3D reconstruction of murine mitochondria reveals changes in structure during aging linked to the MICOS complex

Abstract During aging, muscle gradually undergoes sarcopenia, the loss of function associated with loss of mass, strength, endurance, and oxidative capacity. However, the 3D structural alterations of mitochondria associated with aging in skeletal muscle and cardiac tissues are not well described. Although mitochondrial aging is associated with decreased mitochondrial capacity, the genes responsible for the morphological changes in mitochondria during aging are poorly characterized. We measured changes in mitochondrial morphology in aged murine gastrocnemius, soleus, and cardiac tissues using serial block‐face scanning electron microscopy and 3D reconstructions. We also used reverse transcriptase‐quantitative PCR, transmission electron microscopy quantification, Seahorse analysis, and metabolomics and lipidomics to measure changes in mitochondrial morphology and function after loss of mitochondria contact site and cristae organizing system (MICOS) complex genes, Chchd3, Chchd6, and Mitofilin. We identified significant changes in mitochondrial size in aged murine gastrocnemius, soleus, and cardiac tissues. We found that both age‐related loss of the MICOS complex and knockouts of MICOS genes in mice altered mitochondrial morphology. Given the critical role of mitochondria in maintaining cellular metabolism, we characterized the metabolomes and lipidomes of young and aged mouse tissues, which showed profound alterations consistent with changes in membrane integrity, supporting our observations of age‐related changes in muscle tissues. We found a relationship between changes in the MICOS complex and aging. Thus, it is important to understand the mechanisms that underlie the tissue‐dependent 3D mitochondrial phenotypic changes that occur in aging and the evolutionary conservation of these mechanisms between Drosophila and mammals.


| INTRODUC TI ON
Sarcopenia, the loss of muscle mass associated with aging and decreased quality of life, affects primarily type II muscle fibers but also type I fibers.With age, sarcopenia in skeletal muscle leads to a body mass-independent loss of skeletal function (Miller et al., 2019).
Mitochondrial dysfunction and alterations in mitochondrial structure are also hallmarks of aging (Haas, 2019).The decreased expression of genes associated with mitochondrial dynamics and the loss of function contribute to sarcopenia and other age-related diseases (Coen et al., 2019).Thus, mitochondria are a key target for the development of therapeutics for age-related pathologies (Coen et al., 2019;Haas, 2019).Mitochondria change dynamically, using fission and fusion to tightly regulate structures that are critical to their function (Anand et al., 2014;Cogliati et al., 2016;Kühlbrandt, 2015); therefore, it is important to understand changes in mitochondrial structure over time.The cristae, the inner folds of the mitochondrial membrane, carry out oxidative phosphorylation and contain various transporters (Cogliati et al., 2016).To test our hypothesis that age-related alterations in metabolism and lipids increase mitochondrial fragmentation and loss of cristae integrity, we determined how mitochondrial structure changes during aging.
Disruption of optic atrophy 1 (OPA-1), an inner membrane protein that regulates mitochondrial fusion, causes mitochondrial fragmentation and affects the dimensions, shapes, and sizes of the cristae (Cogliati et al., 2016), and disruption of dynamin-related protein-1 (DRP1), which is associated with mitochondrial fission, causes elongated mitochondria and resistance to cristae remodeling (Favaro et al., 2019;Otera et al., 2013).Nanotunnels, or "mitochondria-on-a-string," are thin, double-membrane protrusions that allow mitochondria to communicate across distances.Nanotunnels, which may increase in mitochondrial disease (Vincent et al., 2017;Zhang et al., 2016), may also be associated with mitochondrial dysfunction during aging.Thus, the concomitant changes in mitochondrial structure and bioenergetics may drive pathologies.
Mutations in genes that regulate the morphology of cristae are associated with aging cardiomyocytes (Zhang, He, et al., 2021).These proteins, located at the crista junctions in the inner membrane, are part of the mitochondrial contact site and cristae organizing system (MICOS) complex, which is important for maintaining mitochondrial shape and size (Kozjak-Pavlovic, 2017).Loss of DRP1 or OPA-1 affects mitochondrial morphology similarly (Garza-Lopez et al., 2022;Lam et al., 2021).Cristae membranes contain the electron transport chain complexes and ATP synthases for oxidative phosphorylation (Friedman et al., 2015;Hu et al., 2020;Rampelt et al., 2017).
Because mitochondrial morphology affects function, altering the structure by knocking out MICOS-associated genes or OPA-1, a GTPase, may affect mitochondrial function during aging (Friedman et al., 2015;Hu et al., 2020;Rampelt et al., 2017).We hypothesize that MICOS-associated proteins are lost during aging and that loss of MICOS-associated genes can mimic the age-associated changes in mitochondrial morphology.Therefore, we determined how the MICOS complex affects gross mitochondrial structure, beyond the cristae, as well as how mitochondrial structure changes in aging.
Mitochondria have a tissue-dependent response to the environment (Holmström et al., 2012), which may be due to heterogeneity in mitochondrial DNA (mtDNA) quality check mechanisms across different tissues (Herbers et al., 2019).The 3D reconstruction of tissues using manual contour tracings provides information on mitochondrial phenotypes and how they differ across tissue types.To better understand age-related changes in mitochondrial structure, we used 3D reconstructions of aged gastrocnemius, soleus, and cardiac tissue in 3-month-old and 2-year-old mice to compare the size, shape, quantity, complexity, and branching of mitochondria.We observed an age-related loss of transcripts of the MICOS complex.We also used contact site and cristae organizing system (MICOS) complex genes, Chchd3, Chchd6, and Mitofilin.We identified significant changes in mitochondrial size in aged murine gastrocnemius, soleus, and cardiac tissues.We found that both age-related loss of the MICOS complex and knockouts of MICOS genes in mice altered mitochondrial morphology.Given the critical role of mitochondria in maintaining cellular metabolism, we characterized the metabolomes and lipidomes of young and aged mouse tissues, which showed profound alterations consistent with changes in membrane integrity, supporting our observations of age-related changes in muscle tissues.We found a relationship between changes in the MICOS complex and aging.Thus, it is important to understand the mechanisms that underlie the tissue-dependent 3D mitochondrial phenotypic changes that occur in aging and the evolutionary conservation of these mechanisms between Drosophila and mammals.CRISPR/Cas9 on myotubes to knockout three genes of the MICOS complex, Chchd3 (Mic19), Chchd6 (Mic25), and Mitofilin (Mic60), to determine whether loss of the MICOS complex may be phenotypically similar to aging through modulation of mitochondrial size, morphology, and oxygen consumption rate.To further characterize factors affecting mitochondrial structure in aging, we used multivariate analysis to identify changes in metabolites.We also characterize the lipidome during aging to identify possible commonalities in metabolic changes that occurred with the loss of the MICOS complex.
Finally, we also used a Drosophila model to better understand the evolutionarily conserved role of the MICOS complex in aging.

