Mitochondrial DAMPs-dependent inflammasome activation during aging induces vascular smooth muscle cell dysfunction and aortic stiffness in low aerobic capacity rats

Mitochondrial DAMPs-dependent inflammasome activation during


INTRODUCTION
Both aerobic exercise capacity and activity are well-established predictors of age-adjusted cardiovascular and overall mortality in both healthy subjects and those with cardiovascular disease (CVD) [1][2][3][4][5] . Peak aerobic exercise capacity declines with age, and the rate of decline increases markedly with each successive decade [6] . A decrease in the peak heart rate, peripheral oxygen consumption, cardiac β-adrenergic receptor density, and postsynaptic β-adrenergic signaling underlies the decline in aerobic capacity associated with aging, which lowers the threshold for CVD development.
Most of the population studies on exercise capacity focus on correlation-based assessment of patients without attempting to infer a causal relationship between the variables. Experimental validation of causal inference, if any, in correlation-based population studies was performed using unique rat models with high and low intrinsic aerobic capacities (High-Capacity Runners, HCR and Low-Capacity Runners, LCR) [7] . Specifically, this was accomplished by genetic segregation via two-way artificial selection based on treadmill exercise tests to investigate the influence of intrinsic aerobic capacity on all-cause mortality and morbidity [8] . The median lifespan for LCR is 28%-45% shorter than that for HCR, and maximal oxygen uptake serves as a reliable predictor of lifespan across adulthood [9] . In addition, for every 10 g gain in body weight during adulthood, lifespan decreased by approximately 1 month in a combined population of male and female rats, indicating that gain in body weight over time is also an important predictor of longevity. Early deaths in LCR may be attributed to pathophysiological mechanisms associated with aging, such as the failure of one or more organs. The LCR rats also exhibit a number of cardiovascular risk factors and impaired mitochondrial oxidative pathways [10] .
The hardening of the central arteries is a hallmark of advancing age, which impacts cardiac function, microcirculation, and organ function [30,31] and is a major cause of age-related CVD [32] . Lower cardiorespiratory fitness is correlated with arterial stiffening in coronary artery disease patients with myocardial infarction [33] , and higher cardiorespiratory fitness is correlated with a reduction in age-related arterial stiffness in adults [34] . There is also evidence that obesity accelerates the process of artery stiffening earlier in life [35,36] . It has been proposed that defective mitochondrial quality control contributes to the thickening of large elastic arteries and stiffening with age via mitochondrial stress/dysfunction [30] . However, the molecular mechanisms regulating aortic stiffening are yet to be fully elucidated.
This study examines VSMC function and metabolism in aging at the intersection of low aerobic capacity and obesity phenotype. We hypothesized that reduced aerobic capability in older LCR rats would be associated with impaired vascular function, increased aortic stiffening, elevated mitochondrial oxidative stress, mtDNA damage, impaired mitophagy, and inflammasome activation, along with vascular remodeling.

MATERIALS AND METHODS
Rats: All animal procedures were performed in compliance with the protocols approved by the Institutional Animal Care and Use Committee of the University of Michigan in accordance with NIH OLAW guidelines. The Low-Capacity Runner (LCR) and High-Capacity Runner (HCR) rat models have been described previously [7] . Exercise capacity was the basis for artificial selection in a genetically heterogeneous population of rats (N:NIH outcrossed stock) to isolate LCR and HCR rats. At 11 weeks of age, every rat underwent endurance running tests on a treadmill, and the time and maximal distance it was able to run before exhaustion were monitored. The top 20 percent and the bottom 20 percent of the runners were bred for genetic heterogeneity for 40 generations in order to segregate LCR and HCR rats [9] . We used young (4-month-old, 37th generation) and old (27-month-old, 34th generation) male rats in this study. All rats used in this study were also subjected to a one-time running test at 11 weeks of age to verify whether they had low-or high-running capacities. The data showed that LCR rats were able to run for 11.6 min before exhaustion, while HCR rats were able to run for 75 min before exhaustion, making it a relevant model of aerobic capacity.
RNASeq: RNA was extracted from thoracic aortas of young and old HCR and LCR rats (n = 3) using the Qiagen RNeasy Kit. RNA quality was assessed using an Agilent 2100 Bioanalyzer. All RNA samples had an RNA Integrity Number of ≥ 9.5. Sequencing libraries were prepared at University of Michigan Advanced Genomics Core with TruSeq RNA Library Preparation Kit (Illumina). RNA sequencing was performed on an Illumina HiSeq 4000 platform with three lanes per sample generating 125 bp paired-end reads with an average of 80 million reads per sample. Quality control and alignment for RNA sequencing were performed at the University of Michigan Bioinformatics Core. Raw reads were concatenated into a single FASTQ file for each sample. The quality of the raw reads was checked with FastQC. Reads were aligned to the reference genome (UCSC rn5.fa) using Tuxedo Suite: TopHat and Bowtie2 [37] . Cufflinks was used for post-alignment quality control, expression quantification, and normalization. iDEP.91 was used to perform hierarchical clustering, principal component analysis, and differential gene expression analysis with default parameters (Available from: http://bioinformatics.sdstate.edu/idep/) [38] . We identified differentially expressed transcripts and genes using the FDR ≤ 0.05 and fold change ≥ ±1.5. Differentially expressed genes were annotated using gene ontology terms, and pathway enrichment analyses were performed using the g:Profiler g:GOSt web application (Available from: http://biit.cs.ut.ee/gprofiler/gost) [39] . To identify enriched signaling pathways, we used the KEGG Pathway database (Available from: https://www.genome.jp/kegg/pathway.html).
VSMC culture: Four separate VSMC isolates were made from the aortas of young (4-month-old) and old (27-month-old) HCR and LCR rats. Details of the cell isolation protocol were published previously [40] . Briefly, aortic VSMCs were isolated by stripping off the adventitial layer and using only the media for cell isolation. The adventitial and medial layer separation was confirmed by immunostaining with α-SM actin antibody of random samples. Before digestion of the media to obtain VSMCs, calcium alginate swabs were used to remove the endothelium. Aortic rings (< 3 millimeters) were then incubated in collagenase type II and elastase mixture at 37 °C for 1.5 h or until complete digestion with intermittent mixing with a glass Pasteur pipet. Released VSMCs were cultured in DMEM supplemented with 10% FBS, antibiotics, and Amphotericin B (Gibco). Experiments were performed with cells from passages 3-10. Gene expression analysis of Acta2 (Actin Alpha 2, Smooth Muscle), Myh11 (Myosin heavy chain 11), and Cnn2 (Calponin 2) confirmed the phenotype of VSMCs in all isolates. Immunofluorescence: For Immunofluorescent staining, the tissues or cells were fixed in 3.6% formaldehyde for 20 min at room temperature and permeabilized in 0.05% Triton X-100 for 3 min, followed by 3 × 5 min washings in PBS. The fixed tissues or cells were incubated with primary antibodies [Supplementary Table 1] overnight at 4 °C according to the manufacturer's instructions. Thereafter, the tissue was washed three times with PBS and incubated with an Alexa Fluor-conjugated secondary antibody for 2 h. The Autofluorescence was quenched with TrueVIEW kit (Vector laboratories). The slides were mounted in mounting media containing DAPI (H100, Vector labs).

