Abstract

   The free radical theory of aging suggests that oxidative stress from
   free radicals contributes to aging. While free radicals cause DNA
   damage and cellular dysfunction, they also regulate essential signaling
   pathways. This duality complicates direct testing of the theory in
   mice, as demonstrated by the minor lifespan impact of Sod1 knockout. We
   hypothesize that base excision repair mechanisms, that involve SIRT2, a
   longevity-associated SIRTUIN family member, may mitigate excessive free
   radical effects. Our study found that Sirt2^−/−Sod1^−/− double-knockout
   mice exhibited significantly reduced lifespan and progeroid phenotypes,
   including spinal curvature and tissue degeneration. These mice
   displayed increased aging-related gene expression, cellular senescence,
   enlarged spleens, elevated cytokines, and immune dysregulation,
   potentially leading to cytokine storm-related deaths. Additionally,
   Sirt2 overexpression rescued genomic instability caused by Sod1
   deficiency in cells. These findings refine free radical theory of aging
   and highlight SIRT2 as a target for enhancing DNA repair and mitigating
   aging-associated phenotypes.

   Keywords: Genomic stability, Progeroid, Oxidative stress, Sirt2

1. Introduction

   In 1956, Denham Harman proposed the free radical theory of aging, which
   posits that the aging process is fundamentally driven by the
   accumulation of oxidative stress due to free radicals generated during
   aerobic respiration [[45]1]. This theory suggests that the intrinsic
   process of aging is caused by cumulative oxidative damage to cells.
   Numerous studies have indeed demonstrated that the levels of free
   radicals increase with age across different species as reviewed in
   Refs. [[46][2], [47][3], [48][4], [49][5]], potentially leading to DNA
   damage and genomic instability, which in turn can cause cellular
   senescence.

   However, the relationship between free radical scavengers and aging in
   mice is complex. For instance, mice lacking Nrf2, a key regulator of
   cellular antioxidant response and detoxification pathways, show little
   to no effect on aging phenotype median lifespan [[50][6], [51][7],
   [52][8]]. Similarly, mice deficient in Sod1, an enzyme crucial for
   scavenging superoxide radicals, only experience an approximately 30 %
   reduction in lifespan [[53]9]. In the case of catalase-deficient mice,
   premature aging phenotypes are observed, but these are attributed more
   to metabolic changes than to the accumulation of free radicals
   [[54]10]. These in vivo studies have sparked debate and further
   investigation into the role of free radicals in aging, suggesting that
   the relationship between free radical accumulation and aging may not be
   as straightforward as initially proposed by the free radical theory of
   aging. The complexity of free radicals in cellular biology is
   underscored by their dual role as integral components of cellular
   signaling pathways [[55][11], [56][12], [57][13]]. Excessive
   elimination of free radicals can lead to cellular dysfunction and the
   disruption of tissue homeostasis. Indeed, transgenic mice
   overexpressing Sod1 do not exhibit extended lifespans or changes in
   aging phenotypes; on the contrary, their lifespans are shortened
   [[58]14]. Another potential reason we do not observe the phenotype of
   severely accelerated aging due to free radical accumulation is that
   after inflicting oxidative damage to macromolecules such as DNA, cells
   possess a secondary defense mechanism involving base excision repair
   (BER) to maintain normal cellular functions [[59]15], thereby
   maintaining tissue homeostasis and preventing the emergence of aging
   phenotypes. Providing in vivo data on the accelerated aging of animals
   with simultaneous knockout of key genes involved in free radical
   scavenging, such as Sod1, and genes regulating BER pathway will help to
   substantiate the free radical theory of aging and lay the foundation
   for the development of new intervention strategies.

   The BER pathway is a vital mechanism for repairing minor DNA base
   damage induced by free radicals caused oxidative stress, chemical
   modifications, or spontaneous hydrolysis [[60]16]. The process begins
   with DNA glycosylases, such as OGG1 (8-oxoguanine glycosylase), which
   identifies and excises the damaged base [[61]17]. Following this, the
   enzyme APE1 (apurinic/apyrimidinic endonuclease) acts as an AP
   endonuclease, cutting the DNA backbone at the abasic site left by the
   glycosylase [[62]18]. Subsequently, the DNA polymerase β or other
   polymerases fill in the missing base, aligning with the undamaged
   complementary strand [[63]19,[64]20]. Lastly, the DNA ligase complex
   seals the remaining nick, thereby finalizing the repair and preserving
   the genome integrity, which is crucial for preventing mutations
   [[65][21], [66][22], [67][23]]. Knocking out critical BER factors
   including Ape1, Xrcc1 or DNA polymerase β all leads to early-stage
   embryonic or postnatal lethality [[68][24], [69][25], [70][26],
   [71][27]], indicating that it is probably not a feasible approach to
   simultaneously knock out Sod1 and these essential BER factors to
   validate the free radical theory of aging.

   Our earlier research has shown that SIRT2, one member of the Sirtuin
   protein family, modulates the BER pathway by enhancing the binding of
   OGG1 to its promoter, thereby upregulating its expression [[72]28].
   This regulatory mechanism is contingent upon stress, as the DNA damage
   sensors ATM and ATR trigger the role of SIRT2 role in BER. This
   indicates that SIRT2 is activated to modulate BER specifically under
   conditions of oxidative stress-induced DNA damage. Significantly, mice
   deficient in Sirt2 have not been reported to exhibit alterations in
   aging phenotypes, which provides us with the opportunity to investigate
   the combined impact of free radical accumulation and partial BER
   deficiency on aging phenotypes in mice.

   In this study, we found that mice lacking both Sirt2 and Sod1 had a
   reduced lifespan and showed severe premature aging features, including
   spinal curvature and skin ulcers. These mice had higher levels of
   aging-related gene expression and cellular senescence in various organs
   compared to controls. They also displayed signs of inflammation and
   immune cell activation, possibly leading to a cytokine storm and early
   death. Sirt2 overexpression mitigated the genomic instability and
   senescence effects caused by the absence of Sod1 in mouse cells. This
   indicates that the loss of both antioxidant defense and DNA repair
   accelerates aging, reinforcing the free radical theory of aging, and
   positions SIRT2 as a potential target for anti-aging interventions.

