Increased remodeling and impaired adaption to endurance exercise in desminopathy

Desminopathy the most common intermediate filament disease in humans. Desmin is an essential part of the filamentous network that aligns myofibrils, anchors nuclei and mitochondria, and connects the z-discs and the sarcolemma. We created a rat model with a mutation in R349P DES, analog to the most frequent R350P DES missense mutation in humans. To examine the effects of a chronic, physiological exercise stimulus on desminopathic muscle, we subjected R349P DES rats and their wildtype (WT) and heterozygous littermates to a treadmill running regime. We saw significantly lower running capacity in DES rats that worsened over the course of the study. We found indicators of increased autophagic and proteasome activity with running in DES compared to WT. Stable isotope labeling and LC-MS analysis displayed distinct adaptations of the proteomes of WT and DES animals at baseline as well as with exercise: While key proteins of glycolysis, mitochondria and thick filaments increased their synthetic activity with running in WT, these proteins were higher at baseline in DES and did not change with running. The results suggest an impairment in adaption to chronic exercise in DES muscle and a subsequent exacerbation in the functional and histopathological phenotype.


Cytoskeleton, membrane repair and metabolism
As expected, biochemical analysis of desmin protein levels demonstrated lower desmin in sedentary HETs (46%) and DES mutant animals (69%) compared to sedentary WT (Fig3A). Running decreased desmin protein levels by 43% compared to the sedentary condition in WT rats. However, desmin protein levels increased by 51% in HETs and by 68% in DES compared to the sedentary animals (p for genotype <0.01, p for running = 0.83, p for interaction <0.01, Fig3A). Dysferlin, a protein mainly associated with membrane repair and fusion of repair vesicles to the plasma membrane, tended to be higher in DES rats. There was no effect of running or an interaction between running and the genotype (p for genotype =0.06, p for running = 0.33, p for interaction = 0.52, Fig3B). Annexin A2, which interacts with dysferlin during membrane repair, was not different between genotypes, but increased with running (p for genotype =0.96, p for running < 0.01, p for interaction = 0.66, Fig3C). Muscle LIM (mLIM), a protein prominent in the establishment and maintenance of the myocyte cytoskeleton, was on average 1.4-fold higher in the HET and 2.2-fold higher in the DES compared to the WT rats in sedentary animals. Running increased mLIM levels in all genotypes, but while mLIM increased 2.1-fold in WT and 1.7-fold in HET, DES only increased marginally by 1.03-fold (p for genotype < 0.001, p for running < Fig4G). The ubiquitination of total proteins, however, went down with running in WT by 23%, stayed the same in HETs and increased by 26% in DES compared to the respective sedentary groups (p for genotype = 0.86, p for running = 0.53, p for interaction <0.01, Fig4H). PTEN-induced kinase 1 (PINK1), a protein involved in autophagy of damaged mitochondria, did not differ significantly between the groups (p for genotype = 0.49, p for running = 0.92 and p for interaction = 0.41, Fig4I). Levels of the autophagy substrate p62 were similar for all genotypes in the sedentary group. Running increased p62 levels in all groups (p for genotype = 0.25, p for running < 0.01, p for interaction = 0.12, Fig4J). Levels of Caspase 3, a cysteine protease that is associated with apoptosis, were significantly higher in the running groups, and elevated by 1.6-fold in WT, 1.7-fold in HET and 2-fold in DES running animals compared to sedentary controls (p for genotype = 0.26, p for running < 0.0001 and p for interaction = 0.42, Fig4K). NOS (pan) levels were reduced 1.6-fold in WT and 1.4-fold in HETs after running, while NOS levels in DES stayed the same after running (p for genotype = 0.09, p for running < 0.001 and p for interaction = 0.12, Fig4L). We observed a similar effect in total OxPhos-levels (p for genotype = 0.91, p for running = 0.04 and p for interaction = 0.44, Fig4M).

