Sirtuin 5 levels are limiting in preserving cardiac function and suppressing fibrosis in response to pressure overload | Scientific Reports -…

Generation and characterization of SIRT5-overexpressing mice

To investigate the effects of SIRT5 on cardiac function, we generated a transgenic SIRT5 overexpressing (SIRT5OE) mouse strain by inserting a LoxP-STOP-LoxP-SIRT5 cassette into the 3UTR of the Col1A1 locus, using a well-characterized transgene insertion system (Fig.1A)19. Constitutive, whole-body SIRT5 overexpression is driven by the CAGGS promoter (Supplemental Fig.1A)20. SIRT5OE mice were born at normal Mendelian and sex ratios, and grossly indistinguishable from their WT littermates, with no obvious differences in weight gain with age in either sex (Supplemental Fig.1B, Fig.1B). We tested the molecular effects of SIRT5 overexpression on bulk protein acylation in the heart and found modest decreases in total lysine succinylation (Ksucc) and lysine malonylation (Kmal), which were significant for Kmal (Fig.1C-E).

SIRT5OE mice exhibit decreased mildly decreased cardiac lysine malonylation. (A) Immunoblot analysis of SIRT5 expression in heart samples presented in c-e. Original immunoblot images are presented in Supplemental Fig.7A. (B) Weight (g) of WT and SIRT5OE male and female mice with age. (C-E) Representative immunoblot analysis of succinyl-lysine (Ksucc), and malonyl-lysine (Kmal) levels in WT and SIRTOE hearts, with quantification. Original immunoblot images are presented in Supplemental Fig.7B-C.

To evaluate sensitivity to cardiac stress, male WT and SIRT5OE littermates between 48months of age were randomly assigned to a sham or TAC operation (Fig.2A)21. TAC surgically narrows the transverse aorta, generating a pressure gradient that reflects the severity of pressure overload on the heart. Sham animals undergo an identical surgical procedure but without aortic banding. Echocardiograms (echos) to assess cardiac structure and function were performed prior to surgery, and then repeated at 4-weeks, at which point tissue was collected for downstream analyses (Fig.2A). WT and SIRT5OE sham animals did not exhibit any notable differences by echo, indicating that global SIRT5 overexpression does not alter cardiac structure or function under basal conditions (Table 1). Additionally, the magnitude of the aortic pressure gradient was comparable between WT and SIRT5OE TAC mice (Fig.2B). Thus, any differences observed between genotypes following TAC were not due to discrepancies in the degree of cardiac outflow obstruction. Pressure overload first induces concentric cardiac hypertrophy, a temporary compensatory mechanism to relieve stress on the heart22. Echo-based measurement of the interventricular septum (IVS) and the posterior wall (PW) thickness during both systole and diastole showed comparable LV hypertrophy at four weeks after TAC in all mice irrespective of genotype (Fig.2C-D, Table 1). Consistent with other studies, the TAC procedure itself was associated with a 1518% mortality rate, with 3 mice and 4 mice dying in the WT and SIRT5OE groups respectively (Supplemental Table 1). Therefore, SIRT5 overexpression does not affect cardiac concentric hypertrophy in response to TAC.

SIRT5OE mice are protected against TAC-induced heart failure. Echocardiography was performed on WT sham (n=16), WT TAC (n=12), SIRT5OE sham (n=10) and SIRT5OE TAC (n=14) mice to measure changes in cardiac function four weeks post-surgery. (A) Depiction of groups, procedures, and timeline of surgery. (B) Aortic pressure gradient in mice after TAC. Echo measurements for (C) systolic interventricular septum (IVS) thickness; (D) systolic posterior wall thickness (PWT); (E) LV end-diastolic diameter; (F) left ventricle mass normalized to body weight; (G) fractional shortening; (H) ejection fraction. (I) Quantification of CM cell area, normalized to WT sham four weeks after surgery [WT sham (n=4), WT TAC (n=4), SIRT5OE sham (n=4) and SIRT5OE TAC (n=10)]. (J) Representative wheat germ agglutinin-stained cardiac sections of the indicated genotypes and treatments four weeks post-surgery. Scale bar=50 um. (K) qRT-PCR for Acta1 (n=3 for all groups); Nppa and Myh6 [WT sham (n=5), WT TAC (n=5), SIRT5OE sham (n=3) and SIRT5OE TAC (n=6)]; and Myh7 [WT sham (n=5), WT TAC (n=4), SIRT5OE sham (n=3) and SIRT5OE TAC (n=5)] expression normalized to GAPDH. Statistical significance was determined using Students t-test for 2-group analysis or two-way ANOVA followed by Sidaks correction for multiple comparisons for 4-group analyses.

