Primary human umbilical vein endothelial cells (HUVECs, Modern Cell & Tissue Technologies, Seoul, Korea) were cultured in the EC basal medium provided in the EBM-2 Bullet Kit (Lonza, Basel, Switzerland), supplemented with 10% FBS. Cells were grown to confluence at 37°C in a humidified atmosphere of 5% CO₂/95% air. Cells were used in the experiments after passage 4.

HUVECs were plated in 60-mm culture dishes at a density of 1×10⁶ cells/dish. After 16 to 24 hours, the cells were washed with EBM-2 basal medium and antimycotics but no growth factors and then the cells were preexposed to 20 μM of gemigliptin (LG Chem Ltd., Seoul, Korea). The pretreated HUVEC cells were cultured in the EBM-2 medium containing 2% FBS, 10 ng/mL IL-1β for 24 hours.

Cells cultured in a 4-well chamber were used for von Willebrand factor (vWF) and α-smooth muscle actin (α-SMA) immunofluorescence. Briefly, after being washed three times in Tris-buffered saline (TBS; 25 mM Tris-HCl, pH 7.6, and 150 mM NaCl), the cells were placed in methanol:acetone (1:1) for 20 minutes at −20°C. They were washed three times in TBS for 5 minutes and blocked 5% horse serum/TBS for 30 minutes at room temperature. The cells were incubated in primary antibody for overnight at 4°C and washed three times in TBS. Subsequently, the cells were incubated with the biotynated secondary antibodies for 1 hour, washed three times with TBS and mounted with mounting medium with 4′,6-diamidino-2-phenylindole (DAPI, Vector Laboratories Inc., Burlingame, CA, USA). To calculate the relative percentage of α-SMA-positive cells in total HUVECs, the numbers of total HUVECs and α-SMA-stained HUVECs were determined. The rate of vWF positivity was calculated as a percentage of total HUVECs, as described by the following equation: (number of α-SMA-positive cells/number of total HUVECs)×100. On average, 100 HUVECs were counted for each slide of experimental conditions.

Cell migratory assays were performed using Transwell chambers with filter membranes of 8 μm pore size (Corning, Life Sciences, New York, NY, USA). Filter membranes were coated with 0.2% gelatin and chambers were inserted in 24-well cultured plates. HUVECs were seeded into the upper chamber (3×10⁴ cells per well in 2% FBS in EBM-2 medium) and allowed to attach for 3 hours at 37°C. Subsequently, cells were treated with 10 ng/mL IL-1β, 10 ng/mL IL-1β plus 20 μM gemigliptin for 24 hours (medium alone, with gemigliptin were used as controls). Filter membranes were removed, fixed with 70% ethanol and stained with 1:1,000 DAPI (1 mg/mL) for 30 minutes at 37°C. Migration activity was quantified by counting the migrating cells using a florescence microscope (Carl Zeiss Co. Ltd., Oberkochen, Germany). Data are reported as the ratio of migrating versus non-migrating cells expressed as a percentage of control (untreated cells).

HUVECs seeded on 6-well plates were cultured until confluency. They were then scratched with a yellow pipette tip, washed with PBS and further incubated with 10 ng/mL IL-1β, 10 ng/mL IL-1β plus 20 μM gemigliptin in EBM-2 medium supplemented with 2% FBS (medium alone, with gemigliptin were used as controls). Five fields of each of the three wounds analyzed per condition were photographed immediately after the scratch had been made (0 hour) and 24 hours later to monitor cell movement into the wounded area.