| Aging reduces mitochondrial size in murine gastrocnemius, soleus, and cardiac muscles
The gastrocnemius, a mixed muscle with both mitochondria-rich oxidative fibers and mitochondria-poor glycolytic fibers (Mukund & Subramaniam, 2020), is ideal for studying changes in mitochondrial dynamics.In contrast, soleus tissue, with predominantly slow-twitch muscle fibers, relies on mitochondrial oxidative metabolism (Crupi et al., 2018).We characterized mitochondria in cardiac tissue (Vue et al., 2022), which relies on efficient energy transfer to myofibrils and constant ATP production for contractile function (Chaudhary et al., 2011).Because mitochondrial function depends on structure (Cogliati et al., 2016;Kühlbrandt, 2015), it is important to determine how that structure changes over time.We hypothesized that, over time, mitochondrial fragmentation correlates with loss of the integrity of the cristae.
To determine how aging alters mitochondrial networks and individual mitochondrial structures, we imaged gastrocnemius, soleus, and cardiac biopsies from adolescent (3-month-old) and aged (2-year-old) mice by serial block-face scanning electron microscopy (SBF-SEM) with a resolution of 10 nm for the x-and y-planes and 50 nm for the z-plane, to visualize the electron connectome (Vue Zer et al., 2023).Approximately 50 intermyofibrillar (IMF) mitochondria were segmented from each image stack (Figure 1a-f) to generate a 3D surface view (Figure 1a′-f′).We analyzed IMF mitochondria instead of other mitochondrial subpopulations, such as subsarcolemmal, as IMF mitochondria are larger and display more significant age-related changes.We analyzed mitochondrial sub-network volumes from four regions of interest (ROIs) with an average of 175 mitochondria for each mouse (n = 3), for a total of over 500 mitochondria.Mitochondrial networks in skeletal muscle tissue in aged mice showed largely interconnected mitochondria (Figure 1a″-d″).
As in our previous study (Vue et al., 2022), we found cardiac tissue mitochondria remained relatively clumped with no apparent changes in their distribution (Figure 1e″-f″).We found that across all tissue types, the volume, area, and perimeter of mitochondria were significantly lower for 2-year-old versus 3-month-old mouse samples (Figure 1g-x).The mitochondrial volume is a measure of total capacity the area is analogous to surface area and the perimeter represents the boundary pixel count (Figure 1aa).
Although there was some variability among the three mice for each age cohort (Figure S1), this heterogeneity was more pronounced in the gastrocnemius (Figure 1g,h).The gastrocnemius also showed a greater reduction in mitochondrial size and surface area, with much smaller mitochondria that lacked hyperbranching.We found similar heterogeneity between the samples of soleus tissue and reductions in mitochondrial volume with aging (Figure 1m-r).
In contrast to skeletal muscle, the mitochondria were more homogenous in cardiac tissue (Figure 1s-x).Overall, in older mice, we saw a decrease in mitochondrial volume, area, and perimeter that was associated with increased fragmentation and smaller mitochondria.Because the size and length of mitochondria decreased with age, we further characterized the complexity of the mitochondria, which is implicated in mitochondrial communication.

| Aging results in poorly connected mitochondria with decreased branching in murine gastrocnemius, soleus, and cardiac muscles
We hypothesized that fewer networks and simpler shapes would occur with aging and dysfunction; therefore, we measured mitochondrial complexity to identify changes in mitochondrial shape during aging.Because we observed in 3D reconstructions that mitochondrial populations are heterogeneous and diverse, we used mito-otyping, a karyotyping-like method for arranging mitochondria (Vincent et al., 2019), to visualize the diversity of IMF mitochondria (Figure 2a-c).We found, for every volumetric measurement, smaller and less complex mitochondria with age.Soleus and gastrocnemius tissues showed more branched mitochondria in the adolescent mice versus the aged mice, and the latter had smaller volumes.In contrast to cardiac tissue, skeletal muscle showed large phenotypic changes with aging.To validate these changes, we analyzed 3D mitochondrial complexity using 3D form-factor measurements (Koopman et al., 2005;Vincent et al., 2019).We measured the mitochondrial complexity index (MCI) and sphericity to further characterize changes in complexity (Figure 2d-o).MCI and sphericity measure the roundness of mitochondria (Figure 2p,q).In gastrocnemius tissue, MCI increased concomitantly with sphericity in aged mice (Figure 2d-g).Thus, in contrast to their appearance in mito-otyping, in aged mice, mitochondria in the gastrocnemius showed increased sphericity.Among the young and the old mice, there were variations for both metrics.Soleus tissue, which showed a similar heterogeneity, was less complex and more spherical in aged mice compared to gastrocnemius tissue (Figure 2h-k), although this was less significant than in the gastrocnemius.Finally, cardiac tissue showed a significant increase only in MCI (Figure 2l-o).
Together, these data suggest that complexity changes with age but is tissue-type dependent.Therefore, we determined the role of the MICOS complex in age-related changes in mitochondrial structure and function.
The small increases in ATF-6 and Mitofilin expression in the aged flies were not significant.In the cardiac tissues of aged flies, ATF-4, Opa1, MINOS1, and Mitofilin were downregulated compared with controls; however, only the changes in ATF-4 and Mitofilin expression were significant (Table 2; Figure 3n-r).In contrast, FGF-21 expression increased significantly in the aged flies (Table 2).These support the results in mice showing that mitochondrial, MICOS, and MERC mRNA transcripts respond differently in cardiac and skeletal muscle tissue during aging, but in some organisms, the MICOS complex shows greater changes during aging.

| 2D and 3D structural changes in cristae and mitochondria after loss of the MICOS complex and Opa1
To determine the role of OPA-1 and the MICOS complex in mitochondrial structure and networking, we ablated the genes for Opa1 and the MICOS complex proteins in isolated primary skeletal muscle cells from 3-month-old mice.We isolated primary satellite cells and then differentiated myoblasts into myotubes.Using a CRISPR/Cas9 method and a control plasmid, we knocked out the genes for MICOS complex components and Opa1 from skeletal muscle cells.
We measured 1250 mitochondria across 10 cells, with loss of Opa1 as a positive control for mitochondrial morphological changes because in vitro deletion of Opa1 alters mitochondrial morphology (Hinton et al., 2023;Lam et al., 2021;Pereira et al., 2017).Although Opa1 expression decreases with age (Tezze et al., 2017), how the loss of the MICOS complex affects mitochondria 3D morphology is poorly understood.However, knockout of the MICOS subunit Chchd3 results in fragmented mitochondria as the cristae lose their normal structure (Darshi et al., 2011).Similarly, downregulation of Chchd6, which is important in maintaining crista structure, results in hollow cristae that lack an electron-dense matrix, thereby inhibiting ATP production and cell growth (An et al., 2012;Ding et al., 2015;Ott et al., 2015).Using transmission electron microscopy (TEM), we compared mitochondria and cristae in myotubes from wild-type (WT) and Opa1, Mitofilin, Chchd3, and Chchd6 knockout myotubes, which are essential for the organization of mitochondrial cristae (Ding et al., 2015;John et al., 2005; Figure 4a-e).Mitochondrial average area decreased for Opa1, Mitofilin, Chchd3, and Chchd6 knockout myotubes (Figure 4f), whereas the mitochondrial circularity index (the roundness and symmetry of mitochondria) and the number of mitochondria, once normalized, increased (Figure 4g,h).
This suggests that mitochondria become smaller, less complex, and more abundant upon loss of the MICOS complex.For Opa1, Chchd3, Mitofilin, and Chchd6 knockouts, the number of cristae per mitochondrion decreased, as did the cristae score and cristae surface area compared with the WT (Figure 4i-k).Here, the cristae score is defined as follows: 0: No sharply defined cristae are visible.
2: Over 25% of the mitochondrial area lacks cristae.TA B L E 2 RT-qPCR results from Drosophila cardiac tissue.| 9 of 26 Although the loss of the MICOS complex resulted in similar changes for all of the knockouts, the Chchd3 knockout showed the least significant changes.Together, these data showed quantitative and structural changes in both mitochondria and cristae upon loss of MICOS proteins.
TEM provides cristae detail but not 3D morphology; therefore, we used SBF-SEM to analyze the 3D structure of the mitochondria.
We measured a total of 200 mitochondria across 10 cells, comparing WT, Opa1, Mitofilin, Chchd3, and Chchd6 knockout myotubes (Figure 4l-p).We found that unlike the elongated mitochondria in the WT, Opa1 and MICOS protein knockouts had a much shorter 3D length (Figure 4q).Similarly, the mitochondria of Chchd3, Chchd6, Mitofilin, and Opa1 knockouts had smaller volumes than the WT (Figure 4r).The 3D reconstruction data, in combination with the prior TEM results, show that mitochondrial dynamics change with the loss of MICOS subunits and mimic the mitochondrial phenotypes observed during aging.
We also determined the effect of the loss of the MICOS complex and other mitochondrial genes on mitochondrial structure in a Drosophila model.Because QIL1 (Mic13) and CHCHD3/6 (Mic19) were downregulated and Mitofilin was slightly upregulated in aging skeletal tissue, we determined how knockdown (KD) of these proteins affected Drosophila mitochondrial structure as well as loss of mitochondrial fusion proteins OPA1 and MARF and the fission protein DRP1.Mitochondrial-actin staining showed differences in mitochondrial organization and relative myofibrillar density among the strains (Figure 4s).Using TEM, we found that in a DRP1 KD, in vivo flight muscle showed increased mitochondrial area, whereas MARF and OPA1 KDs reduced mitochondrial area (Figure 4t).Muscles from strains with a KD of the MICOS complex proteins (Mic13, Mic19, and Mic60) had reduced mitochondrial volume, although this reduction was less severe in Mic13 (Figure 4t).In considering circularity as a factor in mitochondrial complexity, we found that loss of the MICOS complex, similar to the KD of DRP1, decreased circularity (Figure 4u).
In contrast, the loss of mitochondrial fusion proteins increased the circularity of mitochondria.Also, KD of Mic13 and Mic19 reduced mitochondria, whereas Mic60-deficiency increased mitochondria (Figure 4v).This suggests that the functions of the MICOS complex and related mitochondrial structure are evolutionarily conserved to some degree, but there are organism-dependent alterations in associated mitochondrial dynamics and age-related changes in gene expression.