After prolonged pressure overload, concentric hypertrophy progresses to ventricular dilation and heart failure (HF)5. Four weeks after TAC, WT mice showed significantly increased LV diameter, indicating a transition from adaptive to maladaptive ventricular hypertrophy. In comparison, LV diameter did not significantly increase in SIRT5OE TAC mice (Fig.2E, Table 1). The combination of ventricular hypertrophy and dilation was also reflected in the LV size normalized to body weight. Both genotypes showed increased normalized LV mass after TAC; however, this increase was blunted by SIRT5 overexpression (Fig.2F). Ventricular dilation leads to reduced fractional shortening (FS) and impaired ejection fraction (EF), a measure of systolic function. Both were significantly reduced in response to chronic pressure overload in the WT mice but preserved in the SIRT5OE mice (Fig.2G-H). These patterns were also reflected in CM size, with WT TAC mice showing a significant increase in CM area compared to both WT sham and SIRT5OE TAC groups (Fig.2I-J).

To complement echo-based measurements, we assessed RNA levels of standard markers of cardiac hypertrophy and HF, skeletal muscle Actin (Acta1) and atrial natriuretic peptide (Nppa)23. As expected, expression of both genes increased in response to TAC, but this increase was blunted in SIRT5OE TAC mice compared to WT TAC (Fig.2K). The failing myocardium also shifts contractile protein expression from myosin heavy chain isoform (Myh6) to (Myh7) (Fig.2K)23. The decrease in Myh6 expression by TAC was attenuated in SIRT5OE mice compared to WT controls. Overall, based on these physiological and molecular assays, we conclude that SIRT5 overexpression protects against the transition from adaptive, concentric hypertrophy to maladaptive ventricular dilation and systolic dysfunction.

To understand how SIRT5 overexpression protects against TAC-induced stress, we performed RNA-seq transcriptomic profiling on whole hearts from WT sham, WT TAC, SIRT5OE sham, and SIRT5OE TAC mice four weeks after surgery. Principal component analysis (PCA) showed that SIRT5 overexpression did not appreciably alter baseline gene expression in the heart, as sham mice clustered together regardless of genotype. In contrast, TAC induced a marked transcriptional response in both genotypes. Apart from one animal, SIRT5OE TAC mice clustered between sham animals and WT TAC animals, suggesting that SIRT5OE blunts the overall TAC transcriptional response (Fig.3A). Hierarchical cluster analysis (HCA) of the top 30 genes ranked by variance generated a dendrogram with two major clades (Fig.3B). Sirt5 transcript levels were elevated in the SIRT5OE samples but were not altered by TAC in either genotype (Fig.3A, Supplemental Fig.2A-B). Apart from one animal, all SIRT5OE mice clustered together with WT sham mice in one clade. The second clade contained all WT TAC mice and one SIRT5OE TAC animal. Genes in signaling pathways important to heart failure were evident in this unbiased HCA clustering (Fig.3B). Consistent with the qRT-PCR data (Fig.2K), atrial natriuretic peptide (Nppa), skeletal muscle Actin (Acta1), and myosin heavy chain isoform (Myh6) increased with TAC but these increases were blunted in SIRT5OE TAC mice compared to WT TAC mice (Fig.3B). Fibroblast remodeling of the extracellular matrix (ECM) and increased fibrosis is a key component of maladaptive remodeling in HF. Postn, a key marker of cardiac fibroblast activation and differentiation, was strongly induced in WT TAC samples, but trended less so in the SIRT5OE TAC mice (Fig.3C)24,25. Markers of monocyte-derived macrophages that stimulate fibrosis, Spp1 and Thbs1, were also elevated in TAC samples, and this increase was significantly mitigated in the SIRT5OE mice (Fig.3C)26.Thus, based on both PCA and HCA, SIRT5OE TAC mice are generally more transcriptionally similar to WT sham mice than WT TAC mice.