Picrosirius red staining:
The aortic sections were fixed in acetone and cold methanol for 30 min each. Sections were stained with Picrosirius red kit (Abcam). The excess stain was washed off twice with acetic acid, and the slides were dehydrated with ethanol. Sections were mounted using Vectashield H100 mounting medium without DAPI. The images were acquired under brightfield using a Nikon TI-E camera attached to an upright E600 microscope at equal exposure time for all images. NIH Image J was used to measure the area of picrosirius red staining in the aortic media.
Lipofuscin staining: The VSMCs were cultured in 8-well chambered glass slides until 80% confluency was reached. Cells were washed twice with PBS at room temperature, then incubated for 3 min in cold Glutaraldehyde fixative solution (25 mL acetone and 75 mL glutaraldehyde solution) for 3 min and washed thoroughly with PBS. Next, cells were stained with Sudan Black B staining reagent (Sigma Aldrich) for 5 min. The excess stain was removed by washing three times with 70% ethanol, followed by deionized water. The slides were mounted in VectaMount AQ (Vector Laboratories Inc.). Images were acquired using a Nikon TI-E camera attached to an upright E600 microscope under brightfield at equal exposure times. Perinuclear and cytoplasmic dark blue/black staining was measured using NIH Image J as an integrated density (mean gray value per area).
Mitophagy: VSMCs were cultured in 8-well chambered slides until 60% confluency was reached. Cells were incubated for 30 min at 37 °C with Mtphagy dye (Mitophagy kit, Dojindo), followed by mitophagy-inducing compounds or vehicle for 16 h. The cells were washed twice with PBS and then exposed to Lyso dye (Mitophagy kit, Dojindo) for another 30 min at 37 °C, according to the manufacturer's protocol. The slides were washed twice with 1xPBS and mounted in mounting media containing DAPI (H100, Vector labs). All slides were imaged by Nikon Eclipse Ti-fusion microscope with DAPI, FITC and Texas Red Channels at equal exposure, gain and offset. The colocalization of mtphagy and lyso dyes (yellow color) was quantified by NIH Image J as integrated density per area of interest. An indicator of mitophagy flux is the ratio of colocalization signals in induced samples to those in uninduced samples. We then measured autophagic vacuole formation using the CYTO-ID kit (ENZ-KIT175-0050, ENZO) according to the manufacturer's protocol. Briefly, the VSMCs were cultured until 80% confluency in 8-well chambered slides. The attached cells were incubated with the working solution containing CYTO-ID dye for 30 min at 37 °C. This was followed by two washes with 1xPBS and mounting in mounting media containing DAPI (H100, Vector labs) and imaging.

Measurement of ATP levels:
Cellular ATP levels in both young and old HCR and LCR VSMCs were measured using a bioluminescence ATP determination kit (ThermoFisher), following the manufacturer's instructions. In this assay, ATP reacts with luciferin in the presence of luciferase, converting it into oxyluciferin. As a result, the light produced is proportional to the cellular ATP levels [41] . In brief, equal volumes of VSMC protein lysates were incubated with a reaction buffer containing D-luciferin (500 M), firefly luciferase (1.25 g/mL), and 1 mM (DTT). The luminescence was measured with a luminometer, and ATP levels were determined using a standard curve generated simultaneously.
Western blotting: VSMC and aortic protein lysates were prepared using M-PER and T-PER lysis buffers, respectively, and the protein concentration was determined using BCA assay (ThermoFisher). Equal amounts of proteins were loaded on SDS-PAGE gels and were separated by electrophoresis. The proteins were then transferred onto nitrocellulose membranes using a semi-dry Western transfer apparatus (Trans-Blot Turbo system, BioRad). The blots were blocked with Blocker TM BSA (ThermoFisher) and probed overnight with primary antibodies (AIM2, Caspase1, VCAM, LAMP1 from Santa Cruz; IL18, TOM20 from Abcam; FITC-tagged-ACTA2 from BIOS). Blots were washed 3 × 5 min with PBST (PBS + 0.1% TWEEN 20), then stained with IRDye 800CW (green) or IRDye 680RD (red) secondary antibodies (LI-COR) for 1 h at room temperature. The blots were imaged with a near-infrared fluorescence digital detection scanner, Odyssey CLx (LI-COR).

