Porphyrin‐Based SOD Mimic MnTnBuOE‐2‐PyP⁵⁺ Inhibits Mechanisms of Aortic Valve Remodeling in Human and Murine Models of Aortic Valve Sclerosis

Abstract

Background

Aortic valve sclerosis (AVSc), the early asymptomatic presentation of calcific aortic valve (AV) disease, affects 25% to 30% of patients aged >65 years. In vitro and ex vivo experiments with antioxidant strategies and antagonists of osteogenic differentiation revealed that AVSc is reversible. In this study, we characterized the underlying changes in the extracellular matrix architecture and valve interstitial cell activation in AVSc and tested in vitro and in vivo the activity of a clinically approved SOD (superoxide dismutase) mimic and redox‐active drug MnTnBuOE‐2‐PyP⁵⁺ (BMX‐001).

Methods and Results

After receiving informed consent, samples from patients with AVSc, AV stenosis, and controls were collected. Uniaxial mechanical stimulation and in vitro studies on human valve interstitial cells were performed. An angiotensin II chronic infusion model was used to impose AV thickening and remodeling. We characterized extracellular matrix structures by small‐angle light scattering, scanning electron microscopy, histology, and mass spectrometry. Diseased human valves showed altered collagen fiber alignment and ultrastructural changes in AVSc, accumulation of oxidized cross‐linking products in AV stenosis, and reversible expression of extracellular matrix regulators ex vivo. We demonstrated that MnTnBuOE‐2‐PyP⁵⁺ inhibits human valve interstitial cell activation and extracellular matrix remodeling in a murine model (C57BL/6J) of AVSc by electron microscopy and histology.

Conclusions

AVSc is associated with architectural remodeling despite marginal effects on the mechanical properties in both human and mice. MnTnBuOE‐2‐PyP⁵⁺ controls AV thickening in a murine model of AVSc. Because this compound has been approved recently for clinical use, this work could shift the focus for the treatment of calcific AV disease, moving from AV stenosis to an earlier presentation (AVSc) that could be more responsive to medical therapies.

Clinical Perspective

What Is New?

• Aortic valve sclerosis, the early presentation calcific aortic valve disease, is associated with pronounced architectural remodeling despite marginal effects on mechanical properties in both human and mice.

• MnTnBuOE‐2‐PyP⁵⁺ prevents aortic valve thickening in a murine model of aortic valve sclerosis.

What Are the Clinical Implications?

• Because this compound has been approved recently for clinical use for other indications, this work could shift the focus for the treatment of calcific aortic valve disease, moving from aortic valve stenosis to an earlier presentation (aortic valve sclerosis) that could be more responsive to medical treatment.

Introduction

Aortic valve sclerosis (AVSc), the early presentation calcific aortic valve disease (CAVD), is a hallmark of several cardiovascular conditions including congestive heart failure, myocardial infarction, and severe aortic valve stenosis (AS).1, 2, 3, 4, 5 Several clinical trials have failed to halt the development of CAVD, rendering this condition as treatable only at the end stages, through surgical or percutaneous means.6, 7, 8, 9 Sclerosis presents in 25% to 30% of patients aged >65 years and in up to 40% of those aged >75 years. It is characterized by thickening of the aortic valve (AV) leaflets with marginal effects on the mechanical properties of the valve, making its presentation largely asymptomatic.1, 10, 11 Of these patients, 10% will develop symptomatic AS within a decade. AS inevitably leads to heart failure and death unless patients undergo replacement of the valve.12, 13 Potential therapies are hampered by our poor understanding of the molecular mechanisms leading to AVSc and by the asymptomatic nature of its clinical presentation.

Although sclerotic valves are hemodynamically normal, valve interstitial cells (VICs) have already been activated and remodeling of the leaflet has begun.1, 14, 15, 16 We and others have shown that an early pathological event in the remodeling of the AV is the accumulation of reactive oxygen species (ROS).14, 17 ROS generates an oxidative injury that activates VICs with an impaired response to DNA damage in the sclerotic AV. In both humans and mice, impaired cellular responses to DNA damage are associated with premature aging and valve disease.18 A hallmark of this phenomenon is the sustained formation of subnuclear structures known as DNA‐repair foci.14 We demonstrated that adenoviral delivery of antioxidant enzymes reverses ROS‐induced DNA damage and reduces AVSc‐derived VIC activation and calcification.14

Another key pathway in AVSc is the osteogenic differentiation induced by the TGF‐β (transforming growth factor β) superfamily.15, 19 We and others reported that mechanical stimulation and BMP4 (bone morphogenetic protein 4) are able to further differentiate VICs, with expression of extracellular matrix (ECM) regulators, such as osteopontin, CTGF (connective tissue growth factor), TSP (thrombospondin)‐1 and ‐2, collagen, and osteocalcin, promoting valvular ectopic calcification.15, 20, 21 A BMP4 antagonist, Noggin, reverses this effect. The functional association of osteopontin with one of its receptors, CD44v6, contributes to the regulation of calcium deposition in vitro and ex vivo.16

