Miller, Jordan D.

Increases in vascular calcification (VC) and fibrosis are a near ubiquitous finding with aging. Here, we sought to test the hypothesis that Metformin (MET)—an AMPK activator entering large clinical trials that is thought to attenuate multimorbidity associated with aging—attenuates key molecular processes known to drive vascular calcification (VC) and fibrogenic signaling in vascular smooth muscle cells (VSMCs) in vitro. Given recent reports highlighting the importance of mitochondria and sirtuins with aging, we also tested the hypothesis that MET-driven reductions in osteogenic signaling are associated with upregulation of SIRT3. We used exogenous BMP2 (100 ng/ml) and TGFβ1 (10 ng/ml) as osteogenic and fibrogenic stimuli, respectively. Consistent with prior work from our lab, treatment with BMP2 for 24 hours robustly increased expression of the osteogenic genes Msx2 (3.9±1.5 fold-change, FC), Runx2 (3.5±1.2 FC), and osterix (17.2±5.2 FC) in VSMCs. Critically, co-treatment with Metformin dramatically reduced expression of these genes by 68 ± 22%, 58 ±20%, and 63 ± 4%, respectively. Subsequent experiments treating cells with exogenous TGFβ1 resulted in significant induction of the fibrogenic marker/known TGFβ1 target versican (1.7±0.4 FC), an effect that was largely attenuated by co-treatment with MET (58±3%). Treatment with MET did not, however, impact TGFβ1-induced increases in CTGF or fibronectin. Treatment with either BMP2 or TGFβ1 resulted in significant increases in mRNA levels of SIRT3 (2.5 ±0.11 and 1.4 ±0.17 FC, respectively), whereas treatment with MET resulted in modest and less consistent increases in SIRT3 mRNA (1.3 ± 0.2 FC). Interestingly, however, co-treatment with BMP2+MET or TGFβ1+MET resulted in uniform suppression of BMP2/ TGFβ1-induced increases in SIRT3 (54±19% and 25±12%, respectively). Collectively, these data suggest that treatment with MET may be a viable strategy to attenuate age-associated increases in osteogenic and fibrogenic signaling in conduit vessels, and suggest further in vivo investigation into therapeutic efficacy is warranted.

We recently reported activation of TGF-β superfamily pathways and activation of Wnt/β-catenin signaling in the pathogenesis of myxomatous mitral valve disease (MMVD), which may converge upon increased Runx2 expression. We also demonstrated that chronic in vivo administration of angiotensin II (AngII) recaptures various molecular aberrations characteristic of human disease -including increased TGF-β signaling and activation of Runx2- in mouse mitral valves. Here, we sought to determine whether Runx2 haploinsufficiency can attenuate molecular and/or phenotypic effects of AngII infusion in mouse mitral valves. We implanted littermate matched Runx2+/+ and Runx2+/- mice with osmotic minipumps delivering AngII (1000ng/kg/min) or saline for 2 weeks. We measured changes in mRNA (qRT-PCR) and protein (immunohistochemistry) levels in mitral valves, and assessed mitral valve function by echocardiography. In Runx2+/+ mice, AngII elicited increased mRNA levels of TGF-β3 and the TGF-β target COL3A1, and increased staining for canonical (p-SMAD2/3) and non-canonical (p-AKT) TGF-β signaling proteins. Runx2 haploinsufficiency attenuated AngII-induced changes in COL3A1 expression and p-AKT staining, without impacting AngII-mediated alterations in TGF-β3 and p-SMAD2/3. Evaluating collagen thickness by picrosirius red staining and visualization under circularly polarized light, AngII infusion in Runx2+/+ mice was associated with increased collagen density at the base of posterior mitral valve leaflets, a change that was abrogated by Runx2 haploinsufficiency. Similarly, Runx2-haploinsufficiency was again protective when examining mRNA and protein levels of the proliferation marker KI67 in response to AngII. Echocardiography identified a trend towards increased mitral valve regurgitation following AngII infusion that was unaltered by Runx2 haploinsuffiency. Collectively, these data suggest that increases in Runx2 expression in the mitral valve may serve to orchestrate tissue fibrosis and cellular proliferation that are hallmarks of MMVD, and that future pharmacological strategies aimed at reducing activity of Runx2 in the mitral valve may yield viable non-surgical alternatives for the management of this condition.