2. Results

2.1. The absence of Sirt2 exacerbates the aging phenotypes in Sod1 knockout
mice

   To further validate the free radical theory of aging, we generated
   double knockout (DKO, Sirt2^−/−Sod1^−/−) mice by crossing mice lacking
   Sirt2, which responds to oxidative damage and is involved in regulating
   BER [[73]28], with Sod1 knockout mice. Western blot analysis using the
   protein extracts from brains of different groups of mice confirmed the
   successful breeding ([74]Supplementary Fig. 1A). Using these DKO mice,
   along with separate cohorts of wild-type (WT), Sirt2 knockout
   (Sirt2^−/−), and Sod1 knockout (Sod1^−/−) mice, we conducted subsequent
   analysis on lifespan and aging-related phenotypes.

   Lifespan analysis revealed that DKO mice had significantly shortened
   median lifespan (∼65 weeks) in comparison to either Sirt2^−/− or
   Sod1^−/− mice ([75]Fig. 1A). Previous work has demonstrated that
   overexpressing Sirt2 had no impact on mouse lifespan or healthspan
   [[76]29], and our data also indicates that knocking out Sirt2 had no
   significant effect on mouse lifespan at least during the observation
   window of at least 85 weeks ([77]Fig. 1A), further suggesting that
   Sirt2 might not be the critical factor regulating lifespan in a
   pathogen- and stress-free breeding environment. Additionally, Sod1^−/−
   mice exhibited mild lifespan shortening by ∼30 % compared to WT mice
   [[78]9], and our data revealed that these mice had a median lifespan of
   over 83 weeks ([79]Fig. 1A). Our data reveal that DKO mice had
   significantly shortened lifespan by at least 21.7 % in comparison to
   Sod1^−/− mice ([80]Fig. 1A). In summary, these results reveal that the
   absence of Sirt2 exacerbates progeroid features and further reduces
   lifespan in Sod1-deficient mice, suggesting a synergistic detrimental
   impact on organismal health.

Fig. 1.

   [81]Fig. 1
   [82]Open in a new tab

   Exacerbated aging phenotypes in Sod1^−/−Sirt2^−/− mice in comparison to
   Sod1^−/− or Sirt2^−/− mice. (A) Kaplan-Meier survival curves of mice of
   the indicated genotypes over a 1.63-year period. n = 35–62 mice per
   genotype. (B) Representative X-ray images of 12–14-month-old mice are
   shown. White circle arc indicates curvature angle measurements. (C)
   Kyphosis analysis of angle measurements in B. n = 10 mice per genotype.
   (D) Body weight curves for WT, Sirt2^−/−, Sod1^−/− and DKO mice.
   n = 24–41 mice per genotype. (E) Subcutaneous (s.c.) fat thickness in
   12–14-month-old DKO mice compared to 12-14-month-old WT, Sirt2^−/−,
   Sod1^−/− mice. n = 7–9 mice per genotype. (F) Representative H and
   E-stained skin sections shown in E. (G) Analysis of epidermal thickness
   of 12–14-month-old mice. n = 4 male mice per genotype. (H)
   Representative H and E-stained of skin sections. (I) Kaplan–Meier
   curves depicting lesion-free survival of mice of the indicated
   genotypes. n = 30–48 mice per genotype. (J) Bliss independence analysis
   was performed to evaluate the potential synergistic effect between
   Sirt2 and Sod1 knockout on lesion development over time. (K) Liver
   tumor incidence of 12–14-month-old mice. n = 7–8 male mice per
   genotype. (L) Representative H and E-stained liver tumor sections.
   A.D.I, Data were calculated using log-rank test. C, E, G, Data are
   mean ± s.d. Two-tailed Student's t-test. K, Statistical significance
   was determined using a 2-tailed Fisher's exact test. J, Statistical
   significance was determined by permutation tests (1000 permutations).
   ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. ns, not
   significant.

   Next, we set out to determine whether Sirt2 deficiency could intensify
   the aging phenotypes associated with the absence of Sod1 using these
   mice. Kyphosis, one of the primary aging-related phenotypes in both
   mice and humans, is characterized by an increased anterior curvature of
   the thoracic spine [[83]30]. The analysis of kyphotic angle, which is
   the forward curvature of the spine and is used to quantify the degree
   of kyphosis, revealed that DKO mice had excessive curvature in
   comparison to WT, Sirt2^−/− or Sod1^−/− mice ([84]Fig. 1B and C). The
   body weight analysis demonstrated a significant decrease in DKO mice
   compared to WT, Sirt2^−/− or Sod1^−/− mice ([85]Fig. 1D). This could be
   partly attributed to the significant loss of subcutaneous fat in DKO
   mice ([86]Fig. 1E and F). A distinct aging phenotype observed in the
   epidermis is the pronounced thickening of the cornified layer, as
   supported by a previous study [[87]31]. In fact, the thickness of this
   layer was 3-fold greater in DKO mice compared to their wild type,
   Sirt2^−/−, and Sod1^−/− counterparts ([88]Fig. 1G and H).

   Prior observations indicate that Sod1^−/− mice are susceptible to skin
   lesions [[89]32]. Here we found that DKO mice exhibited early onset of
   mild dermatological pathologies ([90]Fig. 1I). To assess the
   synergistic impact of Sirt2 and Sod1 deficiencies on lesion
   progression, Bliss independence analysis was conducted at intervals
   from 100 to 400 days. Initial time points (100 and 200 days) showed
   negligible synergy (scores: 0.062 and 0.055, respectively), not
   reaching statistical significance (P > 0.05). However, a robust
   synergistic effect was detected at 300 days with a score of 0.360 (95 %
   CI: 0.157–0.549, P = 0.0419), intensifying further at 400 days with a
   score of 0.474 (95 % CI: 0.229–0.714, P = 0.0241). The lesion incidence
   in DKO mice at 400 days (85.4 %) significantly surpassed the Bliss
   model-predicted rate (38.0 %), based on individual knockout effects
   (Sirt2^−/−: 35.0 %, Sod1^−/−: 4.5 %). Fisher's exact tests confirmed a
   significant correlation between genotype and lesion development at
   later stages (P < 0.001 for both 300 and 400 days) ([91]Fig. 1J).
   Subsequent immunostaining assays utilizing an antibody against 8-oxoG
   disclosed heightened oxidative damage in DKO mice ([92]Supplementary
   Fig. 1B). These results underscore a synergistic exacerbation of lesion
   progression attributed to an accumulation of oxidative stress caused by
   compounded Sirt2 and Sod1 deficiencies.

