Extracellular Tuning of Mitochondrial Respiration Leads to Aortic Aneurysm

Supplemental Digital Content is available in the text.

D ifferent inherited disorders affect the structure of the arterial wall and lead to thoracic aortic aneurysms (TAA). 1,2 Most of these diseases are associated with mutations in genes related to the extracellular matrix (ECM), or the vascular smooth muscle cell (VSMC)-contractile apparatus. Marfan syndrome (MFS) is one of the most commonly inherited disorders affecting connective tissue. It is caused by mutations in the gene that encodes for the extracellular protein FBN1 (fibrillin-1) and has a reported incidence of 1 in 5000 individuals. 3,4 MFS patients present extended bones, lens luxation, and decreased life expectancy primarily attributable to TAAs. 5 Mutations in genes that encode proteins involved in TGFβ-signaling cause Loeys-Dietz syndrome, a disease that phenocopies part of the skeletal features of MFS. 6 Vascular-type Ehlers-Danlos syndrome is mainly caused by mutations in COL3A1 gene (collagen 3a1) that produce hyperelasticity in joints and skin, small stature, and aortic dissections. 7 Moreover, the rare condition Cutis laxa syndrome is caused by mutations in elastin fibers genes (ELN [elastin], FBLN4 [fibulin 4], FBLN5 [fibulin 5]) that produce a loss of elastic properties in tissues such as the skin and arteries. 8,9 Finally, although most of the genes involved in familial forms of nonsyndromic thoracic aortic aneurysm and dissections remain unknown, some of those identified (accounting for only about 20% of all patients with familial forms of nonsyndromic thoracic aortic aneurysm and dissections ) are involved in the maintenance of the smooth muscle contractile apparatus, such as PRKG1 (cGMP-dependent protein kinase 1), MYLK (myosin light chain kinase), MYH11 (myosin heavy chain11), and ACTA2 (actin α2, smooth muscle). 10 The major complication of all these connective inherited disorders is the risk to develop TAAs. TAA is a complex vascular pathology characterized by permanent dilation of the thoracic aorta. TAA progression can lead to catastrophic consequences like rupture or dissection of the arterial wall, causing patient death because of extensive hemorrhage. VSMCs are located in the medial layer of arteries and directly drive the contraction of the vascular wall regulating the size of the vessel lumen. VSMCs can shift reversibly from a quiescent or contractile phenotype to a secretory phenotype. [11][12][13] Changes toward this secretory phenotype generate pathologic features such as an elevated proliferation rate and increased ECM accumulation favoring medial degeneration in the aneurysm development. [12][13][14][15] ECM remodeling in the aortic wall results in increased aortic stiffness. [16][17][18] Studies have started to reveal the connection between cellular adhesion and cytoskeletal reorganization and the metabolism of the cell, [19][20][21] supporting that the composition and the stiffness of the ECM regulate cellular metabolism. 22 Here, we investigate the implication of VSMC metabolism in the progression and development of aortic aneurysm in MFS. We provide evidence that mitochondrial metabolism is a key regulator of VSMC phenotype during aortic remodeling and is fine-tuned by ECM composition. To demonstrate the importance of mitochondrial metabolism in the development of aneurysms, we generated a conditional mouse model with mitochondrial dysfunction specifically in VSMCs by depleting Tfam (mitochondrial transcription factor A). As Tfam controls the transcription, replication, and stability of mitochondrial DNA (mtDNA), 23 depleting Tfam has been proved to be an effective approach to induce severe mitochondrial dysfunction in different cells and tissues. [24][25][26][27] Tfam-deficient VSMC mice developed aortic dilation, medial degeneration, aortic aneurysm, and fatal dissections, further supporting that mitochondrial function is a key regulator of the VSMC phenotype during aortic remodeling. Finally, we assessed the therapeutic potential of boosting mitochondrial metabolism in MFS. Our results indicate that mitochondrial function could be an effective target for pharmacologic approaches to managing aortic dilation and preventing aortic dissection associated with genetic disorders.

METHODS
The data, analytic methods, and study materials that support the findings of this study are available from the corresponding author on reasonable request.