Cells were lysed in ice-cold lysis buffer. Cell lysates were centrifuged at 15,000 rpm for 10 minutes at 4°C. Protein concentrations were measured by the BCA method using BSA as the standard. Proteins (20 μg) were separated by 8% to 10% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with 5% fat-free milk for 1 hour in TBS containing 0.1% Tween 20 (TBS-T) and then incubated with the following primary rabbit or mouse antibodies: anti-endothelial nitric oxide synthase (eNOS) and BMP receptor type II (BMP-RII, BD Bioscience, San Jose, CA, USA), vWF, and fibronectin (FN, Dako Korea LLC, Seoul, Korea), glucagon-like peptide 1 receptor (GLP-1R), transgelin (SM22), fibroblast-specific protein 1 (FSP-1), activin receptor type 1A (Act-RIA) and activin receptor type IIB (Act-RIIB), BMP receptor type IA (BMP RIA) and BMP receptor type IB (BMP-RIB), CD26 (DPP-IV), osteogenic-specific transcription factor runt-related transcription factor 2 (RUNX2), osterix, hepcidin (all from Abcam, Cambrige, UK), collagen type I (Col I) and α-SMA (Merck Millipore Co., Darmstadt, Germany), BMP2, Act-RIIA, and TGF-β receptor type II (TGFβ-RII) (all from Novous Biologicals LLC, Littleton, CO, USA), vascular endothelial cadherin (VE-cadherin), BMP4, BMP7, TGF-β receptor type I (TGFβ-RI) (all from Santa Cruz Biotechnology Inc., Dallas, TX, USA), phospho-Smad1/5/9, Smad1, phospho-Smad2, Smad2, phospho-Smad3, Smad3, phospho-extracellular regulated protein kinases (p-Erk), Erk, phospho-p38, p38, phospho-SAPK/JNK, and SAPK/JNK (all from Cell Signaling Biotechnology, Danvers, MA, USA), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH, Abclone, Seoul, Korea). The antibodies were diluted 1:500 to 1:2,000 in 1% BSA in TBS-T and incubated at 4°C overnight. After washing, the membranes were incubated with secondary peroxidase-conjugated anti-mouse and anti-rabbit antibodies (Bio-Rad Laboratories, Hercules, CA, USA) diluted 1:1,000 to 1:2,000 in 1% fat-free milk in TBS-T at room temperature for 1 hours. Detection was achieved using Supersignal West Pico Chemiluminescent Substrate (ThermoFisher Scientific Inc., Rockford, IL, USA). Quantitation of Western blots by densitometry was performed with Genetool version 4.03.05 (SynGene, Cambridge, England).

Data are expressed as mean±standard error of the mean (SEM). Differences between groups were evaluated using SigmaPlot version 13.0 (Systat Software Inc., San Jose, CA, USA). The independent t test and one-way analysis of variance were used to analyze the quantitative variables between groups. A P<0.05 was deemed to indicate significance.

In HUVECs, DPP-IV protein levels were increased in a dose-dependent manner following IL-1β treatment (Supplemental Fig. S1A, B). HUVECs were stimulated with 10 ng/mL of IL-1β for various times (0, 5, 15, 30, 60 minutes, or 24 hours), and DPP-IV protein levels were significantly increased after 15 minutes of treatment (Supplemental Fig. S1C, D). IL-1β-induced GLP-IR protein levels were decreased in a dose-dependent manner, and then decreased at 24 hours after IL-1β treatment (Supplemental Fig. S1A-D). In contrast, DPP-IV protein levels were increased in a dose-dependent manner following gemigliptin treatment (Supplemental Fig. S1E, F). When HUVECs were stimulated with gemigliptin, DPP-IV protein levels decreased in a time-dependent manner (Supplemental Fig. S1G, H). In addition, gemigliptin-induced GLP-1R protein levels were increased in a dose- and time-dependent manner (Supplemental Fig. S1E-H). Gemigliptin treatment significantly reduced IL-1β-induced DPP-IV protein expression in ECs (Supplemental Fig. S2). Furthermore, GLP-IR protein levels were increased following treatment with 20 μM gemigliptin for 24 hours (Supplemental Fig. S2).

We assessed the effects of gemigliptin on IL-1β-mediated EndMT in cultured ECs. Exposure of HUVECs to IL-1β for 1 day triggered an obvious alteration in cellular morphology, from a cobblestone-like shape to a more spindled, fibroblast-like shape (Supplemental Fig. S3A). In contrast, gemigliptin treatment significantly suppressed this transition by inverting the endothelial phenotype (Supplemental Fig. S3A). To examine the effects of gemigliptin on IL-1β-mediated EndMT, we assessed the presence of fibroblast cell marker α-SMA and EC marker vWF by immunofluorescence (Supplemental Fig. S3B). Representative immunostaining revealed that IL-1β-mediated EndMT was associated with reduced vWF levels (Supplemental Fig. S3C) and increased α-SMA levels (Supplemental Fig. S3C). However, upon gemigliptin treatment, most cells were positive for vWF and negative for α-SMA (Supplemental Fig. S3B).

Furthermore, we examined the effects of IL-1β and gemigliptin on chemotactic motility using a transwell chamber. IL-1β treatment resulted in a significantly reduced number of migratory cells in HUVECs (Supplementary Fig. 3D, E). Consistent with the results of the Transwell assays, the migration rates were significantly higher after gemigliptin treatment (Supplemental Fig. S3D, E). We examined the effect of gemigliptin on wound healing using a HUVEC monolayer. Wound healing cell invasion assays revealed that IL-1β also inhibited the migration of fibroblast-like EndMT cells, while gemigliptin induced the invasion of these cells (Supplemental Fig. S3F). Thus, gemigliptin stimulated migration of HUVECs.

