Anatomically specific reactive oxygen species production participates in Marfan syndrome aneurysm formation

Abstract Marfan syndrome (MFS) is a connective tissue disorder that results in aortic root aneurysm formation. Reactive oxygen species (ROS) seem to play a role in aortic wall remodelling in MFS, although the mechanism remains unknown. MFS Fbn1C1039G/+ mouse root/ascending (AS) and descending (DES) aortic samples were examined using DHE staining, lucigenin‐enhanced chemiluminescence (LGCL), Verhoeff's elastin‐Van Gieson staining (elastin breakdown) and in situ zymography for protease activity. Fbn1C1039G/+ AS‐ or DES‐derived smooth muscle cells (SMC) were treated with anti‐TGF‐β antibody, angiotensin II (AngII), anti‐TGF‐β antibody + AngII, or isotype control. ROS were detected during early aneurysm formation in the Fbn1C1039G/+ AS aorta, but absent in normal‐sized DES aorta. Fbn1C1039G/+ mice treated with the unspecific NADPH oxidase inhibitor, apocynin reduced AS aneurysm formation, with attenuated elastin fragmentation. In situ zymography revealed apocynin treatment decreased protease activity. In vitro SMC studies showed Fbn1C1039G/+‐derived AS SMC had increased NADPH activity compared to DES‐derived SMC. AS SMC NADPH activity increased with AngII treatment and appeared TGF‐β dependent. In conclusion, ROS play a role in MFS aneurysm development and correspond anatomically with aneurysmal aortic segments. ROS inhibition via apocynin treatment attenuates MFS aneurysm progression. AngII enhances ROS production in MFS AS SMCs and is likely TGF‐β dependent.


| Echocardiography
Transthoracic echocardiography (TTE) was performed at age 3 weeks (baseline) and then weekly until age 12 weeks (n = 7-8). Mice were sedated with 2% inhaled isoflurane (Baxter Healthcare Corporation) delivered via nose cone. The aorta was imaged in the parasternal long-axis view using a Vevo 770 Imaging Station (VisualSonics Inc) equipped with a Vevo 2100 system and the ultra-high frequency liner array transducer RMV 704. For statistical analysis, measurement of the aortic diameter (from edge to edge, at the largest portion of the AS aorta) was performed using three technical replicates for each animal by two blinded investigators.

| Aortic tissue preparation
Mice were killed by inhalation of an overdose of isoflurane delivered via nose cone. The aorta was dissected and divided into the following segments: (a) aortic root (distal to aortic valve), ascending and arch

| Fbn1 C1039G/+ -derived SMC
Aortic SMCs were isolated from 4-week-old Fbn1 C1039G/+ and littermate WT control mice, cultured in SMC media (SmGM-2 Bullet Kit CC3182; Lonza) and kept at 37°C with 5% CO 2 . Prior to treatment, SMC purity was confirmed by flow cytometry analysis of SMC-specific markers. All experiments were conducted on SMC at passages 3-6. SMC were depleted of growth factors for 24 hours and kept in medium supplemented with 0.2% foetal bovine serum for 24 hours, then cultured for another 24 hours in testing conditions.

| In vitro detection of ROS in SMCs treated with TGF-β neutralizing antibody treatment, in vitro
For ROS detection, a lucigenin chemiluminescence test was used on SMC for 24 hours in medium supplemented with either (a) anti-TGF-β (c) anti-TGF-β antibody + AngII; or (d) isotype control IgG antibody.
Isolated tissue was embedded in Tissue-Tek OCT Compound Histomount, and serial cryosections of 5μm thickness were displayed in groups of three on the same histological slide. The slides were stained with Accustain EVG kit (Sigma-Aldrich, St. Louis, MO, USA) and imaged at 10x and 40x magnification using a Leica DM4000B microscope. 12 The elastic lamina in three AS and DES aortic sections per slide were assessed by a blinded pathologist for the following: (a) average number of breaks per elastic lamina by counting them circumferentially in all lamina; and (b) semi-quantitative assessment of elastic lamina thinning on a scale of: 0 = none, 1 = mild; 2 = moderate; 3 = severe (ie to the point of multi-perforate in areas). 12 Experiments included n = 9 mice for control and n = 10 for apocynin-treated groups, and have been done in triplicate using three consecutive sections of the AS and the DES aorta from each animal (a total of 3 × 3 = 9 tissue sections/animal). Animals allocated for histological evaluation were different from mice assigned for echo measurements. For statistical analysis, each data set consists of nine biological replicates for control and 10 biological replicates for apocynin and nine technical replicates (3 × 3 tissues per mouse) analysed by blinded investigators.