Mouse Strains and Animal Procedures
The Marfan mouse model, which harbors a Fbn1 C1039G/+ allele (JAX stock No. 012885), was previously described. 28 For specific ablation of Tfam in smooth muscle, we crossed Tfam flox/ flox mice 23 with mice carrying the Myh11-Cre ERT2 allele (JAX

Clinical Perspective
What Is New?
• Tfam and mitochondrial-DNA levels decline correlate with aortic dilation in Fbn1 c1039g/+ mice. • Aortic samples and cells from patients with Marfan syndrome present mitochondrial defects. • Mitochondrial respiration is fine-tuned by the extracellular matrix. • Mitochondrial dysfunction in vascular smooth muscle cells promotes aortic dilation, aneurysms, and premature death. • Boosting mitochondrial function by a NAD + precursor rapidly reverses transcriptional signature and aortic dilation in Fbn1 c1039g/+ mice.
What Are the Clinical Implications?
• Boosting mitochondrial function with NAD + precursors is a new therapeutic opportunity to manage aortic aneurysms associated with connective genetic disorders and to prevent aortic dissection.

Blood Pressure Measurements and In Vivo Imaging
Arterial blood pressure (BP) was measured by the mouse tailcuff method using the automated BP-2000 Blood Pressure Analysis System (Visitech Systems, Apex, NC). Mice were trained for BP measurements every day for 5 consecutive days. After the training period, BP was measured 1 day before treatment to determine the baseline BP values in each mouse cohort. BP measurements were recorded in mice located in a tail-cuff restrainer over a warmed surface (37°C). Fifteen consecutive systolic and diastolic BP measurements were made, and the last 10 readings per mouse were recorded and averaged. For in vivo ultrasound images, the aortic diameter was monitored in isoflurane-anesthetized mice (2% isoflurane) by high-frequency ultrasound with a VEVO 2100 echography device (VisualSonics, Toronto, Canada) at 30-μ resolution. Maximal internal aortic diameters were measured at diastole using VEVO 2100 software, version 1.5.0.

Cell Procedures
Isolation and culture of primary mouse VSMCs were described in Esteban et al. 30 Tissue was digested with a solution of collagenase and elastase (Worthington Biochem) until a singlecell suspension was obtained. All experiments with primary VSMCs were performed during passages 2 to 7. Lentiviral transduction was performed in VSMCs during 5 hours at a multiplicity of infection of 3. The medium was then replaced with fresh DMEM supplemented with 10% FBS and cells were cultured for 5 more days, treated with NR for 5 more days, and then serum-starved for 16 hours. The HEK-293T (CRL-1573) and Jurkat (Clone E6-1, TIB-152) cell lines, required for high-titer lentivirus production and lentivirus titration, respectively, were purchased from American Type Culture Collection . Four apparently healthy male controls (GM00024,  GM01717, GM03652, GM23963) and 4 males with agedmismatched fibroblasts from MFS patients with aortic dissection, 2 with FBN1 point mutations (GM21499, GM21946), and 2 with haploinsufficient FBN1 mutations (GM21983, GM21978), were purchased to Coriell Cell Repositories. The experiments were performed during passages 5 to 10. All cells tested negative for Mycoplasma.

Lentivirus Production and Infection
The Cre and green fluorescent protein coding sequence was obtained by polymerase chain reaction (PCR) amplification and cloned into the pHRSIN lentiviral vector. 31

Extracellular Flux Analysis and Extracellular L-Lactate Determination
Oxygen consumption rates (OCR) were measured in XF-96 Extracellular Flux Analyzers (Seahorse Bioscience) in 25 000 mouse aortic VSMCs or 50 000 human fibroblasts. Cells were seeded in nonbuffered DMEM medium containing either 25 mM glucose or 1 mM CaCl 2 . Three measurements were obtained under basal conditions and on addition of oligomycin (1 mM), fluoro carbonyl cyanide phenylhydrazone (1.5 mM), and rotenone (100 nmol/L) + antimycin A (1 mM). OCR measurements were normalized to protein cell extracts. Extracellular lactate determination was performed with single-use reagent strips for Accutrend Plus Lactate Pro (Roche), based on an enzymatic spectrophotometry system by lactate oxidase layer. We analyzed 20 µL of conditioned medium after 24 hours of culture. L-lactate was normalized to total protein cell extracts.

β-Galactosidase Activity Assay
For the β-galactosidase quantitative assay, tissues were lysed with T-PER Tissue Protein Extraction Reagent (78510; Thermo Scientific). Lysates were centrifuged at 10 000g for 5 minutes and the supernatant was collected. Fifty microliters of protein lysates were mixed with 50 µL of Pierce β-galactosidase Assay Reagent (75705; Thermo Scientific) for the assay. The reaction was incubated for 1 hour and the absorbance was measured at 415 nm. Values were normalized to total protein extracts.