| Changes in oxygen respiration rate and metabolites after loss of the MICOS complex and Opa1
Loss of Opa1 induces bioenergetic stress and decreases electron transport chain function (Pereira et al., 2017), and ablation of the MICOS complex alters mitochondrial capacity (Kondadi et al., 2019;Stephan et al., 2020).To determine the effect of the loss of the MICOS complex on mitochondrial function, we measured the oxygen consumption rate (OCR) using an XF24 Seahorse analyzer.We showed no significant differences compared to the control (Figure 5f).
To determine the global effects of loss of Opa1 or the MICOS complex in skeletal muscle myotubes, we analyzed the metabolome to identify changes in metabolites that occurred with changes in mitochondria and cristae.Principal component analysis (PCA) revealed distinct populations in the control versus the Mitofilin knockout strains, which suggested that their genotypes contributed to the clustering (Figure 5g).
To identify the metabolites that best discriminated between true versus false positives, we constructed a model using analysis of variance (ANOVA) to determine the key metabolites that were statistically significant (Figure 5h).This unique metabolite signature revealed that  we observed in Mitofilin (Figure 5q).We constructed a model using ANOVA to determine which metabolite changes in Chchd3 and Chchd6 knockouts were statistically significant (Figure 5r).There was a loss of protein synthesis and changes in carbohydrate me-  et al., 2020;Sarzi et al., 2016;Wasilewski et al., 2012).This atypical response was evident in the increase in lactose and starch synthesis, but there was poor protein turnover, as seen in methionine metabolism (Figure 5t).Because a loss of MICOS complex proteins caused a change in metabolism, we determined whether this metabolic change parallels age-related changes in gastrocnemius, soleus, and cardiac metabolism.

| Metabolomics/lipidomic profiles in gastrocnemius, soleus, and cardiac tissue exhibit altered metabolism during aging
Because loss of the MICOS complex was implicated in altered ster- decreased with age (Figure S2J).CAR, which transports acyl groups from the cytosol into the mitochondrial matrix for beta-oxidation (Dambrova et al., 2022), was reduced in aged gastrocnemius tissues (Figure S2J).Conversely, Cer levels increased with age in these tissues (Figure S2J), supporting a previous report that elevated Cer levels in replicative senescent cells contribute to senescence by inducing cell cycle arrest (Stith et al., 2019).
MGs, which are linked to a lipotoxicity that triggers immune senescence (Feng et al., 2022), accumulate in the aged soleus, whereas TGs, which are important for storing and transporting fatty acids in cells and in the circulation, decreased in the aged soleus (Figure S2K).These findings align with previous reports of decreased TG levels in plasma and increased levels of fatty acids during aging (Johnson & Stolzing, 2019).
Notably, CL, which plays a critical role in regulating mitochondrial proteins and maintaining mitochondrial structures such as cristae and contact sites (Paradies et al., 2019), was reduced in aged cardiac muscles (Figure S2L).This lipid class has been implicated in age-related alterations in mitochondrial bioenergetics (Paradies et al., 2019;Semba et al., 2019;Shen et al., 2015).We also observed an accumulation of TGO in the heart, which has been linked to inflammation, endothelial dysfunction, oxidative stress, atherosclerotic plaques, and, ultimately, cardiovascular disease (Figure S2L; Singh & Singh, 2016).TGOs are also closely associated with aging (Johnson & Stolzing, 2019).
Among other metabolic changes, there was an accumulation of nicotinic acid riboside (NaR) in the soleus, cardiac, and gastrocnemius tissues (Figure 6g), indicating potential compensatory mechanisms to maintain NAD + levels, which decline with age in muscle tissues (McReynolds et al., 2021).Linoleic acid, a fatty acid essential for cell membrane integrity and synthesis of inflammation-related eicosanoids, also increased with age in these three tissues (Figure 6h).In contrast, aminolevulinic acid, which is essential for the production of the heme required for mitochondrial oxygen and energy production (Sawicki et al., 2015), decreased in all three tissues from aged mice (Figure 6i).In summary, our metabolic and lipid profiling analyses revealed significant metabolic alterations in cardiac, soleus, and gastrocnemius muscles with age, some of which may parallel the metabolic alterations resulting from the loss of the MICOS complex.

