Loss of HtrA2/Omi protease activity induces mitonuclear imbalance and sarcopenia via differential regulation of mitochondrial biogenesis

Cellular homeostasis requires tight coordination between nucleus and mitochondria, organelles that each possess their own genomes. Disrupted mitonuclear communication has been found to be implicated in many aging processes. However little is known about mitonuclear signaling regulator in sarcopenia which is a major contributor to the risk of poor health-related quality of life, disability and premature death in older people. HtrA2/Omi is a mitochondrial protease and play an important role in mitochondrial proteostasis. HtrA2 mnd2(-/-) mice harboring protease-decient HtrA2/Omi Ser276Cys missense mutants exhibit premature aging phenotype. Additionally, HtrA2/Omi has been established as a signaling regulator in nervous system and tumors. We therefore asked whether HtrA2/Omi participates in mitonuclear signaling regulation in aging muscle. Using We employed bioinformatics and as gene differentially we gastrocnemius and determined , mitochondrial

Herein, we set out to investigate the role of HtrA2/Omi protease activity in skeletal muscle using HtrA2 mnd2(-/-) mice. The data demonstrate denervation-independent sarcopenia induced by HtrA2/Omi protease de ciency. We also tested whether HtrA2/Omi protease de ciency impacts mitochondrial function by destroying mitochondrial proteostasis and promoting ROS production. Contrary to the expectation, we failed to observe neither the upregulation of UPR mt and mitohormesis related genes, nor increased ROS level in HtrA2 mnd2(-/-) muscle. Thus we followed a data-driven approach using a sarcopenia gene expression microarray dataset, which provides an unbiased view of biological functions and genes expression correlated with HtrA2/Omi in sarcopenia without a previous de ned hypothesis.
The analysis were then biologically validated in HtrA2 mnd2(-/-) mice. The results demonstrated that loss of HtrA2/Omi protease activity results in mitonuclear imbalance and differential regulation of mitochondrial biogenesis genes, suggesting a role of HtrA2/Omi protease activity in mitonuclear coordination in sarcopenia.

Animals
The heterozygous mnd2 (HtrA2 mnd2(+/-) ) mice (Stock Number:004608) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). HtrA2 mnd2(+/-) mice were interbred to generate populations containing all three genotypes (wild type, heterozygous, and homozygous). All animals were maintained under a 12hour light/dark cycle with freely available food and water. 30 days homozygous male mice gastrocnemius muscles were used in this study. All procedures were carried out under the Guideline of National Institutes of Health, and approved by the Institutional animal Care and Use Committee of Jilin University.

Genotyping
Experimental offspring mice were identi ed by PCR-restriction enzyme analysis according to the protocol provided by JAX mice. Brie y, the Ser276Cys mutant allele of HtrA2 is detected by ampli cation of a 500 base pair (bp) fragment containing the exon-3 from nuclear DNA with the primers mnd2 (see supplementary table S1). Digestion genomic DNA with AluI produces a 244 bp product from the homozygous mice instead of 171 bp and 73 bp fragments from the wild-type. PCR of the digestion fragments was carried out and the products were separated on agarose gels and stained with GeneGreen (Tiangen, Beijing, China) [34].

Muscle contractile force measurements
Ex-vivo gastrocnemius muscle force measurements were performed as described previously [35]. Brie y, mice were anesthetized, and electrodes wires were placed on the sciatic nerve. The distal tendon of gastrocnemius muscle was mounted on a force transducer (Techman BL-420N, China) and kept in moisture with oxygenized (95% O2 and 5% CO2) Krebs-Henseleit solution (pH 7.6) at 30 °C. After optimizing the stimulation conditions and muscle length showing maximal isometric twitch tension, the muscle was allowed to rest for 5 min. Peak twitch tension were measured in single twitch (stimulus), and maximal tetanic tension was measured under series of stimuli. 30 s pause between stimuli was performed to avoid effects due to fatigue. After force measurements, animals were killed by cervical dislocation and muscles were dissected, weighted and stored for further experiments. Muscle contractile force was expressed in absolute and values normalized to physiological cross-sectional area (pCSA), which was calculated using the equation (1) [36]: In the equation M is muscle mass, θ is pennation angle (default as zero), ρ is fresh muscle density (0.001056 g/mm 3 ) [37][38][39], and L f is fascicle length. The fascicle length of 6.6 mm was used in the calculation [38, 40,41].