Transcriptomic analysis of heart tissue four weeks after TAC identifies a dampened transcriptional response to pressure overload in SIRT5OE mice. (A) Principal component analysis (PCA) of RNA-sequencing data of the four groups of mice [WT sham (n=4), WT TAC (n=4), SIRT5OE sham (n=3) and SIRT5OE TAC (n=4)]. (B) Hierarchical clustering analysis (HCA) of the RNA-seq data based on the top 30 genes, determined by greatest variance from the mean. Genes with higher expression compared to the mean skew red; genes with lower expression compared to the mean skew blue. (C) Postn, Spp1, and Thbs1 mRNA levels measured by RNA-seq. (D) Venn Diagram comparing exclusive and overlapping differentially expressed genes in the WT (sham vs TAC) mice and the SIRT5OE (sham vs TAC) mice. (EG) Top five positively and negatively enriched Gene Ontology (GO) pathways from the indicated comparisons, determined by gene set enrichment analysis (GSEA). NES: normalized enrichment score. All listed pathways were significantly enriched with a false discovery rate (FDR)<0.0001.

We defined the criteria for differentially expressed genes (DEGs) between two groups as a false discovery rate (FDR)<0.05 and fold-change of |1|. Strikingly, there were only 8 DEGs, including SIRT5, between sham WT and SIRT5OE mice, consistent with the observation that SIRT5 overexpression does not significantly alter baseline cardiac parameters. Following TAC, the number of DEGs increased dramatically in both genotypes. There were 2,606 unique DEGs between WT sham and TAC hearts, but only 177 in SIRT5OE sham and SIRT5OE TAC. There were 870 DEGs in common between the WT and SIRT5OE sham to TAC comparisons, representing a core group of genes that respond to chronic pressure overload across both genotypes (Fig.3D). The lower number of DEGs in SIRT5OE mice after TAC suggests that transcriptional responses induced by pressure overload are partially mitigated by SIRT5 overexpression.

Unbiased analysis of gene signatures enriched by the DEGs of each group was performed using Ingenuity Pathway Analysis (IPA) and Gene Set Enrichment Analysis (GSEA)27,28. The top 5 most positively and negatively enriched Gene Ontology (GO) pathways based on normalized enrichment score (NES) were extracellular matrix (ECM) remodeling and metabolic pathways, respectively in sham to TAC comparisons of both genotypes, albeit to different magnitudes (Fig.3E-F). The most enriched GO terms between TAC samples revealed that SIRT5OE TAC hearts show reduced activation of the immune system and suppressed ECM organization while maintaining expression of respiratory genes compared to WT TAC hearts (Fig.3G). IPA analysis of these samples showed similar patterns of pathway enrichment (Supplemental Fig.2C-E). In summary, based on PCA, HCA, and function enrichment analysis, transcriptomic analysis supports the conclusion that SIRT5 overexpression protects against pressure overload-induced HF partially through the suppression of molecular signaling pathways responsible for HF progression.

Under normal conditions, the heart utilizes fatty acids as its main substrate for ATP generation. The failing heart increases glycolysis, without a concomitant increase in glucose oxidation, leading to overall decreased mitochondrial oxidative metabolism, resulting in an energy deficit29. Gene ontology analysis of the RNA-seq data revealed that WT TAC mice showed significant downregulation of cellular respiration and mitochondrial matrix composition pathways (Fig.3E). Similarly, SIRT5OE mice post-TAC also showed reduced expression of genes in various catabolic processes, including fatty acid and small molecule metabolite breakdown (Fig.3F). Comparison of the WT and SIRT5OE TAC transcriptomes revealed up-regulation of hallmark glycolytic genes in the WT TAC samples. Conversely, genes for fatty acid metabolism and components of the respiratory electron transport chain (ETC) were enriched in the SIRT5OE TAC samples (Fig.4A). Overall, SIRT5OE TAC hearts exhibits an intermediate metabolic state between healthy (sham) hearts and the failing, WT TAC hearts 4weeks after TAC.