| DISCUSS ION
We demonstrated that either aging or loss of MICOS proteins in skeletal muscle resulted in tissue-dependent, suboptimal mitochondrial morphology, suggesting a correlation between aging and MICOS protein expression.Previous studies used 3D-focused ion beam scanning electron microscopy (FIB-SEM) to characterize the networking of the mitochondria in human (Dahl et al., 2015) and mouse skeletal muscle (Glancy et al., 2015).Quantitative 3D reconstructions used SBF-SEM to define the morphological differences in the skeletal muscles of humans versus mice and compared patients with primary mitochondrial DNA diseases versus healthy controls (Vincent et al., 2019).However, our current study is the first to use 3D reconstruction to characterize mitochondrial phenotypes in aged skeletal muscles.We used manual contour tracing rather than machine learning techniques to ensure the accuracy of these highly variable mitochondrial phenotypes.Future research is needed to determine whether the human gastrocnemius and soleus muscles have a similar phenotype to murine skeletal muscles.
Although the murine and human soleus have similar transcriptomes (Kho et al., 2006), we need further characterization of these muscles as well as oxidative muscle types that may vary in structure and function.
Skeletal muscles depend on mitochondria, comprising ~6% of the cell volume, that change during aging (Garnier et al., 2003).The gastrocnemius muscle has both type I slow-twitch muscle fibers and type II fast-twitch muscle fibers; type I fibers are more effective for endurance, whereas type II fibers better support short bursts of muscle activity (Garnier et al., 2003;Lin et al., 2018;Mukund & Subramaniam, 2020).In sarcopenia, the size and frequency of both types of fibers decrease, and type II fibers transition to type I (Romanick et al., 2013).In contrast, muscle atrophy from disuse does not change the fiber number, and there is a shift from type I fibers to type II (Romanick et al., 2013).Thus, age-related changes in mitochondrial shape may be due to sarcopenia-dependent alterations in fiber frequency, producing different mitochondrial phenotypes.
In our 3D morphologic data, we observed many variable muscle fibers both within a sample and between samples from animals of different ages; however, we could not distinguish the two fiber types, which requires quantitating primary myosin heavy chain proteins after separating them by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Galpin et al., 2012).In chicken muscle fiber subtypes, there are reportedly large differences in mitochondrial content and morphology; type I does not contain lipid droplets, so the presence of lipid droplets identifies type II fibers (Hosotani et al., 2021;Makida et al., 2022).However, methods for identifying fiber types have not been used with 3D-reconstructed aged murine skeletal muscles.
Using 3D reconstructions, we found that mitochondria in aged gastrocnemius muscles were smaller in volume, area, and perimeter (Figure 1a-x), and the mitochondria were less interconnected; however, in other tissues, we saw a less significant decrease in volume.
Increased fragmentation suggested decreased mitochondrial fusion, which is likely associated with the age-dependent decrease of OPA-1, a regulator of mitochondrial fusion.We also saw a decrease in the MCI in aged tissue, suggesting a reduction in mitochondrial networking; however, the mitochondrial shape may not change greatly because of the increased sphericity as they age (Figure 2).
MICOS proteins play key regulatory roles in mitochondrial structure and function (Hu et al., 2020;Li et al., 2016;Wang et al., 2020).By TEM 3D reconstructions, we found that the KD of Mitofilin, Chchd3, and smaller mitochondria (Figure 4), similar to the loss of Opa1, which results in changes in oxidative phosphorylation (Hu et al., 2020;Pereira et al., 2017;Zheng et al., 2019).Overall, mitochondria lacking the MICOS genes had characteristics similar to those of aged mouse skeletal muscle (Figures 1 and 2), and the similarity of the phenotypes suggests an association.Thus, changes in mitochondrial morphology due to aging may be caused by a lack of MICOS protein expression.This is supported by decreased Chchd3, Chchd6, Mitofilin, and Opa1 transcripts in aged muscle (Figure 3).However, it is possible that despite both skeletal and having lower transcript levels, only skeletal muscle shows reduced MICOS complex and OPA1 protein levels, demonstrating the need to measure age-dependent protein expression in the future.We found differences in the metabolome and lipidome across different tissue types, which suggests that the MICOS-complex-dependent response to aging may differ across cell types with different metabolomes.The tissue-specific loss of the MICOS complex requires further characterization.
The slight upregulation of Mitofilin in Drosophila (Figure 3j) suggested that its expression may mitigate the loss of other MICOS complex proteins during aging.In a Drosophila Mitofilin KD, there was an inverse correlation between the number of mitochondria and the loss of other MICOS complex components.This suggests different roles for Mitofilin in Drosophila versus mice and indicates the need for additional studies on Mitofilin.Furthermore, the different roles of the MICOS complex proteins need further study.Although there were some common phenotypes associated with all MICOS complex KD mice, Mitofilin and Chchd6 KDs did not reduce proton leak as did the Chchd3 KD (Figure 5f,p).Although all three of these components are necessary for the stabilization of the MICOS and SAM complexes, and, thus, cristae architecture (An et al., 2012;Darshi et al., 2011;Darshi et al., 2011;Ding et al., 2015;Li et al., 2016), further research is needed to determine whether compensatory increases for specific components of the complex, such as Mitofilin, prevent dysfunctional phenotypes after the loss of certain MICOS complex components.
To better understand the effect of the loss of the MICOS complex, we characterized the metabolome and the lipidome in aged mice and found differences not only between skeletal and cardiac tissue but also between soleus and gastrocnemius (Figure 6).This indicates the importance of understanding tissue-specific differences in mitochondrial structures we observed.Metabolomics revealed pathways that led to muscular dystrophy, whereas other pathways, such as NaR, rescued the muscles.Cardiac tissue had the greatest change in lipids in aged muscle.This suggests that lipid accumulation may be particularly important in cardiac tissue, consistent with the evidence that lipid metabolism is closely linked with pathology (Chung, 2021).
Conversely, in the soleus muscle, there was a decrease in sphingolipids, which may indicate that lipotoxicity is important in the aging soleus muscle.Sphingolipids are also associated with MERC regions prior to and during apoptosis (Mignard et al., 2020).Thus, further studies are needed to characterize the association of mitochondria with lipid droplets or interactions with the endoplasmic reticulum (ER) (Ilacqua et al., 2021) to determine whether the contact sites change during aging to protect against lipotoxicity.In addition, mass spectrometry imaging may reveal changes in the spatial distribution of these metabolites during aging (Hogan et al., 2023).In gastrocnemius tissue, phospholipids decreased (Figure 6), as did linoleic acid, which is important for membrane integrity (Cury-Boaventura et al., 2004).This suggests altered mitochondrial membrane viscosity in aged tissue, consistent with the age-dependent increase in viscosity that affects oxidative phosphorylation through modulation of supercomplexes (Dencher et al., 2007).Decreased membrane viscosity, which is associated with age-related pathologies, including Alzheimer's disease, may be a component of a pathomechanism (Kuter et al., 2016).One potential target of the decreased viscosity may be the MICOS complex, because lipids, including cardiolipin, interact with cristae (Ikon & Ryan, 2017).We only saw a change in cardiolipin in cardiac tissue; however, as this region had the least change in mitochondrial structure, cardiolipin may protect against MICOS-dependent loss of structure.Thus, the loss of cristae morphology may be associated with an increase in cardiolipin.
Therefore, aging may be associated with lipid-mediated membrane changes that affect the MICOS complex and modulate mitochondrial structure and function.
Age-related lipidomic and metabolomic changes may be due to age-dependent alterations in the MICOS complex.Many studies have analyzed the mitochondrial metabolome using mouse skeletal muscles (Bocca et al., 2018;de la Barca et al., 2017de la Barca et al., , 2019;;Garcia-Cazarin et al., 2011;Garnier et al., 2003;Wortel et al., 2017).We found that loss of Mitofilin affected cristae morphology (Figure 4ik), decreased oxidative phosphorylation (Figure 5a), and may have increased lipid and steroid synthesis, which may be important for the regulation of MERCs and cristae formation.We found an increase in tryptophan and methylhistidine metabolism (Figure 5j) and an increase in taurine metabolism and hypotaurine, a key sulfur-containing amino acid for fat metabolism.Loss of Opa1 also changes amino acid and lipid metabolism, similar to the loss of Mitofilin (Chao de la Barca et al., 2020;Sarzi et al., 2016;Wasilewski et al., 2012).Steroidogenesis, which makes the membrane less rigid, increased.The loss of Mitofilin, Chchd6, or Chchd3 resulted in a decrease in oxidative capacity (Figure 5a-f,k-p).However, increased steroid synthesis may allow the cell to recover bioenergetic functions, as steroids such as estrogen decrease membrane viscosity (Torres et al., 2018).Mitofilin is critical for maintaining cristae (Hessenberger et al., 2017;Tarasenko et al., 2017), as cristae junctions and contact sites fall apart with the loss of Mitofilin (Kondadi et al., 2020).Cells lacking Mitofilin may make steroids to help the membrane reconstitute broken cristae.Just as the loss of Opa1 results in more MERCs (Rowland & Voeltz, 2012), the loss of Mitofilin may increase phospholipids (Figure 5j) as a result of increased smooth MERCs, which are associated with lipid changes (Rieusset, 2018).This is supported by the fact that the biosynthesis of phosphatidylethanolamine and phosphatidylcholine, and the metabolism of arachidonic acid and sphingolipids increased with the loss of Mitofilin (Figure 5j).Because these phospholipids aggregate around MERCs and may shuffle into the ER, Mitofilin may be a key gene for regulating cristae morphology, with a novel role in regulating mitochondrial metabolism.
Mitofilin may be an important target to restore energy production.Loss of Mitofilin may lead to ER stress, which, via ATF4, activates amino acid transporters (Han et al., 2013) that then activate mTORC1.ER stress activates mTORC as a result of a decrease in glucose (Wortel et al., 2017).Critically, mTORC1 affects glucose homeostasis (Zhang, Wang, et al., 2021), which may lead to inefficient energy use and result in changes in autophagy.Therefore, if loss of Mitofilin increases mTORC1, this may explain why deletion of MICOS in Drosophila increases autophagy (Wang et al., 2020).Similarly, loss of Opa1 increases ER stress (Pereira et al., 2017), and loss of Mitofilin may increase amino acid catabolism.If ER stress activates amino acid transporters, branched-chain amino acids could increase ER stress, resulting in a positive feedback loop that affects the health of the cell, cellular energy, metabolism, and antioxidants.ER stress may also result in the poor performance and fragmentation of mitochondria (Figures 4 and 5), and loss of Mitofilin may result in the breakdown of protein pathways that regulate ER stress.Other amino acid pathways, such as homocysteine (Figure 5j), are involved in triglyceride uptake and increased intracellular cholesterol, suggesting that proteins like ATF4 (Wortel et al., 2017) and the MICOS complex Kozjak-Pavlovic, 2017;Li et al., 2016) are important during aging.In particular, the MICOS components may prevent mitochondrial fragmentation by blocking ER stress pathways in aging.We showed that genes for several MERC proteins were differentially regulated concomitantly with MICOS complex proteins during aging in Drosophila (Tables 1 and 2).Further studies are needed to better understand the role of MICOS in MERC formation and the relationship between smooth MERC and lipid synthesis.
Downregulation of Chchd3 is linked to type 2 diabetes (Eramo et al., 2020).In our metabolomics enrichment dataset (Figure 6t), loss of Chchd3 or Chchd6 in mouse myotubes resulted in a preference for alternative energy sources, such as lactate, lactose, and starches.Supplementation of healthy myotubes with galactose leads to a 30% increase in oxidative capacity (i.e., OCR) due to an increase in AMPK phosphorylation and cytochrome c oxidase (COX) activity, thereby forcing cells to become more oxidative to maintain ATP levels (Martin et al., 2021).In our tissues, as oxidative metabolism decreased, anaerobic metabolism and lactate levels increased, forcing cells to produce ATP by anaerobic glycolysis.
However, long and high-level exposure to D-galactose generates free radicals, which alter MERCs that result in mitochondrial dysfunction and induce aging (Barja, 2014;Kandlur et al., 2020).This is the likely explanation for mitochondrial fragmentation in aged samples and loss of the MICOS complex, which should be investigated further.
In conclusion, we present a quantitative evaluation of mitochondrial morphology in mouse skeletal muscle and cardiac tissue using 3D reconstructions, with TEM studies of cell lines to characterize other factors such as cristae architecture.We found structural changes, including changes in gross 3D mitochondrial structure, which produced functional differences upon loss of MICOS proteins.
Similar changes in mitochondrial morphology were observed in aged muscles and after the loss of MICOS proteins in mouse skeletal muscle, and we found that MICOS mRNA transcripts decreased with age.We also found that the metabolome and lipidome were heavily altered in aged muscles, suggesting a role for the MICOS complex in membrane integrity during aging.In vivo Drosophila models demonstrated the importance of understanding the tissue-specific responses to aging, the roles of individual components in the MICOS complex, and potential MICOS complex-MERC pathway interactions that may regulate mitochondrial structure and function.Despite a link between aging and the loss of Opa1 (Tezze et al., 2017;Varanita et al., 2015), little is known about the role of the MICOS complex in aging.A reduction in MICOS proteins could result in changes in mitochondrial architecture and loss of integrity; thus, we need therapies to restore MICOS proteins and Opa1 lost during aging to mitigate the deleterious effects of mitochondrial dysfunction.Although knockouts can determine the role of specific proteins in mitochondrial dynamics, few studies have attempted to restore MICOS proteins in mitochondria (Liu et al., 2020;Zheng et al., 2019).Our results established a relationship between the MICOS complex and aging; thus, further studies using 3D reconstruction could elucidate the link between sarcopenia, the MICOS complex, and the role of mitochondria in aging and certain diseases.

