Miller, Jordan D.

Introduction: Myxomatous mitral valve disease (MMVD) is a degenerative condition characterized by tissue fibrosis and matrix remodeling which ultimately progresses to leaflet prolapse. While recent work has shown that transforming growth factor beta (TGF-β) signaling contributes to the pathogenesis of MMVD, upstream regulators of this and other pathways remain poorly understood. Hypothesis: We sought to use high-throughput RNA sequencing coupled with microRNA (miRNA) sequencing to identify novel molecular targets as well as upstream regulators contributing to MMVD. Methods: We conducted mRNA and miRNA sequencing on normal (n = 10) and myxomatous human mitral valve samples (n = 10). Differential expression was identified using linear modeling and parallel random forest analyses. Canonical pathways were identified by Ingenuity Pathway Analyses (IPA). Predicted miRNA targets were identified using TargetScanHuman 6.2. Results: We found 2784 mRNAs that were differentially expressed between normal and myxomatous mitral valves, which IPA largely categorized in to pro-fibrotic, matrix remodeling and cellular proliferation signaling. In miRNA sequenced from the same samples, 67 miRNAs were differentially-expressed between normal and myxomatous mitral valves. Increased expression of TGF-β ligands, collagen isoforms, and matrix metalloproteinases were associated with reductions in miRNAs predicted to target them. Conversely, mRNA levels of the “protective” genes TGFβ-induced factor homeobox 1, salt-inducible kinase 1, TIMP metallopeptidase inhibitor 4, and cyclin-dependent kinase inhibitor 1C mRNA levels were decreased in myxomatous tissue, and miRNAs predicted to target these genes (e.g., miR-656, miR-379-3p, miR-664a-3p, and miR-34c-5p) were significantly increased. Conclusions: Collectively, these data not only identify novel genes that are differentially regulated in MMVD, but also suggest miRNAs may play an active role in suppressing key protective molecules in MMVD. Thus, anti-mIRs therapy may be a viable therapeutic target to restore anti-fibrotic and anti-proliferative molecules in the valve and slow progression of MMVD. Future mechanistic studies will lay a critical foundation for translational work in these areas.

Increasing age is a major risk factor for calcific aortic valve disease (CAVD). Interestingly, SIRT6 knockout mice have a marked progeroid phenotype, and we recently reported that sirtuin enzyme expression is dramatically reduced with aging and that SIRT6 expression is reduced further in valves from patients with end-stage CAVD. It is unknown, however, whether experimentally reducing SIRT6 promotes osteogenic signaling in the valve and ultimately accelerates progression of CAVD. Thus, we used cultured mouse aortic valve interstitial cells and ldlr-deficient, apolipoprotein B100-only mice (LA) that were SIRT6 wild-type (LA-SIRT6 +/+ ) or heterozygous (LA-SIRT6 +/- ) and fed a Western diet for 3 or 12 months to determine the role of SIRT6 in valve calcification. In vitro , reduction of SIRT6 increased histone acetylation and significantly increased mRNA and protein levels of the osteogenic genes Runx2 and Sp7 in response to bone morphogenetic protein 2 (100ng/ml BMP2 for 18 hours), and siRNA knockdown of SIRT6 increased mRNA levels of Sp7 even in the absence of exogenous bone morphogens. Using high-resolution ultrasound to evaluate aortic valve function in vivo, we found that 3 month old LA-SIRT6 +/- mice did not have significant impairments in valve function compared to LA-SIRT6 +/+ mice. In contrast, 12 month old LA-SIRT6 +/- had dramatically worsened aortic valve dysfunction and stenosis compared to LA-SIRT6 +/+ mice, which was also associated with reductions in left ventricular ejection fraction. Collectively, our data strongly suggest SIRT6 plays a critical role in the tonic repression of osteogenic signaling in the aortic valve, and that age-related reductions in SIRT6 are likely to increase susceptibility to valve calcification in response to risk factors for CAVD such as hypercholesterolemia. Collectively, increasing activity of SIRT6 or reducing acetylation of its targets may serve as viable therapeutic strategies to slow progression of age-related valvular calcification and stenosis.