ECM Assays
Experiments of ECM were based on Castelló-Cros et al. 33

ORIGINAL RESEARCH ARTICLE
and 0.01% collagen I (Gibco) for 1 hour at 37ºC. The gelatin was cross-linked by adding 1% glutaraldehyde in PBS and incubated for 30 minutes at 37ºC. Then, the plates were washed 3 times for 5 minutes with PBS and 1 additional wash with DMEM. Then, VSMCs were seeded at a confluence of 50% and cultured for 5 days to allow production of the ECM. For the decellularization step, ice-cold extraction buffer (0.2% of Triton X-100, 20 mM NH 4 OH in DMEM) was added to the cells and observed under a tissue culture microscope until the VSMCs were completely lysed (5-10 minutes). Total cell lysis was tested by analyzing 4′,6-diamidino-2-phenylindole staining in a coverslip. Then, control VSMCs were seeded in the ECMs at 100% of confluence and the experiments were analyzed after 24 hours.

Real-Time and Quantitative PCR
Aortas were extracted after perfusion with 5 mL saline solution, and the adventitia layer was discarded. Liquid nitrogen frozen tissue was homogenized using a cold mortar and an automatic bead homogenizer (MagNA Lyzer, Roche). Total RNA was isolated with Trizol (Life Technologies). Total RNA (1 μg) was first digested with DNAse and reverse-transcribed with Maxima First Strand cDNA Synthesis Kit (ThermoFisher). For analysis of mtDNA levels, total DNA from cells and tissues was extracted with the SurePrep kit (Fisher Scientific) or Trizol, respectively, according to the manufacture's guidelines. DNA was amplified using primers specific for cytochrome mt-Co1 (mitochondrially-encoded cytochrome c oxidase 1) and mitochondrial 16S rRNA, then normalized to B2M (β-2 macroglobulin) and H2K nuclear-encoded gene controls. Real-time quantitative PCR (qPCR) was performed with the primers indicated in Table I in the Data Supplement. qPCR reactions were performed in triplicate with SYBR master mix (Promega), according to the manufacturer's guidelines. To examine probe specificity, we conducted a postamplification melting curve analysis. For each reaction, only 1 melting temperature peak was produced. The amount of target mRNA in samples was estimated by the 2−CT relative quantification method, using B2M, YWHAZ (tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein ζ), and PP1A (protein phosphatase 1 catalytic α) for normalization. Fold ratios were calculated relative to mRNA expression levels from controls.

Library Preparation and Illumina Sequencing
Aortas were extracted after perfusion of cold saline solution and the adventitia layer was discarded. Frozen tissue was homogenized, and total RNA was isolated with Trizol (Roche). RNA from Libraries were prepared according to the instructions of the NEBNext Ultra Directional RNA Library prep kit for Illumina (New England Biolabs), following the Poly(A) mRNA Magnetic Isolation Module protocol. The input yield of total RNA to start the protocol was >300 ng quantified by an Agilent 2100 Bioanalyzer using an RNA 6000 nano LabChip kit. The obtained libraries were validated and quantified by an Agilent 2100 Bioanalyzer using a DNA7500 LabChip kit and an equimolecular pool of libraries was titrated by qPCR using the KAPA SYBR FAST qPCR Kit for LightCycler 480 (Kapa BioSystems) and a reference standard for quantification. After processing on the Illumina HiSeq 2500 instrument, FastQ files were generated containing nucleotide data and quality scores for each position. RNA-sequencing reads were mapped to the Mus musculus reference genome, GRCm38.p6, using Hisat2 v2.1.0 software. Reads were then preprocessed with SAMtools v1.7 to transform Sequence Alignment/Map files into Binary Alignment/Map files and sorted. These files were used as an input for the HTSeq v0.6.1 package that produces a file containing the mapped reads per gene (as defined by the Mus musculus, GRCm38.96 version, gff file) for each sample. Once obtained HTSeq read counts, the Bioconductor RNA-sequencing workflow was followed to detect the expression differences of genes using the DESeq2 statistical package. Ingenuity pathway analysis was used to identify cellular biologic functions, gene clusters, and upstream regulators.

Vascular Reactivity
Aorta and mesenteric resistance arteries were dissected and 2-mm length segments were mounted in a wire myograph for isometric tension recording. After a 30-minute equilibration period in oxygenated Krebs−Henseleit solution at 37°C and pH 7.4, segments were stretched to their optimal lumen diameter for active tension development. Contractility of the segments was tested by an initial exposure to a high K+ solution (120 mmol/L). After washing, a concentration−response curve to increasing concentration of phenylephrine (1 nmol/L-50 µmol/L) and to U46619 (0.1 nmol/L-1 µmol/L) were carried out.