Increased BMP signaling and markers of osteogenic differentiation are strongly associated with regions of valvular calcification in aortic valves from humans with end-stage calcific aortic valve stenosis. While oxidative stress is also markedly elevated in these regions, it is unknown whether increasing antioxidant capacity can reduce pro-osteogenic signaling and valvular calcification. Therefore, we hypothesized that overexpression of manganese superoxide dismutase (MnSOD) can reduce pro-osteogenic signaling in the aortic valves from hypercholesterolemic mice, ultimately resulting in reduced valvular calcification and improved valvular function. We used Ldlr -/- /ApoB 100/100 (LA) mice that were wild-type for MnSOD (LA-MnSOD 0/0 ) or overexpressing MnSOD (LA-MnSOD Tg/0 ) and fed a western diet (TD88137) for six months. We first measured mRNA levels of the BMP ligands BMP2 and BMP4 (qRT-PCR), both of which were unchanged between LA-MnSOD 0/0 and LA-MNSOD Tg/0 mice. While there was a small but significant increase in Runx2 mRNA levels in LA-MnSOD Tg/0 mice compared to wild-type littermates (1.22±0.09, 1.0±0.09, qRT-PCR, p < 0.05.) gene expression levels of late stage osteoblast marker Sp7 were reduced by approximately 50% in LA-MnSOD Tg/0 compared to wild-type littermates (0.53±0.15, 1.0±0.24, qRT-PCR, p < 0.05). Sp7 protein levels (immunohistochemistry) were also significantly reduced in LA-MnSOD Tg/0 compared to wild-type littermates (164.5±12.3 RFU vs 259.6±10.4 RFU, p < 0.05). Critically, aortic valve calcification (Alizarin Red) was also significantly reduced in LA-MnSOD Tg/0 mice compared to their LA-MnSOD 0/0 littermates. Functionally, aortic valve cusp separation distance (high resolution echocardiography) was slightly improved in LA-MnSOD Tg/0 mice compared to LA-MnSOD 0/0 mice (0.98±0.04 mm vs 0.91±0.03 mm, p = n.s.). Collectively, these data suggest that increasing mitochondrial antioxidant capacity can reduce pro-osteogenic signaling and attenuate aortic valve calcification, and may be a potential therapeutic target to slow the progression of valvular calcification and dysfunction in patients with calcific aortic valve stenosis.

Increases in local angiotensin II (ANGII) signaling have been implicated in a number of diseases, including atherosclerosis, hypertension, and aneurysm. Endothelial dysfunction is strongly associated with increased ANGII signaling and the pathogenesis of these diseases. While increases in Runx2 have been implicated in cardiovascular calcification and vascular stiffening in diabetes, little is known regarding the role of Runx2 in the regulation of endothelial function. In the present study, we hypothesized that Runx2 is a key mediator of impaired vascular function and increased vascular wall stiffness provoked by ANGII. Wild-type mice (WT) and littermates that were deficient in one copy of Runx2 (HET) were implanted with osmotic minipumps containing saline or a pressor dose of ANGII (1000 ng/kg/min). Blood pressure was measured non-invasively using tail-cuff methods (CODA) and mice were sacrificed following 14 days of treatment. Changes in endothelium-dependent relaxation (relaxation to acetylcholine, R ACH ) and passive pressure-diameter relationships were measured in vitro (Living Systems arteriograph). In WT mice, treatment with ANGII for 14 days resulted in profound impairment in R ACH (85 ± 4% vs 47 ± 14% p<0.05) and slight reductions in vascular diameter at high distending pressures (~10% reduction in maximal diameter). While Runx2 deficiency did not alter R ACH or vascular distensibility/compliance in saline-treated mice, loss of one copy of Runx2 in ANGII-treated mice dramatically improved R ACH compared to their wild-type, ANGII-treated littermates. Interestingly, impairments in the pressure-diameter relationship with ANGII infusion were not affected by loss of one copy of Runx2. Critically, reductions in Runx2 did not significantly alter the pressor response to ANGII. In conclusion, our data suggest that Runx2 may play an integral role in the pathogenesis of ANGII-induced endothelial dysfunction, with these changes occurring independently of alterations in extracellular matrix remodeling and vascular stiffness. These data suggest that Runx2 may play a key, context-dependent role mediating vascular responses to ANGII.