Endothelial cell senescence promotes inflammation and impairs endothelium-dependent vasodilation. While aging, hypercholesterolemia, and caloric excess are known to drive cellular senescence in part through activation of mTOR, it is unknown whether reducing mTOR signaling can improve vasomotor function in aging mice. Here, we used hypercholesterolemic mice (Ldlr -/- /apoB 100/100 , or LA) carrying intact copies of mTOR (LA-mTOR +/+ ) or were deficient in one copy of mTOR (LA- mTOR +/- ). Mice were fed Western Diet for 6 or 12 months, and vasomotor function in aorta was evaluated using isolated organ chamber baths. Maximum relaxation to acetylcholine (MR ACH ) was progressively impaired from 6 months to 12 months in LA-mTOR +/+ mice (55±4% and 30±4%, respectively). In contrast to our hypothesis, however, MR ACH was not significantly improved in LA-mTOR +/- mice at either time point (62±3% and 35±5%, respectively). Interestingly, a subgroup analysis showed significant improvement in MR ACH in male mice at the 6-month point (male LA-mTOR +/+ mice = 30 ± 9% vs male LA-mTOR +/- mice = 55±4%; p < 0.01), whereas vasomotor function in female mice was nearly identical across groups. There was no significant difference at the 12-month point (male LA-mTOR +/+ mice = 19 ± 3% vs male LA-mTOR +/- mice = 22 ± 5%). Relaxation to acetylcholine was attenuated by L-NAME at 6 months and 12 months (but identical between genotypes) and relaxation to nitroprusside was identical between strains at each time point, suggesting that changes in endothelial function were not masked by compensatory mechanisms (e.g., endothelium-derived hyperpolarizing factors). Furthermore, reduction of mTOR did not elicit changes in maximal tension generated in response to Prostaglandin F 2 α at either time point. Collectively, despite compelling evidence from in vitro model systems, genetic reduction of mTOR signaling in vivo does not appear to be sufficient to improve vasomotor function in our model of aging hypercholesterolemic mice. Our data do, however, shed light on a potential role of sex in dictating phenotypic changes in early to moderate stages of disease, which warrants further investigation.

Marfan syndrome (MFS) is a genetic disease with a mutation for the microfibrillar constituent protein fibrillin-1 being the most prevalent. MFS is often associated with progressive aortic root dilation, ultimately progressing to aortic aneurysm and dissection. While recent work has shown that increased angiotensin-II receptor type-1 and transforming growth factor beta (TGF-β) signaling contributes to aneurysm formation in aorta, efficacious therapeutic targets remain elusive. Given previous reports of progeriod phenotypes in a subset of Marfan patients, we sought to determine whether there are molecular changes that are consistent with accelerated aging in a mouse model of Marfan syndrome. In mice carrying a loss-of-function mutation in fibrillin-1, we assessed aortic root dimensions by echocardiography and gene expression levels of TGF-β1-3, runt related transcription factor 2 (RUNX2), and the cellular senescent marker CDKN2A by quantitative real time PCR at 3 and 14 months of age. As expected, aortic sinus dimensions did not change significantly with aging in wild type mice, but increased dramatically in fibrillin-1 mutant mice compared to wild-type littermate controls and with age (p < 0.05 for both). TGF-β1 ligand expression paralleled age and disease-dependent changes in aortic dimensions, however TGF-β2 and TGF-β3 mRNA levels did not. Aortic dilatation was associated with increased gene expression of RUNX2 with aging and in marfanoid mice. Interestingly, fibrillin-1 mutant mice demonstrated marked increases in expression of the anti-proliferative cell-cycle checkpoint protein CDKN2A at both time points, and correlated with changes in TGF-β1 (R 2 =0.54) and RUNX2 (R 2 =0.69) mRNA. CDNK2A gene expression patterns, however, demonstrated a poor correlation with expression of TGF-β2 and TGF-β3 (R 2 =0.04 and 0.06. respectively). Collectively, these data lend insight into novel mechanisms that may regulate development of aortic root dilation in patients MFS and are the first to implicate increased senescent cell burden in Marfan syndrome. Furthermore, we propose that clearance of senescent cells could be a viable therapeutic intervention to slow progression aortic root-dilation and aneurysm in patients with Marfan syndrome.

Increasing age is the greatest risk factor for development and progression of calcific aortic valve disease (CAVD), and mitochondrial dysfunction has been implicated in the pathogenesis of cardiovascular calcification. Previous findings by our group suggested a significant reduction in expression of multiple sirtuin (SIRT) isoforms in normocholesterolemic aged mouse aortic valves, however, it is unclear if losses in SIRT3—a major mitochondrial SIRT isoform—modulates progression of aortic valve calcification. Therefore, we hypothesized that loss of SIRT3 in a hypercholesterolemic mouse model of CAVD will result in augmented calcium burden in aortic valve, increased osteogenic signaling, and impaired aortic valve function. To test this we used mice that were Ldlr -/- /ApoB 100/100 mice that were either wild-type (LA-SIRT3 +/+ ) or null for SIRT3 (LA-SIRT3 -/- ) and fed a western diet (TD88137) for 12 months. Alizarin red was used to quantitate calcium burden, qRT-PCR was used to measure changes in mRNA levels, and high-resolution echocardiography was used to assess aortic valve function (cusp separation distance). In line with our hypothesis, we observed a substantial increase in calcium burden in LA-SIRT3 -/- mice compared to their LA-SIRT3 +/+ littermates(11.7±4.0, 4.8±1.8; respectively). Interestingly, expression of Runx2 and osterix—classic markers of osteogenic differentiation—were decreased in LA-SIRT3 -/- mice compared to their LA-SIRT3 +/+ littermates. Expression of Msx2, were markedly increased in aortic valve tissue from LA-SIRT3-deficient mice. Despite these histological and molecular changes, SIRT3 deletion did not alter cusp separation distance (LA-SIRT3 +/+ = 0.81±0.04; LA-SIRT3 -/- = 0.81±0.02). Collectively, this data suggest losses in SIRT3 can contribute to accelerated valve calcification through unclear molecular mechanisms, but these changes are not sufficient to drive reductions in cusp separation distance. Additional experiments delineating the histopathological and molecular sequelae of SIRT3 deletion will be critical to understanding its role in the pathogenesis of CAVD.