Mitochondrial bioenergetics:
The mitochondrial bioenergetics of rat VSMCs were measured using the Seahorse XFp and XF96 machines (Agilent technologies). Approximately 20,000 cells were cultured in DMEM supplemented with 10% FBS and antibiotics in cell culture microplates supplied with Flexpack (Agilent technologies) for 24 h. The next day, the media were changed to phenol-red free basal DMEM (Agilent technologies) supplemented with 25 mM glucose, 2 mM pyruvate, and 1 mM glutamine, and incubated in a CO 2 -free incubator at 37 °C for least 1 h before analysis. Cellular oxygen consumption rate (OCR) at basal level and in the presence of 2 M oligomycin, 500 nM trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP) and 500 nM rotenone + antimycin A was quantified in three consecutive cycles for 3 min each. Pre-treatment of cells with 20 nM rapamycin or 10 nM MitoTEMPO were performed where indicated for 16-18 h. In order to normalize the OCR values to protein levels, the cells were lysed with M-PER buffer and the protein content was determined by BCA assay.
Mitochondrial DNA damage analysis: The total genomic DNA of rat aortas was isolated with a Qiagen blood and tissue kit. The DNA concentration was determined using two different methods, the Nano-Drop ND-1000 spectrophotometer and the Quant-iT PicoGreen dsDNA Assay Kit (Thermo Fisher Scientific). Equal amounts of DNA were used as a template to amplify a 13.4 kb mitochondrial DNA fragment. We used rat-specific primers (Forward: AAA ATC CCC GCA AAC AAT GAC CAC CC, Reverse: GGC AAT TAA GAG TGG GAT GGA GCC AA) [42] and Takara LA Taq TM long DNA polymerase (Takara Bio USA) for amplification of mitochondrial DNA in a polymerase chain reaction. An equal aliquot of the PCR product was run on an agarose gel to determine the quality and densitometry to determine the quantity of the expected size band. The amount of amplification product from the long-fragment PCR was normalized by measuring the levels of MTCO1 in the DNA template in a separate quantitative PCR reaction using a Taqman probe.
Cell-free DNA isolation and amplification: An equal number of VSMCs from young and old LCR and HCR rats were plated in 6-well culture dishes. After 30 h of incubation, the media were spun down at 2000 xg for 5 min to remove floating cells and debris. The DNA was extracted in 500 µL of cell-free medium from each sample using a blood and tissue kit (Qiagen). The exuded DNA was measured using a Nano-Drop ND-1000 spectrophotometer, and equal volumes of DNA were used as templates in a quantitative PCR (ABI 7500 Real-time PCR machine). A fragment spanning 80 base pairs from the MTCO1 gene was amplified using the Taqman probe (Applied Biosystems). The experiment was repeated at least three times, and the average cycle number (Ct) was plotted. The Ct value is inversely proportional to the amount of circulating mtDNA. For in vivo circulating mtDNA analysis, we used 300 μL of plasma from at least three different rats.
DAMPs preparation: MtDNA was isolated from rat liver mitochondria using DNA blood and tissue kit (Qiagen). Firstly, mitochondria were isolated with mitochondrial isolation kit (Pierce), and then mitochondrial pellets were used to isolate DNA. DNA was sonicated up to 5 cycles on ice, at a low setting, each for 30 s, spaced by 1-min intervals. The DNA samples before and after the sonication were run on 1.5% agarose gel to determine the size of the fragmented DNA samples. A length of 150 to 250 base pairs was considered optimal to induce inflammasome activation in cells.

Pressure myography:
The descending thoracic aorta segments were isolated from LCR and HCR rats (n = 7). The tissues were mounted with 6-0 nylon sutures so that aortic segments (approximately two to three millimeters in length) between intercostal arteries could be pressurized in a large vessel chamber of a pressure myograph system (Living Systems Instrumentation, Burlington, VT). Aortic segments were incubated in Ca 2+ -free PSS (130 mM NaCl, 4.7 mM KCl, 1.18 mM KH 2 PO 4 , 1.17 mM MgSO 4 ·7H 2 O, 14.9 mM NaHCO 3 , 5.5 mM dextrose, 0.026 mM EDTA, 1.6 mM EGTA, pH~7.4) for 20 min at 37 °C before testing. Initially, vessel had an internal pressure of 10 mmHg, and pressure was increased by 20 mmHg incrementally while allowing time for the vessel to stabilize (< 1 min). The external diameter of each vessel (opacity prevented visualization of internal diameters) was recorded at each increment up to 190 mmHg pressure using a video dimension analyzer calibrated to a 2X objective with a 0.35X c-mount lens (Nikon). Strain (ε) of the vessel was calculated using the formula ε = (D i -D o )/D o , where D i represents the external diameter at a given pressure and D o represents the initial lumen diameter at 10 mmHg pressure. Pressure vs strain data was fitted to a simple exponential formula P = α*e (βε) , where α is a constant, and β represents a measure of relative vascular stiffness [43,44] .

Statistics:
All biochemical experiments were repeated at least three times. All analyses were performed with Prism 8 (GraphPad Software). All data were examined for normality using the Shapiro-Wilk test. Where applicable, data were analyzed using an unpaired 2-tailed t-test. For experiments involving more than two groups, one-way ANOVA with Tukey post-hoc tests were used. Differences were considered significant at P < 0.05.

