Disturbed Flow Increases UBE2C (Ubiquitin E2 Ligase C) via Loss of miR-483-3p, Inducing Aortic Valve Calcification by the pVHL (von Hippel-Lindau Protein) and HIF-1α (Hypoxia-Inducible Factor-1α) Pathway in Endothelial Cells

Abstract

Objective- Calcific aortic valve (AV) disease, characterized by AV sclerosis and calcification, is a major cause of death in the aging population; however, there are no effective medical therapies other than valve replacement. AV calcification preferentially occurs on the fibrosa side, exposed to disturbed flow (d-flow), whereas the ventricularis side exposed to predominantly stable flow remains protected by unclear mechanisms. Here, we tested the role of novel flow-sensitive UBE2C (ubiquitin E2 ligase C) and microRNA-483-3p (miR-483) in flow-dependent AV endothelial function and AV calcification. Approach and Results- Human AV endothelial cells and fresh porcine AV leaflets were exposed to stable flow or d-flow. We found that UBE2C was upregulated by d-flow in human AV endothelial cells in the miR-483-dependent manner. UBE2C mediated OS-induced endothelial inflammation and endothelial-mesenchymal transition by increasing the HIF-1α (hypoxia-inducible factor-1α) level. UBE2C increased HIF-1α by ubiquitinating and degrading its upstream regulator pVHL (von Hippel-Lindau protein). These in vitro findings were corroborated by immunostaining studies using diseased human AV leaflets. In addition, we found that reduction of miR-483 by d-flow led to increased UBE2C expression in human AV endothelial cells. The miR-483 mimic protected against endothelial inflammation and endothelial-mesenchymal transition in human AV endothelial cells and calcification of porcine AV leaflets by downregulating UBE2C. Moreover, treatment with the HIF-1α inhibitor (PX478) significantly reduced porcine AV calcification in static and d-flow conditions. Conclusions- These results suggest that miR-483 and UBE2C and pVHL are novel flow-sensitive anti- and pro-calcific AV disease molecules, respectively, that regulate the HIF-1α pathway in AV. The miR-483 mimic and HIF-1α pathway inhibitors may serve as potential therapeutics of calcific AV disease.

Keywords: aortic valve; aortic valve stenosis; endothelial cells; endothelial-to-mesenchymal transition; inflammation; microRNAs; ubiquitination.

Figure 1.. miR-483 expression is shear-sensitive and side-specific in ECs, and regulates shear-dependent EndMT and inflammation.

a,b) miR-483 qPCR was performed using a) RNAs from HAVECs exposed to unidirectional laminar (SF) or oscillatory shear (DF) for 24 hours (n=7) and b) endothelial-enriched total RNA from the ventricularis (Ven) and the fibrosa (Fib) side of healthy porcine AVs normalized to U6 (n=9). c-g) HAVECs transfected with miR-483 mimic (483-M), control mimic (Ctrl-M), anti-miR-483 (AM-483) or Anti-miR scramble (AM-Ctrl) for 24 hours were exposed to d-flow (DF) or s-flow (SF) for 24 hours. Following shear, total RNAs were prepared for qPCR to analyze markers of EndMT (n=3–4) (c,d) and inflammation (n=3–7) (f,g) normalized to 18S, or cells were used for THP1 monocyte adhesion assay (n=3–7) (e). Mean±sem, *p<0.05.

Figure 2.. UBE2C is a shear-dependent target of miR-483 and regulates DF- and miR-483-induced inflammation and EndMT in HAVECs.

a,b) UBE2C expression was increased in HAVECs exposed to d-flow (DF) compared to s-flow (SF) for 24 hours, as determined by Western blot and its quantification by ImageJ (n=3). c,d) HAVECs treated with miR-483-mimic (483-M) or control mimic (Ctrl-M) followed by DF were analyzed by UBE2C Western blot and its quantification (n=3). e-g) HAVECs were co-transfected with siUBE2C or siCtrl and anti-miR-483 (AM-483) or anti-miR-control (AM-Ctrl), followed by monocyte adhesion (n=11) (e) and qPCR analysis for inflammatory markers (n=3) (f). g) HAVECs transfected with UBE2C overexpression plasmid or GFP plasmid were used for monocyte adhesion assay (n=6). h-i) HAVECs treated with siUBE2C or siRNA control (SiCtrl) for 24 hours, followed by DF or SF conditions for another 24 hours. Then, monocyte adhesion (h) and qPCR analyses for markers of inflammation (i) and EndMT (j) were carried out, normalized to 18S (n=4–6). Mean±sem, *p<0.05.