Marfan syndrome is a connective tissue disorder frequently driven by mutations in the fibrillin-1 gene (Fbn1). This results in multiple aberrant cardiovascular phenotypes, including aortic aneurysm and mitral valve prolapse. While the molecular changes underlying aortic aneurysm formation have been extensively studied, the molecular underpinnings of mitral valve prolapse in this syndrome remain poorly understood. Therefore, we hypothesized that smad2/3 phosphorylation would be significantly increased in both aorta and mitral valve from Fbn1 +/1037G mice with Marfan syndrome and associated with impaired anti-inflammatory signaling in both tissues. We used young (3 months) mice that were wild-type (Fbn1 +/+ ) or fibrillin-1 mutant (Fbn1 +/1037G ) and fed a normal chow diet. qRT-PCR was used to measure mRNA levels in aortic arch and mitral valve tissue, and fluorescent immunohistochemistry to evaluate changes in canonical TGFβ signaling. Consistent with previous reports, TGF-β1 mRNA levels were increased in Fbn1 +/1037G compared to wild-type mice in both aorta and mitral valve. Interestingly, while p-SMAD2/3 protein levels were increased in aorta from Fbn1 +/1037G mice, we were surprised to find that p-SMAD2/3 protein levels were decreased in mitral valve from Fbn1 +/1037G mice compared their wild-type littermates. Given the emerging role of inflammatory signaling in accelerated development of Marfan phenotypes, we sought to determine whether there were compensatory or maladaptive changes in IL10 (a known inhibitor of p-SMAD2/3). In aorta, IL10 mRNA levels were decreased in Fbn1 +/1037G compared to Fbn1 +/+ mice, suggesting a maladaptive response. In contrast, IL10 levels in mitral valve tissue were significantly increased in Fbn1 +/1037G mice compared to their wild-type littermates, suggesting a compensatory/protective response. Collectively, these data suggest that tissue-specific changes in IL10 levels may modulate p-SMAD2/3 signaling in aorta and mitral valve, which may be a key permissive step in the onset of Marfan-related phenotypes. Future work to experimentally determine the role of IL10 in the regulation of canonical and non-canonical TGFβ signaling and the penetrance of Marfan-associated phenotypes is warranted.

Previous studies have reported increases in oxidative stress due to reductions in antioxidant enzymes in calcified regions of stenotic human aortic valves. Whether reductions in antioxidant enzymes are adaptive, maladaptive, or an epiphenomenon in calcifying aortic valves remains unclear. Thus, the purpose of this study was to test whether overexpression of catalase reduces valvular calcification and slows progression of aortic valve stenosis in mice. Wild type (WT) mice with calcific aortic valve disease (ldlr-deficient, apoB100-only) were crossed with transgenic mice that overexpress human catalase (TG). All mice were fed a western diet (TD88137; Harlan Teklad). Functional echocardiogram data was collected and analyzed at early (3 months) and advanced (12 months) stages of valve disease, and valves were OCT embedded, sectioned, and stained for calcium levels (Alizarin Red). The transgenic valves show a slight increase in calcification compared to wild type counterparts. While there were no differences in valve function between genotypes at the early time point, cusp separation was decreased in the transgenic mice (0.74 ± 0.02 mm) compared to their wild-type littermates (0.84 ± 0.03 mm) at the 12 month time point (p < 0.05). In a subset of animals, valve calcification also tended to increase in transgenic mice at both time points compared to their wild-type littermates. In conclusion, overexpression of catalase resulted in a surprising finding of accelerated valvular calcification and stenosis. Interestingly, this finding aligns with the findings from several negative clinical trials testing antioxidant therapies in humans with atherosclerosis, suggests that the net impact of changing redox balance on disease progression are highly context dependent, and that increasing antioxidant capacity may not be a viable strategy to slow progression of calcific aortic valve disease in humans.