The interplay between aerobic exercise capacity and aging affects arterial function, inducing aortic stiffening
The development of aortic stiffening is a key feature of aging and an independent predictor of cardiovascular events in hypertensive patients as well as in the general population [45][46][47] . In this study, we investigated the interaction between exercise capacity, aging, and arterial stiffness. First, we investigated the vasomotor function of mesenteric artery rings by wire myography; however, it did not indicate any significant differences in phenylephrine-induced maximum contraction between old age LCR and HCR rats [ Figure 1A]. We then measured acetylcholine (ACH)-induced relaxation of mesenteric arteries from young and old age rats. The relaxation response to ACH was similar in young LCR and HCR rats [ Figure 1B]. However, relaxation of the mesenteric arteries from old age LCR rats was significantly smaller than that from old age HCR rats at lower concentrations of ACH, with a marked rightward shift in the concentrationresponse curve [ Figure 1C] and a smaller half-maximal relaxation response [-log(EC50); Figure 1D].
We asked whether a decrease in eNOS expression might be responsible for arterial dysfunction in old LCR rats, since nitric oxide derived from eNOS is important for vascular function. The Western blot analysis of aortic lysates of old age LCR rats revealed a 40% reduction in eNOS protein expression in comparison with the old HCR rats [ Figure 1E]. In order to find out if attenuated eNOS phosphorylation and inhibition of its activity contribute to arterial stiffening with low aerobic capacity and aging [48] , we analyzed the expression of phospho-eNOS (Ser1177) in the aortas of old LCR and HCR rats. The phospho-eNOS levels in old HCR rats were higher than in old LCR rats [ Figure 1F], suggesting that endothelial dysfunction in mesenteric arteries and aortas of LCR rats with aging may be due to reduced eNOS protein expression and activity.
To determine the effects of low aerobic capacity on the function of large conduit arteries during aging, we measured the aortic stiffness of thoracic aortic segments using pressure myography. There was no difference in aortic plasticity between young LCR and HCR rats [ Figure 1G]. In contrast, the old LCR rats showed a significant increase in aortic stiffness compared to age-matched HCR rats [ Figure 1H and I, P = 0.0002]. The remodeling of the extracellular matrix is a major contributor to vessel wall stiffness in aging, and collagen deposition and cross-linking are key to this process [16,49,50] . We stained fresh frozen aortic cross sections with picrosirius red to measure the collagen content. Aortic collagen levels were not different in young LCR and Figure 1. Increased vascular dysfunction, fibrosis, and stiffness in LCR rats with aging. The contractile force in mesenteric artery rings of young (4-month old) and aged (27-month old) LCR and HCR rats in response to phenylephrine (PE) was measured by wire myography (mean ± SEM; n = 5) (A). Concentration-response curves for acetylcholine (ACH) induced relaxation of mesenteric arteries from young (B) and old age (C) LCR and HCR rats (mean ± SEM; n = 5). Half-maximum relaxation concentration of ACH [-log(EC50)] was calculated from independent dose-response curves and plotted as mean ± SEM, n = 5 (D). Representative Western blot of eNOS protein expression in aortic lysates. Densitometry analysis of eNOS protein levels normalized to β-tubulin levels (mean ± SEM, n = 3) (E). Representative fluorescence images of aortic cross sections stained for immunoreactive total eNOS (red) or phospho-eNOS Ser1177 (green). Data are presented as mean integrated density/ROI ± SEM where n = 4 [(F), scale = 50 μm)]. Pressure vs strain data of thoracic aorta segments from young (G) and old (H) LCR and HCR rats were analyzed using pressure myography. The relative stiffness coefficient (H) was calculated by fitting pressure vs strain data to an exponential formula (mean ± SEM, n = 5) (I). Representative images of rat aortic cross sections stained with Picrosirius red (PSR) (J). Representative fluorescence images of aortic cross sections co-stained for TGF-β (red) and SM-α-actin (green). Bright yellow fluorescence indicates colocalization of TGF-β and SMα-actin. Data are presented as mean integrated density/ROI ± SEM where n = 4 (K). *P < 0.05; **P < 0.01; ****P < 0.0001. HCR rats [ Figure 1J]. Conversely, old LCR aortas have a 34% higher collagen content than old HCR aortas [ Figure 1J, P < 0.01]. Because eNOS-deficient mice have elevated TGF-1β and collagen I levels in their aortas [51] , we investigated whether a decrease in eNOS expression or activity in old LCR rat aortas was associated with increased TGF-1β expression. Interestingly, TGF-1β levels were 2.7-fold higher in old LCR rat aortas than in old HCR aortas [ Figure 1K]. An increase in TGF-1β expression causes VSMCs to undergo phenotypic changes, upregulating contractile proteins, collagen synthesis and secretion, and enhancement of intrinsic cellular and ECM stiffness, decreasing aortic compliance [52] . The data suggest that low aerobic capacity in old age impairs endothelial NOS and alters the VSMC phenotype, causing dysregulation of vascular wall homeostasis and aortic stiffening.

Aging is characterized by increased inflammation and necroptosis pathways in the aortic transcriptome of LCR rats
In order to examine how aging and low aerobic capacity impact vascular phenotype at the global level, we performed genome-wide RNA sequencing analysis and compared the transcriptional profiles and signaling pathways enriched in the aortas of young and old HCR and LCR rats. In a hierarchical cluster analysis of the top 1000 most variable genes, there was a distinct separation between young and aged rats, as well as between HCR and LCR rat aortas [ Figure 2A]. A principal component analysis showed that old LCR aortas have different transcription profiles from old HCR aortas, besides the obvious differences between young and old rat aortas [ Figure 2B]. In the comparison of significantly differentially expressed genes (FDR ≥ 0.05 and fold change ± 1.5) among all groups, the largest differences were found between old HCR and old LCR aortas [ Figure 2C]. Transcription analysis revealed a unique expression pattern between old LCR and old HCR aortas with over 1100 genes differentially expressed [ Figure 2D].
Compared to old LCR aortas, HCR aortas had 555 upregulated genes and 560 downregulated genes. By contrast, the differentially expressed genes in young HCR aortas were 168 upregulated and 160 downregulated compared with those in young LCR aortas. A gene ontology pathway enrichment analysis focused on selected signaling pathways defined in the KEGG database was conducted to better understand phenotypic changes defined by transcriptional profiles. Significantly differentially upregulated transcripts in old HCR aortas were related to longevity regulating pathways, growth factor signaling pathways (insulin, AMPK, glucagon, and neurotrophin signaling), oxidative phosphorylation, DNA/RNA sensing (RIG-I receptor signaling), and mitophagy/autophagy pathways[ Figure 2E]. In contrast, genes significantly upregulated in old LCR aortas were associated with cell cycle regulation (p53 and FoxO signaling) and proinflammatory pathways (AGE-RAGE, TGF-β, Toll-like receptor, NF-kappa B, and chemokine signaling) and apoptosis/necroptosis/ferroptosis regulating pathways [ Figure 2F]. Based on pathway enrichment analysis, old LCR aortas have reduced mitochondrial biogenesis and mitophagy, increased damage and cytosolic accumulation of DNA/RNA, pro-inflammatory pathway activation, and enhanced apoptosis and necroptosis.