Proteomics
Fractional synthesis of individual proteins was measured via LC-MS and clustered through the NIH DAVID database. Because we did not see relevant differences between WT and HET animals in histopathology and the protein levels we investigated, we only analyzed proteomics in WT and DES samples. Data filtering and calculations were performed according to previous reports 28 . Baseline protein synthesis across all proteins tended to be lower in WT than in DES in the control condition (0.151 ± 0.174 to 0.168 ± 0.167) but did not differ significantly (p=0.97). However, with running global protein synthesis increased to a greater extent in WT (on average 36%) (p<0.05) compared to DES (on average 20%) (p=0. 19). Running increased the synthesis of 32 out of 45 proteins in WT and decreased protein synthesis in 8 out of 45 proteins, with 5 remaining unchanged. In DES an equal number of proteins increased their synthesis with running (32 out of 45). However, out of the 32 proteins that increased in WT with running only 26 also increased in DES. Of the proteins that showed the strongest difference in synthesis between WT and DES, many were associated with mitochondrial function (llactate dehydrogenase A chain (Ldha), 3-mercaptopyruvate sulfurtransferase (Mpst)), glycolysis (glucose-6-phosphate isomerase (Gpi), phosphoglycerate kinase 1 (Pgk1)) and muscle contraction (myosin light chain 1/3, skeletal muscle isoform (Myl1), myosin light chain 3 (Myl3), myosin-4 (Myh4)).
Interestingly, among those proteins that showed a strong increase in DES but not in WT were proteins that are closely associated with muscle contraction such as Tropomyosin beta chain (Tpm2), Troponin T, fast skeletal muscle (Tnnt3) and Tropomyosin alpha-1 chain (Tpm1).
Out of 22 potential annotation clusters, mitochondrial-and glycolytic protein clusters were among the most prominent in our samples (Fig5A and B). We found 17 mitochondrial proteins and 12 glycolytic proteins that passed our quality control criteria (selected via GOTERM_CC_DIRECT and KEGG_PATHWAY, respectively). Four out of 17 mitochondrial protein synthesis values were significantly altered in WT through running, while no statistical difference could be found for DES animals ( Figure 5A). The protein showing the strongest increase with running in the WT group was cytochrome c, somatic (Cycs) (4.3-fold as high) (p<0.001), while peroxiredoxin 1 (Prdx1) was the protein with the strongest albeit insignificant decrease in protein synthesis (27%). For DES, ATP synthase, H+ transporting, mitochondrial F1 complex, beta polypeptide (Atp5b) was the protein that showed the strongest tendency for an increase with running without reaching statistical significance (p=0.14). Another mitochondrial protein that stood out was heat shock protein family D member 1 (Hspd1), which decreased by 11% with running in WT animals but by 67% in DES animals compared to their own control condition as well as by 88% compared to the WT control condition.
For the synthesis of glycolysis associated proteins, four out of 12 changed significantly (post-hoc analysis) with running in WT while only two changed in DES (Fig5B). The most robust changes with running occurred in Pgk1 (8.6-fold as high) (p<0.0001), Gpi (5.6-fold as high) (p<0.0001) and Pgm1 (3.8fold as high) (p<0.05) in the WT animals. In the DES animals, the proteins showing the strongest increase with running were Pgk1 (p<0.001) and Pgm1 (p<0.01) when compared to their own control condition.
Finally, looking at proteins directly involved in muscle contraction we found substantial differences ( Figure 6). Particularly WT and DES showed different adaptive responses to the running protocol: Myl3 protein synthesis increased to a level 2-fold as high in running WT compared to sedentary WT (Fig6A).
In contrast, in DES animals running decreased Myl3 by 7%. This discrepancy can be partially attributed to the fact that baseline Myl3 levels were 2.2-fold higher in DES compared to WT. Following a similar pattern, Myh4 increased by 23% with running in the WT group but decreased by 16% in the DES group (Fig6B). For myosin-6 (Myh6), running increased protein synthesis by 39% in WT and only by 6% in DES (Fig6C). Similarly, myosin-7 (Myh7) protein synthesis increased by 39% with running in WT and by 7% in DES (Fig6D).