Histological analysis
The gastrocnemius muscles were xed with 4% paraformaldehyde and embedded into para n. After dehydration and rehydration, 4 µm coronal and sagittal sections were stained with haematoxylin and eosin (H&E). For Sirius red staining, the coronal sections was stained by Sirius red for 1 h, the other steps were the same as the H&E procedure. Finally, the morphology characteristics of the muscles were observed via microscope (ECLIPSE Ci-L, Nikon, Japan).

Muscle bers morphological quantitative analysis
Morphological Quantitative Analysis of muscle bers were performed with ImageJ software 1.52a (http://rsb.info.nih.gov/ij/) using images of H&E stained cross-sections. Brie y, Open-CSAM, an ImageJ macro supporting quantitative analysis of muscle bers, was built in ImageJ according to a recently published work by Thibaut et al. [42]. H&E images were converted to grayscale in 8 bit with Photoshop software (Adobe Inc., San Jose, CA) before imported into Open-CSAM. Finally, the parameters of muscle bers analyzed by Open-CSAM were exported in excel. Equations of Round score and Aspect Ratio were as follow (2-3): In the equation (2), A is cross-section area (CSA) of muscle ber, P is perimeter of muscle ber, and in the equation (3)

Mitochondrial copy numbers
Mitochondrial copy numbers were measured by absolute quanti cation RT-PCR as previously described [43]. Brie y, genomic DNA from gastrocnemius muscle tissues was isolated using the DNeasy Blood and Tissue Kit (QIAGEN, Hilden, Germany), following the manufacturer's instructions. Quanti cation of mtDNA copy number was performed in triplicates by qRT-PCR. mNADH1, mCYTB, mATP6 and mCOX2 were used as mtDNA markers, and β-globin nuclear intron was used as nDNA marker. Relative gene expression was normalized to that of theβ-globin gene (ΔCT) in each sample, and was normalized to the WT group. The primer sequences are listed in in Additional le 1: Table S1.

Measurement of ATP generation
Muscle tissue ATP content was measured using a bioluminescent assay kit (Beyotime, China) according to the manufacturer's instruction. Brie y, fresh tissue lysates were collected and then centrifuged at 12,000 g at 4℃ for 10 min, the supernatant were then added into detection reagent. ATP content was measured using a multimode microplate reader (FLUOstar Omega, BMG Labtech, Germany), the relative luminescence unit (RLU) obtained were normalized to the protein concentration. The value of the WT control was set to 1.

Measurement of ROS generation
Fresh muscles were used to prepare a 10% (w/v) PBS homogenate. After centrifuging at 12,000 g for 10 min at 4 °C, the supernatant was collected and used to detect ROS and protein content. 90 μL of the supernatant and 10 μL of 1 mM DCFHDA (Beyotime, China) were added to each well of a 96-well plate, After incubating at 37 °C for 30 min in the incubator, the uorescence was measured at 488 nm for excitation and 525 nm for emission using a multimode microplate reader (FLUOstar Omega, BMG Labtech, Germany). The results were calculated as the relative uorescence unit (RFU)/μg protein. The value of the WT control was set to 1.

Western blot analysis
The gastrocnemius muscles were segregated and placed in liquid nitrogen for rapid freezing. Frozen samples were pulverized and lysed with 500 µl of the RIPA buffer (Beyotime, China). The lysates were ultrasonicated for 6*3 sec on ice and placed on ice for 45 min. Then the lysates were centrifuged at 4,500 g for 15 min at 4°C and the precipitate was discarded. Protein concentrations in the supernatants were determined using the Bradford reagent (Bio-Rad, Hercules, CA). The protein samples (10 µg) were resolved by 10%-12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto PVDF membrane (Millipore, Billerica, MA, USA). Finally, immunodetection was performed using an enhanced chemiluminescence detection kit (DW101, TransGen Biotech, Beijing, China), then images were captured by Syngene Bio Imaging (Synoptics, Cambridge, UK). The following primary antibodies were