Perhaps a major challenge in advancing our knowledge of AV disease is the lack of a reliable animal model to translate basic and clinical science data into an in vivo experiment. Available models can develop calcification of the AV structures, but they rarely develop the large drop in pressure across the valve and other hemodynamically significant changes seen in human patients. We have shown that a lesion similar to human AVSc can be elicited in a mouse model given a hypercholesterolemic diet in combination with chronic infusion of Ang II (angiotensin II). Analyses of the aortic valve in these animals demonstrate ROS accumulation, cusp thickening, and ECM remodeling. This model does not mimic AS but rather resembles the early initiating phases of AVSc.

Oxidative damage to DNA is thought to contribute to a number of human pathologies. To modulate the oxidative response in vitro and ex vivo, a class of manganese porphyrin‐based compounds have been identified. They are potent anti‐inflammatory agents that catalytically inactivate a range of ROS and interact with transcription factors.22, 23, 24, 25 An analog, MnTnBuOE‐2‐PyP⁵⁺ (MnBuOE),24 is currently in early phase clinical trials (NCT02655601, NCT0338650, and NCT02990468) to mitigate the oxidative damage caused by radiation therapy.

We hypothesized that sclerotic AV leaflets, although normally functioning, are already biologically altered at both extracellular and cellular levels. To investigate this hypothesis, we pursued an experimental design that involved human aortic leaflet specimens, a murine model of AVSc, and a biomechanical simulator. The human biobanking efforts of 2 different institutions allowed us to compare explanted human AV leaflets from across the spectrum, from healthy to AVSc to AS. This work intends to provide an in‐depth characterization of the early changes in the AV architecture, using histological, structural, and biomechanical analysis. We also tested the ability of the SOD (superoxide dismutase) mimic MnBuOE to control activation of human VICs in vitro and to reduce thickening of AV leaflets in a murine model. Because our SOD mimic MnBuOE has been approved recently for clinical trials, our work represents one of the first steps toward better understanding and eventually mitigating the pathogenesis of aortic valve degeneration.

Method

The authors declare that all supporting data are available within the article (and its online supplementary files).

Study Population and Patient Adjudication

Sclerotic aortic valves were collected from the explanted hearts of patients undergoing heart transplantation and from deceased donor hearts that were considered unsuitable for transplantation following institutional review board–approved guidelines of the University of Pennsylvania Perelman School of Medicine (protocol no. 809349), Columbia University (AAAR6796), and the Valley Hospital (protocol no. 20132309). Informed consent was obtained for participant enrollment, and clinical information was obtained by patient interview and chart review. AVs from AS patients were collected from patients undergoing AV replacement, as described previously.14, 15, 16 Surgical patients were assessed by transesophageal echocardiography, computed tomography, or both and confirmed by intraoperative observation if applicable. AVSc has been defined by the American Heart Association and American College of Cardiology guidelines14 as focal areas of leaflet calcification with leaflet thickening. For this study, we defined AVSc as irregular, nonuniform thickening of portions of the AV leaflets or commissures or both, thickened portions of the AV with/or without an appearance suggesting calcification (ie, bright echoes), nonrestricted or minimally restricted opening of the aortic cusps, and peak continuous‐wave Doppler velocity across the valve <2 m/s.

In Vitro MnBuOE, MnE (MnTE‐2‐PyP5+), and TGF‐β Treatment of VICs

Patients’ derived cells were treated with 30 μmol/L MnBuOE, 30 μmol/L MnE (MnTE‐2‐PyP5+), and 10 ng/mL TGF‐β and harvested 72 hours later for RNA and protein analysis. Gene expression level was standardized to actin B, and fold changes were calculated using untreated cells as basal.

Collagen Fiber Alignment

Small‐angle light scattering was used to quantify collagen fiber architecture at different disease stages. Samples used for small‐angle light scattering were shipped to the University of Texas (UT) Austin in PBS. After 2 PBS washes, the samples underwent a gradient glycerolization, as described previously.1 Briefly, the fixed samples were dehydrated, cleared in a graded glycerol/water solution, and stored in 100% glycerol. A 5‐mW nonpolarized continuous HeNe laser (JDS Uniphase Corp) was passed through the tissue samples, and the normalized orientation index (NOI) was used to quantify tissue deformation. The NOIs for each sample (nondiseased, sclerotic, and stenotic) were averaged and used to compute an average NOI representative of the collagen fiber architecture at each CAVD stage.