Gelatin Zymography
Supernatants form cell culture were prepared as described. 32 Extracts (15 μg) were fractioned under nonreducing conditions on 10% SDS-polyacrylamide gels containing 1% gelatin (Sigma). Gels were washed 3 times in 2.5% Triton X-100 for 2 hours at room temperature, incubated overnight at 37°C in 50 mM Tris-HCl (pH 7.5), 10 mM CaCl 2 , and 200 mM NaCl, and stained with Coomassie blue. The areas of gelatinolytic matrix metalloproteinase (MMP) activity were visualized as transparent bands. Images were analyzed with Quantity One software (Bio-Rad).

Human Samples
The study was approved by the Ethics and Clinical Research Committee of Instituto de Salud Carlos III and Cantabria University (B2017/BMD-3676, AORTASANA-CM, ref. 27/2013; respectively). Aortas for use as controls were obtained anonymously from multiorgan transplant donors after written informed consent was obtained from their families. During preparation of the heart for transplantation, excess ascending aortic tissue was harvested for the study. Samples from patients were obtained during elective or emergency aortic root surgery for aortic aneurysm dissection. Patient clinical data were retrieved while maintaining anonymity. Tissues were immediately fixed, kept at room temperature for 48 hours, and embedded in paraffin. DNA and RNA extraction from paraffin sections were performed with All Prep DNA/ RNA FFPE [formalin-fixed, paraffin-embedded] kit (Qiagen).

Statistical Analysis
Normality of data were checked using a Shapiro-Wilk test. F-test was used to assess the equality of variance assumption.
Differences between 2 groups were assessed using the unpaired Student t test, t test with Welch correction for unequal variances, or Mann-Whitney U test, where appropriate. Differences in experiments with ≥3 groups were analyzed by 1-way, 2-way, or repeated-measurement 2-way ANOVA, or a mixed-effects linear model and Newman post hoc test, as appropriate. For the statistical analysis of RNA-sequencing data, the P values were corrected using the false discovery rate method (false discovery rate <0.05). For survival curves, differences were analyzed with the log-rank (Mantel-Cox) test. Multiple linear regression analysis adjusted for age was performed with R software. For all other analysis GraphPad Prism software 9 was used. Statistical significance was indicated as *P<0.05, **P<0.01, ***P<0.001, and ****P<0.0001. Sample size was chosen empirically based on our previous experiences in the calculation of experimental variability; no statistical method was used to predetermine sample size. Before data analysis, outliers were identified and excluded by using the Robust Outlier removal method (Q value =5%) to identify outliers provided with GraphPad Prism 9.
The numbers of animals used are described in the corresponding figure legends. All experiments were done with at least 3 biologic replicates. Experimental groups were balanced in terms of animal age, sex, and weight. Animals were genotyped before experiments, and were all caged together and treated in the same way. Sex and age are indicated in the figure legends. Appropriate tests were chosen according to the data distribution. Variance was comparable between groups in experiments described throughout the article. For the rest of experiments, no randomization was used to allocate animals to experimental groups, and investigators were not blinded to group allocation during experiments or to outcome assessments.

Aortas From Marfan Syndrome Mice Present Features of Mitochondrial Decline
Fbn1 C1039G/+ mice reproduce the aortic dilation, aneurysms, and histologic features of aortic medial degeneration found in MFS patients. 28 To identify novel molecular mechanisms underlying TAA formation, we performed transcriptional analysis in aortas from 24-week-old Fbn1 C1039G/+ mice, showing an intermediate stage of the aortic disease ( Figure 1A). 34,35 Ingenuity pathway analysis revealed that 6 of the 10 most altered canonical pathways were related to metabolism, including oxidative phosphorylation, mitochondrial dysfunction, fatty acid β-oxidation, and the tricarboxylic acid cycle ( Figure 1B). Ingenuity pathway analysis prediction of activity of upstream regulators identified canonical MFS regulators such as inducible Nos2 (nitric oxide synthase) and Tgfb1 (transforming growth factor β1). 28,32 This analysis pointed to increased activity of the glycolytic regulator Hif1α (hypoxia-inducible factor α) and reduced activity of several mitochondrial biogenesis and function regulators, including Tfam ( Figure 1C). Fbn1 C1039G/+ aortas showed reduced ORIGINAL RESEARCH ARTICLE   Figure 1E). Tfam is a nuclear-encoded mitochondrial transcription factor that controls the transcription, replication, and stability of mtDNA. 23 Analysis of relative mtDNA content in Fbn1 C1039G/+ aortas revealed below-normal mtDNA content ( Figure 1E). To model MFS in vitro, we silenced Fbn1 with lentiviral vectors in primary murine VSMCs ( Figure IA in the Data Supplement). Confirming the transcriptomic data from MFS mice, shFbn1 VSMCs displayed increased levels of Hif1α and its targets Pdk1 (pyruvate dehydrogenase kinase-1) and Slc2a1 (glucose-transporter 1; Figure 1F), and significantly reduced expression of Pgc1α (Ppargc1a), as well as Tfam and its targets, MT-Nd1 (mitochondrially-encoded NADH-ubiquinone oxidoreductase chain 1) and MT-Co1 ( Figure 1G), correlating with a reduction in mtDNA content ( Figure 1H). Flux analysis of the OCR, an index of mitochondrial oxidative phosphorylation, revealed reduced mitochondrial respiration in Fbn1-silenced cells ( Figure 1I). This was accompanied by increased production of extracellular lactate, an indicator of glycolysis ( Figure 1J). As mitochondrial decline induces cellular senescence through p53, 36 and our ingenuity pathway analysis predicts p53 (Tp53) as an active upstream regulator ( Figure 1C), we analyzed senescence features in shFbn1 VSMCs. Both protein and mRNA p53 levels were upregulated in shFbn1 VSMCs ( Figure IB in the Data Supplement) together with increased SABG (senescence-associated β-galactosidase) activity ( Figure IC in the Data Supplement) and increased expression of proinflammatory cytokines related to SASPs (senescence-associated secretory phenotype) such as Tnfa, Il1b (interleukin 1β), and Il6 (interleukin 6; Figure ID in the Data Supplement). These data support that aortas from Fbn1 C1039G/+ mice or VSMCs carrying MFS mutations present mitochondrial respiration decline, rewire their metabolism toward glycolysis, and display senescence and inflammation features.