Immunohistochemistry and staining
Fresh gastrocnemius muscles were xed in 4% paraformaldehyde for 24 h and embedded in para n. Para n sections were cut in section at 4 μm thickness, then depara nized and rehydrated before antigen retrieval. Sections were blocked with 10% bovine serumal bumin (BSA) in TBS-Tween 20 (Sigma Aldrich) for 1 h at room temperature. Sections was incubated overnight at 4 °C with the respective following primary antibodies against PGC-1α (ab191838, Abcam, UK), NRF-1 (A14190,
Construction of PPI network with an HtrA2 core Protein-protein interaction (PPI) network were construct by uploaded HtrA2 and its neighbor genes of rst three-layer in DEGs networks to the STRING database (https://string-db.org/). Then, the PPI network was further analyzed in Cytoscape software (www.cytoscape.org/). CytoHubba, a Cytoscape plug-in, sorts the genes by analyzing 12 parameters, including MCC, DMNC, MNC, DEGREE, EPC, BOTTLENECK, EcCentricity, CLOSENESS, RADIALITY, BETWEENNESS, STRESS and ClusteringCoe cient. We explore the genes sorting by 8 or more parameters as the hub genes with more essential in the functional network.

GSEA analysis
Gene Set Enrichment Analysis (GSEA) was performed using GSEA v4.0 software. Specimens were divided into low and high expression groups using HtrA2 expression quartile level as a cut-off points. In order to identify the potential function of HtrA2 in sarcopenia, GSEA was used to determine which gene sets were enriched in both groups. Number of permutation was set as 1000, and (c5.all.v6.2.symbols.gmt and c2.cp.kegg.v6.2.symbols.gmt, respectively) was used as reference gene-sets. A false discovery rate (FDR) q value<0.25 and nominal p value<0.05 was set as the signi cance cut-off threshold.

Correlation analysis
The genes of electron transport chain (ETC) subunits were obtained from the Molecular Signature Database (http://software.broadinstitute.org/gsea/msigdb/). Expression correlation among genes were calculated using "dplyr" and "tidyr" packages in R. The results were represented as Spearman's rank correlation coe cient. "ggplot2", "circlize" and "ggstatsplot" package in R were used to draw bubbles plot, chordal graph and scatter plot, respectively.

Statistical analysis
All results are expressed as the mean ± SD. Two-tailed Student's t test was used to perform comparisons between two groups. * p < 0.05, ** p < 0.005, *** p < 0.001. Comparisons between groups that do not follow a Gaussian distribution were performed via Mann Whitney U test. Statistical analyses were performed using GraphPad Prism software 8.0. Each replicate is 4 individual samples pooled.

Results
Overt symptoms of HtrA2 mnd2(-/-) mice mnd2 was identi ed as a spontaneous and recessively inherited mutation that arose on a C57BL/6J background in 1990. Herein, wild type, heterozygous and homozygous mnd2 mice ( Figure 1a) were identi ed presymptomatically using the closely linked primers anking microsatellite sequences in the C57BL/6J genome [34]. The earliest symptoms of HtrA2mnd2(-/-) mice begin around 3 weeks of age, displayed as cessation of normal weight gain and progressive movement disorder. By 30 days of age, the mice become completely akinetic, and their weight was less than half that of WT and heterozygous littermates ( Figure 1b). Details of the body weight, gastrocnemius muscle weight, and skeletal muscle mass index (SMI) are shown in table 1. The above data demonstrates that HtrA2/Omi protease de ciency results in decreased muscle mass. The data represent average ± SD (n=4). Two-tailed unpaired Student's t test were used. Statistical significance: HtrA2 mnd2(+/+) vs. HtrA2 mnd2(-/-) . Abbreviations: SMI, skeletal muscle mass index; pCSA, calculated physiological crosssectional area.

Muscle performance
To understand whether muscle performance was affected, the strength of gastrocnemius muscle was measured in living mice of 30-day-old when only sporadic degeneration of motor neurons was observed in cervical spinal cord of HtrA2 mnd2(-/-) mice [44,45]. The absolute single twitch and maximal tetanic force were reduced in HtrA2 mnd2(-/-) mice compared to WT controls ( Figure 1c). The forces normalized for calculated physiological cross-sectional area (pCSA) of gastrocnemius muscle, were also decreased in Overall, HtrA2 mnd2(-/-) muscle showed decreased muscle mass, impaired muscle performance, decreased cross-section area, presentation of morphological characteristics of aging and consistent decrease in ber numbers of all types. Considering the diagnostic criteria of sarcopenia [2], these data support the notion that HtrA2 mnd2(-/-) mice exhibit sarcopenia.