Low aerobic capacity in old age is associated with increased mitochondrial oxidative stress, mitochondrial DNA (mtDNA) damage and dysfunction
We and others have demonstrated that mitochondria-derived oxidative stress increases with aging and affects vascular function and arterial stiffness in animal models and humans [16,40,[53][54][55] . VSMC plasticity plays a pivotal role in vascular remodeling in aging [56,57] , which is why we investigated many mechanisms of mitochondrial dysfunction in both the aorta and VSMCs to determine how low aerobic exercise capacity impacts VSMC function. VSMCs isolated from 4-month-old rats did not show any differences in proliferation or senescence between LCRs and HCRs. Aortic ROS levels and mitochondrial ROS levels (mtROS) were measured in aortas as well as in the aortic VSMCs of young and old LCR and HCR rats. Aortic ROS levels measured by DHE fluorescence were not significantly different between both young and old age LCR and HCR rats [Supplementary Figure 1]. However, a higher percentage of VSMCs from old LCR rats had comet tails than VSMCs from old HCR rats, indicating increased nuclear DNA damage [ Supplementary Figure 2A]. This was supported by increased nuclear 8-OHdG levels in VSMCs from old LCR rats [Supplementary Figure 2B]. Increased nuclear DNA damage in VSMCs from old LCR rats was accompanied by decreased levels of nuclear γ-H2AX [Supplementary Figure 2C]. An RNA-Seq analysis showed significantly lower superoxide dismutase 2 expression levels in VSMCs from old LCR rats compared to old HCR rats (P = 0.0378). The MitoSOX Red fluorescence measurements of the aortic mitochondria of young LCR and HCR rats did not reveal any differences [ Figure 3A]. However, mtROS levels in aortic cross sections were found to be three times higher in aged LCR rats than in HCR rats [ Figure 3A; P = 0.0013]. In agreement with this, the aortic VSMCs of old age LCR rats have a 2-fold increased mtROS level compared to that of old age HCR rats [ Figure 3B; P < 0.05; Supplementary Figure 3], suggesting a link between low aerobic capacity and increased mitochondrial oxidative stress during old age.
Since increased mitochondrial oxidative stress can cause mtDNA damage [58,59] , we investigated mtDNA integrity by long PCR amplification of a 13.4 kb mtDNA fragment. Consistent with the increase in mtROS . The amount of mtDNA was estimated by qPCR of cell-free DNA isolated from a fixed volume of VSMC cell-free culture medium (E). Data presented as cycle number ± SEM, n = 3-4. Mitochondrial membrane potential was measured by flow cytometry and the data presented as percentage of cells scored positive for high tetramethylrhodamine ethyl ester (TMRE) fluorescence. Data are mean ± SEM, n = 3-4 (F). ATP levels were measured using a colorimetric assay kit and the data presented are arbitrary units (mean ± SEM, n = 4 (G). Oxygen consumption rate (OCR) of aortic VSMCs isolated from young and old age rats. VSMCs were plated at a density of 15000 cells/well in the XF96 culture plates and OCR was measured at the basal level and in the presence of 2 μM oligomycin, 500 nM trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP), and 0.5 μM antimycin A + 500 nM rotenone in VSMCs from young (H, left panel) and old age (I, left panel) LCR and HCR rats. Mitochondrial bioenergetic parameters were derived from VSMC OCR measurements (H and I, right panels). Data are mean ± SEM, n = 4. *P < 0.05; **P < 0.01; ****P < 0.0001. levels, an 8.6-fold lower mtDNA amplification was observed in aortas from aged LCR rats compared to aged HCR rats [ Figure 3C]. We wondered if this dramatic reduction in mtDNA amplification was due to an additive effect of mitochondrial ROS and the loss of protective mechanisms. Western analysis for TFAM in the VSMC lysates showed a 3.3-fold decrease in old LCRs compared to old HCRs [ Figure 3D]. Based on these results, low aerobic capacity-induced mtROS contribute to mtDNA damage in old age, impairing the replication of mtDNA.
Because aging can lead to increased production of damage-associated molecular patterns (DAMPs) [25] , we investigated the effect of the interaction of low aerobic capacity and aging on mtDNA damage as measured by QPCR with a short Taqman probe specific for MTCO1 gene amplification in VSMC culture medium of aged LCR and HCR rats. The Ct value for MTCO1 amplification decreased significantly in DNA amplified from VSMC culture media from old age LCR rats compared with that in HCR rats, which is inversely related to the amount of extracellular MTCO1 DNA present in the culture media [> 2 Ct value decrease, P < 0.01, Figure 3E].
Because extrusion of cellular mtDNA occurs following mitochondrial depolarization [23] , we measured mitochondrial membrane potential in rat aortic VSMCs from old age LCR and HCR rats by flow cytometric analysis of TMRE retention [60] . A 30% decrease in TMRE high cells was observed in old age LCR rat VSMCs compared with HCR rat cells [ Figure 3F]. Mitochondrial membrane potential harnesses proton motive force to generate ATP. A colorimetric assay was used to compare levels of ATP in VSMCs from young and old age LCR and HCR rats. The VSMCs from older LCR rats exhibited 2.24-fold lower ATP levels (P < 0.05) than VMSCs from younger HCR rats, while cells from young LCR and HCR rats exhibited equal ATP levels [ Figure 3G].
The mtDNA-encoded proteins are essential components of respiratory complex subunits, and mtDNA damage compromises cellular respiration [61] . We measured oxygen consumption rate (OCR) to examine the effects of low aerobic exercise capacity as well as aging on VSMC respiration. The OCR of VSMCs from young LCR and HCR rats was not significantly different [ Figure 3H]. However, the maximal OCR by complex IV in the presence of FCCP, which collapses the proton gradient and disrupts the mitochondrial membrane potential and enables an uninhibited electron transport chain, was significantly lower in VSMCs from old age LCR compared to HCR rats [ Figure 3I]. In addition, the reserve respiratory capacity, defined as the difference between maximal and basal respiration, was significantly lower in VSMCs from old age LCR versus HCR rats [ Figure 3I]. Together, these data confirm that low aerobic capacity negatively affects the ability of conduit arteries in LCR rats to respond to increased energy demands or stresses, particularly as the animals age.