| Animal care and maintenance
All procedures for the care of mice were in accordance with humane and ethical protocols approved by the University of Iowa Animal Care and Use Committee (IACUC), or the University of Washington IACUC, following the National Institute of Health (NIH) Guide for the Care and Use of Laboratory Animals, as described previously (Pereira et al., 2017).Therefore, all studies are performed in accordance with the ethical standards established in the 1964 Declaration of Helsinki and its later amendments.All experiments used WT male C57Bl/6J mice housed at 22°C on a 12-h light, 12-h dark cycle with free access to water and standard chow.Mice were anesthetized with 5% isoflurane/95% oxygen and followed by cervical dislocation.

| RNA extraction and RT-qPCR
Total RNA was extracted from tissue using TRIzol reagent (Invitrogen; cat #15596026), purified with the RNeasy kit (Qiagen Inc; cat #74004), and quantitated by the absorbance at 260 and 280 nm using a NanoDrop 1000 (NanoDrop products) spectrophotometer.Total RNA (~1 μg) was reverse transcribed using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosciences; cat #4368814), followed by real-time quantitative PCR (qPCR) reactions using SYBR Green (Life Technologies; cat #S7563) (Boudina et al., 2007).Triplicate technical replicates for qPCR (~50 ng) in a 384-well plate were placed into ABI Prism 7900HT instrument (Applied Biosystems) programmed as follows: 1 cycle at 95°C for 10 min; 40 cycles of 95°C for 15 s; 59°C for 15 s, 72°C for 30 s, and 78°C for 10 s; 1 cycle of 95°C for 15 s; 1 cycle of 60°C for 15 s; and 1 cycle of 95°C for 15 s.Data were normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh), and results are shown as fold changes.qPCR primers were designed using Primer Blast or were previously published sequences (Pereira et al., 2017), as shown in Table 3.

| Experimentally evolved Drosophila populations, RNA extraction, RT-qPCR
Groups of experimentally derived Drosophila melanogaster were selected using different generation cycles for hundreds of generations to produce populations with different patterns of aging and longevity.In the control populations, termed CO 1-5 , genetically diverse populations (census ~2000 per were maintained on a 28-day generation cycle.From these CO populations, a new population was maintained on a 9-day cycle, ACO 1-5, for hundreds of generations, as described in Chippindale et al. (1997).This resulted in an ACO population that evolved to reproduce earlier, develop more rapidly and die much earlier than the CO population (Burke et al., 2016).The accelerated aging in the ACO flies is associated with differences in genetics (Graves et al., 2017), patterns of gene expression (Barter et al., 2019), and the metabolome (Phillips et al., 2022).
We refer to the ACO populations as "aged flies" and the CO population as "control flies." We used RT-qPCR to compare gene expression for the genes as shown in Table 2 in cardiac tissue from 21-day-old flies from the ACO and CO populations.We chose Day 21 based on demographic data (Burke et al., 2016) and whole-body transcriptomic data (Barter et al., 2019) showing large differences between the two populations at that age.For each group, we collected heart tissue using the following protocols.On Day 21 from eggs, female fruit flies from each of the 10 ACO and CO populations were anesthetized using Fly Nap (Carolina), a triethylamine-based anesthetic, for about 1 min or until no movement was detected.Flies were dissected in oxygenated artificial hemolymph to expose the cardiac tubes, and the abdomens were opened to remove guts/intestines, fat, and ovaries.Although it was not possible to fully remove fat and pericardial cells from the cardiac tube without damaging it, excess fat and pericardial cells were carefully suctioned away from the cardiac tube in the exposed hearts (Vogler & Ocorr, 2009).
Three biological replicates were collected for each ACO and CO

| Drosophila strains and genetics
For genetic crosses, flies were grown on yeast corn medium (Katti et al., 2017(Katti et al., , 2022) ) at 25°C.The Mef2-Gal4 strain served as a control within their respective genetic backgrounds.Mef2-Gal4 was crossed to the genetic background w1118 to generate the UAS-RNAi knockdown lines per previous protocols (Ranganayakulu et al., 1996).Mef2 encodes the transcription factor myocyte enhancer factor-2, which regulates muscle development.Gal4 is a transcriptional activator from yeast commonly used to drive gene