WT TAC mice exhibit a metabolic shift by RNA-seq, but are functionally comparable to SIRT5OE TAC mice four weeks post-surgery. (A) GSEA of SIRT5OE TAC compared to WT TAC RNA-seq samples. NES: normalized enrichment score. FDR: false discovery rate. (B-C) Immunoblot analysis of electron transport chain subunits and quantification of CV-ATP5A. Original immunoblot images are presented in Supplemental Fig.8A. (D) Agilent/Seahorse analysis of electron flow through the ETC in mitochondria isolated from WT (n=5) and SIRT5OE (n=6) hearts 4weeks after TAC. y-axis indicates oxygen consumption rate (OCR) in pmol/min. (E) qPCR of mitochondrially encoded (mtDNA) genes 16S and ND1, normalized to HK2, encoded by the nuclear genome (nuDNA). (F) Sod2 mRNA levels measured by RNA-seq. (G-H) SOD2 immunoblot from heart lysates and quantification. Original immunoblot images are presented in Supplemental Fig.9A.

To investigate whether RNA expression differences in these metabolic pathways manifested at the protein level, we immunoblotted for components of the ETC, and found similar reductions in expression in both TAC groups (Fig.4B-C, Supplemental Fig.3A). We then directly assessed ETC activity using mitochondria isolated from hearts four weeks post-TAC using the Agilent/Seahorse XFe96 analyzer. Both groups exhibited similar rates of election flow (Fig.4D).

With comparable protein expression levels of ETC components, and electron flow, we then asked whether differences in total mitochondrial content might contribute to the discrepancies in LV function between the two TAC groups at four weeks post-surgery. Mitochondrial mass was quantified using two distinct approaches. First, qPCR was performed for mitochondrial protein-encoding genes 16S and Nd1, and normalized to Hk2. Then, each sample was normalized to WT sham (Fig.4E)30. Second, we assessed mRNA and protein levels of the mitochondrial protein SOD2 (Fig.4F-H). In both assays, TAC resulted in a notable decrease in mitochondrial content, particularly in WT mice. However, only modest differences were evident between genotypes in the TAC groups. From these data, overall we conclude that while WT and SIRT5OE TAC hearts show transcriptional differences related to electron transport and other metabolic pathways, these expression differences do not robustly manifest at the level of electron transport chain activity, levels of respiratory complexes proteins, or mitochondrial content.

To test whether the cardio-protective properties of SIRT5 overexpression might be associated with alterations in the metabolome, we performed unbiased metabolomics on cardiac samples across all groups at four weeks post-surgery using LC/MS-based mass spectrometry. PCA showed minimal variance between samples on the first two principal components (Fig.5A). TAC was the main influence on sample clustering, with minimal contribution by genotype. Analysis of metabolite differences between genotypes also supported this conclusion, with only 6 and 4 metabolites significantly altered in WT sham to SIRT5OE sham and WT TAC to SIRT5OE TAC comparisons, respectively. In contrast, the WT sham to TAC comparison showed the greatest number of significantly different metabolites (44), followed by the SIRT5OE sham to TAC comparison (25).

TAC alters the cardiac metabolomic landscape. (A) PCA of the metabolomics data from hearts four weeks post-surgery across all groups [WT sham (n=10), WT TAC (n=6), SIRT5OE sham (n=6) and SIRT5OE TAC (n=4)]. (B) Succinate levels. (C-D) Immunoblot analysis of Ksucc in sham and TAC mice, with quantification. Statistical significance was determined using two-way ANOVA followed by Sidaks correction for multiple comparisons for 4-group analyses. Original immunoblot images are presented in Supplemental Fig.10A.