LCR rats exhibit impaired mitophagy in aging
Inhibition of mitophagy, the autophagy of mitochondria, causes the accumulation of dysfunctional mitochondria and extrusion of mitochondrial components [25] . As we observed increased mitochondrial damage and higher levels of mtDNA DAMPs, we investigated whether LCR rats suffer from impaired mitophagy in old age. Mitophagy was measured as integrated density from the colocalization of immunofluorescent red mitochondrial marker protein TOM20 with green lysosomal marker protein LAMP1 and blue SM α-actin in aortas and from the colocalization of red TOM20 and green LAMP1 in VSMCs. Older LCR rats displayed a 54% decrease in mitophagy in medial SMCs of aortas [ Figure 4A], which was also supported by a 76% reduction in cultured VSMCs [ Figure 4B; Supplementary Figure 4] in comparison to older HCR rats. The mitophagy levels did not differ significantly between young age LCR and HCR rats.
Mitophagy induction in live VMSCs was observed using Mtphagy Dye, which emits a high signal when damaged mitochondria fuse with lysosomes (Mitophagy kit, Dojindo) in the presence of rapamycin, a wellknown mitophagy inducer [62] . The fusion of Mtphagy Dye-labeled mitochondria and lysosomes was confirmed with Lyso Dye. The rapamycin-induced mitophagy in VSMCs of both young and old LCR and HCR rats [ Figure 4C]. However, mitophagy induction was higher in VSMCs of young HCR rats than in old LCR rats, and in VSMCs of old HCR rats than in old LCR rats.
The autophagic vacuole formation is one of the essential steps in transporting damaged mitochondria into lysosomal compartments. We measured the formation of autophagic vacuoles in VSMCs by measuring the green-fluorescent puncta that form when CYTO-ID (ENZO Inc.) is specifically localized to autophagic vacuoles. In old LCR VSMCs, green fluorescence area of puncta was 12.0-fold lower than in old HCR VSMCs, indicating impaired autophagy flux, but this difference was not found in young HCR and LCR VSMCs [ Figure 4D]. Western blot analysis of VSMC lysates also revealed a decreased level of LC3B-II, a marker of autophagosomes, in old LCR rats compared to old HCR rats [ Figure 4E, P < 0.05]. Additionally, we tested the effects of autophagic flux inhibitor Bafilomycin A1 (50 nM for 16 h) on LC3B-II and autophagic adapter p62 levels in VSMCs. The levels of p62 [Supplementary Figure 5A] and LC3B-II [Supplementary Figure 5B] in VSMCs from young LCR and HCR rats as well as old HCR rats were significantly increased by bafilomycin A1. In contrast, p62 and LC3B-II levels in VSMCs from old LCR rats were lower than those from old HCR rats, indicating lower synthesis of these proteins and impaired autophagy and mitophagy. The results of Western blot analysis failed to show differences between VSMCs from old LCR and HCR rats in p62 or LC3B-II levels in response to Bafilomycin A1 treatment, possibly due to heterogeneity of HCR/LCR rats and difference in the sensitivity of the assays [Supplementary Figure 5C]. Additionally, the accumulation of lipofuscin, an indicator of lysosomal dysfunction and a biomarker of impaired molecular and organellar recycling, was significantly greater in VSMCs of old LCR rats compared with those of old HCR rats [ Figure 4F, P < 0.05]. Based on these findings, it appears impaired autophagy/mitophagy underlies mitochondrial dysfunction with aging in LCR rats.

Inflammasome activation in aging is associated with diminished aerobic capacity
The inhibition of mitophagy activates inflammasomes, cytosolic danger sensing platforms via increased mitochondrial ROS generation and DAMPs [27,63] . Damaged mtDNA is actively extruded from cells via exosomes, and cytosolic DNA activates the AIM2 inflammasome [64] . Hence, we investigated whether impaired mitophagy and elevated DAMP levels in old LCR rat VSMCs increase the expression of AIM2, caspase-1 and IL-18, which are downstream effectors of AIM2 inflammasome activation. As shown in Figure 5A, immunofluorescence studies indicated a marked increase in AIM2 (4.1-fold), caspase-1 (5.6-fold), and IL-18 expression (18.3-fold) within the medial layer of aortic cross sections from old age LCR rats compared with HCR rats.
In consonance with inflammasome activation, VCAM-1, a major regulator of leukocyte adhesion and transendothelial migration that is upregulated in chronic inflammatory conditions [65] , also increased by 40-fold in old age LCR rat aortic medial layer compared with HCR rats [ Figure 5A]. Using Western blotting, we investigated the expression of AIM2 and its downstream effectors in aortic VSMC lysates in order to corroborate inflammasome expression and activation in aortic media at the intersection between aging and low oxidative capacity. Similar to the findings of aortic immunofluorescence, the expression levels of AIM2, IL-18, and cleaved caspase-1 were 3.0, 2.0, and 1.8-fold greater, respectively, in aged LCR rat VMSCs compared to HCR rat VSMCs [ Figure 5B, Supplementary Figure 6A]. VCAM-1 expression increased 1.9-fold in aged LCR rat VMSCs compared to aged HCR rat cells, confirming increased inflammation [ Figure 5B]. The levels of inflammasome expression/activity or effector protein expression were not significantly different between the young LCR and HCR rat aortas [ Figure 5A] or VMSC lysates [ Figure 5B]. In addition, we examined NLRP3 inflammasome expression in young and old rat aortic VSMCs. No significant differences in NLRP3 protein levels were observed between young and old LCR and HCR rat aortic VSMC lysates [Supplementary Figure 6B].

Rapamycin and MitoTEMPO attenuate inflammasome activation in aortic VSMCs from old age LCR rats
To determine whether mtDNA damage is sufficient to induce inflammasome expression and activation, VSMCs from young LCR rats were treated with liver mtDNA DAMPs at a concentration of 10 ng/mL for 24 h. The levels of inflammasome expression and activation markers in VSMCs treated with DAMPs were comparable to those in untreated VSMCs from old age LCR rats, and DAMP treatment increased AIM2, cleaved caspase-1, and cleaved IL-18 levels by 18.6, 3.3, and 9.9-fold, respectively [ Figure 6A and B, Supplementary Figure 7A]. The high degree of variability in inflammasome marker expression in response to DAMPs treatment might reflect the polygenic backgrounds of rat strains.
Since rapamycin inhibits AIM2 activation [66] and reduces aortic stiffness with aging [67] , we investigated whether rapamycin inhibits AIM2 pathway activation in aged LCR rat VSMCs. Rapamycin decreased cleaved caspase-1 and cleaved IL-18 levels by 2 and 7.5-fold, respectively [ Figure 6C and D, Supplementary Figure 7B]. However, rapamycin had no effect on AIM2 protein levels, suggesting that inhibition occurs at the activity level rather than at the protein level. We also studied the effects of Figure 5. Inflammasome is activated in aortic medial VSMCs and in cultured aortic VSMCs of old age LCR rats. Representative confocal images of aortic cross sections stained for inflammasome activation markers (AIM2, caspase-1, IL-18, and VCAM1) in young and old age rats (A, left panels, red). Aortas were co-stained for ACTA2 (green). The bright yellow fluorescence indicates activation of specific inflammasome markers in medial VSMCs. Data are presented as average integrated density/ROI ± SEM where n = 4 (A, right panels). Representative Western blots showing expression of the inflammasome markers in VSMC lysates (B, left panel). Data are presented as fold change, mean ± SEM, n = 4 (B). Scale is 50 μm. *P < 0.05; **P < 0.01; ***P < 0.001.
Next, we tested whether inhibition of inflammasome activation resulted from decreased mtDNA DAMP generation. From rapamycin-treated and untreated VSMC culture media, cell-free DNA was isolated and mtDNA DAMP levels were determined by analyzing the MTCO1 gene by quantitative PCR. The rapamycin treatment reduced mtDNA amplification by 4.5 ± 1.7 cycles (P < 0.05) [ Figure 6G]. MitoTEMPO also reduced mtDNA DAMP levels in the culture medium of old age LCR rats [ Figure 6G], supporting the notion that inflammasome activation increases mtDNA DAMPS levels, exacerbating inflammation in a feedforward manner.
The hematopoietic interferon-inducible nuclear protein p202 serves as a dominant-negative antagonist of AIM2, blocking downstream signal transduction by binding to cytoplasmic dsDNA that act as ligands to AIM2 [68] . Because none of the antagonists we tested had an effect on AIM2 protein expression, we examined p202 levels in response to antagonist treatment in VSMCs from old LCR rats. Rapamycin increased p202 levels by 3.7-fold [ Figure 6H and I]. MitoTEMPO increased p202 levels 2.4-fold, but the effect was not statistically significant. Together, these data support the notion that targeting mtDNA damage and decreasing mtDNA DAMPs prevents inflammasome activation at the intersection of low oxidative capacity and aging.