Based on the distinct pattern in protein synthesis between WT and DES, we grouped the myosin isoforms from Figure 6 A-D together and normalized all values to the WT control group (Fig6E). This overview revealed a significant interaction effect between the genotype and running (p<0.03): Clustered myosin protein synthesis values increased by 47% with running in the WT group. In DES, however, protein synthesis at baseline were already 49% higher than in WT control animals and slightly decreased (6%) from there through running.

DISCUSSION
We recently created a novel CRISPR-Cas9 rat model for the most common type of desminopathy 26 .
While the age of onset in the human disease ranges from infancy to late adulthood, most patients first develop symptoms after their third decade [4][5][6]15 . In previous studies we investigated histopathological and molecular changes in our model at a preclinical age after functional overload and in old rats after acute eccentric loading. We showed that mutant animals at an age corresponding to a preclinical age in humans had a mild myopathy and increased markers for muscle injury and repair 26 . To see whether ageing exacerbates the phenotype, in our next study we looked at 400-500-day old animals in conjunction with exercise. Interestingly, while signs of chronic remodeling appeared similarly pronounced in old, compared to young, desmin mutants, DES animals appeared to be protected from exercise induced muscle damage unlike their WT littermates 27 . We concluded that ageing alone does not manifestly aggravate the disease progression, and hypothesized that our rats, having spent their lives in cages in a mostly sedentary manner, lack an adequate stimulus to trigger the pathophysiology.
In this study, we set out to examine the effects of chronic injury on DES muscle. We chose downhill treadmill running to provide repeated mechanic stress and stimulate eccentric-induced injury and regeneration in the muscles of the hindlimbs while minimizing the metabolic cost of the exercise.
During downhill running, plantar flexor muscles (gastrocnemius muscle, soleus muscle and plantaris muscle) primarily perform eccentric contractions to decelerate the animal's center of mass in order to maintain a constant running velocity 29 . These eccentric contractions cause injury consisting of disruptions of sarcoplasmic reticulum and other structures followed by local inflammatory and regenerative processes in the activated muscles 29-31 . At the same time, running down an incline leads to lower whole body oxygen consumption, blood lactate concentrations and presumably cardiovascular strain than during level locomotion at the same speed 32-34 .
From previous reports on downhill treadmill exercise that had 400-to 500-g wildtype Sprague-Dawley rats run continuously at 16 m/min for 90 min either on the level or down a 16° incline 29 we expected much longer running times in our animals. The longest time our rats ran was 22 min at a top velocity of 17m/min, while the average maximal running time in our WT was 17.5 min at a median top velocity of 13 m/min. Our animals were heavier, at an average body weight of 618g, but we saw no correlation between body weight and running time (R squared = 0.25; data not shown). Even though the study by Armstrong et al. did not report the age of their animals, given our rats were substantially heavier it appears likely that our animals may have been older. Other reports have shown a decrease in running endurance with increasing age as well as a pronounced stress response and poorer compliance in treadmill running in older animals compared to younger rats 35 . The difference between previously reported running performance and the running endurance of our rats might therefore be age-related.
However, we did not see signs of an increased stress response as monitored through weekly body weight, and compliance was good after two weeks of familiarization.
One DES animal had to be terminated during the course of the running experiments in week 2 due to severe bowel obstruction. Bowel obstruction or intestinal pseudo-obstruction have been reported in rare human cases of desminopathy, and histopathologic assessment of the affected intestines showed subsarcolemmal aggregates typical for the disease in smooth muscle 36 .
We saw distinct differences between the genotypes in their running performance. Performance in DES rats tended to be lower at the beginning of protocol, the running endurance in DES rats dropped significantly by the fourth ILT, and performance continued to decrease for the whole 24-day training period. Given this stark decrease in muscle function, it was somewhat surprising to see only moderate histopathological changes after running in the DES animals. The R349P animals demonstrated a higher percentage of fibers with central nuclei in DES compared to WT rats in the sedentary condition, which is in line with our previous findings in this strain 26 . However, this percentage did not go up in DES animals after the downhill running protocol, but rather tended to be reduced. In our previous study on acute bouts of eccentric loading in R349P rats, we also did not see a significant increase in fibers with central nuclei both in WT and DES, which we attributed to the time course of our experiment, as the centralization of myonuclei is thought to be a sign of regeneration following ~7 days after injury 37 and we collected the tissue 24 hours after loading. This explanation obviously does not hold up in the current chronic injury study. To assess myofibrillar injury via membrane damage, we stained for intracellular IgG. Except for one DES animal with 0.13% IgG-positive fibers, no animals showed any signs of membrane permeability at the time of collection. In our previous studies, we saw no IgGpositive fibers at baseline, but an increase in IgG-positive fibers after eccentric loading that was 5-fold higher in older WT than in age-matched DES animals.