| Superoxide identification by dihydroethidium (DHE) staining
Freshly dissected mouse aortic segments were embedded in Tissue- Tek OCT Compound Histomount and frozen on dry ice. The aortic segments were sectioned (7 μm). A single sample was incubated with superoxide dismutase-polyethylene glycol (SOD-PEG) (Sigma-Aldrich) and stained with DHE (50 μL per sample) (AnaSpec Inc) 29 to confirm no autofluorescence. One sample per slide was left unstained as negative control for the staining procedure. Samples were placed in a dark humidity chamber at 37°C for 30 minutes. After washing with PBS 1×, a solution of 70% glycerol was used to mount a glass coverslip onto the tissues. Samples were kept at 4°C until imaging. Specimens were imaged at 10× and 40× magnification using a Leica DM4000B fluorescent microscope at excitation and emission wavelengths of 520 and 610 nm, respectively. ImageJ software (NIH open source) was used to outline cells manually in the region of interest (ROI) and quantify integrated pixel intensity (mean intensity/ area). Reciprocal ROI from unstained samples was used to subtract the background noise. Triplicate experiments included n = 6 mice per group, using three consecutive sections of the AS and the DES aorta from each animal. Statistical analysis has been done on data sets including five biological replicates (n = 6 mice) and nine technical replicates (3 × 3 tissue sections).

| Evaluation of ROS production in mouse aortic tissue using lucigenin chemiluminescence assay
Mice were killed, and the aorta was dissected and prepared, as noted above. All specimens were kept in cold PBS solution (PBS pH 7.4, Gibco, Life Technologies Corporation) until use. Superoxide production in whole mouse aortic tissue was measured by lucigeninenhanced chemiluminescence using a single-tube luminometer (Berthold FB12, Titertek-Berthold, Berthold Detection Systems GmbH) modified to maintain the sample temperature at 37°C. Basal chemiluminescence from whole AS and DES mouse aortic tissue (n = 6 per group) was measured in PBS buffer (2 mL) containing 4 μL lucigenin (5 µmol/L) after reaching equilibrium (7 minutes). After application of 20 µL NADPH (100 µmol/L), the chemiluminescence signal in each AS and DES aorta reached a plateau within 5 minutes.
Therefore, the measurements in this preparation do not reflect the maximal NADPH-stimulated superoxide production. 30

| Measurement of ROS production in murine vascular SMC using lucigenin chemiluminescence assay
Superoxide production in mouse vascular SMC isolated from the AS and DES aortas, respectively, was measured by lucigenin-enhanced chemiluminescence in a 96-well plate system and multi-mode microplate BioTek Synergy H1 Hybrid Reader (BioTek Instruments Inc) at 37°C. Investigators have reported that the lucigenin chemiluminescence assay signal is not generated by NADPH oxidase activity alone, but also via the cytochrome system (add reference). SMC were cultivated in a 6-well plate until they reached 90%-95% confluence, then starved 24 hours in medium with 0.2% FBS and finally submitted to different treatments for another 24 hours. 30 The SMC monolayer was disrupted with Gibco TrypLE TM (Thermo Fisher Scientific), and (from 0 to 540 seconds). 30 Light detection was possible after the first 30 seconds, and the signal was stabilized after 60-300 seconds.
Time-point measurements between 90 and 300 seconds were used for statistical analysis.