Gene Primers
TA B L E 3 qPCR primers used.
expression in Drosophila when cloned upstream of a promoter region.In this study, Mef2-Gal4 refers to a transgenic fly line expressing Gal4 under the control of the Mef2 promoter (Ranganayakulu et al., 1996).Male and female flies were analyzed collectively, as there were no discernible sex differences in mitochondrial morphology in WT muscles.Mef2-Gal4 (III) was utilized for muscle-specific knockdown of MICOS genes.The mitochondrial network was visualized using UAS-mito-GFP, located on the second chromosome (BS# 8442).For muscle-specific knockdown of MICOS genes, UAS-RNAi transgenic RNAi project (TRiP) lines were used for UAS-Chchd3 RNAi (BS#51157), UAS-Mitofilin RNAi (BS# 63994), UAS-QIL1 RNAi

| Mitochondrial measurements
Mitochondrial measurements have been described previously (Haas, 2019;Lam et al., 2021), and images were analyzed using the NIH ImageJ software (https:// imagej.net).Individual mitochondria were outlined on 2D light microscopic images using the freehand tool provided by the software, and their area and aspect ratio (the ratio of the major axis to the minor axis) were calculated using ImageJ.For each data set, three animals were analyzed, and these analyses were part of three independent experiments conducted to gather quantifiable data.The number of mitochondria was counted across every three-sarcomere segment.
We incubated 2.5% relevant CRISPR, 2.5% RNAiMax (ThermoFisher Scientific; cat #13778075), and 95% Opti-MEM (Gibco; cat #31985070) in a tube for 20 min.Cells were washed twice with PBS after removal of the medium; then, 800 μL of OPT-MEM (Gibco; cat #31985062) and 200 μL of the CRISPR mixture were added to each well and were run in triplicate.Cells were incubated for 4 h at 37°C, 1.0 mL of DMEM medium was added, and cells were incubated overnight.The myotubes were then washed with PBS, and the medium was replaced.Experiments were performed between 3 and 7 days after knockout for a total of 6 days of differentiation.
Male mice were anesthetized with 5% isoflurane, the hair and skin were removed, and the hindlimbs were incubated in 2% glutaraldehyde with 100 mM phosphate buffer for 30 min.Gastrocnemius muscles were dissected, cut into 1 mm 3 cubes, and incubated in 2.5% glutaraldehyde, 1% paraformaldehyde, and 120 mM sodium cacodylate solution for 1 h.Tissues were washed three times with 100 mM cacodylate buffer at room temperature before immersion in 3% potassium ferrocyanide and 2% osmium tetroxide for 1 h at 4°C, then treated with 0.1% thiocarbohydrazide, 2% filtered osmium tetroxide for 30 min, and left overnight in 1% uranyl acetate at 4°C.
Between each step, three deionized water washes were performed.
The following day, samples were immersed in 0.6% lead aspartate solution for 30 min at 60°C and dehydrated in graded concentrations of acetone.Dehydration was for 5 min each in 20%, 50%, 70%, 90%, 95%, and 100% acetone.Tissues were impregnated in Epoxy Taab 812 hard resin, then embedded in fresh resin, and polymerized at 60°C for 36-48 h.Once polymerization was complete, blocks were sectioned for TEM to identify areas of interest, trimmed to 0.5 mm × 0.5 mm, and glued to aluminum pins.The pins were run on an FEI/Thermo Scientific Volumescope 2 SEM, a state-of-the-art SBF imaging system, yielding 300-400 10 μm by 10 μm ultrathin (90 nm) serial sections, as per previous techniques (Garza-Lopez et al., 2022).All sections were collected onto formvar-coated slot grids (Pella), stained, and imaged as described previously (Garza-Lopez et al., 2022;Hinton et al., 2023;Neikirk et al., 2021).

| Quantification of TEM micrographs and parameters using ImageJ
Quantification of TEM images was performed as described previously using NIH ImageJ software (Hinton et al., 2023;Lam et al., 2021).Cells were divided into four quadrants, and two quadrants were selected randomly for complete analysis.Individuals blinded to the experimental design measured a minimum of 10 cells using three analyses to obtain accurate and reproducible values.
When there was variability, we assigned 30 cells per individual to reduce the variability.

| Measurement of OCR using seahorse
We used the Seahorse XF24 extracellular flux (XF) bioanalyzer (Agilent Technologies/Seahorse Bioscience) to measure cellular respiration.Cells were plated at a density of 2 × 10 4 per well and differentiated.After 3 days of differentiation, Opa-1, CHCHD3, CHCHD6, or Mitofilin genes were knocked out as described above.
Three days after knockout, the medium was changed to XF-DMEM (Agilent-103680), and cells were kept in a non-CO 2 incubator for 60 min.The basal OCR was measured in XF-DMEM.Oxygen consumption was measured after the addition of oligomycin (1 μg/mL), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; 1 μM), rotenone (1 μM), and antimycin A (10 μM) (Pereira et al., 2017;Wende et al., 2015).Cells were then switched to glucose-free XF-DMEM and kept in a non-CO 2 incubator for 60 min for the glycolysis stress test.Seahorse experimental data used triplicate Seahorse plates.Three independent experiments were performed with four to six replicates for each time and each condition, and representative data from the replicates are shown.

| Segmentation and quantification of 3D SBF-SEM images using Amira
The mitochondria analyzed were the IMF mitochondria located between myofibrils that are arranged in pairs at the z-band of each sarcomere, with 2D elongated tubular shapes (Vendelin  , 2005).For each ROI across the two age groups, we analyzed 300 slices at 50 μm intervals in transverse intervals.For 3D reconstruction, SBF-SEM images were segmented manually using Amira software (Thermo Scientific) as described previously (Garza-Lopez et al., 2022;Hinton et al., 2023).All serial sections (300-400 slices) were loaded onto Amira, and structural features were traced manually on sequential slices of micrograph blocks.Structures in mice were collected from 30 to 50 serial sections that were then stacked, aligned, and visualized using Amira to make videos and quantify volumetric structures.An average of 500 total mitochondria across four ROIs from three mice was collected for quantification.For the 3D reconstruction of myotubes, approximately 20 mitochondria from a minimum of 10 cells were collected.Quantification of SBF-SEM images was performed as described previously (Garza-Lopez et al., 2022) using the Amira software.
Samples were extracted in −80°C 2:2:1 methanol/acetonitrile/water that contained a mixture of nine internal standards (d 4 -citric acid, 13 C 5 -glutamine, 13 C 5 -glutamic acid, 13 C 6 -lysine, 13 C 5 -methionine, 13 C 3 -serine, d 4 -succinic acid, 13 C 11 -tryptophan, standard verified peaks and retention times, to profile the metabolites and to compare metabolite peaks in each sample against an in-house library of standards.For these standards, we analyzed retention times and fragment ions for each, with fragment ions for both the target peak and two confirming ions.For the samples, we identified metabolites that matched both retention times and the three fragment ions.TraceFinder was also used for GC-MS peak integration to obtain peak areas for each metabolite.After this analysis, we used previously described protocols (Li et al., 2017)  The mobile phase gradient started at 80% solvent B, decreased to 20% solvent B over 20 min, returned to 80% solvent B in 0.5 min, and was held at 80% for 7 min (Cantor et al., 2017).From there, the mass spectrometer was operated in the full-scan, polarity-switching mode for 1-20 min, spray voltage set to 3.0 kV, capillary heated at 275°C, and HESI probe heated at 350°C.The sheath gas flow, auxiliary gas flow, and sweep gas flow were 40 units, 15 units, and 1 unit, respectively.We examined an m/z range of 70-1000, the resolution was set at 70,000, the automatic gain control (AGC) target at 1 × 10 6 , and the maximum injection time was set to 200 ms (Li et al., 2017).
TraceFinder 4.1 software was used for analysis, and metabolites were identified based on an in-house library.Drift was corrected for as described above (Li et al., 2017).Data were normalized, and further visualization and analysis were performed on MetaboAnalyst 5.0 (Chong et al., 2018).

| Analyzing metabolomic data for MICOS KO
Metabolomic analysis was performed as described previously (Phillips et al., 2022)
Frozen tissues from aged mice were weighed, ground in liquid nitrogen in a cryo-mill (Retsch) at 25 Hz for 45 s, extracted in 40:40:20 acetonitrile:methanol:water +0.5% formic acid +15% NH 4 HCO 3 (Lu et al., 2018) in 40 μL of solvent per 1 mg of tissue, vortexed for 15 s, and incubated on dry ice for 10 min.Samples were centrifuged at 16,000 × g for 30 min, transferred to new microcentrifuge tubes, and then centrifuged again at 16,000 × g for 25 min to remove residual debris.
The supernatants were transferred to clean tubes and centrifuged again for 5 min at 15,000 × g at 4°C to remove any remaining particulates.For LC-MS lipidomic analysis, 60 μL of the sample extracts was transferred to mass spectrometry vials.