Thoracic aortic aneurysm (TAA) is a degenerative condition of the aorta characterized by aortic dilatation secondary to tissue inflammation and matrix remodeling, which ultimately progresses to aortic dissection and rupture. While recent work suggests that in vitro exposure to pro-inflammatory stimuli promotes DNMT3b-dependent DNA methylation in skeletal muscle, epigenetic mechanisms regulating gene expression in TAA remain largely uncharacterized. We therefore tested the hypothesis that expression of tumor necrosis factor α (TNF-α) is increased in human TAA tissue and that such increases are associated with upregulation of DNA methyltransferase (DNMT) levels. Normal aortic samples (n=13) were acquired from non-aneurysmal hearts/aorta not suitable for cardiac transplant, whereas TAA samples from patients with tricuspid aortic valves (n=29) were acquired from patients undergoing elective surgery for aortic aneurysm. We measured relative expression of TNF-α, DNMT1, DNMT3a and DNMT3b mRNA levels in normal and TAA tissue by qRT-PCR, and DNMT3b protein levels by western blot. Marked increases in expression of the inflammatory marker TNF-α were noted in TAA tissue (12.3± 2.34), and were associated with significantly increased DNMT1 (1.4±0.13), DNMT3a (1.44±0.09) and DNMT3b (1.94±0.27) mRNA levels (p<0.05 for all). DNMT3b protein levels were also dramatically increased in human TAA samples (2.99±0.58-fold, n = 10/group, p<0.05) when compared to normal aortic tissue. Collectively, these data suggest a novel potential role for DNA methyltransferases as major regulators of altered gene expression in TAA, and support our working hypothesis that inflammation-induced increases in DNMT3b levels may be critical repressors of protective gene expression in TAA.

Mitochondrial antioxidant genes are major regulators of oxidative stress in the vasculature. Previous data has shown that reducing antioxidant capacity accelerates atherosclerotic plaque size, however, it is unclear whether reducing mitochondrial antioxidant capacity alters plaque composition. Therefore, we tested the hypothesis that reductions of mitochondrial antioxidant capacity will increase osteogenic markers and intimal plaque calcification in aorta from hypercholesterolemic mice. We used Ldlr -/- ApoB 100/100 (LA) mice that were either wild-type (LA-MnSOD +/+ ) or heterozygous (LA-MnSOD +/- ) for manganese superoxide dismutase that were fed a Western Diet (TD88137) for 3 or 6 months. Changes in changes gene expression were assessed in aortic arch using qRT-PCR, plaque calcium levels were examined using histological evaluation from aortic cryosections, and endothelial function was measured using isolated organ bath chambers ( ex vivo ). While expression of SP7 (an osteogenic transcription factor) was unchanged between LA-MnSOD +/+ (1.0±0.6) and LA-MnSOD +/- (0.95±0.3) at 3 months, SP7 levels were dramatically increased in LA-MnSOD +/+ (8.0±2.8) at 6 months and increased further in LA-MnSOD +/+ mice (17.5±5.9; p<0.05). Alizarin Red staining for calcium showed increased calcium burden in plaques of 6 month LA-MnSOD +/- compared to LA-MnSOD +/+ (3636±669 pixels versus 649±206 pixels; p<0.05). Vasomotor function in aorta was unchanged between LA-MnSOD +/+ and LA-MnSOD +/- mice, suggesting that the transcriptional and functional changes related to calcification in LA-MnSOD +/- mice occur independently of nitric oxide signaling. In conclusion, our data strongly support our working hypothesis that reductions in mitochondrial antioxidant capacity significantly accelerate osteogenic calcification in aorta from hypercholesterolemic mice, and suggest that targeting mitochondrial oxidative stress may be a viable therapeutic strategy to slow vascular calcification.