Figure 3.. UBE2C mediates shear-dependent expression of pVHL and HIF1α in HAVECs, and UBE2C, pVHL and HIF1α are expressed in a side-dependent manner in human AV leaflets.

a) Depicts a hypothesis that overexpression of UBE2C ubiquitinates pVHL, leading to increased HIF1α level, endothelial inflammation and EndMT. b) HAVECs sheared for 72 hours by s-flow (SF) or d-flow (DF) were immunostained with antibodies for pVHL or HIF1α (b) and quantification of the image was done using ImageJ (c), or by qPCR (d) normalized to 18S (n=3–4). e) HAVECs treated with siCtrl or siUBE2C for 24 hours, followed by exposure to DF for 24 hours, were analyzed by Western blot (e) and Image J quantification (f) for pVHL (n=5). For the HIF1α study, HAVECs exposed to DF for 72 hours were immunostained with HIF1α (g) and Image J quantified (h) (n=6). i) Human AVs with sclerosis were stained with antibodies to UBE2C, pVHL, HIF1α, Twist1 and Runx2 with DAPI nuclear staining, and j) shows quantification of the fluorescent intensities of each staining in the endothelial layer on each side using Image J (n=5–8). Mean±sem, *p<0.05. F: fibrosa, V: ventricularis.

Figure 4.. UBE2C binds and ubiquitinates pVHL, leading to its degradation.

a,b) HAVECs transfected with siUBE2C or siCtrl were immunoprecipitated with an antibody for ubiquitin (Ub) (a) or pVHL (b) and Western blotted with the pVHL (a) or Ub (b) antibody (n=5). c) HEK cells co-transfected with myc-UBE2C or RFP plasmid (1μg) and HA-pVHL-WT or HA-pVHL-RRR mutant plasmids (0.5μg) for 48 hours were immunoprecipitated using the antibody to HA-tag and Western blotted with the antibody to UBE2C or pVHL (exposed using High-intensity ECL and Low-intensity ECL). Untreated (Unt) HEK cells were used as a control. d-g) HEK cells co-transfected with HA-pVHL-WT or HA-pVHL-RRR mutant plasmids and increasing doses of UBE2C (0.5–2μg) or RFP plasmids for 48 hours were lysed and Western blotted with antibodies to pVHL and β-actin as an internal control (d,f), and ImageJ quantified (e,g) (n=3–6). Mean±sem, *p<0.05.

Figure 5.. pVHL and HIF1α mediate UBE2C-induced endothelial inflammation and EndMT.

a,b) HAVECs were transfected either individually with sipVHL, siHIF1α, and siCtrl (a-d) or co-transfected with siUBE2C (e,f) for 48 hours. Treated cells were then used for monocyte adhesion assay (a,c,e) or qPCR analyses for markers of inflammation and EndMT (b,d,f) normalized to 18S (n=3–6). Mean±sem. *p<0.05.

Figure 6.. The miR-483-mimic or HIF1α inhibitor inhibits AV calcification.

a) Freshly-harvested porcine AV leaflets were transfected with either miR-483 mimic or Ctrl mimic (a-f) or PX478 (20 μM) or HBSS vehicle (g-i) every 3 days for 2 weeks in osteogenic media (OM). AV leaflets were then divided for total RNA isolation and qPCR assay for miR-483 (a) (n=3), immunohistochemical assay using Alizarin Red (b,c) (n=6) and antibodies for UBE2C, pVHL and HIF1α (e,f) (n=6–12) and Arsenazo calcium assay (d) (n=5–6). Alizarin images were quantified using MATLAB (c,h) while the fluorescence images (e) were quantified by ImageJ (n=3–12). j) Porcine AV leaflets were exposed to fibrosa-(Fib.) or ventricularis-(Ven.) flow profiles in osteogenic media for 7 days with PX478 (20μM) or vehicle. Total calcium was quantified via Arsenazo assay and normalized to dry tissue weight (n=13–15). Mean±sem, *p<0.05.

Figure 7.. Working hypothesis