   A previous report indicates that Sod1 deficiency leads to the
   development of liver cancer at late age of mice (∼20 months) [[93]9],
   which might attribute to persistent oxidative damage to DNA in livers.
   Given the fact that SIRT2 participates in the regulation of BER
   pathway, one would expect that knocking out Sirt2 would accelerate the
   hepatocarcinogenesis in mice lacking Sod1. Indeed, DKO mice had
   significant increased liver tumor incidence at age of ∼12–14 months
   ([94]Fig. 1K and L). Over 85 % of DKO mice developed liver tumor, while
   less than 20 % of Sod1^−/− mice had liver tumor and no
   hepatocarcinogenesis was observed in wild type or Sirt2^−/− mice
   ([95]Fig. 1K and L).

   Given the elevated hepatocarcinogenesis rate in DKO mice ([96]Fig. 1K
   and L), we proceeded to investigate potential shifts in transcriptomic
   profiles associated with liver cancer. We established an expression
   profile based on the top 10 upregulated homologous genes in the livers
   of DKO mice ([97]Supplementary Table 1), and compared their expression
   levels between human hepatocellular carcinoma and normal liver tissues
   in GEPIA2 database. Our analysis revealed a significant upregulation of
   the DKO-associated gene signature in human liver cancers relative to
   normal tissues ([98]Supplementary Fig. 1C). Furthermore, survival
   analysis demonstrated that an elevated expression profile of this gene
   signature correlates with poorer prognosis in liver cancer patients
   ([99]Supplementary Fig. 1D).

2.2. DKO mice exhibit greater transcriptomic alterations and an
aging-associated transcriptome profile

   To determine whether the transcriptome profiles reflect
   aging-associated changes, we performed RNA sequencing on the liver,
   skin, and spleen of female WT, Sirt2^−/−, Sod1^−/−, and DKO mice. Our
   transcriptomic analysis confirmed the genotype of different cohort of
   mice ([100]Supplementary Fig. 2A). Principal component analysis showed
   that the samples segregated primarily by tissue rather than genotype
   ([101]Supplementary Fig. 2B), while samples of different genotypes were
   clearly distinguishable within the same tissue ([102]Supplementary Fig.
   2C). First, we utilized two gene sets: pos-MLS, comprising genes
   positively correlated with rodent maximum lifespan, and neg-MLS, which
   contains genes negatively correlated with lifespan [[103]33]. The
   pos-MLS gene set was downregulated in the skin and liver of DKO mice,
   while, conversely, the neg-MLS gene set was significantly upregulated
   in the skin and spleen of DKO mice compared to Sirt2^−/− and Sod1^−/−
   mice ([104]Fig. 2A and B, and [105]S2D). Thus, the double knockout of
   Sirt2 and Sod1 results in activation of negative-longevity genes and
   suppression of positive-longevity genes.

Fig. 2.

   [106]Fig. 2
   [107]Open in a new tab

   DKO mice exhibit an aging-associated transcriptome signature and
   undergo more transcriptomic changes compared to Sirt2^−/− and Sod1^−/−
   mice. (A) GSEA analysis demonstrates significant downregulation of
   pos-MLS genes and upregulation of neg-MLS genes in the skin of DKO
   mice. NES, normalized enrichment score. Significant represents
   False-discovery rate (FDR)-adjusted P < 0.05. (B) GSEA enrichment plots
   illustrate enrichment of pos-MLS genes among upregulated genes and
   neg-MLS genes among downregulated genes in the skin of DKO mice. (C)
   The transcript length distribution of differentially expressed genes in
   DKO mice resembles that of aging mice. Statistical significance was
   assessed using a two-tailed Wilcoxon test. ns, not significant;
   ∗P < 0.05; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001. (D) The number of
   differentially expressed genes in DKO mice is significantly greater
   than that in Sirt2^−/− and Sod1^−/− mice. (E) Venn diagrams depict the
   overlap of differentially expressed genes in Sirt2^−/−, Sod1^−/− and
   DKO mice.

   Aged mice also display length-associated transcriptome imbalances
   [[108]34]. In DKO mice, the significantly upregulated transcripts were
   markedly shorter than the significantly downregulated transcripts
   across all the analyzed tissues. However, this transcriptome imbalance
   was not observed in Sirt2^−/− mice, and for Sod1^−/− mice, the
   difference was not observed in liver and skin, and mild in spleen
   ([109]Fig. 2C).

   In some organs, Sirt2^−/− or Sod1^−/− mice exhibited no significant
   transcriptome changes, whereas DKO mice displayed the highest number of
   differentially expressed genes across all examined organs. The number
   of differentially expressed genes in DKO mice exceeded the combined
   total of differentially expressed genes observed in Sirt2^−/− and
   Sod1^−/− mice ([110]Fig. 2D). Moreover, the differentially expressed
   genes in DKO mice overlapped substantially with those of Sirt2^−/− and
   Sod1^−/− mice, including most of their respective differentially
   expressed genes ([111]Fig. 2E).

   These results indicate that DKO mice exhibit an aging- and
   negative-longevity-associated transcriptome profile, which is
   extensively perturbed, in comparison to Sirt2^−/− or Sod1^−/− mice.

2.3. The simultaneous depletion of Sirt2 and Sod1 accelerates accumulation of
senescent cells in tissues and precipitates immune system dysregulation

   To quantify senescent cell burden, we performed the
   senescence-associated (β-galactosidase) SA-β-gal staining on tissue
   sections from DKO mice and their controls. Our results showed that the
   spleen, liver, and lung from DKO mice displayed signs of advanced aging
   compared to those from WT, Sirt2^−/−, and Sod1^−/− mice ([112]Fig.
   3A–F). Additionally, the expression of p21, also a marker for senescent
   cells [[113][35], [114][36], [115][37]], was increased in the skin and
   lung tissues of DKO mice ([116]Supplementary Fig. 3A–D), indicating a
   greater accumulation of senescent cells in these tissues from DKO mice.

Fig. 3.