Cardiovascular calcification is often an active process involving re-differentiation of cells to an osteoblast-like phenotype, and is strongly associated with aging. Sirtuins are histone and protein deacetylases that play a critical role in a number of age-related diseases, and SIRT6-deficient mice have the most profound progeroid phenotype. The role of SIRT6 in the regulation of cardiovascular calcification, however, is poorly understood. We hypothesized that age-related reductions in sirtuins promote osteogenic differentiation in cells from aortic valve and aorta. We used aortic valve interstitial cells isolated from SIRT6-deficient mice, pigs, or humans, and vascular smooth muscle cells from mouse aorta to determine whether osteogenic differentiation of cardiovascular cells is regulated by SIRT6 levels. In line with our hypothesis, we found increased p-SMAD1/5/8 levels in cells from SIRT6-deficient aortic valve and aorta. Second, global SIRT inhibition with NAD+-depletion increased histone acetylation and protein levels of the osteogenic proteins Sp7 and Runx2. Third, inhibition of SIRTs enhanced the transcriptional responses of Sp7 and Runx2 when cells were treated with BMP2. Fourth, knockdown of SIRT6 in mouse and human aorta and aortic valve cells by RNA interference increased Sp7 expression. Finally, chronic pharmacological inhibition of SIRTs promoted calcified nodule formation and calcium deposition in cultured mouse/swine aortic valve cells. Collectively, our data suggest that reductions in SIRT6 levels favor amplification of pro-osteogenic signaling and osteogenic differentiation of cardiovascular cells and promote the formation of calcified nodules in vitro . These data also suggest that sirtuin re-activation may be a useful therapy to slow initiation and/or progression of cardiovascular calcification in vivo.

Thrombin plays a major role in coagulation cascade and participates in regulation of vascular function through its effects on protease-activated receptors (PARs). The effect of thrombin inhibition on vascular and aortic valve function during prolonged hypercholesterolemia, however, remains unknown. In the present study, we tested the hypothesis that direct thrombin inhibition (DTI, dabigatran etexilate) slows progression of endothelial and aortic valve dysfunction in ldlr-deficient, apolipoprotein B100-only (LA) mice. Carotid arteries from LA mice were used for studies of endothelial function in isolated organ baths, and aortic valve function was evaluated using echocardiography. In placebo-treated mice, maximum relaxation to acetylcholine was reduced following 6 and 9 months of hypercholesterolemia (maximum relaxation to acetylcholine (MR ACH ) =75%± 3.7, and 67%± 3.6, respectively), but not affected by DTI (75%± 2.5, and 62%± 3.2, respectively). Relaxation to exogenous NO donors was also unaffected by DTI. Aortic valve function progressively deteriorated at 3, 6, and 9 months in placebo-treated mice (2.1 m/s± 0.04, 2.5 m/s± 0.1, and 2.9m/s ± 0.1, respectively), and was not affected by DTI (2.0m/s± 0.1, 2.5m/s± 0.1 and 3.0m/s ± 0.4, respectively). While PAR receptor expression was detected in aorta and aortic valve tissue, expression of SPHK1 (an indicator of PAR1 activity) was not altered by DTI. Expression of CTGF in aortic valves, however, tended to be reduced by DTI. In conclusion, our data suggest that DTI does not prevent vasomotor or aortic valve dysfunction in mice with prolonged hypercholesterolemia, but may attenuate pro-fibrotic signaling in aortic valves.