Increased glucose usage and glycolytic flux are commonly observed during HF31. Levels of glycolytic metabolites did not show any significant changes between groups (Supplemental Fig.4A-B). In contrast, several TCA cycle metabolites were significantly affected by TAC (Supplemental Table 2, Supplemental Fig.4C-D). Succinate levels exhibited an interesting pattern, showing decreased levels with both TAC and SIRT5 overexpression (Fig.5B). Since succinate is a substrate of SDH, a known SIRT5 target, we directly measured SDH activity but found no differences between the groups (Supplemental Fig.5A). This result was consistent with electron flow data, where no differences in complex II/SDH activity were observed between mitochondria isolated from TAC hearts (Fig.4C). Succinate is also linked to SIRT5 through SIRT5s desuccinylase activity, as succinyl-CoA donates the succinyl group used to generate the succinyl-lysine modification (Supplemental Fig.5B). Ksucc levels were lower in both SIRT5OE and TAC samples (Fig.5C-D).

To complement these data, we obtained heart tissue samples from human HF patients. Similar to the mouse hearts, Ksucc levels trended lower in the human failing hearts compared to the control samples (p=0.07) (Supplemental Fig.5C-D). Therefore, cardiac stress leads to decreased levels of Ksucc, independent of SIRT5 genotype, in both mice and humans. Taken together, transcriptomic and metabolomic analyses reflect a primarily TAC-driven, rather than genotype-associated, shift in the metabolic landscape in both genotypes following TAC.

To further investigate the underpinnings of the improved response of SIRT5OE mice to TAC, we extracted a list of pathways IPA categorized as cardiovascular disease-related pathways (Fig.6A). These pathways play critical roles in the development and progression of pathologic hypertrophy and HF, and all were more significantly enriched in WT mice with TAC compared to SIRT5OE mice. Fibrosis-related pathways (i.e., Hepatic fibrosis signaling pathway, Apelin cardiac fibroblast signaling pathway, TGF- signaling) were readily apparent (Figs. 6A and 3E-G). Hepatic Fibrosis Signaling was the most significant pathway identified in the IPA analysis, and includes signaling occurring during fibroblast activation, an essential process during adaptive hypertrophy by which resident fibroblasts differentiate into myofibroblasts (Fig.6A). However, sustained myofibroblast activity results in excessive cardiac remodeling and is also a major contributor to LV dysfunction32.

SIRT5OE protects against TAC-induced fibrosis four weeks after surgery. (A) Heat map of gene ontology analysis of WT (sham vs TAC) and SIRT5OE (sham vs TAC) mapping to cardiovascular disease related pathways using IPA. Blue color scale represents -log(p-value) of each pathway. (B) GSEA of SIRT5OE TAC compared to WT TAC RNA-seq samples. NES: normalized enrichment score. FDR: false discovery rate. (C) Volcano plot highlighting significant genes in from the GO: collagen containing extracellular matrix gene list. (D) Quantification of the amount of fibrosis in each heart sample, expressed in percentage [WT sham (n=3), WT TAC (n=4), SIRT5OE sham (n=4) and SIRT5OE TAC (n=10)]. Two-way ANOVA interaction term: p-value=0.0012. (E) Representative images of heart sections stained with Picrosirius red for collagen. (F) Summary figure of effects of SIRT5 overexpression during pressure overload.

GSEA confirmed that WT TAC samples were significantly enriched with genes upregulated by TGF1, a cytokine central to the fibrotic activation response (Fig.6B). TGF1 triggers expression of a fibrotic gene program that stimulates proliferation, cell migration, and increased secretion various macromolecules that restructure the ECM and promote wound healing33. ECM-related genes were highly enriched in WT TAC samples, including ECM components (Col1a1, Col4a2, Fbn1), ECM modifying genes (Timp1, Adamts4, Anxa2, Plod3), and fibrotic signaling cytokines (Tgf1, Gdf15, Icam1) (Fig.6B-C)33,34. mRNA levels of these genes were significantly lower in the SIRT5OE hearts after TAC. Thus, SIRT5 overexpression suppresses pressure overload-induced fibrosis signaling. These differences in mRNA expression were also reflected in tissue histology. Staining for collagen using picrosirius red showed a dramatic increase in ECM accumulation in WT TAC but not SIRT5OE TAC tissue (Fig.6D-E). Thus, gene ontology and histologic analyses show that SIRT5 overexpression blunts TAC-induced cardiac fibrosis.

Read the original:
Sirtuin 5 levels are limiting in preserving cardiac function and suppressing fibrosis in response to pressure overload | Scientific Reports -...