   [117]Fig. 3
   [118]Open in a new tab

   The simultaneous depletion of Sirt2 and Sod1 accelerates accumulation
   of senescent cells in tissues and precipitates immune system
   dysregulation. (A–F) Senescence-associated β-galactosidase (SA-β-gal)
   staining analysis of spleens (A–B), livers (C–D), lungs (E–F) of
   12–14-month-old mice. Representative SA-β-gal images are shown in (A,
   C, E), and the analysis of SA-β-gal positive area are shown in (B, D,
   F). n = 4–7 mice per genotype. Scale bar: 100 μm. (G) Spleen weight in
   WT, Sirt2^−/−, Sod1^−/− and DKO mice. n = 10–13 mice per genotype. (H)
   Spleen/Body weight ratio in WT, Sirt2^−/−, Sod1^−/− and DKO mice.
   n = 10–13 mice per genotype. (I) Median fluorescence intensity (MFI)
   was determined by flow cytometry to evaluate the level of cytokines in
   skin tissues of 12–14-month-old mice. n = 3 mice per genotype. B, D, F,
   Data are mean ± s.e.m. Mann-Whitney U test. G, H, I, Data are
   mean ± s.d. Two-tailed Student's t-test. ∗P < 0.05; ∗∗∗P < 0.001;
   ∗∗∗∗P < 0.0001.

   We extended our investigation to determine if cells deficient in both
   genes would exhibit increased genomic instability. Despite our efforts,
   we were unable to successfully isolate DKO mouse fibroblasts, likely
   due to the vulnerability of DKO cells during in vitro culture. As an
   alternative approach, we isolated MEFs from WT mice and utilized siRNAs
   to knock down Sirt2 and/or Sod1. We observed increased genomic
   instability, as measured by both comet assay and immunostaining with an
   anti-γH2AX antibody ([119]Supplementary Fig. 3E–J), when compared to
   control cells or those with single gene knockdown of Sirt2 or Sod1.
   Additionally, knocking down Sirt2 and Sod1 using siRNA greatly
   decreased the expression of Ki67, whose absence is also a molecular
   marker for cellular senescence ([120]Supplementary Fig. 3K–L).
   Simultaneously knocking down Sirt2 and Sod1 increased the expression of
   p21 in the presence of H[2]O[2] treatment in mouse cells
   ([121]Supplementary Fig. 3M). These in vitro findings suggest that the
   absence of both Sirt2 and Sod1 contributes to enhanced DNA damage,
   genomic instability and cellular senescence at cellular level.

   When analyzing the aging status of different tissues, we surprised
   observed enlarged spleens in DKO mice ([122]Fig. 3G), and statistical
   analysis indicates this enlargement was significant ([123]Fig. 3H).
   Since spleens are very critical immune organs, we therefore further
   examined cytokine profiling. Examination of key cytokines by FACS
   revealed that five cytokines up-regulated in DKO mice ([124]Fig. 3I),
   further highlighting the radical changes in the matrix microenvironment
   in the remodeled spleen.

2.4. Transcriptome analysis suggests increased inflammation in DKO mice

   Since we observed the enlarged phenotype of the immune organ spleen, we
   speculated that there might be changes in immune-related pathways in
   DKO mice. We therefore conducted Gene Ontology (GO) term biological
   process and Reactome pathway enrichment analysis on the genes
   specifically upregulated in the Sirt2^−/−, Sod1^−/−, DKO mice, and
   commonly upregulated in them across all three organs. The results
   indicated a significant activation of inflammation-associated pathways
   in DKO mice, including enhanced activation and migration of myeloid
   cells and leukocytes, neutrophil activation and degranulation,
   regulation of Toll-like receptor (TLR) cascades, acute inflammatory
   responses, and immune complex activation, compared to the WT,
   Sirt2^−/−, and Sod1^−/− mice. Furthermore, hemostasis, complement
   cascade activation and fibrin clot formation were upregulated in DKO
   mice, indicating increased tissue damage and bleeding, which likely
   contribute to the activation of these pathways. Additionally,
   senescence-associated pathways, such as the activation of matrix
   metalloproteinases, along with positive regulation of lipid metabolic
   process, were also upregulated in the spleen of DKO mice ([125]Fig.
   4A–C). Interestingly, in specific tissues such as the liver and skin,
   the genes uniquely upregulated in Sirt2^−/− or Sod1^−/− mice showed no
   enrichment in any particular pathway. This suggests that the
   transcriptome of DKO mice undergoes more extensive and systemic
   alterations ([126]Fig. 4A and B).

Fig. 4.

   [127]Fig. 4
   [128]Open in a new tab

   The transcriptome of DKO mice shows increased inflammation. (A) Pathway
   enrichment analysis of genes specifically upregulated in the liver of
   Sod1^−/− mice, specifically upregulated in the liver of DKO mice, and
   commonly upregulated in both. (B) Pathway enrichment analysis of genes
   that were specifically upregulated in the skin of Sirt2^−/− mice,
   specifically upregulated in the skin of DKO mice, and commonly
   upregulated in both. (C) Pathway enrichment analysis of genes
   specifically upregulated in the spleen of Sod1^−/− mice, specifically
   upregulated in the spleen of DKO mice, and commonly upregulated in
   both. (D) Scatter plots illustrate that inflammatory-associated genes
   are more highly upregulated in DKO mice compared to Sirt2^−/− and
   Sod1^−/− mice. Red dots indicate inflammatory-associated genes.

   Although certain immune-associated pathways were also upregulated in
   specific tissues of Sirt2^−/− or Sod1^−/− mice, we hypothesized that
   the inflammatory response is broader and more pronounced in DKO mice
   compared to Sirt2^−/− and Sod1^−/− mice. To further clarify this
   observation, we compared the fold change of commonly upregulated genes
   between DKO and Sirt2^−/− or Sod1^−/− mice. Immune-associated genes
   exhibited greater upregulation in DKO mice ([129]Fig. 4D). In contrast,
   the pathways enriched by downregulated genes did not display unique
   commonalities specific to DKO mice ([130]Fig. S4). In DKO mice, beyond
   the commonly upregulated genes, a substantial number of genes uniquely
   upregulated in DKO mice were also involved in immune-related pathways
   ([131]Fig. 4A–C). These findings reveal a systemic activation of the
   immune profile in DKO samples, reflecting a coordinated response likely
   driven by the synergistic disruption of Sod1 and Sirt2.