These observations were further confirmed by Western blotting. In the presence of IL-1β, eNOS, VE-cadherin, and vWF protein levels were reduced, and α-SMA and SM22 protein levels, SMC markers, were increased (Fig. 1). The mesenchymal markers FSP-1, FN, and Col I were increased after IL-1β treatment (Fig. 1). However, gemigliptin treatment reversed these IL-1β-mediated effects. Quantitative analysis of the Western blots verified our results. These findings demonstrate that gemigliptin suppresses EndMT induced by IL-1β.

To investigate the mechanism by which IL-1β is associated with BMP activity, key changes in BMP signaling were detected by Western blotting. We investigated BMP signaling during IL-1β-mediated EndMT. After IL-1β treatment of HUVECs for 24 hours, BMP2 and BMP7 were significantly increased, whereas BMP4 was suppressed (Fig. 2A, B). Because the biological functions of BMP2 and 7 are mediated through BMP/activin signaling via the six BMP- and activin-specific receptors, Act-RIA, Act-RIIA and RIIB, BMP-RIA and RIB, and BMP-RII, we analyzed the expression of these receptors after 24-hour IL-1β treatment in HUVECs. The expression of Act-RIA and BMP-RIB, including BMP receptor type I, significantly increased, and Act-RIIA, Act-RIIB, and BMP-RII, including BMP receptor type II, were activated by IL-1β (Fig. 2A, B). Gemigliptin significantly inhibited the effects of IL-1β on Act-RIA and BMP-RIA protein levels and Act-RIIB and BMP-RII protein expression (Fig. 2A, B). Furthermore to investigate the mechanism by which IL-1β is associated with TGF-β1, key marker in EndMT, we were detected by Western blotting. After IL-1β treatment of HUVECs for 24 hours, TGF-β1 and TGFβ-RI protein levels were increased, whereas TGF β-RII protein levels were decreased (Fig. 2C, D). Gemigliptin significantly inhibited the effects of IL-1β on TGFβ-RII protein levels. Therefore, gemigliptin blocks the IL-1β-induced BMP signaling pathway.

The binding of BMP to BMP receptors induces heterodimeric complexes and subsequently activates Smad or mitogen-activated protein kinases (MAPKs) by phosphorylation [14]. We found that the Smad1/5/9, Smad2, and Smad3 signaling pathways were activated by IL-1β. IL-1β significantly increased the phosphorylation of Smad1/5/9, Smad2, Smad3, and endogenous Smad2 and Smad3 (Fig. 3). Gemigliptin-treated cells showed a significant reduction in the phosphorylation of Smad1/5/9, Smad2, and Smad3 induced by IL-1β. Densitometric analyses of band intensities showed that gemigliptin partially increased endogenous Smad1/5/9 in contrast to Smad2 and 3 (Fig. 3). These results indicate that gemigliptin blocks IL-1β-induced EndMT in part by inhibiting the BMP/Smad signaling pathway.

Because inflammatory cytokines are primarily associated with the activation of MAPK signaling, we tested whether gemigliptin regulates IL-1β-induced EndMT through this pathway. We treated HUVECs with IL-1β or gemigliptin for 0, 15, 30, 60 minutes, and 24 hours. We observed that the addition of IL-1β to the culture medium induced a rapid and strong activation of MAPK signaling, including Erk, p38, and JNK in HUVECs (Supplemental Fig. S4). After the addition of gemigliptin, only Erk was phosphorylated (Supplemental Fig. S4). The Erk, p38, and JNK phosphorylation via IL-1β-induced EndMT was strongly inhibited in the cells treated with gemigliptin (Fig. 4). Therefore, our results indicated that IL-1β upregulated MAPK activity, but the effect was prevented by gemigliptin.

Because of the positive role of BMP/Smad signaling in regulating Runx2 expression, we examined the effects of IL-1β-activated BMP2 on osteogenic-specific transcription factors Runx2, osterix, and hepcidin in HUVECs. The osterix, Runx2, and hepcidin transcriptional activities tended to increase when HUVECs were treated with 10 ng/mL IL-1β, even though in the presence of gemigliptin their significantly suppressed (Fig. 5).