Microstructural Analysis

Scanning electron microscopy was used to characterize the microstructure of diseased AV leaflets in humans and mice. Samples were fixed in grade II EM‐grade glutaraldehyde (Electron Microscopy Sciences) and shipped to UT Austin in fixative. Samples were stained with osmium tetroxide for 4 hours, critically point dried, and then cut to image the transmural/circumferential cross‐section en face. Samples were mounted on aluminum stubs, coated with carbon paint, sputter coated with 15 nm of Platinum/Palladium using a Cressington 208 Benchtop Sputter Coater, and imaged with a Zeiss Supra 40V scanning electron microscope (Carl Zeiss AG).

Quantification of Oxidized Amino Acid in the AV Explants

Thick paraffin sections (50 μm) were used for oxidized amino acid (OxAA) analyses, as described previously.26, 27 The content of each OxAA in paraffin‐embedded AV leaflet samples was determined by stable isotope dilution liquid chromatography tandem mass spectrometry methods with an AB SCIEX API 5000 triple quadrupole mass spectrometer interfaced with an Aria LX Series HPLC multiplexing system (Cohesive Technologies Inc) in the Mass Spectroscopy Core Facilities of the Cleveland Clinic laboratories, as described.26, 27 Briefly, samples were deparaffinized with xylene and mixed with a single‐phase mixture of H₂O/methanol/H₂O‐saturated diethylether (1:3:8 vol/vol/vol) after protein delipidation and desalting. The mixture was then supplemented with synthetic [¹³C₆]‐labeled OxAA standards and isotope‐labeled precursor amino acids ([¹³C₉, ¹⁵N₁]‐tyrosine and [¹³C₉, ¹⁵N₁]‐phenylalanine). The mixtures were then hydrolyzed under argon gas using methane sulfonic acid and then passed through a C₁₈ solid‐phase extraction column (Discovery DSC₁₈ mini‐column, 3 mL; Supelco). During liquid chromatography tandem mass spectrometry, each OxAA and its precursor, tyrosine or phenylalanine, were quantitated, with results presented as a ratio of OxAA:tyrosine, the precursor amino acid.26, 27

Ang II Chronic Infusion Model

Seven‐week‐old wild‐type male mice (C57BL/6J) were fed with a hypercholesterolemic diet and infused with saline or Ang II (1000 ng/kg per minute) using osmotic pumps (Alzet 2004) for 28 days, as described previously.14, 15, 16 In total, 28 treated and 12 untreated mice were used. Blood pressure measurements were performed in conscious mice using a computerized tail‐cuff method (Coda 8; Kent Scientific Corp), as reported previously. Mice were euthanized with an inhalation overdose of isoflurane. AVs were harvested after 28 days. Experiments were conducted under the approved institutional animal care and use committee protocols 804440 (University of Pennsylvania) and AC‐AAAU6474 (Columbia University).

SOD Mimics Preparation

MnBuOE solution was diluted in water to obtain a concentration of 1 mmol/L. The solution was filtered and stored in a sterile container. MnBuOE was administered by daily intraperitoneal injections at a concentration of 1 mg/kg for 2 days and then 0.5 mg/kg for up to 26 days. Another SOD mimic, MnE, was administered by daily intraperitoneal injections at a concentration of 10 mg/kg per day for the first 2 days and then with 5 mg/kg per day for 26 days.

Bioreactor Design and Tissue Preparation

The tension bioreactor used in this study is similar to the previously described flexure bioreactor used for flexural stimulation of engineered valve tissue, as described earlier.28, 29 The bioreactor consists of 2 chambers, each with multiple media baths; each bath has stationary pins press‐fit into the bottom for tissue anchorage. The opposing pins are fixed to an actuating arm that is attached to a cross‐arm, which is connected to a central motorized piston. Intact human AVs excised in operating rooms were collected in Advanced DMEM and maintained at 4°C to ensure maximal cell survival during transport. Each leaflet was trimmed to form a tissue strip measuring 20 mm circumferentially and 8 mm radially. An attempt was made to keep endothelial cells viable through limited handling. The tissue strip was threaded into the springs and inserted into the tension bioreactor and used for the experiments.

Statistical Analysis

Quantitative data including age at surgery, calcium levels, and OxAA levels in AV leaflets were presented as mean±SE. Given the small sample size, nonparametric tests such as the Mann–Whitney test and the Kruskal–Wallis test (depending on the number of levels within the independent variable) were used to test differences between groups. Multiple comparisons were adjusted using Bonferroni correction. Statistical analyses were performed with R (version 3.4.2; R Foundation for Statistical Computing). P≤0.05 was considered significant.