2.5. The DKO mouse transcriptome exhibits changes in the immune
microenvironment

   We identified 26 commonly upregulated genes across all organs,
   constituting the core signature of DKO mice ([132]Fig. 5A, [133]S5A-B).
   We also confirmed the expression by quantitative PCR
   ([134]Supplementary Fig. 5C). Interestingly, S100a8 and S100a9, which
   have recently been identified as novel aging biomarkers, were among
   these genes [[135]38]. GO term biological process and Reactome pathway
   enrichment analysis revealed that these genes are primarily enriched in
   inflammatory pathways ([136]Fig. 5B). Specifically, pathways such as
   neutrophil degranulation, migration, and neutrophil chemotaxis were
   highlighted, indicating activation and directed movement of
   neutrophils, key components of innate immunity. The enrichment of
   granulocyte migration and granulocyte chemotaxis pathways further
   underscores the increased mobilization of immune cells involved in
   inflammatory and immune responses. Additionally, the activation of RHO
   GTPases and NADPH oxidases, which are associated with the production of
   reactive oxygen species (ROS) and cytoskeletal rearrangements, was
   observed. These processes are critical for immune cell movement and
   function, particularly for macrophage activity [[137][39], [138][40],
   [139][41]]. Collectively, these findings underscore a coordinated and
   systemic activation of specific immune cell responses that are unique
   to the DKO condition across multiple tissues.

Fig. 5.

   [140]Fig. 5
   [141]Open in a new tab

   DKO mice transcriptome reveal the immune microenvironment changes. (A)
   Common up-regulated genes across all three tissues in DKO mice. (B)
   Pathway enrichment analysis of common up-regulated genes. (C)
   mMCP-counter scores of all samples demonstrate common changes in the
   immune microenvironment of DKO mice. (D) GSEA analysis demonstrates
   significant upregulation of core inflammation program gene set of
   neutrophils in the liver and skin of DKO mice. NES, normalized
   enrichment score.

   To further examine immune composition changes, we utilized
   mMCP-counter, a tool for estimating the immune cell composition in
   murine heterogeneous tissue using transcriptomic data, to compare
   Sirt2^−/−, Sod1^−/− and DKO mice. The results demonstrated an increase
   in neutrophils and macrophages in DKO mice ([142]Fig. 5C). Recent
   research has found that CD14^+ macrophages and neutrophils are closely
   associated with the induction and progression of cytokine storms driven
   by TNF responses in a mouse sepsis model, significantly contributing to
   lethality [[143]42]. We observed a significant upregulation of Cd14 in
   the liver and skin of DKO mice ([144]Supplementary Fig. 5D).
   Additionally, the core inflammation gene set associated with activated
   neutrophils [[145]43] was also upregulated in the liver and skin of DKO
   mice ([146]Fig. 5D). These findings suggest that the absence of Sirt2
   in Sod1 knockout mice may alter the immune microenvironment,
   contributing to the aging phenotypes and shortened lifespan of DKO
   mice. This effect is likely driven by secreted immune factors and
   chronic inflammation induced by activated neutrophils and macrophages.

2.6. Overexpression of Sirt2 rescues Sod1-deficiency mediated rise in genomic
instability in MEFs

   To test whether Sirt2 rescues Sod1-deficiency mediated rise in genomic
   instability, we performed comet assay and immunostaining with an
   anti-γH2AX antibody to analyze the changes in genomic stability by
   overexpressing Sirt2 in Sod1-depleted MEFs. As expected, we observed
   that depleting Sod1 using siRNA impaired genomic stability ([147]Fig.
   6A–D). Overexpressing Sirt2 significantly rescued the Sod1-depletion
   mediated rise in genomic instability ([148]Fig. 6A–D). Then, we
   determined whether Sirt2 rescues Sod1-deficiency mediated rise in
   cellular senescence, we examined the level of p21, pRPS6, Ki67 and the
   activity of SA-β-gal in mouse cells. We found that knocking down Sod1
   using siRNA greatly promoted the expression of p21 in the absence or
   presence of X-ray irradiation or H[2]O[2] treatment cells, while
   overexpressing Sirt2 attenuated Sod1-deficiency induced increase in p21
   level ([149]Fig. 6E, [150]Supplementary Fig. 6A–C). Overexpressing
   Sirt2 also inhibited the SA-β-galactosidase activity, increased the
   level of pRPS6 and promoted the expression of Ki67 in H[2]O[2] induced
   senescent MEFs ([151]Fig. 6F–H, [152]Supplementary Fig. 6D–E). To
   delineate the transcriptional impact of Sirt2 on Sod1-deficient MEFs,
   we performed bulk RNA-seq on Sod1-deficient MEF cells with and without
   Sirt2 overexpression. GSEA analysis against the complete Reactome and
   GO biological process libraries revealed that Sirt2 overexpression
   significantly attenuated immune- and senescence-associated pathways
   ([153]Fig. 6I, [154]Supplementary Fig. 6F). In particular, pathways
   including neutrophil degranulation, myeloid leukocyte activation, acute
   inflammatory response, and degradation of the extracellular matrix,
   which were significantly enriched in Sod1^−/− or DKO mice, were
   markedly reduced in Sod1-deficient MEFs after Sirt2 overexpression
   ([155]Fig. 6J). These findings indicate that targeting SIRT2 could be a
   promising strategy to counteract DNA damage, inflammation and cellular
   senescence triggered by free radicals.

Fig. 6.