F
Decreased mitochondrial size and volume in the gastrocnemius, soleus, and cardiac muscle of aged mice in SBF-SEM 3D reconstructions.(a, b) Representative SBF-SEM orthoslices for male murine gastrocnemius, (c, d) soleus, and (e, f) cardiac tissues.(a′, b′) 3D reconstructions of mitochondria (various colors) in gastrocnemius, (c′, d′) soleus, and (e′, f′) cardiac tissues from 3-month-old and 2-year-old mice overlaid on ortho slices.(a″, b″) Pseudo-colored individual mitochondria in gastrocnemius, (c″, d″) soleus, and (e″, f″) cardiac tissues identify micro-level changes.(g-x) Quantification of 3D reconstructions, with each dot representing the average for all mitochondria quantified for one mouse.(g) Mitochondrial volume in the gastrocnemius muscle from 3-month-old and 2-year-old mice and (h) mitochondrial volume distributed as the percent of total mitochondria to visualize relative heterogeneity.(i) Mitochondrial 3D area in gastrocnemius muscle from 3-month-old and 2-year-old mice and (j) mitochondrial area distributed as the percent of total mitochondria to visualize relative heterogeneity.(k) Mitochondrial perimeter in gastrocnemius muscle from 3-month-old and 2-year-old mice and (l) mitochondrial perimeter distributed as the percent of total mitochondria to visualize relative heterogeneity.These quantifications are also displayed in (m-r) soleus and (s-x) cardiac tissues.Cartoon representations of metrics to calculate (y) mitochondrial volume, perimeter, and perimeter.Approximately 550 mitochondria were analyzed for each tissue type and age cohort (n = 3 mice per age cohort).Significance values: **** represents p ≤ 0.0001.

F
I G U R E 2 SBF-SEM 3D reconstruction in gastrocnemius, soleus, and cardiac muscle of aged mice showed altered mitochondrial networks.Representative examples of 3D reconstruction of mitochondria in (a) gastrocnemius, (b) soleus, and (c) cardiac tissue of 3-monthold and 2-year-old mice organized by volume to show the mitochondrial phenotypes.(d) Mitochondrial complexity index (MCI), analogous to sphericity, in the gastrocnemius muscle from 3-month-old and 2-year-old mice, and (e) MCI distributed as the percent of total mitochondria to visualize relative heterogeneity.(f) Sphericity in the gastrocnemius muscle from 3-month-old and 2-year-old mice and (g) mitochondrial sphericity distributed as the percent of total mitochondria to visualize relative heterogeneity.These quantifications are also displayed in (h-k) soleus and (l-o) cardiac tissues.Cartoon representations of metrics to calculate (p) MCI and (q) sphericity.Approximately 550 mitochondria were analyzed for each tissue type and age cohort (n = 3 mice per age cohort).Significance values: *p ≤ 0.05; ****p ≤ 0.0001.reticulum contact (MERC) genes in experimentally evolved Drosophila populations subjected to hundreds of generations of accelerated aging versus control flies

Note:
The fold change compares expression in the ACO population versus the CO population.Fold-change values of >1.0 indicate that the gene is expressed more in the CO populations (28-day cycle) than in the ACO populations (9-day cycle) at Day 21.Fold-change values of <1.0 indicate that the gene is expressed more in the ACO populations than in the CO populations at Day 21.Values of ~1.0 indicate no difference between the populations.Statistically significant fold-change values are denoted by an asterisk (*).
aging.The fold change compares expression in the ACO population versus the CO population.Fold-change values of >1.0 indicate that the gene is expressed more in the CO populations (28-day cycle) than in the ACO populations (9-day cycle) at Day 21.Fold-change values of <1.0 indicate that the gene is expressed more in the ACO populations than in the CO populations at Day 21.Values of ~1.0 indicate no difference between the populations.Statistically significant fold-change values are denoted by an asterisk (*).
found that loss of Opa1 or Mitofilin in myotubes decreased basal OCR (Figure5a,b) and decreased ATP-linked, maximum, and reserve capacity OCR (Figure5c-e).Although Opa1 knockout myotubes exhibited a decrease in proton leak, which represents protons that go from the mitochondria to the matrix without producing ATP, Mitofilin knockouts Mitofilin plays a critical role in regulating amino acid metabolism and steroidogenesis (Figure 5i,j).Upregulation of steroidogenesis pathways may result from the increased fluidity of membranes caused by Mitofilin (Kitajima & Ono, 2016; Torres et al., 2018).

F
Knockout of Opa1, Mitofilin, Chchd3, or Chchd6 in myotubes resulted in structural changes of mitochondria and cristae in TEM and 3D reconstruction.(a-e) Representative images of mitochondria and cristae from myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (f) Mitochondrial area in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (g) Circularity index, measuring the roundness and symmetry of mitochondria, in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (h) The number of mitochondria in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (i) The number of individual cristae in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (j) Cristae scores measuring the uniformity and idealness of cristae in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (k) The surface area of the average cristae in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (l-p) Representative images showing 3D reconstructions of mitochondria in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (q) Mitochondrial 3D length in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (r) Mitochondrial volume on a log scale in myotubes of Opa1, Mitofilin, Chchd3, and Chchd6 knockout mice compared to WT. (s-v) Altered Drosophila mitochondrial structure resulting from loss of the MICOS complex and mitochondrial proteins.(s) Actin-mitochondria staining for Drosophila flight tissue in knockouts of MICOS complex and mitochondrial proteins.TEM quantification of mitochondrial changes in Drosophila flight tissue for (t) mitochondrial area, (u) circularity, and (v) quantity per sarcomere upon knockout of MICOS complex and mitochondrial proteins.Significance values: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001.Dots represent the number of mitochondria quantified.In Opa1, Chchd3, and Chchd6 knockouts, there was a decrease in basal, ATP-linked, maximum, and reserve capacity OCR compared with the control (Figure5k-o).Although proton leak OCR decreased in Opa1 and Chchd3 knockout myotubes (Figure5p), there was no significant difference between the control and Chchd6.The decrease in OCR may be attributed to smaller, fragmented mitochondria; mitochondrial density decreases as fragmentation targets them for autophagy(Tezze et al., 2017;Wang et al., 2020).Together, these results showed that MICOS and Opa1 are essential for the normal respiration of muscle tissue.We also measured the effect of knocking out the genes for Chchd3 and Chchd6 in skeletal muscle myotubes on bioenergetic metabolism.PCA revealed distinct populations in | 11 of 26 VUE et al. the control and the Chchd3 and Chchd6 knockouts, similar to what tabolism (Figure 5s,t).Loss of Opa1 typically favors fatty acid synthesis, so the results showing increased carbohydrate metabolism differ from previous Opa1 knockout responses (Chao de la Barca FigureS2), we found significant metabolic changes in all three tissue types.These changes included various processes, including NAD + metabolism, linolenic acid metabolism, porphyrin synthesis, heme biosynthesis, and glycine and lysine metabolism (FigureS2D-I).Notably, we observed an accumulation of cholic acid in aged soleus and gastrocnemius muscles (FigureS2A,B,D,E), an inducer of muscle atrophy(Abrigo et al., 2021).Amino acid metabolism was dysregulated in aged tissues across all types.Lipidomic profiling of young versus aged tissues (Figure6b,d,f; TablesS1-S4) revealed changes in lipid classes (FigureS2J-L) and