Mitochondrial Decline Appears in the Onset of Aortic Disease in Fbn1 C1039G/+ Mice
To assess the stage during the aortic disease in which mitochondrial decline appears, we performed analysis of aortic diameter (Figure 2A) and histologic features ( Figure 2B and 2C) alongside Tfam and mtDNA levels ( Figure 2D and 2E) at different stages of TAA disease in Fbn1 C1039G/+ mice. Aortas from 8-week-old mice already present significant below-normal Tfam mRNA levels, mtDNA content, and Mt-Nd1 expression ( Figure 2D) together with the appearance of a significant increase in elastin breaks, proteoglycan deposition, and aortic medial thickness ( Figure 2B and 2C). Similarly, some well-known MFS aortopathy-related genes and metabolic genes already showed altered levels in aortas from 8-week-old Fbn1 C1039G/+ mice ( Figure II in the Data Supplement). In addition, the diameter of the aorta negatively correlated with mtDNA levels in MFS mice ( Figure 2F).
Reduction in Tfam levels induces mtDNA leakage into the cytoplasm and activates the innate immune pathway stimulator of interferon genes (cGAS-STING). 24,37 Interestingly, both Irf7 (interferon regulatory factor 7) and Isg15 (interferon-stimulated gene 15), which are cGAS-STING response target genes, were up-regulated in Fbn1 C1039G/+ aortas at early stages of the disease ( Figure II in the Data Supplement), supporting a previous report on the role of this route on aortic degeneration. 38 Together these data suggest that mitochondrial decline appears with the first signs of aortic remodeling in Fbn1 C1039G/+ mice.

ORIGINAL RESEARCH ARTICLE
upregulation of genes involved in mitochondrial uncoupling and glycolytic rewiring (UCP2, HIF1A, and MYC; Figure 3D). Histologic analysis of TFAM, the mitochondrial proteins MT-ND1, SDHA and the glycolysis-related genes HIF1A and MYC supported glycolytic rewiring in MFS aortic samples ( Figure 3E and 3F). Coimmunostaining for smooth muscle actin confirmed that mitochondrial gene down-regulation was specific to VSMCs ( Figure  III in the Data Supplement). Moreover, we used primary skin fibroblasts from 4 individuals with MFS and 4 healthy donors to analyze the mitochondrial respiratory capacity in vitro. MFS fibroblast presented a reduction in the total mtDNA content, TFAM and PPARGC1a mRNA, and protein expression ( Figure 3G and 3H). In line with human MFS aortas, primary MFS skin fibroblasts showed a decrease of the mRNA levels of MT-ND1, SDHA, MT-CO1, and MT-AP6 and increased expression of UCP2, HIF1A, and MYC ( Figure 3I and 3J). Moreover, MFS fibroblasts showed a reduction in the OCR and increased lactate production ( Figure 3K). Altogether, these data support that aortas from MFS present metabolic rewiring toward glycolysis because of mitochondrial function decline.