Molecular markers for injury and repair followed a similar pattern as we previously saw in preclinical animals and after acute eccentric loading. Most notably, mLIM was upregulated in sedentary HET and more so in DES animals compared to WT, and while mLIM levels were increased in WT and HET after running, they did not change significantly in DES. mLIM has been found to interact with many different proteins both in the cytoplasm and the nucleus, and therefore serves a variety of functions. One of its integral roles is at the Z-disk, where it has been suggested to act as a scaffold protein that promotes the assembly of macromolecular complexes along sarcomeres and the actin-based cytoskeleton as well as a mediator of stretch signaling 38 . An upregulation of this Z-band associated protein after eccentric stress would be in line with previous findings that eccentric exercise has been found to lead to muscular adaptation specifically around the Z-band 39 . Dysferlin, by contrast, is a sarcolemmal protein that is mainly involved in membrane repair by facilitating the fusion of repair vesicles with a damaged membrane 40-42 . Therefore, it is usually upregulated after membrane injury. We saw a tendency for Dysferlin to be higher in DES animals compared to WT and this difference we accentuated after 6 weeks of running. As we do not see any indication for membrane injury as assessed via IgG-staining, we assume that this increase is due to an accumulation of Dysferlin in protein aggregates. Subsarcolemmal protein aggregates are a main histopathological hallmark in Desminopathy, and while the aggregates are initiated through the accumulation of misfolded desmin, they also contain debris of other proteins, including dysferlin 23 . We saw a similar trend towards higher desmin levels, both histologically as an increase in desmin aggregates and biochemically as an increase in western blots after running in the running, but not, as we expected, through mechanical stress-induced membrane damage.
There is conflicting data on the effects of mechanical stress on desminopathic muscle. While most reports point towards an increased mechanical vulnerability of mutated myofibers 43,44 , there is also evidence that loss of desmin might render muscle fibers less susceptible to membrane damage [45][46][47] .
However, to our knowledge there are no studies investigating the influence of chronic mechanical stress in desminopathy. There is anecdotal evidence that desminopathic patients who exercise suffer from a worse disease progression 48 , but this is difficult to interpret since the same mutation in desmin can present varied severity and progression.
This led some investigators to hypothesize that stiffness and mechanical stress is probably not the most important kind of stress that leads to disease progression in desminopathy 48 . The fact that creatine kinase (CK) serum levels, a classic, albeit crude, marker of muscle damage 49-51 , is of limited diagnostic value might be another indicator that membrane damage is not prominent in desminopathy. Fewer than 60 % of mutation carriers have elevated CK levels, and a third of patients with manifest skeletal muscle disease were reported to have normal CK levels 52 .
It is important to consider the fact that given their earlier exhaustion, the DES animals ran a significantly shorter distance than WT and HET over the 6-week training period and therefore received a lower absolute stimulus. While WT rats ran for an average of 171 m per ILT, HET rats ran 133 m and DES only 82 m. Still, this stimulus was enough to almost double the amount of desmin aggregates.
Desmin-positive aggregates are thought to be a central disease-driving mechanism. They are caused by misfolded desmin that cannot be degraded by the PQC. Over time, other cellular debris gets entangled in the aggregates that are then thought to "clog" autophagic and regenerative mechanisms.
We therefore investigated proteins involved in autophagy.