| Treatment with the inhibitor apocynin reduces AS aneurysm formation in the Fbn1 C1039G/+ mouse model
As biological proof of concept that ROS contribute to aneurysm formation in MFS mice, Fbn1 C1039G/+ and WT control mice were treated with the unspecific NADPH oxidase inhibitor apocynin from ages 3 to 12 weeks. Apocynin significantly reduced aneurysm formation in 1.57 ± 0.05 versus 1.84 ± 0.24, P = .015 at 12 weeks, n = 7 for apocynin treated and n = 8 for untreated Fbn1 C1039G/+ mice) (Figure 2A,B).

| D ISCUSS I ON
The specific role TGF-β plays during aneurysm formation in MFS murine models remains complicated, although most animal studies support a pathologic function through aortic wall remodelling in two different Marfan mouse models (Fbn1 mgR/mgR and Fbn1 C1039G/+ ). [5][6][7][8][9][10][11][12][13][14]32,33 Identifying key components within the molecular pathway(s) that lead to aneurysm formation may translate into innovative medical therapies directed at preventing or slowing aneurysm growth 33,34 . Herein, our laboratory reports that ROS contribute to aneurysm formation in MFS. Under normal physiological conditions, ROS play an important role as signalling molecules in endothelial function and vascular tone. [35][36][37] In contrast, excessive ROS production (oxidative stress) can induce a pathologic environment within the aortic wall, including endothelial dysfunction, inflammation and ECM remodelling. 21,38 The novel findings in this study are as follows: (a) ROS are elevated in MFS aortic aneurysms in the In this work, we identified enhanced ROS in murine aortic aneurysm samples. As biological proof of concept that ROS contribute to MFS aneurysm formation, Fbn1 C1039G/+ mice treated with the unspecific NADPH oxidase inhibitor apocynin exhibited reduced ROS production, which directly correlated with decreased protease activity/ECM destruction and diminished aneurysm formation.
Needless to say, it remains possible that apocynin has other beneficial off-target affects, including acting as a ROS-scavenger, blocking other elastases or increasing SMC proliferation. 39 Several investigators have studied the involvement of ROS in thoracic aortic root/ F I G U R E 5 A, In vitro quantification of NADPH oxidase activity by lucigenin chemiluminescence (LGCL) in Fbn1 C1039G/+ (Fbn1) aortic AS and Des-derived smooth muscle cells (SMC) (n = 6, *P = .02). B, In vitro quantification of NADPH oxidase activity by LGCL in Fbn1 aortic AS and WT AS SMC (n = 6, *P = .03). C, In vitro quantification of NADPH oxidase activity by LGCL in Fbn1 AS and Des SMCs treated with AngII or vehicle (n = 6, *P = .03). D, In vitro quantification of NADPH oxidase activity by LGCL in Fbn1 AS SMCs treated with AngII and with either TGF-β neutralizing antibody (TGFb NA) or vehicle (n = 6, *P = .009) ascending aortic aneurysm formation, including individuals with MFS, 24,27,[40][41][42] as well as patients with bicuspid aortic valves. [43][44][45] The hypothesized important role of ROS in MFS aneurysm formation presented here agrees with the findings of Jimenez-Altayo et al who also reported increased nitrotyrosine residues in both aortic aneurysms and cultured SMC from MFS patients. 28 This group identified several possible redox stress mechanistic targets, including alpha-smooth muscle actin (α-SMA), encoded by the ACTA2 gene and the cytoskeleton protein, annexin A2. Intriguingly, although we know that patients with ACTA2 genetic mutations can develop familial thoracic aortic aneurysms, this finding suggests that redox modifications to the normal α-SMA protein may potentially promote aneurysm formation. 28 42,46 Here, we describe the important relationship between ROS and protease activity, both in vitro and in vivo. Our laboratory previously presented the contribution of SMC apoptosis to MFS aneurysm formation, although not directly tested here. 11,12 Of note, because aneurysms still develop despite ROS reduction, other factors must contribute to the pathophysiology triggering MFS aneurysm development.