Extracellular Regulation of Mitochondrial Respiration in MFS
Because elevated arterial stiffness is a characteristic of TAA, 16,17,39 and recent evidence supports that ECM stiffness and composition affect cellular metabolism, 22 we hypothesized that VSMC metabolism could be regulated by the microenvironment. To investigate this hypothesis, we transferred control VSMCs to plates coated with ECM-produced by Fbn1-deficient VSMCs ( Figure 4A). Importantly, ECM from TAA cells induced a reduction in the OCR and increased lactate production in control VSMCs ( Figure 4B). In addition, TAA ECM induced a decrease in the levels of Tfam mRNA and in the mtDNA content and reduced the transcription of the mitochondrial genes Mt-Co1 and Ppargc1a while increasing the transcription of the glycolytic transcription factor Hif1a ( Figure 4C). Moreover, control VSMCs cultured on ECM produced by Fbn1-deficient VSMCs up-regulated the expression of synthetic genes (Tgfb1, Spp1 [osteopontin], andCcn2 [cellular communication network factor 2]; Figure 4D), suggesting that both the metabolism and the phenotype of VSMCs depends on signals coming from the extracellular matrix.

Mitochondrial Dysfunction in VSMCs Causes Aortic Aneurysm
To investigate whether the mitochondrial decline plays a role in VSMC function and in the progression of aortic diseases, we induced mitochondrial dysfunction in VSMCs by targeting Tfam. For in vitro experiments, we transduced aortic VSMCs from Tfam flox/flox mice with lentiviral vectors encoding the Cre recombinase. Tfam deletion was confirmed by analyzing Tfam mRNA and protein levels ( Figure 5A). In accordance with its role in regulating mtDNA levels, Tfam deletion in VSMCs led ORIGINAL RESEARCH ARTICLE to a reduction in mtDNA content that correlated with a decreased expression of the mtDNA-encoded genes Mt-Co1 and Mt-Nd1 and a reduced OCR, alongside with an increase in Slc2a1 expression and lactate production ( Figure 5A-5D). We also observed a sharp decrease in the mRNA expression of genes involved in contractility, such as smooth muscle Myh11, Acta2, Tagln ( transgelin), Cnn1 (calponin), and Smtn (smoothelin; Figure 5E). Moreover, this was accompanied by increased mRNA expression of genes related to the secretory phenotype, such as Spp1, the metalloproteinases Mmp2 and Mmp9, and the inducible Nos2 ( Figure 5F). The increased Mmp9 expression was accompanied by a marked increase of Mmp9 enzymatic activity in cell supernatants ( Figure 5F). As observed in Fbn-1-deficient VSMCs, Tfam-depleted VSMCs showed an increase of p53 expression, senescent-associated β-galactosidase activity, and the proinflammatory cytokines Tnfa and Il6 mRNA levels ( Figure Figure 5H). Lifespan analysis showed a significant decrease in survival rate, with 100% of SM-Tfam −/− mice dying before 33 weeks after tamoxifen ( Figure 5I). Longitudinal analysis of vascular phenotype revealed a decrease in blood pressure alongside with an increase in aortic diameter ( Figure 5J and 5K). Histologic analysis of SM-Tfam −/− mice 28 weeks after tamoxifen revealed aortic dissections, intramural hematomas, and medial degeneration, with elastin lamina degradation and proteoglycan accumulation in the ascending aorta ( Figure 5L). Aortas from SM-Tfam −/− mice also showed defective vascular contractility responses to high K + solution, the α1-adrenergic agonist phenylephrine, and the thromboxane receptor A 2 agonist U46619 ( Figure 5M). Similar to MFS, aortas from SM-Tfam −/− mice showed an increase of the cGAS-STING pathway genes (Irf7, Isg15, Tmem173 [transmembrane protein 173], Mb21d1 [mab-21 domain-containing protein 1]), proinflammatory cytokines, such as Il1b, Tnfa, and Il6, and the senescent marker p21 (Cdkn1a; Figure IVC in the Data Supplement). Consistently, Tfam deficiency in VSMCs greatly affects the vascular contractility and function.
To test how SM-Tfam −/− mice respond to a hypertensive challenge, we infused SM-Tfam +/+ and SM-Tfam −/− mice with Ang II 56 days after tamoxifen ( Figure 6A). Analysis of blood pressure showed a modest increase in SM-Tfam −/− mice ( Figure 6B) compared with SM-Tfam +/+ . Ultrasonography revealed a rapid increase in the diameter of the ascending and abdominal aortas of SM-Tfam −/− mice, supporting the predisposition of these mice to development of aortic aneurysms (Figure 6B-6E). Most importantly, treatment of SM-Tfam −/− mice with Ang II triggered aortic aneurysms and lethal aortic dissections, reducing mean survival ( Figure 6E). Postmortem analysis revealed the presence of intramural hematomas and aortic ruptures with hemothorax or hemoabdomen ( Figure 6F). Histologic analysis of thoracic and abdominal aortic sections showed increased aortic diameter, aortic dissections, intramural hematomas, false lumen formation, and features of medial degeneration, including elastic fiber fragmentation and disorganization, medial thickening, and accumulation of proteoglycans ( Figure 6G). These data support that mitochondrial function in VSMCs is an important regulator of aortic function, and demonstrate that mitochondrial dysfunction in VSMCs induces aortic aneurysm and lethal dissections.