F
I G U R E 5 Knockout of the MICOS complex in myotubes resulted in changes in oxygen consumption rates and metabolomics.(a) OCR in myotubes of Opa1 and Mitofilin knockout mice compared to WT.(b) Basal OCR, the net respiration rate once non-mitochondrial respiration has been removed, in myotubes of Opa1 and Mitofilin knockout mice compared to WT. (c) ATP-dependent respiration, shown from intervals 4-7 in the OCR graphs, was determined by the addition of oligomycin (an inhibitor of respiration) in myotubes of Opa1 and Mitofilin knockout mice compared to WT.(d) Maximum OCR represented by the peak from intervals 7-11 once non-mitochondrial respiration was deducted, in myotubes of Opa1 and Mitofilin knockout mice compared to WT. (e) The reserve capacity, the difference between basal OCR and maximum OCR, in myotubes of Opa1 and Mitofilin knockout mice compared to WT. (f) Proton leak, representing non-phosphorylating electron transfer, in myotubes of Opa1 and Mitofilin knockout mice compared to WT. (g-j) Metabolomic analysis in Mitofilin knockout mice.(g) Metabolite PCA and (h) T-test comparing myotubes for control versus Mitofilin knockout mice.(i) Heatmap showing the relative abundance of ions and (j) enrichment analysis of metabolites, which links similarly functioning metabolites with the relative abundance for the Mitofilin knockout.(k) OCR in myotubes of Chchd3, Chchd6, and Opa1 knockout mice compared to WT. (l) Basal OCR in myotubes of Chchd3, Chchd6, and Opa1 knockout mice compared to WT. (m) ATP-linked respiration in myotubes of Chchd3, Cchchd6, and Opa1 knockout mice compared to WT. (n) Maximum OCR in myotubes of Chchd3, Chchd6, and Opa1 knockout mice compared to WT. (o) The reserve capacity in myotubes of Chchd3, Chchd6, and Opa1 knockout mice compared to WT. (p) Proton leak in myotubes of Opa1, Chchd3, and Chchd6, knockout mice compared to WT. (q-t) Metabolomic analysis in Chchd3 or Chchd6 knockout mice.(q) Metabolite PCA and (r) ANOVA test comparing control to myotubes of Chchd3 and Chchd6 knockout mice (s) Heatmap showing the relative abundance of ions for control and (t) enrichment analysis metabolites for Chchd3 and Chchd6 knockout mice.Significance values: *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; ****p ≤ 0.0001.For Seahorse analysis, n = 6 plates for experimental knockouts and n = 16 for controls.| 13 of 26 VUE et al.
Chchd6 in muscles resulted in fragmentation, disrupted cristae, F I G U R E 6 Metabolomics analysis and lipidomic profiling revealed metabolic dysregulation and disruptions in lipid classes with age in gastrocnemius, soleus, and cardiac muscles.(a) Metabolic heatmap showing the relative abundance of metabolites and (b) the lipidome in young and aged gastrocnemius, (c, d) soleus, and (e, f) cardiac samples.For each tissue and metabolite in the heatmaps, the aged samples were normalized to the median of the young samples and then log 2 transformed.Significantly different lipid classes represented in the figures are those with adjusted p-values < 0.05 (note: p-values were adjusted to correct for multiple comparisons using an FDR procedure) and log fold changes greater than 1 or less than −1.Young, n = 4; aged, n = 4.For all panels, error bars indicate SEM, ** indicates p < 0.01; and *p < 0.05, calculated with Student's t-test.

Fisher's least
significant difference multiple comparison tests were also used.PCA uses score plots to provide an overview of variance for the principal components.Heatmaps separate hierarchical clusters leading to progressively larger clusters.Clusters are based on similarity using Euclidean distance and Ward's linkage to minimize the clustering needed.Metabolite set enrichment analysis, which determines whether a set of functionally related metabolites is altered, identifies consistent changes across many metabolites with similar roles.Overrepresentation analysis determines whether a group of compounds is overrepresented compared to chance and whether a group of metabolites has similar changes.In this analysis, the fold enrichment was calculated by dividing the observed hits by the expected metabolites.The expected number of hits was calculated by MetaboAnalyst 5.0.GraphPad Prism software was used for statistical analysis with data expressed as mean ± standard deviation, and one-tailed p-values ≤ 0.01 were considered significant.
Extracts were analyzed within 24 h by LC-MS, based on hydrophilic interaction chromatography (HILIC) coupled to the Orbitrap Exploris 240 mass spectrometer (Thermo Scientific)(Wang et al., 2019).The LC separation was performed on an XBridge BEH Amide column (2.1 × 150 mm, 3.5 μm particle size; Waters).Solvent A was 95%:5% H 2 O:acetonitrile with 20 mM ammonium acetate and 20 mM ammonium hydroxide, and solvent B was 90%:10% acetonitrile:H 2 O with 20 mM ammonium acetate and 20 mM ammonium hydroxide.The gradient was 0 min, 90% B; 2 min, 90% B; 3 min, 75% B; 5 min, 75% B; 6 min, 75% B; 7 min, 75% B; 8 min, 70% B; 9 min, 70% B; 10 min, 50% B; 12 min, 50% B; 13 min, 25% B; 14 min, 25% B; 16 min, 0% B; 18 min, 0% B; 20 min, 0% B; 21 min, 90% B; and 25 min, 90% B. The parameters for the LC analysis were a flow rate of 150 mL/min, column temperature of 25°C, injection volume of 5 μL, and autosampler temperature of 5°C.For the detection of metabolites, the mass spectrometer was operated in both negative and positive ion modes.The parameters for the MS analysis were a resolution of 180,000 at m/z 200, AGC target at 3 × 10 6 , maximum injection time of 30 ms, and a m/z scan range of 70-1000.Raw LC/ MS data were converted to mzXML format using the command line "msconvert" utility(Adusumilli & Mallick, 2017).Data were analyzed via the EL-MAVEN software version 12.4.18| Lipidomics of aged samples4.18.1 | Tissue homogenization and extraction of lipidsTissues were ground as described in the section above.The homogenate was mixed with 1 mL of extraction buffer containing isopropyl alcohol (IPA)/H 2 O/ethyl acetate (30:10:60, v/v/v) and Avanti Lipidomix Internal Standard (diluted 1:1000) (Avanti Polar Lipids, Inc.).Samples were vortexed and homogenized twice in a VWR Bead Mill at 6000 × g for 30 s.The samples were then sonicated for 5 min and centrifuged at 15,000 × g for 5 min at 4°C.The upper phase was transferred to a new tube and kept at 4°C.The tissue pellet was again extracted using 1 mL of IPA/H 2 O/ethyl acetate extraction buffer, vortexed, homogenized, sonicated, and centrifuged as described above.The supernatants from both extractions were combined, and the organic phase was dried under liquid nitrogen gas.
Drosophila RT-qPCR flight tissue data for knockdowns of the indicated genes for MICOS, mitochondrial, and MERCs proteins during aging.
Guide RNA and plasmids used.
using the web service MetaboAnalyst 5.0 (https:// www.metab oanal yst.ca/ Metab oAnal yst/ Modul eView.xhtml , last accessed on 8 February 2022) that combines machine learning methods and statistics to group data using PCA, heat mapping, metabolite set enrichment analysis, and statistical analysis.One-way ANOVA and