Results

Human AV Sclerosis Is a Pathological But Reversible Stage of CAVD

Under approved institutional review board protocols, we created a collection of human AV cusps from patients undergoing open heart surgery or heart transplant (Table 1). Alizarin red was used to stain calcific deposits in histological sections and revealed minimal calcification of AVSc compared with AS (Figure 1A). Because altered ECM has been reported in AS, we used real‐time reverse transcriptase–polymerase chain reaction and quantitative polymerase chain reaction to measure the expression levels of functional regulators of valve architecture such as OPN (osteopontin), CTGF (connective tissue growth factor), Cyr61 (cysteine rich angiogenic inducer 61), osteonectin, TSP (thrombospondin)‐1 and ‐2, and tenascin in human cusps from healthy control, AVSc, and AS tissues (n=5 per group). Figure 1B shows that the expression of these proteins increases in the asymptomatic stages of CAVD compared with healthy controls. We then used uniaxial biomechanical testing to confirm the plasticity of AVSc phase of CAVD with data showing that the expression of CTGF, OPN, and Cyr61 is induced by a combination of mechanical stimuli (15% stretch at 1 Hz, resembling the physiological cusp deformation) and osteogenic media (10 ng/mL of BMP4; Figure 1C through 1F), confirming our previously published data.10, 14, 16

Table 1.

Patient Demographics

Demographic	Control (n=16)	AVSc (n=18)	AS (n=23)
Age, y, mean±SE	62.5±1.9	65.6±1.3	73.6±5.18
Male	8 (50.0)	10 (55.5)	14 (60.8)
Smoker	2 (12.5)	7 (38.8)	6 (26.0)
Diabetes mellitus	3 (18.7)	2 (11.1)	3 (13)
Hypertension	5 (31.2)	10 (55.5)	15 (65.2)
Cerebrovascular accident	···	···	1 (4.4)
Peripheral vascular disease	1 (6.23)	1 (5.5)	2 (8.7)
Hyperlipidemia	8 (50.0)	7 (38.8)	7 (30.4)
CAD	4 (25.0)	6 (33.3)	5 (21.7)

Data are shown as n (%) except as noted. AS indicates aortic valve stenosis; AVSc, aortic valve sclerosis; CAD, coronary artery disease.

Figure 1.

A, Alizarin red staining shows increased calcification in human sclerotic and stenotic aortic valves (n=5 per group). B, Bar graph shows fold‐change gene expression of CTGF (connective tissue growth factor), Cyr61 (cysteine‐rich angiogenic inducer 61), SPARC (secreted protein acidic and cysteine rich (osteonectin)), TSP (thrombospondin) ‐1 and ‐2, OPN (osteopontin) transcripts by reverse transcriptase–quantitative polymerase chain reaction (RT‐qPCR). Data were normalized against 18S (actin B) gene expression and represented as fold change+SE. *P, 0.01 (n=5 per group). C and D, Tissue preparation and representation of tensile uniaxial bioreactor, respectively. E, Immunofluorescence staining showing expression of Cyr61 and CTGF in sclerotic aortic valve tissue cultured in static condition in the presence of OS (oscillatory shear stress) + BMP4 (bone morphogenetic protein 4) (100 ng/mL; n=3 per group). F, RT‐PCR of OPN, Cyr61, and CTGF from RNA extracts collected from aortic valve cusps exposed to biomechanical stimulation (15% stretch at 1 Hz) for 6 days (n=5 per group). Osteogenic media enriched with BMP4 (100 ng/mL) was used.

Architectural Remodeling of Collagen Fiber Orientation in Human AVSc

Histological analysis showed that AVSc is characterized by thickening of the cusps (Figure 2A). We measured the collagen fiber networks by picrosirius red staining, which confirms architectural remodeling of ECM in AVSc (Figure 2B). Collagen not only provides the AV tissue with its structural integrity but also affects cellular processes through matricellular, matricrine, and mechanical mechanisms.30, 31 To examine how collagen fiber architecture is organized in different stages of the disease, human aortic valve specimens were fixed, optically cleared with glycerol, and examined with small‐angle light scattering to measure the NOI of the collagen fibers at the different disease states (Figure 2C and 2D). As we predicted, collagen fibers are less aligned in the AVSc and AVS samples compared with the healthy control (nondiseased) tissue, as indicated by the average NOI, which drops from 34% in the healthy valve to 31% and 21% in the valves with AVSc and AS, respectively. Ultrastructural characterization of the VIC microenvironment via electron microscopy further highlights the disorganization of collagen fibrils already visible in AVSc (Figure 2E).

Figure 2.