NR Normalizes Mitochondrial Respiration in VSMCs Carrying TAA Mutations
NAD + is a cofactor for several enzymes with a critical role in the maintenance of cellular metabolism and mitochondrial function. 40 NAD + precursor supplementation has been proposed as a strategy to improve mitochondrial function conditions related to mitochondrial decline, [41][42][43] including senescence. 44 NR is a NAD precursor that improves mitochondrial metabolism by increasing Pgc1α and Tfam expression through sirtuin activity. 42,45 Importantly, cellular production of NAD + via Nampt (nicotinamide phosphoribosyltransferase) protects against DNA damage and premature senescence in VSMCs. 46 To investigate the therapeutic potential of NAD + boosting mitochondrial metabolism in familial aortic aneurysms, we treated shFbn1 VSMCs with NR. Exposure of shFbn1 VSMCs to NR for 5 days increased the expression of Pparg1a and Tfam, correlating with increased mtDNA content and the expression of the mtDNA-encoded Mt-Co1 transcript ( Figure 7A and 7B). NR increased OCR and decreased lactate production in shFbn1 VSMCs to levels observed in ShControl VSMCs ( Figure 7C and 7D). NR treatment of shFbn1 VSMCs decreased the expression and activity of the proremodeling matrix metalloproteinases Mmp9 and Mmp2 and the profibrotic genes Spp1 and Col1a1 (collagen, type 1, α1; Figure 7E and 7F).
To confirm the potential of boosting mitochondrial metabolism to revert MFS features in humans, we performed in vitro analysis in primary skin fibroblast cultures from 4 healthy donors and 4 MFS patients. NR treatment restored the mRNA levels of TFAM and mtDNA ( Figure 7G). NR increased the mRNA levels of MT-CO1, MT-ND6, and HIF1A ( Figure 7G and 7H), restored OCR, and decreased lactate production in primary skin fibroblasts from MFS patients ( Figure 7I and 7J). Furthermore, NR reduced the expression of the ECM genes COL1A1 and ACAN (aggrecan) and the profibrotic factors TGFB3 and CCN2 (connective tissue growth factor; Figure 7K). These data indicate that incubation with NR restores TFAM levels and improves mitochondrial respiration in vitro in cells from MFS patients.

NR Treatment Reverts Aortic Aneurysm in a Mouse Model of MFS
We next tested the therapeutic potential of NR to prevent or even revert the development of aneurysm in MFS mice by modulating mitochondrial metabolism. Starting at 20 weeks old, Fbn1 C1039G/+ mice received IP injections of NR every second day for 28 days ( To characterize the effect of NR treatment at the molecular level, we performed RNA-sequencing on aortas from MFS and control mice. Hierarchical clustering of the transcriptomic analysis revealed that NR reverted the transcriptional changes observed in Fbn1 C1039G/+ aortas, bringing expression levels closer to controls than to Fbn1 C1039G/+ mice ( Figure 8F). NR increased the expression levels of genes related to mitochondrial complexes, fatty acid β-oxidation, and mitochondrial function, including Tfam and Ppargc1a, and reduced the expression of genes related to glycolytic rewiring, such as Hif1a and Myc ( Figure 8F). Notably, NR also decreased the expression of classical MFS-affected genes such as transcripts involved in the TGFβ pathway and those encoding ECM-related proteins ( Figure 8G). Additionally, NR brought the expression of some transcription factors involved in the maintenance of the contractile phenotype and genes related to the smooth muscle contraction apparatus, such as smooth muscle Acta2 and Cnn1, down to Fbn1 +/+ ( Figure 8H). Thus, boosting mitochondrial function with NR rapidly restores the transcriptional signature in aortas from MFS mice and fully reverts the associated aortic wall remodeling, aortic dilation, and medial degeneration. Taken together, our data highlight VSMC metabolism as a critical mediator of hereditable TAAs and suggest that the use of mitochondrial-boosting compounds is a promising pharmacologic strategy for treating patients with these hereditary disorders.