Introduction While progression of calcific aortic valve stenosis (CAVS) is known to be strongly associated with activation of pro-inflammatory and pro-osteogenic pathways, effective treatments for CAVS have remained elusive. TNF-α is a pleiotropic inflammatory cytokine and is upregulated in patients with CAVS. Hence, we hypothesized that TNF-α inhibition, using the neutralizing antibody infliximab, can slow the progression of CAVS in vivo . Methods We used Ldlr -/- Apob 100/100 mice, which develop robust CAVS in an age-dependent manner, as a model. Mice were fed a Western-type diet for 6 months, then randomized into one of two groups: 1) twice-weekly infliximab injections (10 mg/kg intraperitoneally; n=10), and 2) saline injections (n=9). Aortic valve function was evaluated using echocardiography, and gene expression in aortic valves was measured using quantitative real-time PCR. Data are presented as mean ± SEM. Results During the 3 months preceding treatment initiation, aortic valve cusp separation distance decreased progressively in both groups (reduction by 0.082±0.011 mm in control group, 0.095±0.016 mm in infliximab group; p =0.27). Three months after initiation of treatment, cusp separation continued to decrease in control mice (by 0.108±0.021 mm), while progression of valve dysfunction was slower by 42% in the infliximab group (reduction of 0.063±0.009 mm; p =0.03 vs control). Expression of the pro-inflammatory genes Vcam1 and Nos2 was significantly reduced in valves of the infliximab group (by 37% and 60%, respectively). Moreover, endothelium-dependent vasodilation was improved by infliximab (improvement in maximum relaxation to acetylcholine in infliximab group 18±6% compared to control). Conclusions This study is the first to report that in vivo inhibition of TNF-α slows progression of CAVS in mice with established valve disease. This beneficial effect was accompanied by reduction in pro-inflammatory gene expression and improvement of systemic endothelial function, suggesting infliximab may exert its effects in part by increasing nitric oxide bioavailability. The findings warrant further investigation into the role of TNF-α in initiation and progression of CAVS, as well as its potential as a therapeutic target in human CAVS.

Increases in mitochondrial reactive oxygen species (ROS) can contribute to endothelial dysfunction in conduit arteries by reducing nitric oxide bioavailability. SIRT3 is a protein deacetylase localized to mitochondria and regulates activity of numerous electron transport chain enzymes, and can be inactivated by increases in mitochondrial ROS. Whether reductions in SIRT3 are an independent contributor to endothelial dysfunction, however, are unknown. Therefore, we hypothesized that SIRT3 null mice (SIRT3 -/- ) will have impaired relaxation to acetylcholine (MR Ach ) compared to wild-type littermates (SIRT3 +/+ ), and that this difference will be exacerbated by Western diet feeding for 3 months (TD88137, 42% of kcal from fat, 0.2% cholesterol). Endothelial function was measured using isolated organ bath chambers, and endothelium-dependent (MR Ach ) and endothelium-independent (sodium nitroprusside, MR SNP ) relaxation were measured. In contrast to our hypothesis, MR ACH was virtually identical between SIRT3 +/+ and SIRT3 -/- fed a standard chow diet (MR ACH = 68±2% and 69±3%; respectively), and MR SNP also remained unchanged between the two genotypes. In groups fed a Western diet, however, endothelium-dependent relaxation was significantly impaired in SIRT3 -/- mice compared to SIRT3 +/+ littermates (MR ACH = 49±4% and 58±3%, respectively; p<0.05). MR SNP was not altered by either Western diet feeding or SIRT3 deficiency. Interestingly, impairments in MR Ach in SIRT3 -/- mice fed a Western diet were not rescued with acute incubation of apocynin (an NAD(P)H oxidase inhibitor) or mito-tempol (a mitochondria-targeted antioxidant), suggesting that chronic antioxidant and/or acetyltransferase inhibition may be required to improve vasomotor function in this model. Collectively, our data suggest that SIRT3 plays a major role in regulating endothelial function with dietary stress, and may be a useful therapeutic target to reduce cardiovascular risk.