   [156]Fig. 6
   [157]Open in a new tab

   Overexpression of Sirt2 rescues Sod1-deficiency mediated-rise in
   genomic instability in MEFs. (A) The genomic instability analysis in
   control MEFs, Sod1-depleted MEFs with or without Sirt2 overexpressed
   using the comet assay. The cells were first transfected with siSod1.
   Forty-eight hours later, the siSod1 and Sirt2 were co-transfected into
   cells, and 16 h post-transfection, the cells were collected for genomic
   instability analysis. The tail moment was employed as the measure of
   genomic instability, and at least 100 cells were analyzed using the
   software CometScore. (B) Representative images of the comet assay. (C)
   Immunofluorescence analysis of γH2AX foci in control MEFs,
   Sod1-depleted MEFs with or without Sirt2 overexpressed. The MEFs were
   first transfected with siSod1. Forty-eight hours later, the siSod1 and
   Sirt2 were co-transfected into MEFs, and 16 h post-transfection, the
   MEFs were collected for Immunofluorescence analysis. (D) Representative
   images of γH2AX foci. Scale bars, 10 μm. (E) Western blot analysis of
   p21 in control MEFs, Sod1-depleted MEFs with or without Sirt2
   overexpressed upon H[2]O[2] treatment. The MEFs were first transfected
   with siSod1. Forty-eight hours later, the siSod1 and Sirt2 were
   co-transfected into MEFs, and 16 h post-transfection, different groups
   of MEFs were treated with H[2]O[2] at a dosage of 200 μM. 3 days later,
   cells were harvested for protein extraction and Western blot analysis.
   (F) Senescence-associated β-galactosidase (SA-β-gal) staining analysis
   in control MEFs, Sod1-depleted MEFs with or without Sirt2
   overexpressed. Different groups of MEFs were treated with H[2]O[2] at a
   dosage of 200 μM. 10 days later cells were harvested for SA-β-gal
   staining analysis. (G) Representative images of the SA-β-gal images.
   Scale bars, 50 μm. (H) Representative images of the Immunofluorescence
   analysis. Immunofluorescence analysis of the Ki67 positive cell
   analysis in control MEFs, Sod1-depleted MEFs with or without Sirt2
   overexpressed. Different groups of MEFs were treated with H[2]O[2] at a
   dosage of 200 μM. 5 days later cells were harvested for
   Immunofluorescence analysis. Scale bars, 25 μm. (I) GSEA analysis of
   the full GO biological process collections, comparing Sod1-deficient
   MEFs overexpressing Sirt2 with Sod1-deficient MEFs. (J) Representative
   enrichment plots for four pathways that are strongly up-regulated in
   Sod1^−/− or DKO mice—neutrophil degranulation, myeloid leukocyte
   activation, acute inflammatory response, and degradation of the
   extracellular matrix—illustrating their attenuation in Sirt2
   overexpressing Sod1-deficient MEFs. NES, normalized enrichment score.
   A, C, F, Data are mean ± s.e.m. Mann-Whitney U test. ∗P < 0.05,
   ∗∗P < 0.01 and ∗∗∗∗P < 0.0001.

3. Discussion

   The free radical theory of aging, which posits that the accumulation of
   ROS contributes to aging, has been a subject of debate in the
   scientific community. A primary contention is the lack of robust
   evidence demonstrating that ROS accumulation directly leads to
   premature aging. This may be attributed to the fact that cells possess
   highly efficient DNA repair mechanisms that prevent excessive
   accumulation of ROS. To further substantiate the role of ROS in aging,
   it is necessary to disable the DNA repair pathways, thereby allowing
   for the accumulation of ROS-induced DNA damage.

   In our study, we have provided novel evidence supporting the free
   radical theory of aging by showing that the knockout of Sirt2, a
   regulator of BER pathway, significantly accelerates aging in Sod1
   knockout mice. However, since SIRT2 is involved in the regulation of
   multiple pathways, to further confirm that the observed aging is indeed
   due to unrepaired ROS-induced DNA damage, we propose the targeted
   knockout of specific non-essential BER pathway genes, such as Ogg1, a
   DNA glycosylase with no apparent phenotype when knocked out in mice.
   Crossing these Ogg1 knockout mice with Sod1 knockout mice could provide
   additional evidence for the role of ROS in aging.

   Furthermore, Sod1 knockout mice exhibit mild premature aging phenotypes
   and mutations in the Sod1 are associated with certain neurodegenerative
   diseases [[158]44,[159]45]. Our study found overexpressing Sirt2
   rescued Sod1-deficiency mediated cellular senescence. Although cellular
   senescence can serve beneficial roles, such as tumor suppression and
   wound healing, chronic accumulation of senescent cells is widely
   recognized as a contributor to age-related functional decline [36, 37].
   Targeting the activation of Sirt2 expression, for instance, by
   generating a transgenic mouse overexpressing Sirt2 and crossing it with
   Sod1 knockout mice, could test whether this approach alleviates the
   premature aging and age-related neurodegenerative disease phenotypes
   observed in Sod1 knockout mice. This would offer new insights and
   strategies for the development of novel lifespan-extending drugs.

   Additionally, directly activating the BER pathway by overexpressing
   multiple factors within the pathway, such as Ape1, Xrcc1, and Pol β,
   and crossing these transgenic mice with Sod1 knockout mice, could
   explore whether activating the BER pathway can mitigate aging and
   related diseases caused by Sod1 deficiency. This approach would provide
   new perspectives and directions in the field of aging and ROS, as well
   as in the broader area of DNA repair.

   Our study analyzed the transcriptomic effects of Sirt2 and Sod1 DKO in
   mice, revealing more pronounced aging-like changes compared to single
   knockouts. DKO mice showed increased inflammatory responses, with
   upregulation of pro-inflammatory genes and pathways, aligning with the
   "inflammaging" concept [[160]46,[161]47]. A set of 26 upregulated
   genes, including S100a8 and S100a9 discovered by a previous report
   [[162]38], indicated potential as aging biomarkers and participants in
   inflammatory and age-related diseases. The activation of RHO GTPases
   and NADPH oxidases pointed to heightened ROS production and
   cytoskeletal rearrangements related pathways changes due to the loss of
   Sod1 and Sirt2. Immune cell analysis confirmed increased neutrophils
   and macrophages, consistent with the upregulated immune pathways,
   potentially driving tissue damage and aging. The simultaneous loss of
   Sirt2 and Sod1 in DKO mice accelerated age-associated immune changes,
   contributing to inflammation, tissue damage, and aging. These findings
   suggest that targeting oxidative stress and immune dynamics could
   combat age-related inflammation and its consequences.