ULK1 plays a pivotal role in the initiation of autophagy. Its activity is positively regulated through phosphorylation at Ser 555 and negatively regulated through phosphorylation at Ser 757 53-55 . Single bouts of exercise have been shown to increase ULK1 phosphorylation at Ser 555 both in mice and humans 56,57 , but phosphorylation both at Ser 555 and Ser 757 did not change after a long-term resistance exercise protocol in rats 58 . We saw lower phosphorylation of ULK1 both at Ser 555 and Ser 757 in DES animals compared to WT, and phosphorylation went down in all genotypes after running. It is not clear whether this is a long-term adaptive process, possibly to reduce autophagy to sustain anabolism and muscle growth in a response to training, or a short-term effect due to collection time. Autophagy responses after exercise peak at about 2 h post exercise and dwindle after 3-4 h post exercise 59 . We collected muscles 24 h after their last running session making it unlikely that ULK1 levels reflected the last exercise bout. Downstream from ULK1, we also did not identify a stark impairment in autophagy in DES animals. However, total ubiquitination of proteins as a marker of ubiquitin proteasome-system (UPS) activity, went down in WT and up in DES animals, which might indicate impaired protein degradation or an overactive UPS response.
Our previous investigation of R349P mutant rats showed histological and molecular signs of increased muscle remodeling at baseline compared to WT. However, acute exercise did not seem to exacerbate this effect and DES muscle turned out to be less vulnerable to injury than WT muscle 26 . In the study at hand, we exposed rats to a chronic running stimulus and combined this with stable isotope labeling to investigate whether repeated exercise can 1) challenge the protection of R349P muscle to injury, 2) elicit a distinct, adaptive response in WT compared to DES animals, and 3) determine which proteins are synthesized in WT and DES animals.
In line with our previous study, we found that global protein synthesis values in the sedentary condition were slightly higher in DES compared to WT rats. The chronic exercise stimulus elevated global protein synthesis values in the WT group to a greater level than in the DES group. Consequently, the fact that most protein synthesis values showed a substantially higher fold-change with exercise in WT can be explained two ways: lower baseline levels and a greater adaptive response to running in WT. Part of the latter could be interpreted as a function of the aforementioned greater absolute training stimulus in WT. However, in healthy individuals relative exercise intensity and volume are likely more relevant for eliciting molecular responses and adaptations to exercise than absolute ones. This becomes evident when looking at the fact that beginners commonly show amplified molecular signaling after exercise compared to more experienced athletes, despite substantially lower training intensities 60 . Since the relative intensity and the running volume for the mutant animals was identical to their WT littermates, a lack of stimulus appears unlikely to explain the differences between the genotypes. In addition, many of the protein synthesis values in DES at baseline were close to or at the same level as in WT with chronic exercise. That DES did not show a further increase with exercise could therefore be a sign of a physiological ceiling of protein synthesis rates rather than a lack of adaptation.
Investigating proteins individually or as part of a cluster revealed that DES mutant animals tended to synthesize different proteins than WT controls both at baseline and following chronic exercise. For example, running increased cytochrome c 4.3-fold in WT, whereas in DES mutants the same protein was elevated in the control condition (2.2-fold compared to WT control) and stayed virtually unchanged with exercise (1.9-fold). In contrast, peroxiredoxin 1 was synthesized almost 2-fold more at baseline in DES compared to WT, but decreased with running to a level equivalent to WT.
Peroxiredoxin is thought to be involved in the protection of cells from oxidative stress and is elevated in inflammatory scenarios and cancer 61,62 . Higher synthesis values in DES at baseline could therefore indicate a greater need for coping with reactive oxygen species induced by the mutation, that is partially relieved through exercise.
For glycolytic proteins, synthesis values at baseline were more comparable between WT and DES than mitochondrial proteins. However, in response to the chronic exercise stimulus we could still see distinct responses between the two genotypes. One of the most pronounced discrepancies was the synthesis of glucose-6-phosphate isomerase. While baseline values were similar between WT and DES, running elevated synthesis to levels 5.6-fold in WT while in DES synthesis of the same protein rose only to levels 1.6-fold as high (largely due to a single individual). In our acute study of R349P mutants after exercise, we found a diminished glucose tolerance in DES animals (accepted for publication).