DISCUSSION
We found that aortas from Fbn1 C1039G/+ mice and human MFS patients present low TFAM expression, belownormal mtDNA levels, and a decline in mitochondrial respiration. This decline in oxidative phosphorylation was compensated by increased glycolytic metabolism and appears at the onset of the aortic disease. In line with our data, aortas from the Cutis laxa mouse model showed decreased mitochondrial respiration and increased glycolysis, 47 suggesting that mitochondrial decline could be a common driver and hallmark of different hereditary TAAs.
Next, we tried to elucidate how aneurysms caused by mutations related to ECM and VSMC contractile function converge on the impairment of TFAM expression and mitochondrial dysfunction. Recent reports in epithelial cells show that stiff matrices produce a mechanical regulation of glycolysis via remodeling of the cytoskeleton architecture. 22 Because VSMCs and aortas carrying TAA mutations show an increase of ECM deposition, active peptides and growth factors, stiffness, and mechanotransduction, 17,18,48 we hypothesized that the ECM could drive this metabolic rewiring toward glycolysis. In vitro experiments in VSMCs seeded in matrices produced by Fbn1-deficient VSMCs support that metabolism is controlled by the ECM. Hence, changes in the composition of the ECM during aneurysm development drives metabolic rewiring toward glycolysis.   In epithelial cells, PI3K (phosphoinositide 3-kinase) signaling coordinates glycolytic metabolism and actin remodeling. Activation of PI3K mobilizes aldolase from F-actin, increasing glycolysis. 20,49 In endothelial cells, another glycolytic pathway enzyme, PFKFB (phospho-fructokinase 2/ fructose-2,6 bisphosphatase), regulates vessel sprouting through the coordination of glycolytic metabolism and the actin cytoskeleton. 50 In dermal fibroblasts, downregulation of fatty acid β-oxidation and an increased glycolysis rate contribute to ECM alterations, promoting fibrosis. 51 In VSMCs from pulmonary hypertension patients, inhibiting glycolysis by silencing α-enolase prevents the malignant secretory phenotype. 52 Altogether, these data support that there is a bidirectional connection between the ECM composition and the cellular metabolism with a critical implication in aortic tissue homeostasis ( Figure VI in the Data Supplement) To demonstrate the importance of VSMC metabolism in the development of aneurysms, we generated a conditional mouse model with specific deletion of Tfam in VSMCs. Tfam deficiency in VSMC promotes mitochondrial dysfunction and a metabolic rewiring toward glycolysis. Tfam-depleted VSMCs acquire a senescent and proinflammatory phenotype together with an impairment in their contractile function. Moreover, these mice develop aortic aneurysm, medial degeneration, and lethal dissections. Similarly, silencing of Tfam in fibroblast prevents differentiation toward myofibroblasts and the acquisition of a contractile phenotype. 53 During the last decade, it became evident that signals from stressed mitochondria regulate senescence and inflammatory cascades in multiple cell types. 36,37,54 The release of mtDNA from the mitochondria to the cytosol activates the cGAS-STING response and contributes to the progression of the aneurysm. 37,38 In the same line, we have observed the activation of cGAS-STING route in both in aortas from Fbn1 C1039G/+ mice and in SM-Tfam −/− mice, further confirming the involvement of this route in the pathogenesis of the aortic diseases.
Besides the high incidence and mortality risk of ruptured aneurysms in inherited TAA patients, there are still limited therapeutic options to delay TAA progression to dissection and none to prevent it. Current pharmacologic standard treatments for TAA are based on BP control with β-adrenergic blockers or the Ang II receptor-1 antagonist losartan, which slow down aortic dilation but do not prevent dissection. 55,56 Therefore, the current management of aortic aneurysms relies on surgical prophylactic repair, which shows significantly higher perioperative mortality and morbidity. 57 Consequently, it is imperative to identify novel molecular mediators involved in hereditable TAA pathophysiology to design new pharmacologic strategies. Since the use of NAD + precursors has emerged as an effective approach to boost mitochondrial dysfunction in different pathologies, 40,41,45 we investigated the pharmacologic potential of NR in TAAs associated with genetic disorders. We observed that NR rapidly raised Tfam levels, improved mitochondrial metabolism, and normalized aortic function and diameter in the Marfan mouse model of TAA. Our results indicate that glycolytic metabolism is a common driver of hereditary TAAs and NR could be trialed for the treatment of different types of genetic disorders that cause TAA. Further research would be required to determine if this approach could also be effective against other vasculopathies, such as cerebral aneurysm or pulmonary hypertension.