A, Hematoxylin and eosin (H&E) staining of human aortic valves (AVs) explanted from controls and from AV sclerotic and stenotic patients. B, Picrosirius red staining of control, sclerotic, and stenotic leaflets was visualized using brightfield microscopy and polarized light (n=5 per group). Collagen accumulation was detected in spongiosa and ventricularis areas in both sclerotic and stenotic valves. C, Collagen fiber alignment maps at control, sclerotic, and stenotic stages. Yellow and red represent highly aligned tissue (n=5 per group) (D). E, Scanning electron microscopy was used to show alteration in the microstructure of diseased AV leaflets (n=3 per group). NOI indicates normalized orientation index.

Products of Oxidative Modifications in AVSc and AS

Several pathological factors linked with abnormal mechanical stimuli have been shown to affect VICs and tissue structure. We investigated how the initiating events of AV remodeling affect AV cusp structure and mobility. We and others have reported extensively that ROS affects both collagen architecture and VIC activation.14 We showed that ROS accumulates in the early asymptomatic stages of CAVD (Figure 3A). With stable isotope dilution liquid chromatography tandem mass spectrometry, we measured the accumulation of oxidized products in the explanted leaflets of 15 patients (n=5 per group; Tables 2 and 3). We found accumulation of dityrosine in end‐stage disease (221±41 μmol/mol tyrosine), whereas other oxidized products were not significantly changed among the groups (Figure 3B and 3C). Univariate analysis of OxAA expression with the patients’ comorbidities was performed. No significant association was noted except for dityrosine and comorbidities such as diabetes mellitus (P=0.0014) and hyperlipidemia (P=0.019; Table 4). Healthy and sclerotic AVs had no detectable levels of dityrosine. It is known that dityrosine, besides being an oxidation product, is also a crosslinker. Therefore, dityrosine crosslinking could increase cuspal stiffness and obstruction and thus contribute to ECM remodeling and calcium accumulation leading to impaired mobility typical of late stages of AV disease.

Figure 3.

A, Immunohistochemical staining of control, sclerotic, and stenotic leaflets shows distribution of nitrotyrosine (NO2Tyr) throughout leaflet, with areas of localized high‐intensity staining in aortic valve (AV) sclerotic and stenotic tissues (n=5 per group). B, Oxidized amino acid (OxAA) quantification in explanted leaflets measured by liquid chromatography–tandem mass spectrometry and expressed as μm/mol. C, Dityrosine (diTyr) quantification and calcium scores in control and AV sclerotic and stenotic leaflets. We conducted a Mann–Whitney U test to compare diTyr/tyrosine (diTyr/Tyr) in stenotic samples and controls (or sclerotic samples) and obtained a P value of 0.0075. BrTyr/Tyr indicates bromotyrosine/tyrosine; ClTyr/Tyr, Chlorotyrosine/tyrosine; diTyr/Tyr, dityrosine/tyrosine; m‐Tyr/Phe, meta‐tyrosine/phenylalanine; NO2Tyr/Tyr, nitrotyrosine/tyrosine; o‐Tyr/Phe, ortho‐tyrosine/phenylalanine.

Table 2.

Demographics of Patients for Stable Isotope Dilution LC‐MS/MS

	Control	AVSc	AS
Patients, n	5	5	5
Age, y, mean±SE	36.17±9.44	63.67±6.17	77.91±5.17
Sex
Male	1 (20)	2 (40)	1 (20)
Female	3 (60)	3 (60)	4 (80)
Unknown	1 (20)
Ethnicity
White	3 (60)	5 (100)	5 (100)
Black	1 (20)
Not disclosed	1 (20)
History of smoking	4 (80)	0	0
CAD	1 (20)	1 (20)	4 (80)
Diabetes mellitus	0	0	4 (80)
Hyperlipidemia	1 (20)	1 (20)	5 (100)
Hypertension	1 (20)	3 (60)	4 (80)

Data are shown as n (%) except as noted. AS indicates aortic valve stenosis; AVSc, aortic valve sclerosis; CAD, coronary artery disease; LC‐MS/MS, liquid chromatography tandem mass spectrometry.

Table 3.

OxAA Products From Stable Isotope Dilution LC‐MS/MS

Diagnosis	m‐Tyr/Phe	o‐Tyr/Phe	diTyr/Tyr	NO2Tyr/Tyr	BrTyr/Tyr	ClTyr/Tyr
Control	0	649.5	0	0	2348.6	132.7
Control	210.7	704.8	0	115.5	2546	48.5
Control	215.6	602.6	0	63.3	3392.7	0
Control	224.7	859.1	0	127.2	1725.2	0
Control	249	656.3	0	147	3521.5	0
Sclerosis	75.2	596	0	87.4	3852.2	26.7
Sclerosis	137.6	707.8	0	93.7	3749.8	0
Sclerosis	270.4	519.2	0	178.2	4149.9	0
Sclerosis	249.9	774.7	0	189.9	2811.5	0
Sclerosis	172.4	538.4	0	114.3	2862.3	114
Stenosis	162.5	535.1	310.8	25.2	1468.5	13.4
Stenosis	435.4	959.1	630.2	28.5	1267.1	50.2
Stenosis	268.9	755.1	58.9	34.2	3041.2	44.3
Stenosis	236.1	677.6	93.4	56.9	3223.7	55.6
Stenosis	134.8	661.5	42.6	154.6	2887.8	38.4

LC‐MS/MS indicates liquid chromatography tandem mass spectrometry; OxAA, oxidized amino acid.