Denervation-independent muscle degeneration
To determine whether the sarcopenia phenotype in HtrA2 mnd2(-/-) muscle is a result of neurodegeneration, denervation-sensitive genes were examined, there was no change in MyoD and Myogenin mRNA expression, and further, the mRNA expression level of acetylcholine receptor α (AChRα) and AChRγ signi cantly decreased in HtrA2 mnd2(-/-) mice, indicating that the sarcopenia phenotype observed in HtrA2 mnd2(-/-) muscle results from processes that originate in muscle rather than neurodegeneration.
HtrA2/Omi protease de ciency causes mitochondrial hypofunction HtrA2 is a mitochondrial serine protease that is involved in the degradation of unwanted proteins to maintain mitochondrial proteostasis, to assess the effects of HtrA2/Omi protease de ciency on muscle mitochondria, we investigated the protein expression levels of mitochondrial membrane protein voltagedependent anion channel 1 (VDAC1) which has been widely used as mitochondrial reference marker, as expected, VDAC1 protein expression level decreased signi cantly in HtrA2/Omi protease de cient muscles (Figure 3a-c). VDAC1 staining presented regularly striped arrangement in WT muscles which is an characteristic of functional and physiological muscles [47], while, in HtrA2 mnd2(-/-) muscles, diffused distribution of VDAC1 was observed, besides, "alopecia areata" of VDAC1 staining was also noticed ( Figure 3a,b), which suggest abnormity in mitochondrial function in HtrA2 mnd2(-/-) muscles. Consistently, mtDNA copy number (Figure 3d) and ATP production (Figure 3e) also decreased signi cantly in HtrA2 mnd2(-/-) muscles. Interestingly, ROS level decreased but not increased in muscle of the homozygote mnd2 mice (Figure 3f). ROS is a byproduct of mitochondrial respiration, generally speaking, mitochondrial dysfunction results in an increased ROS production, while mitochondrial hypofunction leads to the decline of ROS production as a result of low level of electron transportation [48,49].
Absence of mitonuclear crosstalk in HtrA2 mnd2(-/-) muscles A prevailing hypothesis of HtrA2/Omi suggests that HtrA2/Omi protease de ciency would destroy mitochondrial proteostasis and promoting ROS production [29,30,32]. To test this hypothesis, we HtrA2/Omi protease de ciency leads to mitonuclear imbalance To predict biological functions of HtrA2/Omi in sarcopenia, analysis of GSEA, a powerful tool to infer the biological function, was performed using the sarcopenia microarray dataset. The results showed that biological processes associated with cellular respiration, mitochondrial ETC and oxidative phosphorylation were signi cantly enriched in HtrA2/Omi-low group (Figure 5a,b), all data is shown in Additional le 3: Table S3. These results suggest that HtrA2/Omi may participate in the process of sarcopenia via the regulation of ETC.
To further explore the regulation targets of HtrA2/Omi in ETC, 131 genes in KEGG_OXIDATIVE_PHOSPHORYLATION gene set, including all the core subunits of ETC complex I-V, were obtained from the Molecular Signature Database. Expression correlation analysis between these genes and HtrA2/Omi was conducted using the sarcopenia microarray dataset. Interestingly, the results showed that all the subunits that have strong negative expression correlation with HtrA2/Omi (γ<-0.7) are nDNA encoded, while, all the mtDNA encoded ETC subunits revealed either no correlation or strong positive correlation with HtrA2/Omi (γ>0.7) (Figure 5c,d)(Additional le 4: Table S4). To validate the analysis results, we examined the mRNA expression levels of these genes. It turned out that most of the nDNA encoded ETC subunits showed decreased mRNA expression level in HtrA2 mnd2(-/-) muscles compared to WT controls (Figure 5e-i), while the mRNA expression levels of mtDNA encoded subunits either unchanged or increased in HtrA2 mnd2(-/-) muscles (Figure 5j). It should be noted that the sarcopenia groups in the microarray dataset showed increased expression level of HtrA2/Omi, which suggests that HtrA2/Omi may act as a compensatory regulator of mitochondrial function in sarcopenia. Taken together, the above data suggests a differential regulation of mtDNA/nDNA encoded ETC subunits by HtrA2/Omi protease activity.
HtrA2/Omi protease de ciency may affect PGC-1α via Akt1 As we found that PGC-1a is decreased in HtrA2 mnd2(-/-) muscles, we wonder whether some factors that regulate PGC-1a are changed. Therefore potential genes were screened through bioinformatics analysis using the sarcopenia microarray dataset. HtrA2/Omi was identi ed as the differentially expressed genes (DEGs), comparison between groups of 12 and 27 months was chosen for further analysis. PPI network of DEGs with an HtrA2/Omi core was constructed, among which eight genes were also identi ed as hub genes, including Tp53, Casp3, Akt1, Sod2, Vegfa, Myc, CS and Stat3 (Figure 7a). We noticed that among the hub genes Akt1 has been found to in uence PGC-1a abundance or activity through posttranslational modi cation, Akt1 phosphorylate PGC-1a and promote its ubiquitination degradation [52]. Thus, Akt1 is probably responsible for the regulation of PGC-1α by HtrA2/Omi.
In addition to Akt1, we screened for potential genes through literature review. Xu et al. reported that HtrA2/Omi promotes PGC-1α degradation by cleaving glycogen synthase kinase 3β (GSK3β) in HtrA2 mnd2(-/-) brain. Therefore both Akt1 and GSK3β were chosen for further evaluation. Firstly, expression correlation analysis was conducted using the sarcopenia microarray dataset. The results showed that Akt1 but not GSK3β was negatively correlated with PGC-1α, NRF-1 and its target ETC genes, while positively correlated with NRF-2 and its target ETC genes, and mtDNA encoding ETC genes (Figure 7b,c) (Additional le 4: Table S4). Secondly, the mRNA and protein expression of Akt1 and GSK3β were also detected. Consistently, Akt1 but not GSK3β showed increased expression level in HtrA2 mnd2(-/-) muscles, indicating an involvement of Akt1 in PGC-1α regulation response to HtrA2/Omi protease de ciency.