4. Materials and methods

4.1. Mice

   Mice used in this study included WT, Sirt2^−/−, Sod1^−/−, and Sirt2^−/−
   Sod1^−/− mice in a genetic background of C57BL/6J. All mice were housed
   in SPF facilities of the Animal Resource Center at Tongji University.
   All procedures were approved by the Tongji University Animal Care and
   Use Committee (TJAB04022106). Sirt2 knockout mice (stock #012772) were
   obtained from JAX lab. Sod1 knockout mice were generated with
   conventional CRISPR-Cas9 technique. Sirt2^+/−, Sod1^+/− mice were
   crossed to obtain WT, Sirt2^−/−, Sod1^−/−, and Sirt2^−/− Sod1^−/− mice
   for further lifespan and other types of phenotypical analysis in this
   study. Genomic DNA was extracted from mouse tails using lysis buffer
   (50 mM pH 8.0 Tris–HCl, 100 mM EDTA, 1 % SDS, 20 mg/ml Proteinase K)
   for genotyping. The primers for PCR involved in this process include
   the following: 5′-GACTGGAAGTGATCAAAGCTC-3′,
   5′-CAGGGTCTCACGAGTCTCATG-3′, 5′-TCA AAT CTGGCCAGAACTTATG-3’ (Sirt2);
   5′-AACTTTCTCAGTCCGCACGCT-3′, 5′-GCAGCAGCCCCAGAAGGATAA-3′,
   5′-ATGCTGGCCTTCAGTTAATCC-3’ (Sod1).

4.2. H&E staining, analysis of subcutaneous adipose tissue and epidermal
thickness

   The subcutaneous (s.c.) adipose tissue and epidermal thickness of WT,
   Sirt2^−/−, Sod1^−/−, and Sirt2^−/−Sod1^−/− mice was assessed using
   standard H&E staining on skin histology sections (Beyotime, C0105).
   Images were captured at 10 × magnification across the entire section
   using an OLYMPUS BX53 microscope. In these images, the thickness of the
   s.c. adipose tissue was measured at 12 distinct points per section
   using the AxioVision 4.8 software. The average of the 12 measurements
   was then calculated to determine the adipose thickness of each mouse.

4.3. Immunofluorescence staining

   Skin tissues were formalin-fixed and paraffin-embedded. Antigen
   retrieval was conducted in citrate buffer (pH = 6) for skin sections.
   Primary antibody staining was performed using anti-8-oxo-dG
   (4354-mc-050, Trevigen R&D), followed by counterstaining with DAPI.

   For cellular immunofluorescence, cells transfected with the indicated
   siRNA were plated on coverslips in 12-well plates. Twenty hours
   post-transfection, cells were rinsed with cold PBS and fixed with 4 %
   paraformaldehyde (PFA) for 15 min at room temperature. Then, the cells
   were permeabilized with 0.25 % Triton X-100 for 15 min and blocked with
   2 % bovine serum albumin (BSA) for 1 h at room temperature. Cells were
   then incubated with a γH2AX antibody overnight at 4 °C, followed by
   incubation with a secondary antibody for 1 h at room temperature in the
   dark. Cells were mounted in mounting medium containing DAPI, and images
   were acquired with a Nikon laser scanning confocal microscope.

4.4. SA β-gal staining

   Tissues were snap-frozen in OCT compound, fixed in a solution of 2 %
   formaldehyde and 0.2 % glutaraldehyde in PBS for 5 min at room
   temperature, and washed twice with PBS. Staining solution (1 mg/mL
   X-gal in dimethylformamide, 40 mM citric acid/sodium phosphate buffer,
   5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM sodium
   chloride, and 2 mM magnesium chloride) was applied, and the tissues
   were incubated for 24 h at 37 °C. Images of the stained tissues were
   taken using a bright-field microscope.

4.5. Protein preparation from tissues and Western blot

   Western blot analysis was performed on total protein extracts from
   brain, skin and lung tissues, which were homogenized in RIPA buffer
   (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 % NP-40, 1 % sodium
   deoxycholate, 0.1 % SDS, 1 mM PMSF, 5 mM sodium butyrate and 2 mM
   sodium vanadate) containing protease and phosphatase inhibitors.
   Lysates were mixed with 2 × sample buffer, boiled for 10 min, and
   subjected for Western blot analysis using antibodies including
   anti-Sirt2 (09843, Sigma-Aldrich), anti-Sod1 (A12537, Abconal),
   anti-p21 (ab109520, Abcam), anti-Tubulin (Cat# AP0064, Bioworld).

4.6. Alkaline comet assay

   Cells transfected with the indicated siRNA were seeded on coverslips in
   6-well plates. Twenty hours post-transfection, cells were collected,
   resuspended in PBS and diluted to 3 × 10^5 cells per ml before the
   comet assay was performed. The detailed procedure is described in the
   manufacturer's instructions (Trevigen, Cat. # 4250-050-K). Tail moments
   were used to quantify the amount of DNA damage using CometScore
   software (casplab_1.2.3b2).

4.7. Multi-parameter flow cytometry

   Total protein was extracted from skin tissues and stained with
   antibodies against mouse GM-CSF, IL-1β, IL-2, IL-10, CCL3 (RK04397,
   ABconal), and analyzed using Multi-index flow analyzer (ABclonal,
   ABplex-100).

4.8. RNA isolation and RNA sequencing

   Mice that showed any macroscopic lesions at necropsy on histological
   screening were excluded a priori. Total RNA was extracted from tissue
   from tumor-free samples using TRIzol® Reagent (Invitrogen, 15596026),
   following the manufacturer's protocol. Subsequent RNA purification,
   reverse transcription, library construction, and sequencing services
   were provided by Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. in
   Shanghai, China, adhering to Illumina's guidelines (Illumina, San
   Diego, CA).

   For RNA-seq, a transcriptome library was prepared using the Illumina®
   Stranded mRNA Prep, Ligation kit (Illumina, San Diego, CA), starting
   with 1 μg of total RNA. mRNA was enriched through polyA selection with
   oligo(dT) beads, followed by fragmentation using a fragmentation
   buffer. Double-stranded cDNA synthesis was carried out using the
   SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA) and
   random hexamer primers supplied by Illumina. The cDNA underwent
   end-repair, phosphorylation, and the addition of 'A' bases, following
   Illumina's library construction protocol.