Table 4.

Univariate Analysis P Values of Relations Between OxAA Products and Clinical Comorbidities

	m‐Tyr	o‐Tyr	diTyr	NO2Tyr	BrTyr	ClTyr
Sex	0.3037	0.8392	0.9343	0.6354	0.9451	0.4263
Age	0.297	0.871	0.154	0.89	0.672	0.906
Smoking	0.4117	0.7531	0.1397	0.6608	0.2799	0.9462
Congestive heart failure	0.3048	0.4762	0.0856	0.3048	0.6857	0.2542
CAD	0.4894	0.104	0.0873	0.8513	0.2799	0.2002
Diabetes mellitus	0.9495	0.8513	0.0014a	0.2799	0.1773	0.2517
Hyperlipidemia	0.6126	0.7789	0.0192a	0.5358	1	0.2821
Hypertension	0.9551	0.3357	0.1905	0.8665	0.7789	0.232

CAD indicates coronary artery disease; OxAA, oxidized amino acid.

^a P<0.01.

MnBuOE Controls Human VIC Activation and AV Thickening in a Murine Model of AVSc

ROS accumulation, such as superoxide products, can persist in injured tissues for years and have been widely associated with osteogenic differentiation.27, 28, 29 We reported that adenoviral transduction of antioxidant enzymes (SOD [superoxide dismutase] and CAT [catalase]) rescues VICs from an impaired DNA damage response by reducing the expression of early markers of VIC activation and osteoblast‐like differentiation.14 Figure 4A shows enhanced DNA damage response (γ‐H2AX [γ‐H2A histone family member X]) and osteogenic differentiation (RUNX2 [runt related transcription factor 2] and α‐SMA [α‐smooth muscle actin]) of human‐derived aortic VICs from sclerotic patients under H₂O₂ treatment. There are, however, intrinsic limitations to adenoviral delivery in clinical therapy. Manganese porphyrins are a series of drugs that are catalytic scavengers of ${\mathrm{O}}_2^{·-}$ and have been shown to have limited cytotoxicity. In this study, MnBuOE and MnE were used to prevent VICs from osteogenic differentiation in vitro measured by α‐SMA. Figure 4B shows that MnBuOE prevents α‐SMA upregulation induced by TGF‐β, as a pro‐osteogenic factor, on human‐derived VICs.

Figure 4.

A, Western blot showing the expression of H2AX, RUNX2, and SMA (smooth muscle actin) in whole‐cell extract of aortic valve (AV) sclerosis–derived valve interstitial cells treated or not with 500 μmol/L of H₂O₂. B, Bar graph shows fold‐change gene expression of α‐SMA evaluated by reverse transcriptase–quantitative polymerase chain reaction. Data were normalized against actin B and represented as fold change+SE. *Depicts statistically significant difference between Saline and Ang II (P<0.01). **Depicts statistically significant difference between Ang II+MnBuOE and Ang II (P<0.01). C, Relative AV thickness quantified using ImageJ with or without STDEV (n=5 per group) were used to evaluate differences in the thickness. D, Hematoxylin and eosin (H&E) staining of controls and AngII (angiotensin II) infused mouse AV tissues with or without MnBuOE or with MnBuOE alone. E, Collagen accumulation using picrosirius red staining. F, Picrosirius red staining of saline, Ang II, saline plus MnBuOE, and Ang II plus MnBuOE treated mice. G, Scanning electron microscopy used to show alteration in the microstructure of murine aortic valve leaflets.

We next tested the role of MnE and MnBuOE in controlling AV remodeling in vivo. Chronic infusion of Ang II in hypercolesteromic mice resulted in remodeling of the cardiac structures, including thickening of the heart valves.16, 32 The choice of this animal model was driven by growing evidence indicating that Ang II induces its pleiotropic effects through NADPH‐driven generation of ROS. Seven‐week‐old mice were fed a hypercholesterolemic diet and infused with saline or Ang II using osmotic pumps for 28 days in the presence or absence of MnE or MnBuOE, as described in the Method section. Animals receiving chronic Ang II infusion showed increased thickening of the AV (Figure 4C–4G), prevented by cotreatment of ROS scavenger MnBuOE. Figure 4C shows the quantification of AV thickness in murine valves measured with hematoxylin and eosin staining (examples in Figure 4D). In Figure 4E, picrosirus red shows the impact of MnBuOE on collagen fiber. In contrast to these data, another manganese porphyrin, MnE, although a potent SOD mimic, marginally mitigated valve remodeling (Figure S1A–S1C). Such data are likely due to the lower bioavailability of MnE to the tissues, cells, and cellular fragments than of MnBuOE. The log P _OW (the log value of the partition of the drug between n‐octanol and water) was −7.79 for MnE and −4.1025 for MnBuOE.22, 23, 25