Discussion
Disrupted mitonuclear communication is implicated in metabolic diseases, cancer, neurodegeneration, and other aging processes [6, [53][54][55]. While little is known about mitonuclear signaling regulator in sarcopenia. HtrA2/Omi is an IMS protease and play an important role in mitochondrial proteostasis [19,20]. HtrA2 mnd2(-/-) mice harboring protease-de cient HtrA2/Omi Ser276Cys missense mutants exhibit premature aging phenotype [31]. Additionally, HtrA2/Omi has been established as a signaling regulator in nervous system and tumors [11,[21][22][23][24][25][26][27]. To better understand the role of HtrA2/Omi protease activity within skeletal muscle, we utilized HtrA2 mnd2(-/-) mice and found that loss of HtrA2/Omi protease activity induced denervation-independent skeletal muscle degeneration. Interestingly, upregulation of UPR mt and mitohormesis related genes and elevated ROS production were not observed, contrary to previous assumptions that HtrA2/Omi protease de ciency would lead to mitochondrial dysfunction as a result of proteostasis disturbance and ROS burst [31]. Instead, we found that loss of HtrA2/Omi protease activity results in mitonuclear imbalance and differential regulation of mitochondrial biogenesis genes, suggesting a role of HtrA2/Omi protease activity as a regulator of mitonuclear signaling in sarcopenia.

