold, p < 0.05; SLUG 1.38-fold, p < 0.01; TWIST 1.92-fold, p < 0.01, Fig. 4E–H). This indicates that miR-132-3p overexpression ameliorates high glucose-induced EndMT in vitro.

After having established the role of miR-132-3p on high glucose-induced EndMT, we further characterized the role of KLF7 on high glucose-induced EndMT. Therefore, we downregulated KLF7 in high glucose-treated HUVECs. Downregulation of KLF7 decreases the expression of KLF7, SNAIL, SLUG, and TWIST in both high glucose (KLF7 3.86-fold, p < 0.0001; SNAIL 2.9-fold, p < 0.0001; SLUG 2.9-fold, p < 0.0001; TWIST 4.23-fold, p < 0.001, Fig. 5A–D) and KLF7, SLUG, and TWIST in mannitol conditions (KLF7 2.27-fold, p < 0.01; SLUG 1.85-fold, p < 0.001; TWIST 2.27-fold, p < 0.05, Fig. 5A–D).

Altogether, this suggests that KLF7 downregulation ameliorates high glucose-induced EndMT in vitro and that miR-132-3p regulates KLF7 in EndMT-triggered diabetes-related aortic stiffening (Fig. 6).