Finally, we determined the impact of MnBuOE on murine valve ECM remodeling by both histological analysis and electron microscopy. Movat Pentachrome staining was performed on the murine AVs to evaluate the distribution and composition of the ECM. As shown in Figure 4F Ang II infusion induced thickening of the AV, which is characterized by significant ECM remodeling with abundant proteoglycan accumulation (blue staining) and collagen deposition (yellow staining) distributed through the leaflet. Treatment with MnBuOE significantly prevents accumulation of both ECM components, with a stronger effect on collagen fibers deposition.

Ultrastructural characterization was also performed on explanted murine AV leaflets via electron microscopy. As shown in Figure 4G, collagen fibers are well defined and aligned in the murine AV leaflet of saline treated animals. Ang II infusion induced severe rearrangement of the leaflet and loss of fibril microstructure. Fiber disorganization was significantly reduced by treatment with SOD mimetics, and leaflet display improved collagen alignment compared with Ang II–treated AV.

Discussion

Our data show that AVSc is associated with profound architectural remodeling despite marginal effects on the mechanical properties of the valve in vitro. Our data support the theory that AVSc is an intermediate phase of the clinical progression from healthy to calcified AV. At the cellular level, the once‐quiescent VICs found in a sclerotic AV have become activated and, given the right conditions, will continue to transdifferentiate into a new osteogenic phenotype. The tissue collected from AVSc patients was not calcified, which agrees with the normal hemodynamics of the valve. The tissue did, however, display expression patterns of functional regulators of valve architecture that were more consistent with end‐stage CAVD. The central cusp deformation and dityrosine accumulation were suggestive of the additional mechanisms involved in early leaflet remodeling. Calcium deposition has an additive effect causing further cuspal stiffness and obstruction, contributing to ECM leaflet remodeling and impaired mobility. Further analysis of the microregional distribution of dityrosine and other cross‐linkers could inform on the impact of these oxidative products on the ability of the AV to sustain deformation during each cardiac cycle. The in vitro testing in the bioreactor clearly demonstrated that in the presence of mechanical stimuli and osteogenic media, these activated VICs will continue their transdifferentiation into osteogenic‐type cells. Interestingly, we found that the addition of a SOD mimic seemed to mitigate VIC activation. Treatment of human VICs with the manganese porphyrins prevented osteogenic differentiation in vitro. In the murine model, the administration of these SOD mimics prevents the thickening of AV leaflets, likely the result of a reduction in ECM.

AV degeneration is a long process with multiple stages. Clinically, we recognize AS as the most advanced form of the disease, with a clear impact on the hemodynamic performance of the valve. At that stage, the only form of therapy that resolves the problem is replacement of the diseased valve. Previous attempts to intervene in a nonsurgical or nonpercutaneous fashion, including trials of statin therapy, have failed. The likely cause of that failure was not the form of therapy but rather the timing (ie, at the end of a calcific degenerative process). AVSc is a pathology that carries no hemodynamic consequences, but it is far from a benign state. It is a known risk factor for cardiovascular conditions such as congestive heart failure, myocardial infarction, and aortic valve pathology. In addition, our current work contributes to a growing body of literature that AVSc is a precursor to CAVD and suggests a new potential therapeutic window for the treatment of this condition.

Work by our group and others has suggested a possible role of oxidative products in the subclinical stages and progression of AV degeneration and remodeling. A correlation, for example, between PARP1 (poly[ADP‐ribose] polymerase 1), a key regulator of oxidative DNA damage, and the progression of AS has been described. This process begins with events at the level of DNA. Our group has described the process of VIC transdifferentiation from a quiescent cell into an osteogenic‐type cell, secondary to the effects of oxidative damage. This cellular transdifferentiation is accompanied by oxidative DNA damage manifested cytologically by the presence of a sustained DNA repair foci.