Sarcopenia observed in HtrA2 mnd2(-/-) mice is independent of neuromuscular degeneration
Our nding that the denervation-sensitive genes was unchanged in HtrA2 mnd2(-/-) muscle con rmed the prevailing view that physiological denervation is not responsible for the motor abnormalities in HtrA2 mnd2(-/-) mice. mnd2 as its name "motoneuron disease 2" was originally characterized as a spinal muscular atrophy (SMA) because degenerating motoneurons were observed in late stages of the disease. However, Silvia et al. showed that there was no difference in mRNA level of denervation-sensitive gene AChRα between HtrA2 mnd2(-/-) and WT muscle, which distinguish SMAs from primary changes in the muscle, additionally morphological changes diagnostic for motoneuron disease were also not observed, indicating that mnd2 is not a primary motor neuron disease [44]. Indeed, only sporadic degeneration of motor neurons was observed in cervical spinal cord of HtrA2 mnd2(-/-) mice after 30 days postnatal [44,45].
In the present study, the HtrA2 mnd2(-/-) mice were sacri ced at 30 days postnatal. Considering the fact that mRNA levels, structure and weight of muscle did not signi cantly altered until 3-7 days after denervation with control mice, the authors concluded that the abnormalities observed in skeletal muscle of HtrA2 mnd2(-/-) mice are likely due to early onset neurodegeneration. It has been largely documented that organ crosstalk is important in physiological and metabolic processes. Therefore, it cannot exclude that the normal phenotype observed in skeletal muscle of transgenic rescued HtrA2 mnd2(-/-) mice is a result of the function of rescued brain.
HtrA2 mnd2(-/-) mice display a progressive striatal neurodegeneration with parkinsonian features [30]. Considering the motor control dysfunction and fatigability caused by PD, we detected muscle speci c force production via ex-vivo gastrocnemius force measurements, to evaluate the degeneration in skeletal muscle. Neuronal degeneration in the striatum results in motor abnormalities as a consequence of failure of inhibitory inputs. Thus an ex-vivo muscle force test should be more suitable to re ect the function of skeletal muscle itself.
The prevalence of sarcopenia in the population aged 60 years or older ranged from 5% to 50% across studies [46,63], while prevalence of sarcopenia in PD ranged from 6.6% to 55. A prevailing hypothesis of HtrA2/Omi suggests that the loss of protease activity of HtrA2/Omi would lead to premature aging [31] by destroying mitochondrial proteostasis and promoting ROS production [29,30,32]. Interestingly, we observed no change in neither expression level of UPRmt genes nor ROS production in HtrA2 mnd2(-/-) mice. Similarly to our results, in a study by Moisoi et al., CHOP and ATF3 was not upregulated in skeletal muscle or other non-neuronal tissues of HtrA2/Omi KO mice. Although CHOP expression was upregulated in brain, neither HSP60 nor CLPP mRNA were differentially expressed, suggesting that CHOP induction fails to induce any mitochondria protective genes in the brain of HtrA2/Omi KO mice [30]. It is noted that HtrA2/Omi is an IMS protease, Luena et al reported that IMS stress may activate a distinct UPRmt in MCF7 breast cancer cell line by triggering estrogen receptor (ER) activity, which further upregulated the transcription of NRF-1 [73]. However, the transcription of NRF-1 was found to be downregulated in our study. Our results demonstrated that loss of HtrA2/Omi protease activity lead to mitochondrial dysfunction and skeletal muscle premature aging without activation of UPR mt .
Rodrigue et al reported that ROS levels were similar in macrophages of HtrA2 mnd2(-/-) mice compared to WT [33], which is in agreement with our results. Conversely, three independent studies showed signi cantly elevated production of ROS in HtrA2/Omi KO mouse embryonic broblasts (MEFs) [30,74,75]. It was different from our results, as the authors employed HtrA2/Omi KO cells while we based our investigation on Ser276Cys missense mutated protease-de cient mice. Nicole et al. concluded in their study that regulation of mitochondrial morphology appears to depend on HtrA2/Omi protease activity, whereas mitochondrial ROS production could be related to another function of HtrA2/Omi [74]. Indeed, ROS is a by-product of OXPHOS, a normal mitochondrial function is the basis of ROS generation in cells.
This may explain the decreased ROS level observed in our study as we observed that mitochondrial biogenesis genes as well as OXPHOS genes decreased signi cantly in HtrA2 mnd2(-/-) muscle. For further validation, we found no change in mRNA levels of Foxo3, Keap1 and Nrf2 which have been recognized as mitohormesis related genes [5,76]. Mitohormesis is a process in which low, non-cytotoxic concentration of ROS promotes mitochondrial homeostasis. Accumulating evidences revealed that other than a harmful redox product, ROS act as a mediator of skeletal muscle adaptation and associate with improved mitochondrial function [48, [77][78][79]. Meanwhile, mitochondrial-targeted antioxidant prevented the increase in mitochondrial biogenesis induced by caloric restriction (CR), and physical exercise [79][80][81]. Thus this unresponsive levels of ROS observed in our study may be another important cause of sarcopenia in HtrA2 mnd2(-/-) mice.
The disturbed balance between nuclear DNA (nDNA) and mitochondrial DNA (mtDNA) encoded OXPHOS subunits, a state termed mitonuclear protein imbalance has been found to be associated with decreased mitochondrial respiration and UPR mt activation [82,83]. In our study, this imbalance was identi ed through GSEA and expression correlation analysis in which HtrA2/Omi showed strong negatively correlation with nDNA encoded ETC subunits, while no correlation or positively correlation with mtDNA encoded ETC subunits. It was further validated in HtrA2 mnd2(-/-) muscle, which showed different changes between the expression of nDNA and mtDNA encoded ETC subunits. Xu et al. showed that in brain of HtrA2 mnd2(-/-) mice both nDNA and mtDNA encoded ETC subunits decreased in mRNA level [27], this difference compared with our results in skeletal muscle may be due to tissue heterogeneity.
The present study reveals that loss of HtrA2/Omi protease activity per se was capable to induce mitonuclear imbalance. Paradoxically this imbalance in skeletal muscle of HtrA2 mnd2(-/-) mice failed to activate UPRmt. We also noticed that different from other ETC complex, most of the subunits in ETC complex-II showed no change in mRNA expression in HtrA2 mnd2(-/-) mice. Complex-II is the only ETC complex consisting solely of nDNA encoded proteins, therefore does not require a balanced production of proteins from the nDNA and mtDNA.
Collectively, our ndings suggest a role of HtrA2/Omi protease activity in mitonuclear signaling other than proteostasis maintenance. It should be noted that the present study only evaluated the mRNA levels, which may lead to a different conclusion, thus emphasizing the need for further investigation focusing on protein levels of ETC subunits.
PGC-1α may mediate the signaling function of HtrA2 Nucleus regulates mitochondrial adaptations by means of anterograde signaling which is mainly mediated by PGC-1α, a master regulator of nuclear-encoded mitochondrial genes (NEMGs) [4]. PGC-1a has been considered as the master regulator of mitochondrial biogenesis by targeting two key transcription factors, NRF1/2, which then activate nuclear genes encoding the OXPHOS subunits [84,85].
We showed that the expression of PGC-1 and its target antioxidant genes decreased signi cantly in HtrA2 mnd2(-/-) muscle, indicating that HtrA2/Omi protease de ciency lead to impaired expression and cotranscriptional function of PGC-1α. Interestingly, NRF-1, NRF-2 and their target nDNA encoded ETC subunit showed different changes in mRNA expression levels, Additionally, in mouse and rat, NRF-2 but not NRF-1 is responsible for mtDNA transcription and replication [50]. Collectively, PGC-1α may be a downstream target of HtrA2/Omi and mediate the mitonuclear signaling function of HtrA2/Omi. The differential expression of NRF-1/2 is probably the mechanism for mitonuclear imbalance observed in HtrA2 mnd2(-/-) muscle, which call for further investigation.