Rapamycin and MitoTEMPO reduced mitochondrial oxidative stress and increased mitophagy and OCR in VSMCs from old LCR rats
We next examined whether rapamycin or MitoTEMPO had salutary effects on mitophagy. Both rapamycin and MitoTEMPO increased mitophagy by 4.8 and 2.9-fold over respective controls, as measured by colocalization of LAMP1 with TOM20 [ Figure 7A and B; Supplementary Figure 8A]. The mitochondrial ROS levels, as assessed by changes in MitoSOX Red fluorescence, decreased by 2.5 and 3-fold in old LCR rat VSMCs treated with rapamycin and MitoTEMPO compared to respective controls [ Figure 7C and D, Supplementary Figure 8B], suggesting impaired mitophagy promotes ROS generation [69] . In order to determine whether enhanced mitophagy and reduced mitochondrial ROS levels affected mitochondrial function, we measured OCR in VSMCs from old LCR rats treated with and without rapamycin and MitoTEMPO. Both the compounds significantly increased maximal OCR and reserve respiratory capacity in old LCR rat VSMCs [ Figure 7E-H], suggesting the importance of mitochondrial function in vascular health at the intersection of aging and low oxidative capacity.

DISCUSSION
In the present study, we provide experimental evidence that mtDNA damage and dysfunction are induced by low intrinsic aerobic capacity and aging, adversely affecting vascular health. Our results show: (1) increased transcription of genes involved in cell cycle regulation and pro-inflammatory pathways, impaired mitophagy, and impaired oxidative phosphorylation pathways in old age LCR rat aortas versus longevity regulating pathways, increased oxidative phosphorylation and mitophagy/autophagy pathways in old age HCR rat aortas; (2) impaired arterial relaxation and increased aortic stiffness in old age LCR rats compared to the HCR rats; (3) increased mitochondrial oxidative stress and DNA damage in aortas and mitochondrial oxidative stress in aortic VSMCs of old age LCR versus HCR rats; (4) increased expression of AIM2 and its downstream effectors and VCAM1 in the aortic medial layer, as well as AIM2 activation in the aortic VSMCs of old age LCR versus HCR rats; (5) mtDNA DAMP-treated VSMCs from young LCR rats show similar expression and activation of AIM2 inflammasome as do those from old LCR rats; and (6) rapamycin and MitoTEMPO attenuate mtDNA DAMPs-induced inflammasome activation in young LCR rat aortic VSMCs and reduce mitochondrial oxidative stress, improving OCR and mitophagy in old age LCR rat VSMCs. Together, these results demonstrate that mtDNA DAMPs associated with low aerobic capacity contribute to sterile inflammation in aging, resulting in vascular dysfunction and aortic stiffness.
We and others have demonstrated that increased mitochondrial oxidative stress in aging causes arterial stiffening [16,30,40,55] . Seals and colleagues reported that mitochondrial quality control, a balance between Figure 7. Treatment with rapamycin or MitoTEMPO decreases oxidative stress and enhances mitophagy and OCR in old age LCR rat VSMCs. Representative confocal images of VSMCs treated with and without 20 nM rapamycin (A), or MitoTEMPO (B) and co-stained for lysosomal marker LAMP1 (green) and mitochondrial marker TOM20 (red). Bright yellow fluorescence indicates colocalization of LAMP1 and TOM20 and active mitophagy. Data are presented as average integrated density/ROI ± SEM where n = 4. Scale is 10 μm. Representative confocal images of mitochondrial ROS levels with and without rapamycin (C) or MitoTEMPO (D) treatments as measured by MitoSOX Red (MSR) fluorescence. Intensity of fluorescence was presented as integrated density (mean ± SEM, n = 4, Scale is 10 μm). OCR was measured at the basal level and in the presence of 2 μM oligomycin, 500 nM trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP), and 0.5 μM antimycin A + 500 nM rotenone in VMSCs of old age LCR rats treated without and with rapamycin (E and G) or MitoTEMPO (F and H). Mitochondrial bioenergetic parameters were derived from VSMC OCR measurements. mitochondrial biogenesis and mitophagy, is impaired in aging, resulting in higher collagen-1 levels and aortic stiffness [30] . The present data demonstrate that mitochondrial oxidative stress causes VSMC and arterial dysfunction, as evidenced by impaired relaxation of the mesenteric arteries to acetylcholine and aortic stiffening in older LCR rats, extracellular matrix remodeling and fibrosis, and by a leftward shift in the stress-strain curves. The increased production of mitochondrial ROS and inflammation, as evident by increased IL-18 and VCAM1 levels in VSMCs old LCR rats, causes oxidative damage and degradation of elastic laminae, as well as excessive collagen deposition in the vascular wall. Increased vascular stiffness and dysfunction are a direct consequence of these processes [40] . Endothelial dysfunction resulting from increased mitochondrial oxidative stress may also contribute to aortic stiffening in old age LCR rats, as illustrated by decreased eNOS levels, decreased phosphorylation of eNOS (Ser1177), and TGF-β1 upregulation in aortic medial VSMCs induced by eNOS inhibition [49,53,70] . Several mechanisms may have contributed to the downregulation of eNOS in old LCR rats. First, LCR rats exhibit a phenotype of metabolic syndrome, including obesity [71] which is associated with decreased levels of eNOS [72,73] . Second, TNF-α, which is increased in obesity, was shown to downregulate eNOS expression [73] . Interestingly, TNF-α levels were higher in the media of cultured LCR splenic mononuclear cells [74] , suggesting that TNF-α may downregulate eNOS in LCR rats. Lastly, mitochondrial abnormalities due to oxidative stress may have decreased eNOS expression [70] and our data indicate elevated mitochondrial ROS in LCR vasculature.
Reduced cardiorespiratory fitness is associated with arterial stiffness in coronary artery disease patients [33] . Chronic exercise training in aged rats reduces mitochondrial ROS production, preserving aortic mitochondrial function, and reducing aortic stiffening [75] . Old age HCR rats that demonstrate better mitochondrial energetics, lower ROS levels and less DNA damage in VSMCs than do age-matched LCR rats are protected against aortic stiffening and arterial dysfunction. These results are in general agreement with a study in which cardiomyocytes from HCR rats showed higher aerobic capacity, mitochondrial fitness, and health in association with a longer lifespan [76] .
Moreover, our data indicating impaired autophagy and lysosomal dysfunction in aortic VSMCs in old age LCR rats agree with the reports of age-associated decline in proteolytic systems involved in protein quality control, including the autophagy-lysosomal system [77,78] . Aged LCR rats have a significant increase in mitochondrial ROS levels, mtDNA damage and nuclear DNA damage compared with young LCR rats. It is suggested that impaired mitophagy may be caused by reduced ATP levels that inhibit lysosomal function and protein sequestration [79] . When mitochondrial depolarization occurs, PINK1 recruits Parkin to mitochondria and initiates mitophagy. Mitophagy could be compromised by low levels of ATP, which inhibits the upregulation of PINK1 levels and the recruitment of Parkin into mitochondria [80] . As mitophagy genes are encoded by nuclear DNA, impaired mitophagy may reflect nuclear DNA damage in aged LCR rats [81] .
Perhaps lower mitochondrial ROS levels and less nuclear and mtDNA damage can explain why mitophagy is not impaired in aged HCR rats. In fact, our RNA-Seq data show upregulation of mitophagy in old HCR rats. In addition, insulin signaling [82] and AMPK signaling [83,84] pathways that are implicated in mitophagy activation are upregulated in old HCR rats. Further supporting impaired mitophagy in old LCR rats, mitochondrial quality control and biogenesis effectors SIRT1/TFAM are downregulated in aged LCR rats [85] , whereas SIRT3 levels in SIRT3/FOXO3A-PINK1-PARKIN axis [86] linked to higher mitophagy are high in HCR rats [87] .
Additionally, our findings are also supported by the observation that autophagy/mitophagy and mitochondrial fitness are better preserved in HCR rats compared to LCR rats [76] . An impaired mitophagy, especially in a context of pronounced energy deficits, such as in the LCR rat aortas, causes an accumulation of fragments of the mitochondrial genome, such as DAMPs, resulting in sterile inflammation [88] . In fact, higher DAMP levels in aging have been proposed as a potential mechanism for inflammaging [89] . As per conventional wisdom, various pathways of mitochondrial dysfunction converge to activate NLRP3 inflammasomes, but not other types of inflammasomes [90,91] . In contrast, our results indicate higher AIM2 protein levels and AIM2 inflammasome activation in aortic medial VSMCs and isolated aortic VSMCs of old age LCR rats. Interestingly, AIM2 protein expression as well as AIM2 inflammasome activity were both increased in young LCR aortic VSMCs treated with mtDNA DAMPs in culture, supporting the notion that mtDNA DAMPs are inflammaging inducers.
Both rapamycin and MitoTEMPO inhibited AIM2 inflammasome activation but did not inhibit its expression. Rapamycin can reverse age-related vascular dysfunction and increase longevity in rodents [67] . The inhibitory effect of rapamycin [64] on AIM2 activation was attributed to their autophagy/mitophagypromoting actions. Interestingly, MitoTEMPO was able to inhibit mtDNA DAMP-induced AIM2 inflammasome activity and rescue mitophagy, perhaps by reducing mtROS levels. Additionally, both treatments enhanced the mitochondrial respiratory capacity of VSMCs from old-age LCR rats, which might contribute to better vascular function. Furthermore, the treatments inhibited mtDNA damage in aortic VSMCs in aging, possibly disrupting the nexus between DNA damage and inflammasome activation [92] . Rapamycin increased p202 levels, which may block DAMPs binding to AIM2 inflammasomes [68] . These findings are supported by the observation that rapamycin [66] and MitoTEMPO [93] inhibit the activity of AIM2 inflammasomes. We and others have shown that rapamycin and MitoTEMPO are beneficial against aortic stiffening in mouse models of oxidative stress [40,67] .
In summary, our data demonstrate that low aerobic capacity and aging cause mitochondrial dysfunction and increase mtROS levels, impair autophagy and mitophagy, resulting in increased inflammasome activation. Mitochondrial turnover in aging is intricately intertwined with inflammasome activation, affecting mitochondrial fitness, VSMC and endothelial function, and vascular health [ Figure 8]. Mitophagy induction and inflammasome inhibition, as well as the use of mitochondrial antioxidants, may prove to be helpful therapeutic strategies for enhancing vascular health in older individuals with a reduced aerobic capacity.

Authors' contributions
Designed the study and experiments: Canugovi C, Madamanchi NR Performed the experiments: Canugovi C, Stevenson MD, Vendrov AE, Lozhkin A Wrote the manuscript: Canugovi C, Britton SL, Koch LG, Runge MS, Madamanchi NR

Availability of data and materials
Data will be provided on request.

Financial support and sponsorship
None.

Conflicts of interest
Dr. Marschall Runge is a member of the Board of Directors at Eli Lilly and Company. Other authors declared that there are no conflicts of interest.

Ethical approval and consent to participate
All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Michigan (PRO00010178).

Consent for publication
Not applicable.