Our results also suggest that manganese porphyrins may inhibit the osteogenic transdifferentiation of VICs and reverse ECM remodeling. We used a variety of cellular and structural tools to identify and manipulate the mechanisms involved in early AV degeneration. These experiments helped us to better understand the different stages involved in the degeneration of AV leaflets and suggest that some of the structural changes may be reversible. Cationic Mn(III) porphyrins are among the most efficacious SOD mimics and redox‐active experimental agents for the treatment of diseases associated with a disturbed cellular redox environment, commonly described as a state of oxidative stress.22, 23, 25, 33 MnE was identified as a first powerful porphyrin‐based SOD mimic with the safest toxicity profile.22, 23, 34, 35 However, it was too hydrophilic and does not cross the blood–brain barrier to a sufficient extent. Indeed, the insufficient lipophilicity of MnE also resulted in its failure in our study. Further development resulted in the generation of MnBuOE,22, 23, 24, 25, 33, 34, 36 which accumulation in the brain justified its development as a radioprotector of normal brain. MnBuOE has been shown to act as a SOD mimic in several models, including mimicking the MnSOD isoform.22, 23, 25, 35, 37 In addition, it has been shown to upregulate MnSOD via activation of Nrf2 (Nuclear factor [erythroid‐derived 2]‐like 2).24 The notable biological efficiency and the safe toxicity profile (eg, lack of genotoxicity in a rat Comet assay) of MnBuOE in preclinical studies have justified its adoption in clinical trials.37 Such ability of Mn porphyrin‐based SOD mimics may also be important in the field of heart valve biology, as the clinical management of cancer survivors after exposure of mediastinal, thoracic, and breast radiotherapy has been influenced for many decades by major cardiovascular complications. In particular, thoracic radiotherapy is associated with a significant increase in valve pathology, with valve abnormalities appearing in 6% to 40% of Hodgkin lymphoma survivors.36, 38, 39, 40 An initial important use of these compounds in cardiovascular medicine may lie in those patients who undergo radiation therapy for different types of malignancies, with the goal of limiting future radiotherapy‐mediated DNA valvular injuries.

Several limitations are associated with this study. Despite our effort to create homogeneous patient profiles, our approach is based on human retrieved specimens from surgical cases. We have established a protocol to control patient adjudication (Data S1). Despite our efforts, there is intrinsic patient variability and heterogeneous comorbidities that could mask or highlight functional association between our molecular target and the patient profile. Another limitation is that in our bioreactor system, AV leaflets are stretched under uniaxial conditions rather than the more natural‐mimicking biaxial conditions, but this was unavoidable under the current bioreactor setup.

This study helps to further our understanding of mechanisms of AV remodeling in AVSc and may have significant clinical implications. Patients may have AVSc for many years without any hemodynamic compromise of their valve, and this may represent a window of time when cellular changes could be reversed. We demonstrated the way in which one group of substances, a porphyrin‐based SOD mimic recently approved for use in a clinical trial, could potentially prevent some mechanisms of AV remodeling. Additional work is needed, however, before this will be possible: First, we must find ways to measure subtle improvement and degeneration of sclerotic aortic leaflets. Echocardiography is not sensitive enough currently to accurately quantify minute changes in AV leaflets. Second, we must find a way to determine which subpopulation of patients with AVSc will go on to develop AS. The challenge we face now is to identify which patients are at risk for rapid evolution from AVSc to AS, as this will identify the target population for any future interventions. If we cannot determine which patients will develop AS and which patients will not progress beyond AVSc, we will be unable to identify the subpopulation that could most benefit from interventions at an early stage.

It is concluded that AVSc is associated with pronounced architectural remodeling despite marginal effects on the mechanical properties in both human and mice. MnBuOE prevents AV thickening in a murine model of AVSc. Because this compound has been approved recently for clinical use for other indications, this work could shift the focus for the treatment of CAVD, moving from AS to an earlier presentation (AVSc) that could be more responsive to medical treatment.

Sources of Funding

This work was supported by the following research grants and funds: National Institutes of Health (RO1‐HL131872 to Ferrari and Levy, RO1‐HL122805 to Ferrari), American Heart Association (grant 24810002 to Ferrari), the Kibel Fund for Aortic Valve Research (to Ferrari and Levy), the Valley Hospital Foundation “Marjorie C Bunnel” Charitable Fund (to Ferrari and Grau), and both Erin's Fund and the William J. Rashkind Endowment of the Children's Hospital of Philadelphia (to Levy).

Disclosures

Batinic‐Haberle is a consultant with BioMimetix JVLLC and holds equities in BioMimetix JVLLC. IBH and Duke University have patent rights and have licensed technologies to BioMimetix JVLLC. The remaining authors have no disclosures to report.

Supporting information

Data S1. Supplemental material and methods.

Figure S1. Seven‐week‐old wild‐type male mice (CS7BL/6J) were fed with hypercholesterolemic diet and infused with saline or angiotensin II (1000 ng/kg per minute) for 28 days.

(J Am Heart Assoc. 2018;7:e007861 DOI: 10.1161/JAHA.117.007861.)