Limitations
Our results leave open the possibility that some of the alterations in HtrA2 mnd2(-/-) muscle are secondary to effects in satellite cell.

Conclusion
Our study focused on the role of HtrA2/Omi protease activity within skeletal muscle. We nd that HtrA2/Omi protease de ciency induced denervation-independent skeletal muscle degeneration. Loss of HtrA2/Omi protease activity failed to induced UPRmt and mitohormesis related genes in spite of mitochondrial hypofunction. Instead, we showed that HtrA2/Omi protease de ciency results in different changes between the expression of nDNA and mtDNA encoded ETC subunits, which is in consistent with their transcription factors NRF-1/2 and coactivator PGC-1α, suggesting that HtrA2/Omi protease de ciency induces mitonuclear imbalance and sarcopenia via differential regulation of mitochondrial biogenesis. The novel mechanistic insights may be of importance in developing new therapeutic strategies for sarcopenia. nuclear factor (erythroid-derived 2)-like 2; PGC-1α: peroxisome proliferator-activated receptor γ coactivator 1α; NRF-1: nuclear respiratory factor 1; GPX1: glutathione peroxidase 1; SOD1: superoxide dismutase 1; UCP2: Catalase and uncoupling protein 2; GSK3β: glycogen synthase kinase 3β controls, the frequency distribution of ber CSA and Round score in HtrA2mnd2(-/-) muscles showed a leftward shift, while that of Aspect Ratio showed a rightward shift. (h) qRT-PCR analysis of Myh 1, 2, 4 and 7 gene expression in gastrocnemius muscle from mice of three genotypes (n=4). (i) qRT-PCR estimates of mRNA abundance of AChRα, AChRγ, MyoD and Myogenin in gastrocnemius muscles from mice of three genotypes. All estimates (n=4) are differences in mRNA relative to WT controls. In (b-d), Mann Whitney U test were used. In (h,i), two-tailed unpaired Student's t test were used. Statistical signi cance: *p≤0.05; **p≤0.01; ***p≤0.001.