One of the key findings in this study is that ROS are enhanced in the Fbn1 C1039G/+ aneurysmal AS aorta, yet remain at baseline levels in the normal-sized descending aorta. Importantly, this result rules out the theory that aortic root SMCs merely react distinctively to systemically enhanced ROS compared to the rest of the aorta. Lineage studies have shown that vascular SMC in the different aortic anatomic segments have distinct embryologic origins. 47 Recent reviews have suggested that the diversity of SMC origin may explain site-specific location of various vascular diseases, including aneurysm formation. 48,49 Of note, an important limitation of our animal model is that aneurysms involve the aortic root and ascending aorta in Fbn1 C1039G/+ mice, in contrast to the aortic root only in humans. Therefore, when studying the role aortic SMC embryologic lineage plays during aneurysm development in the mouse model, we compare aneurysm to non-dilated aortic segments. Our laboratory's overarching hypothesis is that Fbn1 C1039G/+ AS aorta SMC (second heart field and neural crest) must respond differently to systemically enhanced TGF-β versus the remainder of the aorta (paraxial mesoderm). More specifically, either superoxide-generating enzymes are enhanced or free radical scavengers (superoxide dismutase) are reduced in the specific embryologic aortic segments where aneurysms develop. Although the MFS mouse model allows us to study early aneurysm development, an obvious weakness is this study remains that we did not confirm presence of ROS prior to aneurysm formation. Lastly, although not employed in this study, an alternative strategy to test our overarching "aortic embryologic origin" hypothesis includes differentiating induced-pluripotent stem cells (iPSc) into SMC from each embryologic derivation to study how SMC origin influences ROS production following TGF-β stimulation.
There are several hypothesized triggers for increased ROS production in the MFS aorta, including enhanced (a) TGF-β signalling, (b) aortic root stress and strain (increased biaxial loading unique to the ascending aorta) (add Bellini reference) and (c) AngII signalling. AngII, the major peptide of the renin-angiotensin system, is a potent vasoconstrictor, but also stimulates ROS production via NADPH oxidase (NOX). 50 Through a positive feedback loop, ROS can increase ATR1 expression, further increasing oxidative stress. 51 While seven NOX homologues have been characterized (NOX 1-5 and Duox 1-2), the Egea laboratory elegantly illustrated the importance of NOX4 utilizing a compound mutant Fbn1 C1039G/+ -Nox4 −/− mouse model, demonstrating reduced aneurysm size and decreased elastin breakdown compared to Fbn1 C1039G/+ mice. 28,52 Herein, we report that AngII regionally increases MFS SMC ROS production in AS-derived SMC, but not in DES aorta-derived SMC, in vitro. Previous studies similarly report that AngII can induce ROS in several cell types, including cardiomyocytes, cardiac fibroblasts and vascular SMC. [53][54][55][56] Because TGF-β blockade reduced AngII-dependent MFS AS SMC ROS production, we postulate that TGF-β can increase oxidative stress directly, as well as indirectly via AngII. Kim et al analogously described that AngII increases SMC ROS production via ERK1/2 stimulation concomitantly while ROS induce ERK1/2 phosphorylation, thereby intensifying each other. 57 Finally, we cannot rule out that AngII participates in aneurysm formation through an alternative mechanism, for example direct stimulation of MMP activation. 58,59 It is intriguing that the protective clinical benefit of losartan noted in several MFS clinical trials may be secondary to blockade of AngII-stimulated ROS production. [60][61][62][63][64][65][66][67][68][69][70]  Michael P. Fischbein designed the study, analysed the data and wrote the paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available upon request.