Lifelong SIRT-1 overexpression attenuates large artery stiffening with advancing age

Advanced age is accompanied by aortic stiffening that is associated with decreased vascular expression of sirtuin-1 (SIRT-1). Interventions that increase SIRT-1 expression also lower age-related aortic stiffness. Therefore, we sought to determine if lifelong SIRT-1 overexpression would attenuate age-related aortic stiffening. Aortic pulse wave velocity (PWV) was assessed from 3-24 months in SIRT-1 transgenic overexpressing (SIRTTG) and wild-type (WT) mice. To determine the role of aortic structural changes on aortic stiffening, histological assessment of aortic wall characteristics was performed. Across the age range (3-24 mo), PWV was 8-17% lower in SIRTTG vs. WT (P<0.05). Moreover, the slope of age-related aortic stiffening was lower in SIRTTG vs. WT (2.1±0.2 vs. 3.8±0.3 cm/sec/mo, respectively). Aortic elastin decreased with advancing age in WT (P<0.05 old vs. young WT), but was maintained in SIRTTG mice (P>0.05). There was an age-related increase in aortic collagen, advanced glycation end products, and calcification in WT (P<0.05 old vs. young WT). However, this did not occur in SIRTTG (P>0.05). These findings indicate that lifelong SIRT-1 overexpression attenuates age-related aortic stiffening. These functional data are complemented by histological assessment, demonstrating that the deleterious changes to the aortic wall that normally occur with advancing age are prevented in SIRTTG mice.


INTRODUCTION
Cardiovascular diseases (CVDs) are the leading cause of death in the United States, and there is a progressive increase in the prevalence of CVD and CVD-related death with advancing age [1]. Among the many deleterious changes to the arterial phenotype that occur with advancing age, the stiffening of the large elastic arteries (i.e., aorta and carotid arteries) is one of the most important, as arterial stiffness is present in several CVD states, such as hypertension, stroke, and left ventricular hypertrophy [2]. Moreover, aortic stiffening is associated with an increased risk of cardiovascular events [3]. Thus, identifying treatments that can prevent or reverse agerelated arterial stiffening is of paramount importance.
Recent findings indicate that carotid artery stiffness is lower in old (~20-22 mo) SIRT-1 transgenic overexpressing (SIRT TG ) mice, compared to age-matched wild-type (WT) mice [16], but we have observed a similar magnitude of difference in aortic stiffness between young SIRT TG and WT mice (unpublished observations). Thus, it is unclear whether lifelong SIRT-1 overexpression attenuates the rate of arterial stiffening with age or that these animals simply start with lower arterial stiffness in youth. More importantly, no study has comprehensively examined aortic stiffness and structure across the lifespan in SIRT TG mice. Accordingly, the primary aims of the present study were to measure aortic stiffness, as determined by pulse wave velocity (PWV), and structural changes at specific time points across the normal rodent lifespan (~24 months) in SIRT TG and WT mice in an effort to gain a more complete understanding of the role of SIRT-1 in vascular aging. We hypothesized that in youth, aortic stiffness would be lower in SIRT TG mice and that the rate of aortic stiffening would be attenuated in SIRT TG mice. Moreover, we hypothesized that there would be an absence of the deleterious aortic structural changes in SIRT TG mice that occur with advancing age.

Animal characteristics
Body mass in both young WT and SIRT TG mice were lower than their middle-age counterparts (P<0.05) (Table 1), however, body mass in old SIRT TG mice was similar to young (P>0.05), but lower than age-matched WT mice (P<0.05). In all age groups and in both genotypes, male mice had higher body mass than female mice (data not shown, P<0.05). Heart mass was greater in old WT and SIRT TG mice compared to young mice (P<0.05). Both liver and spleen masses were greater in old WT and SIRT TG mice compared to young mice (P<0.05). Lastly, perigonadal white adipose tissue (WAT) was unchanged across the lifespan and between genotypes (P>0.05).