   Libraries were size-selected to target cDNA fragments of 300 bp using
   2 % Low Range Ultra Agarose, and then amplified via PCR with Phusion
   DNA polymerase (NEB) for 15 cycles. Quantified using a Qubit 4.0
   fluorometer, the paired-end RNA-seq libraries were sequenced on a
   NovaSeq X Plus sequencer, generating 2 × 150bp read lengths. RNA-seq
   analysis was carried out using three biological replicates for each
   tissue.

4.9. RNA extraction and real-time quantitative PCR

   Total RNA was extracted using an RNAsimple Total RNA Kit (TIANGEN) and
   subsequently reverse transcribed into cDNA using the TransScript II
   Reverse Transcriptase kit (TRANS). Briefly, qRT–PCR of S100-a9, Prib,
   Ms4a8a and Ccr1 genes was performed. Real-time PCR was performed with
   FastStart DNA Master SYBR Green Mix (Roche, 4913914001) on a ViiA 7
   Real-Time PCR system (Applied Biosystems). The average threshold cycle
   (Ct) of quadruplicate reactions was determined, and expression was
   analyzed by the ΔΔCt method. The relative expression levels were
   normalized to the level of Gapdh. The primers used to amplify s100-a9
   were as follows: s100-a9-mRNA-F, 5′-GCCAACAAAGCACCTTCTCA-3′;
   s100-a9-mRNA-R, 5′-TGTCAGGGTGTCCTTCCTTC-3′; Prib-mRNA-F,
   5′-AACAATCAGGCTGCCGAATC-3′; Prib-mRNA-R, 5′-CTGGGAGAGAGGAGATGCAG-3′;
   Ms4a8a-mRNA-F, 5′-CGCCCAACAGTTATCCTGTG-3′; Ms4a8a-mRNA-R,
   5′-GGACTTGAGGCTGATTGCTG-3′; Ccr1-mRNA-F, 5′-ACCTGTAGCCCTCATTTCCC-3′;
   Ccr1-mRNA-R, 5′-CTCACTGGGTCTTCTGAGCA-3′.

4.10. RNA-seq analysis

   Total RNA was isolated from liver, skin, and spleen of 12–14-month-old
   mice. Quality control of RNA-seq data was performed using FastQC
   v0.11.9. Adapters and low-quality reads, and reads with length <20 bp
   were removed using Cutadapt v2.8. Quantification of samples was
   conducted at both gene and transcript levels against the mouse GRCm39
   cDNA reference using Salmon v1.10.3 with default parameters. Only genes
   with read counts >10 in at least 3 samples were retained for downstream
   analysis. Differentially expressed genes (DEGs) were identified using
   DESeq2 v1.16.1, with thresholds set at log[2] fold change >1 or < −1
   and adjusted p-value (padj) < 0.05. Genes meeting these criteria were
   defined as DEGs.

4.11. Gene set enrichment analysis (GSEA) and pathway over-representation
analysis

   GSEA was performed on a pre-ranked gene list, ranked by the formula:
   -log10 [padj] × (fold change)/abs(fold change), using fgsea v1.32.0
   with parameters minSize = 15 and maxSize = 500. Gene sets with a FDR
   <0.05 were considered significant. Enrichment analysis for GO
   Biological Processes and Reactome pathways was conducted using the R
   package clusterProfiler v4.12.2. P-values were adjusted using the
   Benjamini–Hochberg correction, with significance thresholds set at
   p-value <0.01 and adjusted p-value <0.05.

4.12. Immune microenvironment evaluation

   Changes in the immune microenvironment were assessed based on gene
   expression profiles of all samples using mMCP-counter, implemented
   through the R package immunedeconv v2.1.3.

4.12.1. Bioinformatics analysis limitations

   Several limitations should be acknowledged regarding our RNA-seq
   analysis approach. First, the relatively modest sample size may limit
   statistical power for detecting genes with smaller effect sizes,
   potentially leading to false negatives despite our stringent
   statistical thresholds. Second, while the Benjamini-Hochberg correction
   effectively controls the false discovery rate, this conservative
   approach may increase the risk of Type II errors, potentially
   overlooking biologically relevant genes with moderate expression
   changes. Third, our differential expression analysis relies on
   predefined fold change and adjusted p-value thresholds (abs(log[2] fold
   change) > 1 and padj <0.05), which, while commonly used, are somewhat
   arbitrary and may not capture all biologically meaningful changes.
   Fourth, gene set enrichment analyses are inherently dependent on the
   completeness and accuracy of existing pathway annotations, which may
   introduce bias toward well-characterized biological processes while
   potentially missing novel or poorly annotated pathways. Fifth,
   regarding immune microenvironment evaluation, the mMCP-counter
   algorithm infers cell-type proportions based on reference gene
   signatures derived primarily from younger, healthy mouse tissues. The
   aged mouse tissues analyzed here may deviate from these reference
   signatures, potentially reducing absolute accuracy. Consequently,
   results from this method were interpreted cautiously and primarily
   considered in relative terms rather than absolute cell-type
   proportions.

CRediT authorship contribution statement

   Anke Geng: Conceptualization, Data curation, Writing – original draft,
   Writing – review & editing. Xiaona Wang: Data curation, Formal
   analysis. Zhenkai Wu: Software. Zhihao Liu: Formal analysis,
   Methodology. Xiao Huang: Formal analysis. Xiyue Wang: Data curation.
   Xiaoxiang Sun: Data curation. Yingjie Wang: Data curation. Jiayu Chen:
   Resources. Ying Jiang: Conceptualization, Writing – original draft,
   Writing – review & editing. Huanyin Tang: Conceptualization, Data
   curation, Writing – original draft, Writing – review & editing. Zhiyong
   Mao: Conceptualization, Data curation, Writing – original draft,
   Writing – review & editing.

Data and materials availability

   All data supporting the conclusions in this manuscript can be found in
   the main text or the supplementary materials. The RNA-seq data
   generated in this study have been deposited in the National Genomics
   Data Center (NGDC) and can be accessed using the provided accession
   number CRA021511.

Funding

   This work was supported by the National Key R&D Program of China
   (2022YFA1103703 and 2021YFA1102003 to Z.M.), the National Natural
   Science Foundation of China (Grant Nos. 32171288 and 32471341 to Y.J.,
   32100605 to A.G.), and the China Postdoctoral Science Foundation
   (2023M732662 to H.T.).

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgments