The longevity effects of reduced IGF-1 signaling depend on the stability of the mitochondrial genome

Abstract

Suppression of insulin-like growth factor-1 (IGF-1) signaling extends mammalian lifespan and protects against a range of age-related diseases. Surprisingly though, we found that reduced IGF-1 signaling fails to extend the lifespan of mitochondrial mutator mice. Accordingly, most of the longevity pathways that are normally initiated by IGF-1 suppression were either blocked or blunted in the mutator mice. These observations suggest that the pro-longevity effects of IGF-1 suppression critically depend on the integrity of the mitochondrial genome and that mitochondrial mutations may impose a hard limit on mammalian lifespan. Together, these findings deepen our understanding of the interactions between the hallmarks of aging and underscore the need for interventions that preserve the integrity of the mitochondrial genome.

INTRODUCTION

Aging is a complex biological process that is characterized by multiple molecular and cellular hallmarks, including genomic instability, mitochondrial dysfunction and deregulated nutrient sensing¹. While significant progress has been made in our understanding of the individual hallmarks of human aging, their interactions remain poorly defined. In addition to obscuring our insight into the basic biology of aging, this gap in our knowledge complicates the development of effective anti-aging interventions, as targeting one hallmark of aging may inadvertently exacerbate others.

One strategy to elucidate the relationships between the hallmarks of aging is to investigate how the disruption of one hallmark affects the trajectory of another. In doing so, it may be possible to assess whether these processes act independently, synergistically, or in opposition of each other as they shape human lifespan. In addition, this strategy may reveal if a hierarchy exists between aging pathways, which could lead to a more integrated and causally ordered model of the aging process. In this study, we apply this strategy to investigate the relationship between two critical drivers of the aging process, mitochondrial mutagenesis² and IGF-1 signaling³.

A large body of evidence supports the idea that instability of the mitochondrial genome (mtDNA) leads to a progressive decline in mitochondrial function, which accelerates the natural aging process and contributes to a wide variety of age-related diseases, including sarcopenia, neurodegeneration and heart failure². A similar body of work describes the role of IGF-1 signaling in the aging process. IGF-1 regulates the growth and metabolism of human tissues, and reduced IGF-1 signaling can not only extend mammalian lifespan, but can also confer resistance against various age-related diseases, including neurodegeneration, metabolic decline, and cardiovascular disease³. However, how mitochondrial mutagenesis and IGF-1 signaling interact with each other to shape mammalian lifespan remains unclear.

It is likely though, that mitochondrial mutagenesis and IGF-1 signaling intersect at multiple decisive junctions during the aging process. At a molecular level, mtDNA mutations are a constant source of oxidative stress⁴, a form of cellular damage that is directly influenced by IGF-1 activity⁵. At a systemic level, mitochondria serve as central hubs of metabolic activity and respond dynamically to changes in nutrient availability and growth signals⁶, processes that are tightly regulated by IGF-1 signaling⁷. And finally, at a macroscopic level, mitochondrial mutations directly contribute to a variety of age-related diseases^2,8,9, many of which can be prevented, or slowed down by reduced IGF-1 signaling^10–14.

Together, these observations suggest that mitochondrial mutagenesis and IGF-1 signaling do not operate in isolation, but instead converge on overlapping biological pathways that shape the rate and quality of aging. If so, it may be possible to counteract the harmful effects of mtDNA mutations by suppressing IGF-1 signaling, which could improve antioxidant defenses⁵, enhance mitochondrial turnover¹⁵, and delay the onset of age-related diseases¹⁶. In support of this idea, we previously found that reduced insulin/IGF-1 signaling can alleviate muscle dysfunction in a C. elegans model of mtDNA disease¹⁷. Alternatively, mitochondrial mutagenesis and IGF-1 signaling may act through parallel, but mechanistically distinct pathways, in which case interventions targeting IGF-1 signaling would fail to overcome the progeroid effects of mtDNA instability.

To distinguish between these possibilities, we reduced IGF-1 signaling in mice that carry an error-prone version of DNA Polymerase γ¹⁸ (PolgA^D257A), the enzyme that replicates the mitochondrial genome. These mice accumulate mtDNA mutations at an accelerated pace^19,20, which reduces their lifespan by 40% and leads to a wide variety of age-associated phenotypes, including sarcopenia, cardiomyopathy, anemia, and inflammation^18,21. We crossed these mice into a background that is deficient for Pappalysin 1 (Pappa), a metalloproteinase that increases IGF-1 bioavailability by cleaving IGF-binding protein 4 (IGFBP-4)²². In the absence of PAPPA, IGFBP-4 persists and sequesters IGF-1 away from receptors, thereby reducing IGF-1 signaling. When Pappa is deleted in WT mice, they exhibit a 35% increase in lifespan and a reduction in age-related pathologies^23–25.

Surprisingly though, we found that Pappa deletion failed to extend the lifespan of PolgA^D257A mice, suggesting that the integrity of the mitochondrial genome is required for lifespan extension by reduced IGF-1 signaling. Moreover, many of the pro-longevity programs normally initiated by Pappa deletion, including improved proteostasis, enhanced DNA repair and better telomere maintenance failed to initiate in PolgA^D257A; Pappa^−/− mice. These observations suggest that mtDNA instability places a hard limit on mammalian lifespan, and that mitochondrial function is required for the successful activation of multiple longevity programs.

RESULTS

Lifespan extension by IGF-1 suppression depends on the integrity of the mitochondrial genome

To examine how IGF-1 signaling and mitochondrial mutagenesis interact to shape mammalian lifespan, we crossed PolgA^D257A mice into a Pappa^−/− background and compared the lifespan of WT mice to PolgA^D257A mice, Pappa^−/− mice, PolgA^D257A; Pappa^−/− mice, Pappa^+/− mice, and PolgA^D257A; Pappa^+/− mice. As expected, we found that PolgA^D257A mice exhibited a marked reduction in both median and maximum lifespan relative to WT controls, with all PolgA^D257A mice expiring before they reached 17 months of age (fig. 1A), a timepoint when all WT mice were still alive. In contrast, the first Pappa^−/− mouse did not expire until it reached 28 months of age, consistent with their extended lifespan¹². Surprisingly though, a complete deletion of the Pappa gene failed to extend the lifespan of PolgA^D257A mice (fig. 1A). Instead, PolgA^D257A; Pappa^−/− mice seemed to exhibit enhanced frailty, characterized by extensive weight loss as they aged (fig. 1B-D, for a weight distribution of young mice and their progression throughout their lifespan, see fig. S1). A heterozygous deletion of the Pappa gene failed to improve the lifespan of PolgA^D257A mice as well (fig. 1A), although this deletion was better tolerated, as indicated by their normalized body weight (fig. 1B,C). Taken together, these results indicate that the pro-longevity effects of IGF-1 suppression are contingent upon the integrity of the mitochondrial genome.

Figure 1|. Lifespan, weight, size and appearance of WT, PolgA^D257A and Pappa mutant mice.

A. The lifespan of PolgA^D257A is not rescued by deletion of the Pappa gene. B-C. A homozygous deletion of the Pappa gene leads to decreased weight and increased frailty in male and female PolgA^D257A mice (male n = 8–22/group, female n = 6–13/group). D. Appearance and size distribution of WT and mutant mice (grid size = 1 cm²).

Deletion of Pappa rescues splenomegaly in PolgA^D257A mice

Next, we tested whether loss of Pappa could attenuate the age-related pathology of the mutator mice. Both male and female PolgA^D257A mice exhibit pronounced splenomegaly at 12 months of age (fig. 2A-C), and surprisingly, this phenotype was rescued in both sexes by homozygous deletion of the Pappa gene, even after we corrected for the reduced size of Pappa^−/− animals (fig. 1D). Interestingly, a heterozygous deletion of Pappa also rescued spleen size in male PolgA^D257A mice, although not in female mice, suggesting that male PolgA^D257A mice are more sensitive to IGF-1 depletion than females. These findings suggest that depletion of Pappa (loss of one copy of the Pappa gene) can partially recapitulate the impact of a full deletion (loss of both copies of the Pappa gene). These findings indicate that while Pappa deletion is insufficient to extend lifespan in PolgA^D257A mice, it is capable of rescuing distinct pathologies that are driven by mtDNA instability.

Figure 2|. The splenomegaly and anemia of PolgA^D257A mice are partially rescued by deletion of Pappa.

A-C.Male and female PolgA^D257A mice display splenomegaly. In both sexes, spleen size is reduced by deletion of two copies of the Pappa gene. In males, deletion of one copy of the Pappa gene is sufficient for spleen reduction as well. Spleen length was normalized to body length (male n = 7–17/group, female n = 6–13/group). D-E. Male and female PolgA^D257A mice exhibit reduced RBC counts, which is rescued in both sexes in PolgA^D257A; Pappa^+/− mice. F-G. Male and female PolgA^D257A mice exhibit reduced hemoglobin content, which is rescued in both sexes in PolgA^D257A; Pappa^+/− mice (male and female n = 4–8/group).

Depletion of Pappa partially rescues anemia in PolgA^D257A mice

The reduced spleen size of PolgA^D257A; Pappa^−/− mice prompted us to investigate whether other phenotypes associated with the PolgA^D257A allele were attenuated as well. A second characteristic phenotype of PolgA^D257A mice is anemia, which is thought to arise from multiple factors, including splenomegaly, which is an accelerant for the destruction of red blood cells (RBCs), impaired hematopoietic stem cell function, and increased damage to circulating erythrocytes²⁶. Accordingly, we observed that old, but not young PolgA^D257A mice displayed a marked reduction in RBCs and hemoglobin content (fig. 2D,E; fig. S2). Excitingly, heterozygous deletion of the Pappa gene partially restored both the RBC levels and hemoglobin content in PolgA^D257A Pappa^+/− mice (fig. 2F,G). In these experiments, as well as others (see below), PolgA^D257A Pappa^+/− mice consistently outperformed PolgA^D257A Pappa^−/− mice, suggesting a dose-dependent reduction in IGF-1 signaling is required for PolgA^D257A mice to achieve the greatest benefits.

Depletion of Pappa partially rescues the sex-specific, inflammatory profile of PolgA^D257A mice

Chronic inflammation is a well-established consequence of mitochondrial dysfunction that is frequently driven by increased reactive oxygen species (ROS) production²⁷, the release of damage-associated molecular patterns²⁸, and activation of innate immune pathways such as the STING pathway^29,30. In addition, inflammation is a major cause of anemia in patients with chronic diseases. Accordingly, we examined whether PolgA^D257A mice display elevated levels of inflammation as well and found that male PolgA^D257A mice exhibit significantly elevated serum levels of IL-2, IL-6, TNF-α, and IFN-γ, cytokines indicative of systemic inflammation and heightened splenic immune activity (fig. 3). Heterozygous deletion of Pappa significantly reduced IL-2 and IL-6 levels in these males, attenuating their inflammatory phenotype. Notably, PolgA^D257A; Pappa^+/− mice showed greater improvement than PolgA^D257A; Pappa^−/− mice, reinforcing the importance of IGF-1 dosage in ameliorating mtDNA disease. In addition, we noted that female PolgA^D257A mice displayed a muted inflammatory profile compared to male PolgA^D257A mice, indicating a sex-specific inflammatory response to mitochondrial mutagenesis. Most likely, this sexual dimorphism is due to the anti-inflammatory³¹ and anti-oxidant activity³² of estrogen, and other sex-linked factors³³. Regardless, the limited inflammatory profile of female PolgA^D257A mice remained unchanged upon Pappa deletion (fig. 3h), highlighting the heightened sensitivity of male PolgA^D257A mice to IGF-1 modulation.

Figure 3|. Inflammation in PolgA^D257A mice is partially rescued by depletion of Pappa.

A, C, E, G. Male PolgA^D257A mice display increased levels of serum inflammation markers. IL-2 (A) and IL-6 (C) markers are rescued by deletion of one copy of the Pappa gene. B, D, F, H. Female PolgA^D257A mice only display increased levels of IL-6 among the evaluated cytokines, which is not rescued by Pappa deletion (D). (male n = 4–15/group, female n = 2–8/group).

Depletion of Pappa improves muscle function in male PolgA^D257A mice

The primary consequence of mitochondrial mutagenesis is impaired energy production, which is particularly detrimental to energy-demanding tissues such as skeletal muscle³⁴. As a result, aged (but not young, fig. S3) PolgA^D257A mice exhibit significant reductions in both grip strength and treadmill endurance compared to WT type controls. These assays primarily assess type I (grip strength) and type II (endurance) muscle fiber performance, suggesting functional impairment across both fiber types. Consistent with our previous findings, we observed improved grip strength and endurance in PolgA^D257A; Pappa^+/− males, but not females and saw no improvement in PolgA^D257A; Pappa^−/− mice (fig. 4A-D). Instead, PolgA^D257A; Pappa^−/− mice tended to display reduced, rather than improved grip strength compared to PolgA^D257A mice, providing further evidence of their overall frailty.

Figure 4|. Impaired muscle and cardiac function in PolgA^D257A mice is partially rescued by deletion of Pappa.

A-D. Male and female PolgA^D257A mice display reduced grip strength and endurance. In male mice, these physiological endpoints are improved upon deletion of one Pappa copy in male, but not female mice (male n = 8–32/group, female n = 5–19/group). E-I. Male, but not female, PolgA^D257A mice display increased LVID during diastole and systole. Heart size, LVIDd and LVIDs return to WT levels in male mice upon deletion of one or two copies of the Pappa gene (male n = 7–10/group, female n = 4–9/group).

Deletion and depletion of Pappa improve the sex-specific cardiac dysfunction of PolgA^D257A mice

Like skeletal muscle, cardiomyocytes rely heavily on continuous energy production, making them particularly vulnerable to mitochondrial dysfunction^35,36. In PolgA^D257A mice, extensive mitochondrial mutagenesis leads to significant cardiac hypertrophy and enlargement of the left ventricle—a phenotype that mirrors age-related cardiac pathology in humans. To determine whether Pappa deletion could mitigate these effects, we performed echocardiography on 12-month-old mice and found that male PolgA^D257A mice exhibited a significant increase in the heart size (fig. 4E-I), as well as the internal diameter of the left ventricle in both diastole (LVIDd) and systole (LVIDs), indicating structural remodeling and reduced cardiac efficiency. Similar to the inflammatory profile of the mice, we did not observe overt cardiomyopathy in female PolgA^D257A mice, suggesting a sex-specific impact of mtDNA instability on cardiac health, consistent with previous reports³⁷. Because inflammation can drive cardiomyopathy³⁸, it is possible that the reduced inflammation in female PolgA^D257A mice is mechanistically linked to their cardiomyopathy. We found that both PolgA^D257A; Pappa^+/− and PolgA^D257A; Pappa^−/− mice showed significant improvement in ventricular dimensions compared to PolgA^D257A males. These findings further confirm that depletion of Pappa partially phenocopies a complete deletion. Indeed, out of the 2,950 genes that are significantly altered in the hearts of male Pappa^−/− mice vs WT mice, 2,309 move in the same direction in Pappa^+/− mice, albeit in a dose dependent manner (Spearman correlation 0.46, p=1.5×10⁻¹⁵⁴, fig. S4). Together, these observations indicate that reduced IGF-1 signaling can partially restore cardiac function even in the context of a high mtDNA mutation burden

The mutation frequency of mtDNA is not affected by deletion of the Pappa gene

To investigate the mechanism by which cardiac function is improved in male PolgA^D257A mice following Pappa deletion, we tested whether reduced IGF-1 signaling lowers the mitochondrial mutation burden. Using duplex sequencing³⁹, we generated high-resolution maps of the single base substitutions (fig. 5A), deletions (fig. 5B), and insertions (fig. 5B) that arose across the mitochondrial genome in 12 month old animals. However, the frequency of these mutation classes was unchanged among PolgA^D257A, PolgA^D257A; Pappa^+/−, and PolgA^D257A; Pappa^−/− mice, as was their mutation spectrum (fig. 5C) and the distribution of mutations across the mitochondrial genome (fig. 5D,E). Digital droplet PCR (fig. 5F) and qPCR (fig. S5) further showed that mtDNA copy number was unchanged among these genotypes. Thus, the improvements in cardiac function are not due to reduced mtDNA mutagenesis or altered mtDNA copy number, but must be mediated by processes that occur after mutation generation.

Figure 5|. Mutagenesis in WT and mutant mice.

A-C. Male PolgA^D257A mice, with or without Pappa deletion display identical mutation rates and spectra for all mutation classes. The same is observed for WT mice with or without Pappa deletion. D,E. Mutations were equally distributed across the mitochondrial genome (n = 5–6/group). First track = coverage. Second track = single nucleotide variants. Third track = insertions. Fourth track = deletions. Fifth track = mutation frequency. F. mtDNA copy number was unchanged by deletion of the Pappa gene (n = 5–6/group). G,H. Deletion of the Pappa gene reduces the number of cardiomyocytes with clonally expanded mtDNA mutations. Blue cells depict cells without WT COX activity (n = 3/group).

Deletion of Pappa slows clonal expansion of mtDNA mutations

While the mutation frequency remained unchanged upon Pappa deletion, the pathogenic consequences of these mutations do not manifest themselves until they clonally expand within individual cells. In both aging humans³⁴ and PolgA^D257A mice¹⁹, these clonally expanded mutations disrupt cytochrome oxidase activity, impair ATP production and promote cardiomyopathy. To determine whether Pappa deletion influences this expansion process, we stained heart sections from 12-month-old mice for COX/SDH activity (fig. 5H), a technique that can identify cells in which mtDNA mutations have clonally expanded. While COX-negative cells were absent in WT, Pappa^+/− and Pappa^−/− hearts, PolgA^D257A mice exhibited a substantial accumulation of COX-negative cells, consistent with widespread clonal expansion of mtDNA mutations. Remarkably, this phenotype was partially rescued in both PolgA^D257A; Pappa^+/− (p = 0.0046) and PolgA^D257A; Pappa^−/− (p = 0.0002) mice (fig. 5G). These findings suggest that although IGF-1 signaling does not affect the generation of mutations, it does affect the rate with which harmful mtDNA mutations clonally expand within cells, thereby preserving mitochondrial function and contributing to improved cardiac health in PolgA^D257A mice.

Transcriptomic profiling reveals molecular signatures of cardiomyopathy in male PolgA^D257A mice

To investigate the consequences of reduced homoplasmy in PolgA^D257A; Pappa^−/− mice, we analyzed their heart tissue by RNA-seq. First, we established a baseline, transcriptomic profile of PolgA^D257A hearts (fig. S6) by comparing differentially expressed genes between PolgA^D257A and WT mice. This analysis confirmed that the hearts of PolgA^D257A mice are under substantial metabolic and inflammatory stress. Specifically, PolgA^D257A mice exhibited a 3-fold upregulation of Gdf15 (p = 0.0005), a well-established marker of cardiac pathology, inflammation, and metabolic perturbation and a 7-fold increase in Fgf21 (p = 0.03), which encodes a hormone that is released in response to metabolic stress. Consistent with a failing myocardium we also observed a 16-fold upregulation of Myh7 (p < 0.0001) and a 3-fold upregulation of Nppb (p < 0.0001), two canonical markers of the “fetal gene expression program” that can be reactivated during cardiac dysfunction. These transcriptional changes were accompanied by a significant downregulation of genes that encode key components of the contractile apparatus, including Myl1, Myl4, Myl7, and Mybphl, all of which encode myosin light chains essential for sarcomere stability and contraction in striated muscle. Additionally, critical regulators of cardiac pacemaking and excitation-contraction coupling were downregulated, such as voltage-gated calcium (Atp2b2, Cacna2d2) and potassium channels involved in maintaining cardiac electrophysiology. Consistent with their inflammatory profile (fig. 4), we also observed a marked upregulation of chemokines that recruit innate immune cells, including Ccl2 (3-fold), Ccl8 (5-fold), Cxcl5 (3-fold) and Cxcl13 (>20-fold), potent attractants of monocytes, macrophages and neutrophils. Components of the complement system (e.g., C4b, 2.5-fold increase) and pattern-recognition receptors like Tlr7 (2-fold) that are associated with the JAK-STAT signaling pathway were also elevated. Notably, a subset of these inflammatory genes was linked to the activation of the AIM2 inflammasome (enrichment p= 4×10⁻⁴), a complex that detects cytosolic double-stranded DNA (including mtDNA that leaks out of dysfunctional mitochondria) and drives innate immune activation⁴⁰. Together, these findings establish a molecular signature of cardiomyopathy in PolgA^D257A mice, characterized by metabolic stress, sarcomeric disorganization, electrophysiological instability, and robust activation of innate immune pathways.

Pappa deletion partially normalizes the transcriptomic profile of male PolgA^D257A hearts

To assess the impact of Pappa deletion on the pathological transcriptomic landscape of PolgA^D257A mice, we then compared the changes in expression profile of PolgA^D257A vs. WT mice to PolgA^D257A; Pappa^+/− vs. WT, and PolgA^D257A;Pappa^−/− vs. WT animals. Remarkably, we found that deletion of one or two copies of the Pappa gene shifted the entire transcriptome of PolgA^D257A hearts closer to WT levels (fig. 6A-C). This normalization was most pronounced when we focused on the genes that are most dysregulated in PolgA^D257A animals. Out of the 252 genes that were significantly downregulated >2-fold in PolgA^D257A hearts, 217 moved closer to WT levels in PolgA^D257A; Pappa^+/− mice (p < 0.0001) and 203 in PolgA^D257A; Pappa^−/− mice (p < 0.0001, fig. 6A). A similar trend was observed among upregulated genes. Out of the 244 genes that were significantly upregulated >2-fold in PolgA^D257A hearts, 202 moved closer to WT levels in PolgA^D257A; Pappa^+/− mice (p < 0.0001) and 167 in PolgA^D257A; Pappa^−/− mice (p < 0.01, fig. 6B). In contrast, genes that were not dysregulated in PolgA^D257A mice showed no significant change when Pappa was deleted (fig. 6C).

Figure 6|. Transcriptomic analysis of WT and mutant mice.

A-C. Loss of one or two copies of the Pappa gene normalizes the expression profile of PolgA^D257A mice. Genes that are upregulated (A) or downregulated (B) in PolgA^D257A mice are normalized upon deletion or depletion of Pappa. Genes that are unchanged in PolgA^D257A mice remain unchanged after complete or partial loss of Pappa (C). D. Gene expression patterns suggest that impaired cardiac and metabolic function are rescued by deletion and depletion of the Pappa gene, drastically shifting the expression patterns away from the dysregulation observed in PolgA^D257A mice. Interestingly, loss of one copy of Pappa in PolgA^D257A mice changes the transcriptional profile in the direction of Pappa^−/− compared to Pappa^+/−, suggestive of implementation of pathways to rescue the negative consequences of mtDNA mutagenesis. E. Pro-aging related gene programs are dampened by deletion and depletion of Pappa. F. Complete loss of Pappa does not restore expression patterns of mitochondrial transcripts in the PolgA^D257A background. G. Lipid metabolism related genes are not upregulated in Pappa^+/− mice or PolgA^D257A mice, but are upregulated in PolgA^D257A; Pappa^+/− mice, suggesting context dependent expression patterns (n = 6/group).

Pappa deletion restores metabolic flexibility and cardiac function in male PolgA^D257A hearts

To identify the pathways most responsible for the improved cardiac health of PolgA^D257A mice, we filtered the transcriptome for genes that change ≥8-fold in expression upon deletion of either one, or two copies of the Pappa gene. This analysis showed that cardiac health was improved at multiple levels of scale. At a functional level, we found that key cardiac contractile genes (Myl7, Myl4, Mybphl, Sln) that were strongly downregulated in PolgA^D257A mice, were robustly restored by Pappa deletion, suggesting a rescue of cardiac contractile function (fig. 6D). This functional rescue was accompanied by widespread metabolic reprogramming. Although heart tissue primarily relies on fatty acid oxidation⁴¹, it can switch to glucose utilization under stress⁴². Two genes that enable this metabolic flexibility are PCK1 the rate limiting enzyme for gluconeogenesis, and FBP1⁴³, which are significantly downregulated in PolgA^D257A hearts, indicating metabolic rigidity. Deletion of Pappa reactivates these genes, indicating enhanced gluconeogenesis and glucose mobilization, which improves the metabolic flexibility of the heart. Systemically, PolgA^D257A mice exhibit reduced expression of the adipokines leptin (Lep) and adiponectin (Adipoq), consistent with global metabolic dysfunction. Pappa deletion reversed this trend, restoring expression of both hormones and correcting this endocrine imbalance. Restoration of Adipoq is particularly noteworthy, given its role in insulin sensitivity, mitochondrial biogenesis, and antioxidant defenses⁴⁴, factors that influence organismal resilience at multiple levels of scale. Further transcriptomic shifts supported improved lipid handling. Pappa deletion upregulated Adipogenin (Adig), which promotes adipocyte differentiation⁴⁵, and Plin1, which stabilizes lipid droplets and prevents lipotoxicity⁴⁶. These changes occurred alongside increased expression of Adrb3⁴⁷ and Acvr1c⁴⁸, regulators of lipid mobilization and adipocyte metabolism. Notably, Adrb3 was co-rescued with Ucp1, suggesting improved thermogenesis and mitochondrial uncoupling and the activation of stress-protective metabolic programs. Together, these changes suggest that Pappa deletion enables PolgA^D257A hearts to shift from metabolic exhaustion towards a more flexible, energy and stress-resilient metabolic state.

The PolgA^D257A allele blocks anti-aging programs triggered by Pappa deletion

To understand why Pappa deletion can improve cardiac function, but not extend lifespan in PolgA^D257A mice, we analyzed the expression of genes that are involved in key anti-aging pathways (fig. 6E). One hallmark of lifespan extension is reduced mTOR signaling⁴⁹. Accordingly, Pappa^−/− mice show reduced mTor expression (p= 0.046), while short-lived PolgA^D257A mice exhibit increased expression (p= 0.01) ; however, mTor reduction was partially blocked when Pappa was deleted in PolgA^D257A mice, limiting its impact on longevity. Another hallmark of longevity is improved telomere maintenance¹. Consistent with their extended lifespan, Pappa^−/− mice, show upregulation of Terf1 (p=0.0004) and downregulation of tankyrase (p=0.036), two changes that shield telomeres from damage and recombination and prevent them from triggering cell cycle arrest⁵⁰. Importantly, neither gene changed expression when Pappa was deleted in a PolgA^D257A background. In addition, Pappa^−/− mice exhibit overexpression of multiple DNA repair genes that are involved in double strand break repair and telomere maintenance, a key feature of longevity programs⁵¹. However, these genes were not overexpressed in PolgA^D257A; Pappa^−/− mice, demonstrating that the PolgA^D257A allele blunts, or blocks this longevity interventions as well. Another feature of long-lived animals is their ability to maintain their epigenetic landscape, which is normally eroded by reduced expression of epigenetic regulators like Hdac1. Accordingly, Pappa^−/− mice overexpress multiple chromatin regulators, including Hdac1 (p=0.005), Ezh2 (0.0002), Chd1 (p=0.03), Chd2 (p=0.02), Chd3 (p=0.0007) and Chd6 (p=0.049), none of which were overexpressed in PolgA^D257A; Pappa^−/− mice. As cells age, the structure of their nuclear envelope tends to deteriorate, a change that is frequently attributed to reduced expression of Lamin B (Lmnb1)^52,53. Consistent with their increased longevity, Pappa^−/− mice overexpress Lmnb1 (p=0.002) together with Nemp1 (p=0.006) and 2 (p=0.042), two structural proteins of the nuclear envelope; however, we found no overexpression of these genes in PolgA^D257A; Pappa^−/− mice. As part of a shift from anabolic to catabolic metabolism, deletion of Pappa decreases expression of protein coding genes from the mitochondrial genome, which may reduce mitochondrial dysfunction and ROS production (fig. 6F). This reduction was mediated by reduced expression of the mitochondrial RNA polymerase (Polrmt, p=0.03) and LRPPC (p=0.06), a protein that selectively protects mitochondrial mRNAs, but not tRNAs^54,55. Accordingly, the expression of tRNAs was not systemically reduced in Pappa^−/− mice. None of these changes were observed in PolgA^D257A; Pappa^−/− mice. Another hallmark of long-lived cells is improved proteostasis¹. Again though, we found that although Pappa deletion significantly upregulated numerous primary heat shock proteins, these genes were not upregulated in PolgA^D257A; Pappa^−/− mice. Indeed, we found that out of the 32 primary heat shock proteins (HSPs) we detected, 27 were downregulated in PolgA^D257A mice (fig. S7A) and none of these genes were rescued in PolgA^D257A; Pappa^−/− mice. Notably, this pattern was not observed in the DnaJ-family HSP40 of co-chaperones (fig. S7B), which, unlike the primary chaperones, are ATP-independent⁵⁶. Thus, loss of mtDNA integrity, and reduced ATP production specifically inhibit the expression of ATP-dependent chaperones, contributing to proteostasis failure.

Pappa deletion triggers context-dependent metabolic rewiring

Aging cells frequently exhibit not one, but multiple hallmarks of the aging process. These combinations could lead to complex, context-dependent interactions that result in novel gene expression states that neither hallmark produces alone. To identify such interactions between mitochondrial mutagenesis and IGF-1 signaling, we filtered the transcriptome for genes that remain unchanged upon deletion of Pappa in WT mice (Pappa^+/− mice), but were strongly altered (>4-fold) in PolgA^D257A; Pappa^+/− mice. Of the 13 genes that were identified by this method, 7 were directly involved in lipid metabolism, storage, or mobilization (fig. 6G). These included Adipoq, a protective adipokine in metabolic disease and inflammation; Cidec and Plin1, which regulate lipid droplet formation and adipocyte energy storage^46,57, and Apoc1, a lipid- and immune-regulatory apolipoprotein⁵⁸. Finally, we identified Fasn, an enzyme that is central to fatty acid synthesis⁵⁹; Elovl6, which catalyzes the elongation of long-chain fatty acids⁶⁰, and Lrp2, a multifunctional endocytic receptor that internalizes lipoproteins, insulin, leptin, and amyloid-beta⁶¹. These results reveal a striking lipid-centric response to a partial deletion of Pappa that only emerges in the presence of mitochondrial dysfunction, underscoring how aging hallmarks can interact with each other to produce context-dependent transcriptional programs. Notably, some of these interactions also occur when both copies of Pappa are deleted, demonstrating that mitochondrial mutations can sensitize cells to modulation by Pappa deletion and offering a plausible explanation for why partial Pappa deletion can improve the phenotype of PolgA^D257A mice as well. A similar effect is seen when column 2 and 3 of figure 6D are compared.

DISCUSSION

Aging is a multifactorial process, characterized by the simultaneous emergence of multiple, interconnected molecular and cellular hallmarks. In naturally aging organisms, these hallmarks do arise independently, but instead accumulate and interact within the same tissues and cells over time. As a result, there is a growing recognition that the hallmarks of aging are critically intertwined, and that to fully understand the aging process, it will be essential to tease apart these complex interactions.

In this study, we investigated how mtDNA instability interacts with IGF-1 signaling to shape mammalian lifespan and healthspan. Suppression of IGF-1 signaling through Pappa deletion extends lifespan in WT mice by triggering well-known pro-longevity programs, including enhanced stress resistance, DNA repair, proteostasis, and metabolic reprogramming. However, the same intervention fails to extend lifespan in PolgA^D257A mice, which accumulate mtDNA mutations at an accelerated pace. This observation echoes prior studies that showed that calorie restriction⁶², exercise^63–66, or overexpression of a mitochondrially targeted catalase⁶⁷, did not extend the lifespan of PolgA^D257A mice either. Taken together, these studies support that a hierarchy exists among the hallmarks of aging, where mtDNA integrity overrides the impact of longevity interventions, such as IGF-1 signaling, nutrient sensing, exercise and oxidative stress to impose a hard limit on mammalian lifespan.

Although Pappa deletion did not extend the lifespan of PolgA^D257A mice, it did attenuate a wide range of age-related phenotypes, including splenomegaly, anemia, chronic inflammation, skeletal muscle decline, and cardiomyopathy. These findings echo those with a mitochondrially targeted catalase, which failed to extend lifespan as well, but improved the cardiomyopathy in PolgA^D257A mice^35,67. The benefits we observed were especially pronounced in male mice, which seemed to suffer more from mtDNA instability than female mice, possibly due to the anti-oxidant and anti-inflammatory effects of estrogen⁶⁸. Surprisingly, these improvements were most evident in mice with partial Pappa deletion. This dose-dependent effect suggests that complete loss of Pappa may excessively suppress IGF-1 signaling in the context of high mitochondrial stress, where some degree of anabolic signaling remains necessary to maintain cellular homeostasis. Our transcriptomic analyses demonstrate that the health improvements in PolgA^D257A; Pappa^+/− mice were mediated, at least in part, by context-dependent interactions between mtDNA instability and IGF-1 signaling. In other words, we observed gene expression states in PolgA^D257A; Pappa^+/− mice that were not observed in either genotype alone. Such context-dependent effects underscore how important the interactions between the hallmarks of aging are: the outcome is not merely the sum of two interventions, but a new state shaped by their interplay.

This principle was most clearly illustrated by the failure of Pappa deletion to activate canonical longevity pathways in PolgA^D257A mice. In WT mice, Pappa deletion upregulates a suite of anti-aging pathways that improve DNA repair, telomere maintenance, proteostasis and chromatin integrity. However, in the presence of mtDNA mutations, most of these pathways are either blocked or blunted, demonstrating that the integrity of the mitochondrial genome is pre-requisite for lifespan extension by reduced IGF-1 signaling. Similarly, previous research has shown that proteostasis⁶⁹, DNA repair⁷⁰ and other molecular mechanisms are pre-requisites for lifespan extension by reduced IGF-1 signaling as well. Our observations now suggest that within this framework, mtDNA integrity is not simply one of the many hallmarks of aging, but rather the foundation upon which others are built. And when that platform is broken, downstream hallmarks like proteostasis or DNA repair cannot be engaged by typical means, no matter how favorable the upstream signaling may be.

This insight has major implications for the development of anti-aging therapies. It suggests that interventions that target nutrient-sensing pathways may fail—or even backfire—when applied to organisms or tissues with high levels of mitochondrial damage. As such, the next generation of geroprotective treatments must be tested in diverse models of aging, including those that combine multiple hallmarks, to better understand the scope and boundaries of their efficacy.

Because mtDNA mutations impose a ceiling on mammalian lifespan, it is vital that these treatments also address the stability of the mitochondrial genome. While a reduction in IGF-1 signaling did not alter the frequency of mutations in WT or PolgA^D257A mice, it did slow the pace with which they reached homoplasmy. Thus, although it may not be possible today to reduce mitochondrial mutagenesis in human cells, our data shows that it may already be possible to curtail the impact of mtDNA mutations on mammalian healthspan by slowing their clonal expansion.

While the precise mechanism by which Pappa influences clonal expansion of mtDNA mutations remains uncertain, several plausible explanations can be proposed. First, the progression of mtDNA mutations toward homoplasmy is likely influenced by mitochondrial genome replication and segregation dynamics, two processes that can accelerate genetic drift, especially during cell division when mutant genomes may be unequally partitioned between daughter cells. Because Pappa deletion suppresses growth-related pathways these changes may reduce mitochondrial workload and mtDNA replication frequency, thereby slowing genetic drift in post-mitotic cells and limiting mitochondrial segregation via reduced cell turnover. Second, Pappa deletion enhances a wide variety of maintenance processes, which could lead to improved autophagic and mitochondrial quality control mechanisms that selectively eliminates dysfunctional mitochondria and curbs the expansion of mutant clones. Third, emerging evidence suggests that damaged mtDNA is often degraded rather than repaired⁷¹. In that context, the enhanced antioxidant capacity observed in Pappa-deficient PolgA^D257A mice may lessen oxidative mtDNA damage, decreasing the need for mtDNA turnover and potentially stabilizing the mitochondrial genome by limiting replication-driven amplification of existing mutations.

Regardless, these findings provide a compelling example of how the interplay between distinct hallmarks of the aging process can fundamentally alter the outcome of otherwise beneficial interventions. They reveal that the efficacy of anti-aging strategies like IGF-1 suppression is not absolute, but context-dependent. They are contingent on the integrity of underlying systems, including proteostasis and DNA repair. Without an intact mitochondrial genome though, these pathways cannot be engaged, indicating that mtDNA integrity is upstream from these critical anti-aging pathways. More broadly, our results underscore the need for a more integrated model of aging, one that considers not just individual pathways but their interactions, hierarchies, and points of failure. By mapping these interactions, we can better anticipate the limitations of existing interventions and design next-generation therapies that are robust to the complex biology of aged tissues. In this light, strategies that target the expansion of mtDNA mutations—rather than their origin—may offer a powerful new axis for preserving tissue function and extending healthspan, even when the underlying genomic damage cannot be undone.

METHODS

Transgenic Mice

PolgA^D257A mice were purchased from JAX (strain #:017341), while Pappa^−/− mice were a generous gift from Dr. Cheryl Conover¹². Male PolgA^D257A/+; Pappa^+/− and female PolgA^D257A/+; Pappa^+/− were bred to generate mice the genotypes described in this study. To limit mtDNA mutations from being inherited through the germline, separate PolgA^D257A/+; Pappa^+/− x WT C57Bl6/j crosses were set up to generate “first generation” PolgA^D257A/+; Pappa^+/− females to be used for breeding. At 21 days, mice were weaned, ear-punched and two PCR reactions were run to genotype each mouse. The first reaction genotyped the PolgA^D257A allele, using primers F: 5’-GCCTCGCTTTCTCCGTGACT-3’ and R: 5’-GGATGTGGCCCAGGCTGTAACTCA-3’. To genotype the Pappa allele, primers F_common: 5’- TAAGCAGGGGTGGGTCCTTT-3’, F_neo: 5’- TCGCCTTCTATCGCCTTCTTG-3’, and R: 5’- CACTCCTCAGCTTCGGCTTTCA-3’ were used. PCR reactions were run on agarose gels by gel electrophoresis and imaged before analysis. Mice were assayed at 3 months or 12 months of age for experimental studies, while a subset of mice were aged for lifespan experiments and weekly body weight measurements. Animals were scored by our lab and USC’s Department of Animal Resources veterinarians to establish humane endpoints. Mice were euthanized by CO₂ exposure.

Grip Strength

Grip strength was assessed using a horizontally mounted bar attached to a sensor (TSE-Systems, 303500-M/E1) designed to measure forelimb strength. Each mouse was allowed to securely grasp the bar before being swiftly and steadily pulled backward. Measurements were recorded only if the mouse let go with both forelimbs at the same time. Ten measurements were taken per animal and averaged to obtain a final value.

Endurance

Mice were acclimated to the treadmill instrument (TSE 303401-M-04/C) 24 hours before testing. During acclimation, animals were placed on the stationary treadmill for five minutes, followed by an increase in speed to 2m/min for five additional minutes. Mice were then tested once each day on the two days following acclimation. The treadmill protocol consisted of 1 m/min for 1 minute, followed by a 1m/minute increase every minute until exhaustion. Time of exhaustion was recorded in seconds. Mice that refused to run would feel a mild shock at the back of the treadmill. The values recorded for each mice over both test days was averaged and reported.

Echocardiography

One week prior to dissection, mice were transported to the molecular imaging center at USC for echocardiography using the VisualSonics Vevo 3100, MX 550 transducer 22–55 MHz. Mice were anesthetized with 2% isoflurane and a depilatory cream (Nair) was used to remove fur prior to echocardiography. Mice were placed in the supine position onto the warmed platform to maintain optimal physiological conditions and their limbs were taped onto the metal EKG leads. Heart rates were monitored and generally maintained at 400–500 beats per minute. Warmed echocardiography gel was placed on the shaved chest and the heart was imaged with a 30 MHz transducer. By placing the transducer along the long-axis of LV and directing it to the right side of the neck of the mouse, two-dimensional LV long-axis can be obtained. The transducer was then rotated clockwise by 90°, and the LV short-axis view was obtained. Transmitral inflow Doppler spectra were recorded in an apical 4-chamber view by placing the sample volume at the tip of the mitral valves. After the scans were concluded, the residual gel was removed, and the mouse was returned to the cage for recovery. Images were subsequently analyzed using the VevoLab software.

Tissue Collection and Analysis

Upon euthanasia, mice were measured, photographed, and the heart was punctured for blood draws for blood composition experiments and serum isolation. Tissues were dissected, placed on a grid sheet for measurements and photography and cut into 4 pieces. One piece was stored in formalin, one piece was frozen in Optimal Cutting Temperature (OCT) compound and two pieces were flash frozen in liquid nitrogen for DNA and RNA extractions. Formalin-fixed tissue was left at room temperature overnight before being washed twice with PBS and stored at 4°C in 70% EtOH. Serum was collected by letting blood clot for 30 min, followed by a 10 minute spin at 1,000g at 4°C. Serum was then collected by aspirating the supernatant and stored at −80°C. IFN-γ, IL2, IL6 and TNF-α levels in serum were measured with commercial immunoassays using a custom V-PLEX Mouse Biomarker Kit (Meso Scale Discovery, Rockville, MD). Blood composition was measured using 40 μl of blood collected in EDTA tubes using the Hemavet 950FS instrument.

Duplex-Sequencing

DNA was extracted from the hearts of experimental mice (n= 5–6/group) using the Qiagen DNeasy Blood and Tissue Kit (cat# 69506), quantified by Qubit 1x dsDNA hs assay kit, and stored at −20C. 500ng of DNA per sample was fragmented using the Biorupter Pico (30 seconds ON, 90 seconds OFF, 6 cycles) and size was measured using the Agilent 4200 TapeStation instrument (fragment size ~300bp). Published protocols^39,72 were followed with minor adjustments. Briefly, DNA was end-repaired and adapters (a generous gift from the Kennedy Lab) were ligated to DNA fragments using the Ultra II DNA Library Preparation Kit from NEB (cat# E7103) samples were cleaned up and libraries were quantified by qPCR using SYBR Green iTaq Supermix (Biorad cat#: 1725124) compared to previously generated libraries with target family sizes and on-target efficiencies. This quantification enables us to input the correct number of molecules into the pre-enrichment PCR depending on mtDNA content for each sample to sequence at the target parameters (depth, target raw reads, family size, on-target reads). Libraries were then enriched for mtDNA sequences using the xGen Hybridization Capture Assay (IDT) with a custom Discovery Pool of biotinylated probes (Suppl. File 1), along with their protocol. Final libraries were quantified, pooled and sequenced on a NovaSeq 6000 S2 kit (150PE). The standard duplex-seq pipeline was used to analyze the sequencing data, https://github.com/Kennedy-Lab-UW/Duplex-Seq-Pipeline. Mutation data are attached as supplementary files (Suppl. File 2).

mtDNA Copy Number Measurements

DNA extracted for duplex-sequencing experiments was used for copy number determination. By quantitative PCR, relative mtDNA copy number was measured using 50ng of input DNA and primers targeting mitochondrial ND1 (F: 5’-GCCTGACCCATAGCCATAAT-3’ and R: 5’-TATTCTACGTTAAACCCTGA-3’) and nuclear RPP30 (F: 5’-GCAACCGGAACATAGAGACA-3’ and R: 5’-CTGGCCTTGGAATGGGTAAT-3’). iTaq Universal SYBR Green Supermix was used in these reactions and the protocol was followed accordingly (Biorad cat#: 1725124). By droplet digital PCR (ddPCR), BioRad ddPCR copy number assay primers/probe for Rpp30 (HEX, Assay ID: dMmuCNS822293939) and for ND1 (FAM, Assay ID: dMmuCNS343824284) were used with 2x ddPCR Supermix for Probes (no dUTP) (Biorad cat#: 186–3023).

RNA-seq

Heart tissue from 12 month old, male mice (n = 6/group) was dounced in Trizol/Chloroform and homogenized with with zirconia beads for 20 minutes at 4°C. Samples were then at 10,000g for 5 minutes and supernatant purified according to the RiboPure Yeast RNA Purification Kit (Invitrogen, cat#: AM1926). RNA was DNase treated and mRNA was enriched using the GenElute Direct mRNA miniprep kit (Sigma, cat#: DMN70–1KT) and RNA-seq libraries were generated using the SMARTer Stranded RNA-Seq Kit (Takara Bio, cat#: 634838) and sequenced on a NovaSeq 6000 SP PE150 kit. Sequencing reads were trimmed using fastp⁷³ (version 0.23.2) with the following parameters: “--trim_poly_g --poly_g_min_len 6 --length_required 120”. Trimmed reads were then used to obtain transcripts abundance with Kallisto quant⁷⁴ (version 0.46) against the Ensembl r96 annotation for Mus musculus. Abundance estimates from Kallisto were then analyzed in R with the package DESeq2⁷⁵ following the standard pipeline to identify genes differentially expressed relative to the WT with FDR < 0.05 (correction for multiple testing). Sequencing data were deposited to SRA under accession PRJNA1271490.

COX-SDH Staining

Frozen OCT blocks containing heart tissue were sectioned onto slides using a Leica CM 1860 Cryostat in 10µm sections and stored at −80°C until they were ready to be stained. COX-SDH staining protocol was adapted from Wanagat et al³⁴. Briefly, heart sections placed on a slide were circled using a PAP pen, fully covered with COX-staining solution and incubated at 37°C for 12 minutes. Slides were then rinsed in 0.1M Tris-HCl and incubated at 37°C with SDH-staining solution for 10 minutes. Afterwards, slides were dehydrated in 70% ethanol, 95% ethanol, 100% ethanol (2x) and xylene before being coverslipped with Cytoseal 60. Slides were imaged on an Echo Revolve microscope at 4x, 10x, and 20x magnifications. Images were analyzed and scored for their % COX-negative cells using QuPath2 software.

Statistics

Group differences means (± SEM) were analyzed by one-way ANOVA with Fisher’s LSD test, Kruskal-Wallis with Uncorrected Dunn’s test for nonparametric distributions, or Welch’s ANOVA when variances were unequal across groups. Significance was defined as p < 0.05. Analyses used GraphPad Prism version 10 (GraphPad Software, San Diego, CA).