Endothelial dysfunction and arterial stiffening major risk factors associated with progression of atherosclerosis and increased risk of acute coronary events. Recent data suggest sirtuins may play a role in protecting against a variety age-related diseases, and we recently reported that sirtuins are reduced in the aging vasculature. The roles of SIRT3 (a deacetylase that is largely restricted to the mitochondria) in the regulation of vasomotor function and arterial stiffness, however, remains poorly understood. Therefore, we hypothesized that genetic inactivation of SIRT3 in hypercholesterolemic mice will impair endothelial function and reduce vascular compliance compared to hypercholesterolemic mice with normal levels of SIRT3. We used Ldlr -/- /ApoB 100/100 (LA) mice that were either wild-type for SIRT3 (LA-SIRT3 +/+ ) or null for SIRT3 (LA-SIRT3 -/- ) and fed a western diet for 12 months. At time of sacrifice, aorta was removed for isolated organ chamber baths to measure changes in endothelium-dependent relaxation to acetylcholine (ACH), and carotid arteries were harvested to assess changes in vascular compliance using a pressurized cannulation system. Surprisingly, losses of SIRT3 did not impair relaxation to ACH in aorta compared to LA-SIRT3 +/+ mice (38.2±4.4%, 37.0±4.4%, respectively). Furthermore, reductions in SIRT3 did not impact vascular compliance or distensibility in carotid arteries from the same animals. In conclusion, in contrast to our hypothesis, inactivation of SIRT3 does not impact vasomotor function or conduit vessel compliance in advanced atherosclerosis. Importantly, this and may be explained by a “floor-effect”, and future work examining the effect of SIRT3 overexpression on vascular function and pathology will be instrumental in determining its biological role and ultimate therapeutic utility.

Activation of NADPH oxidase is a major contributor to hypertension, reactive oxygen species generation, endothelial dysfunction and increased vascular distensibility in angiotensin II (AngII)‐induced hypertension. Since exercise has been associated with increased antioxidant enzyme expression in blood vessels, we hypothesized that exercise training would improve endothelial function and vascular distensibility in mice with AngII‐induced hypertension. C57BL/6J mice received saline or AngII (1000 ng/kg/min) via osmotic minipumps, and were assigned to sedentary or treadmill exercise groups (13 m/min for 30 minutes, 5 days/week) for 14 days. Endothelium‐dependent relaxation in response to acetylcholine and arterial distensibility were assessed in carotid arteries using a pressurized arteriograph system. Gene expression was examined by qRT‐PCR. In AngII‐treated sedentary mice, endothelium‐dependent relaxation was impaired, while distensibility was higher than in saline‐infused sedentary mice. Compared to AngII‐infused sedentary mice, however, endothelium‐dependent relaxation was partially improved, and distensibility normalized in AngII‐infused exercised mice. In response to exercise, mRNA levels of SOD1 and SOD2 in aorta were increased, while mRNA levels of NOX2 were reduced. We conclude that exercise training can dramatically improve endothelial dysfunction and reverse derangements in carotid distensibility following prolonged AngII‐induced hypertension. Additional work examining mechanisms underlying our observed transcriptional changes will be critical to development of novel therapeutic targets.

We previously reported that MnSOD‐deficiency does not cause overt vasomotor dysfunction in aged or hypercholesterolemic mice. To determine whether increasing MnSOD protects against vasomotor dysfunction and deleterious molecular changes in atherosclerosis, we crossed hypercholesterolemic mice with mice overexpressing MnSOD to test the hypothesis that lifelong increases in MnSOD protect against vasomotor dysfunction. We used Ldlr ‐/‐ /ApoB 100/100 mice that were wild‐type (MnSOD 0/0 ) or transgenic (MnSOD Tg/0 ) and fed a western diet for 6 months. We used organ bath chambers to measure aortic endothelium‐dependent relaxation (relaxation to acetylcholine, Ach), and mRNA from aortic arch was used for qRT‐PCR. Surprisingly, relaxation to Ach was unchanged between MnSOD 0/0 and MnSOD Tg/0 mice, as was relaxation to sodium nitroprusside. As expected, gene expression of MnSOD was significantly increased in the MnSOD Tg/0 mice compared to MnSOD 0/0 mice. Unlike our findings in MnSOD +/‐ mice, however (where Nox2 was increased), Nox2 expression was unchanged between MnSOD 0/0 and MnSOD Tg/0 mice. Importantly, markers of osteogenesis such as Runx2 and BMP2 were unchanged between genotypes. Furthermore, MnSOD overexpression did not alter cellular senescence markers. Collectively, our data support the surprising conclusion that physiologically relevant increases in MnSOD do not significantly affect endothelial function or molecular regulators of histopathological changes in advanced atherosclerosis in hypercholesterolemic mice.