Age-related aortic stiffening is attenuated with lifelong SIRT-1 overexpression
A total of 206 aortic PWV measurements (WT: 111 measurements; SIRT TG : 95 measurements) were performed in 93 mice (WT: 24 male and 25 female; SIRT TG : 19 male and 25 female). There were no sexrelated differences in PWV, as assessed by 3-factor ANOVA (age X genotype X sex). The slope of increase in aortic PWV with advancing age was greater in WT, compared to SIRT TG mice ( Figure 1B; 3.8±0.3 vs. 2.1±0.2 cm/sec, respectively, P<0.05). In each age group, aortic PWV was greater in WT, compared to SIRT TG mice ( Figure 1D; P<0.05). In WT mice, aortic PWV increases with age at most 3 mo increments ( Figure 1C; P<0.05). Whereas aortic PWV in SIRT TG mice increases from 3 to 9 mo old ( Figure 1D; P<0.05), but does not rise significantly until 24 mo old. Anesthetized heart rate was similar between WT and SIRT TG mice in all age groups ( Table 2; P>0.05).

Deleterious age-related structural changes to the aorta are prevented with lifelong SIRT-1 overexpression
In histological sections of thoracic aortas excised from young, middle-age, and old mice, lumen diameter was greater in both old WT and SIRT TG mice compared with their young and middle-age counterparts (Figure 2A; P<0.05). However, medial cross-sectional area ( Figure  2B) was greater in old WT, compared with young and middle-age WT, as well as old SIRT TG mice (P<0.05). After normalizing lumen area ( Figure 2C), there was a greater media-to-lumen ratio in old WT, compared with young WT and old SIRT TG mice (P<0.05). Conversely, there were no differences in medial cross-sectional area or media-to-lumen ratio between age groups in SIRT TG mice ( Figure 2B, 2C; P>0.05).
Elastin content was lower in old WT, compared with young and middle-age WT, as well as with old SIRT TG mice ( Figure 3A; P<0.05). Aortic collagen content was higher in old WT, compared with young WT, as well as old SIRT TG mice ( Figure 3B; P<0.05). Although there was no difference in aortic advanced glycation end products (AGEs) intensity in any age group between WT and SIRT TG mice, AGEs intensity was higher in old WT, compared with young WT mice ( Figure 3C; P<0.05). Calcified area was higher in middle-age and old WT, compared to young WT, as well as compared AGING

SIRT-1 overexpression augments aortic superoxide dismutase that remains elevated in advanced age
Aortic gene expression of superoxide dismutase (SOD) isoforms, SOD1, SOD2, and SOD3, trended toward being lower with advancing age, but were augmented by lifelong SIRT-1 overexpression ( Figure 5). Specifically, we observed lower SOD3 gene expression (P<0.05) and a trend toward lower SOD1 and SOD2 gene expression (P=0.07-0.11) in old compared to young WT mice. Whereas, gene expression of all three SOD isoforms was lower in old compared to young SIRT TG mice (P<0.05).
We also observed higher SOD1 and SOD2 in young and old SIRT TG mice compared with age-matched WT mice (P<0.05), while SOD3 was only higher in old SIRT TG mice compared with age-matched WT mice (P<0.05).

DISCUSSION
In the present study, we comprehensively examined aortic stiffness and structural changes across the normal rodent lifespan in WT and SIRT TG mice. In comparison to WT, we observed a lower rate of age-related aortic stiffening in SIRT TG mice. Additionally, the deleterious aortic structural changes that occurred in old WT mice, such as medial wall hypertrophy, reduced elastin, accumulation of collagen, AGEs, and aortic calcification were absent in old SIRT TG mice. These findings indicate that lifelong SIRT-1 overexpression attenuates aortic stiffening and prevents many of the deleterious structural changes to the aorta that normally occur with advancing age.

SIRT-1 overexpression and aortic stiffening
Aortic stiffening with advancing age accompanies decreased vascular SIRT-1 expression and activity [10]. Interestingly, elevations in SIRT-1 activity that are achieved via CR or dietary supplementation with NAD + intermediates are capable of decreasing aortic stiffness in AGING  AGING old mice [10][11][12]14], as well as in older humans [15]. In the present study, we measured aortic stiffness in WT and SIRT TG mice at ~3-month intervals from 3-24 mo old. In addition to having lower aortic stiffness in youth, the rate of aortic stiffening across the lifespan in SIRT TG mice was nearly half that of WT mice. Previously, whole-body SIRT-1 transgenic overexpression resulted in lower carotid artery stiffness in old mice, compared to age-matched WT mice [16]. However, the findings from that study provide only a snapshot of arterial stiffness at one time point in advanced age. Whereas in the present study, through our comprehensive examination aortic stiffness and structure across the lifespan, provides a more complete understanding on the role of SIRT-1 in vascular aging.
Despite the aforementioned benefits of SIRT-1 overexpression, we still observed an increase in aortic stiffness between some age groups in SIRT TG mice, although the majority of aortic stiffening in SIRT TG mice occurred during development. Increases in aortic stiffness during development are likely due to increases in body length, as we observed a similar increase in PWV during youth in WT mice. Thus, increases in aortic stiffness during development from 3-9 mo are likely trivial to vascular health. Still, age-related aortic stiffening in SIRT TG mice did occur at 21-24 mo, which could be due to declines in aortic SIRT-1 gene expression in advanced age. However, despite a decrease in aortic SIRT-1 expression in old SIRT TG mice, SIRT-1 expression remained at least 3-fold higher than that of age-matched WT mice, but was 50% that of young WT mice. Increases in SIRT-1 activity via CR or dietary supplementation with NAD + reverse aortic stiffness in old mice [10][11][12]14]. Considering the age-related decline in aortic SIRT-1 expression in old SIRT TG mice, it is possible that either of these interventions could prevent aortic stiffening that occurred beyond development in these mice. Although we cannot pinpoint a direct mechanism in the present study, it is clear that lifelong SIRT-1 overexpression results in a healthier vascular phenotype in advanced age. It is important to note that the benefits of increased SIRT-1 activation do not need to be lifelong in order to exert vasoprotective effects, as reductions in aortic stiffness in humans and mice can occur in the short-term by increasing SIRT-1 activity in advanced age [10][11][12][13][14]. Future studies are warranted to determine if lifelong maintenance of aortic SIRT-1 at youthful levels is capable of preventing the entirety of age-related aortic stiffening that occurs beyond development.

SIRT-1 overexpression and aortic structure
Unlike aortic stiffness, there was no difference in any aortic structural variable measured between young WT and SIRT TG mice. Although with advancing age both groups demonstrated an increase in aortic lumen diameter, only WT mice developed deleterious structural changes to their aortic wall. In WT mice, there was an increase in aortic medial cross-sectional area with advancing age, but this did not occur in SIRT TG mice. Because increased medial cross-sectional area may be a consequence of the normal age-related widening of the aortic lumen, we normalized medial cross-sectional area to lumen area. However, media-to-lumen ratio was also elevated in old WT compared to both young WT and old SIRT TG mice. Resistance to medial hypertrophy /hyperplasia has also been demonstrated in cultured rat aortic vascular smooth cells that overexpress SIRT-1 [17]. Thus, in the context of aging, it is likely that SIRT-1 overexpression provides a similar resistance to agerelated medial hypertrophy/hyperplasia. Although lumen diameter is greater in advancing age in both SIRT TG and AGING WT mice, the increased lumen diameter in SIRT TG mice does not appear to be deleterious because medial hypertrophy does not accompany it.
In addition to aortic wall dimensions, we also examined the structural components of the aorta. In WT mice, there was an age-related reduction in aortic elastin content with a corresponding increase in collagen content. Unlike WT mice, aortic elastin and collagen content were unchanged with advancing age in SIRT TG mice. Changes in elastin and collagen content are a common alteration in aged arteries [18]. The lack of change in aortic elastin and collagen content with advancing age indicates that SIRT-1 overexpression provides resistance to these deleterious structural changes of aging. It is possible that SIRT-1 exerts these beneficial effects by direct acetylation of collagen and/or elastin, but to the best of our knowledge, this remains unknown. Indeed, SIRT-1 is a positive regulator of tissue inhibitor of matrix metalloproteinase-1 (TIMP-1) and a negative regulator of matrix metalloproteinase-9 (MMP-9) [19,20], which might provide insight in the ability of SIRT-1 to maintain aortic elastin content in advanced age [21,22]. Moreover, increased arterial SIRT-1 activity via dietary supplementation with NAD + intermediates that lowers age-related aortic stiffness also normalizes aortic elastin and collagen content in old mice [14]. The present study is supportive of previous work, indicating that elevated SIRT-1 expression attenuates aortic stiffening and changes in aortic elastin and collagen content in advanced age.
In addition to maintaining a youthful elastin and collagen content, accumulation of AGEs and calcified area in the aorta were also unchanged with advancing age in SIRT TG mice. Conversely, aortic AGEs and calcium content were both elevated with aging in old WT mice. Direct activation of SIRT-1 via the small molecule, SRT-1720, preserves glucose tolerance in aged rodents, preventing the accumulation of AGEs in SIRT TG mice via maintenance of normal glucose tolerance in advanced age [23]. In addition to preventing structural changes, maintaining a properly functioning vascular endothelium may also play a pivotal role. SIRT-1 plays a direct role in modulating nitric oxide (NO) and endothelial function via deacetylation and subsequent activation of endothelial NO synthase (eNOS) [24,25]. Moreover, transgenic whole-body SIRT-1 overexpression prevents the agerelated decline in endothelium-dependent vasodilation [16]. Thus, maintaining a properly functioning endothelium in the aged vasculature may also be a benefit of lifelong SIRT-1 overexpression that could impart vasoprotective benefits leading to attenuated aortic stiffness.
Lastly, as mentioned previously, it should be noted that there was a slight, but significant increase in aortic stiffness in old compared to young SIRT TG mice. Although old SIRT TG mice had elevated aortic stiffness, their PWV values were similar to that of young WT mice. Thus, these data indicate that the deleterious structural changes to the aorta we observed in older WT mice may have been driven by aortic stiffness after it surpassed a critical threshold in midlife or later. Notably, these structural changes to the aorta were prevented in SIRT TG mice by maintaining a youthful aortic stiffness across the lifespan. Taken together, these data indicate that lifelong SIRT-1 overexpression prevents deleterious structural changes to the aged aorta, possibly through attenuations in age-related aortic stiffening.

Experimental considerations
This study is not without limitations. It is well established that SIRT-1 activity increases to elevations in intracellular NAD+ that occur in response to energy/nutrient stresses, such as caloric restriction [26] and exercise [27]. Therefore, higher aortic SIRT-1 gene expression in SIRT TG permits a greater increase in SIRT-1 activity in response to a given change in NAD+ levels. However, we did not measure NAD+ or make indirect measures of SIRT-1 activation, thus, we are unable to confirm this assumption. Furthermore, we did not measure food consumption or spontaneous cage activity in this study. Therefore, it is possible that the attenuated slope of age-related aortic stiffening may have been due to lower food consumption and/or higher activity levels in SIRT TG mice. Still, we see no differences in body mass or heart and skeletal muscle masses normalized to body mass between genotypes, suggesting that it is unlikely there were differences in food consumption or activity level that contributed to a lower rate of aortic stiffening in SIRT TG mice.

Summary and future directions
These findings indicate that lifelong SIRT-1 overexpression results in lower aortic stiffness at any age across the lifespan, but also slows the rate of aortic stiffening occuring with advancing age. These functional data are complemented by histological assessement of aorta structural characteristics, demonstrating that the deleterious changes to the aorta that normally occur with advancing age, such as medial wall hypertrophy, reduced elastin, accumulation of collagen and AGEs, and aortic calficification are all prevented in mice that are afforded lifelong SIRT-1 overexpression. Although it is unknown if SIRT-1 directly effects any of these processes, we also demonstrated elevated aortic SOD1, SOD2, and SOD3 gene expression in SIRT TG mice that suggests SIRT-1 augments antioxidant capacity, which likely attenuates fibrosis, indirectly influencing these processes. Future studies are warranted to determine specific mechanisms by which SIRT-1 exerts its anti-stiffening effects and the use of tissue-specific models of SIRT-1 overexpression or deletion may provide greater mechanistic insight into that area. In summary, these findings indicate that lifelong SIRT-1 overexpression negates many of the deleterious alterations to the aorta that occur with advancing age. Thus, the use of SIRT-1 activators to prevent or reverse the age-related aortic stiffness may be a more viable approach compared to other lifestyle interventions, such as CR.

Animals
Male and female SIRT TG and WT littermate control mice on a C57BL/6 background were generated from breeding colonies at the Veteran's Affairs Medical Center-Salt Lake City (VAMC-SLC) [29]. Animals used in this study were housed in the animal care facility at the VAMC-SLC on a 12:

Aortic stiffness
Aortic PWV was determined in vivo at ~3-month increments over the normal lifespan of C57BL/6 mice (~24 months; 206 measurements in total). Briefly, mice were anesthetized with isoflurane (2-3%) in 100% oxygen at 2 L/min flow rate and placed in the supine position on a heated platform (37° C). Blood velocity waveforms at the transverse aortic arch and at the abdominal aorta were measured simultaneously with two 20-MHz Doppler probes (Indus Instruments, Webster, TX, USA) and recorded using WinDAQ Pro + software (DataQ Instruments, Akron, OH, USA). After blood velocities were collected, a precise measurement of the traveled distance between the Doppler probes was recorded using a scientific caliper. The transit time between Doppler sites was determined using the footto-foot method with WinDAQ Waveform Browser (DataQ Instruments). Aortic PWV was calculated as the traveled distance divided by the transit time.

Aortic histology
Young (5-7 months), middle-age (12-13 months), and old (23-24 months) SIRT TG and WT mice were sacrificed by exsanguination via cardiac puncture under isoflurane anesthesia. Thoracic aortas were quickly excised and placed in cold (4°C) physiological salt solution. For each mouse, an 2-3 mm aortic ring with perivascular tissue intact was excised from the thoracic aorta and embedded in Optimal Cutting Temperature medium. Aortic rings were sliced into 8-micron sections. Each mouse aorta had 3 to 4 sections per slide, which were averaged. For measures of medial crosssectional area the lumen border and the outer medial border were traced in ImageJ and internal areas measured. These areas were used to calculate medial cross-sectional area and were calculated as the outer media border area minus the lumen area. Elastin was quantified by Verhoff's Van Geison stain as percentage of the selected area, as described previously [30]. An 8bit grayscale was used for densitometric quantification of elastin content with ImageJ. Collagen was quantified by picrosirius red stain as percentage of the selected area, as described previously [10]. Green channel images from an RGB stack were used for densitometric quantification of collagen content with ImageJ (NIH, Bethesda, MA). AGEs were assessed by immunohistochemical visualization, as described previously [31]. Briefly, sections were washed and incubated in primary antibody (1:200, GeneTex 20055) or negative control (2.5% horse serum, Vector Labs) overnight. AGEs were visualized using the appropriate secondary antibody and Vector Labs NovaRed (SK-4800) Peroxidase substrate kit. Three separate, blinded observers (DRM, YA, VRG) scored images on a zero to three scale (0 = absence of appreciable positive stain, 1 = minimal positive stain, 2 = appreciable positive stain, 3 = highly positive stain). Scores for each section were averaged across observers and normalized to negative control sections. Calcium deposition was investigated AGING using Von Kossa staining, as previously described [32]. Briefly, sections were fixed with acetone (-20°C) for 10 min, washed 3 times with distilled water, incubated with 1% silver nitrate solution, and exposed to ultraviolet light for 20 min. After a final wash and removal of unreacted silver with 5% sodium thiosulfate, sections were dehydrated with ethanol and xylene. Calcium particles were observed in visual fields at 10x magnification. The quantification of von Kossa staining was performed as described previously [33]. The surface of the aorta and calcification modules were manually measured with Image J. The calcified area was defined as calcified surface/tunica media (pixels/pixels).

Statistics
Statistics were performed using SPSS (IBM, Chicago, IL). A 2-and 3-factor ANOVA was employed to evaluate differences between age (grouped at 3-month increments) and genotype (age X genotype), as well as between age, genotype, and sex (age X genotype X sex). Although there were differences in body mass and some tissue masses between male and female mice (data not shown), these data are not discussed because we observed no sex-related differences in our primary outcome, PWV. When a significant ANOVA was present, a least significant difference t-test was conducted to determine differences between values. Due to the number of comparisons, a Bonferroni correction was applied to within and between group comparisons for aortic PWV and heart rate data. Multiple linear regression was performed to model the relationships of genotype and/or sex with the age-related aortic stiffening. Statistical significance was set at P<0.05 for all analyses. Data are presented as mean±SEM.

Ethical approval
All animal procedures conformed to the Guide to the Care and Use of Laboratory Animals: Eighth Edition [28] and were approved by the University of Utah and Veteran's Affairs Medical Center-Salt Lake City Animal Care and Use Committees.

AUTHOR CONTRIBUTIONS
DR Machin, LA Lesniewski, and AJ Donato contributed to all aspects of the study; including the conception and design, data collection and analysis, and manuscript preparation. Y Auduong, VR Gogulamudi, Y Liu, and MT Islam contributed to the collection and analysis of data and revision of the manuscript.