Furthermore, in an investigation of muscle metabolism after nerve damage we found intramuscular glucose to be elevated despite a tendency for lower glucose 6 phosphate levels 63 . Similar shifts in the levels of muscle metabolites have been observed for other congenital muscular dystrophies such as dysferlinopathy and Duchenne muscular dystrophy 64-66 . The disturbance of systemic as well as local muscle glucose metabolism in our desminopathy model and other muscular dystrophies/atrophies could point to a common physiological mechanism by which skeletal muscle copes with challenges to its integrity, despite different genetic and environmental reasons underlying such challenges.
Finally, looking at contractile protein isoforms individually and in a clustered manner we found baseline protein turnover to be elevated in DES compared to WT. This is in line with histological observations in this and our earlier papers, where we found an increased number of central nuclei in DES muscle at baseline. The finding of a higher number of central nuclei in combination and slightly elevated protein synthesis via puromycin, and a decreased susceptibility to exercise induced muscle damage in an acute setting led us to hypothesize that the R349P mutation in desmin might have caused an adaption in DES animals that results in higher contractile protein turnover. Indeed, in this study we found that DES animals have almost 50% higher values of contractile protein turnover at baseline compared to WT animals. While chronic exercise elevated contractile protein turnover in WT (~50%), in DES animals the running did not significantly alter contractile protein turnover.
In summary, in this study we challenged desminopathic muscle with a chronic, physiological exercise stimulus. We found a decreased running capacity in animals affected by the mutation, which was further exacerbated over the course of the study. In accordance with previous studies, we found histological signs of increased fiber damage and remodeling as well as a shift in fiber distribution towards smaller fibers in desminopathy. We found desmin aggregates at baseline in the mutant rats which become more pronounced with exercise training. On a protein level we found indicators of Future studies will have to determine whether there is a healthy dose of exercise for R349P mutant muscle and its human analog, or whether physical stress is to be avoided under any circumstance.

CRISPR-Cas9-mediated knock-in
Using CRISPR-Cas9, the missense mutation DES c.1045-1046 (AGG > CCG) was introduced in exon 6, leading to p.R349P in rats. For a detailed description of the generation and characterization of this rat desminopathy model see Langer et al, 2020 26 . Male rats between 400 to 560 days of age were selected to study functional, histological, and biochemical differences in the phenotype of CRISPR-Cas9 created knock-in animals carrying a mutation in desmin (DES; n = 9), heterozygous (HET; n = 8) or healthy, wild type littermates (WT; n = 9). Animals were housed in 12:12 h light-dark cycles and fed ad libitum.

Downhill running
The downhill running protocol was performed on a motorized treadmill (Exer-3/6, Columbus

Familiarization with treadmill and run to exhaustion
Rats from each genotype were randomly assigned to a "sedentary" (SED) or an "exercise" (RUN) group.
None of the rats had been on a treadmill or subjected to any kind of exercise prior to familiarization.
Rats in the exercise group (n=14) were accustomed to treadmill running over two weeks, starting at 10m/min at a 0° decline. To encourage running, rats were prodded with a soft bristled brush when they fell back onto the back quarter of the treadmill. The exercise session lasted until exhaustion, which was defined as the rat's inability to maintain running speed despite repeated physical prodding and staying on the electrical grid for more than five consecutive seconds 67 . At exhaustion, the electrical grid was immediately switched off and the animal removed from the treadmill. This protocol was performed every other day for eight days. On day nine, the protocol was set to 15m/min at a 0° decline, followed by a day of rest. On day eleven of familiarization, speed was set to 5 m/min and increased by 2 m/min every 3 min until exhaustion. This was followed by three days of rest, after which the downhill running protocol was started ( Figure 1A).

Downhill running protocol
Rats in the exercise group were subjected to downhill running at a decline of -10° for four weeks. For three weeks, rats performed an incremental load test (ILT) on day 1, 3 and 5, starting at a running speed of 5 m/min, which was increased by 2 m/min every 3 min until exhaustion. Upon exhaustion, the maximal speed completed as well as total running time were recorded. On day 2 and 4, to avoid overexercising but maintain a state of chronic physical activity, rats were exposed to walking at 40% of their respective maximal completed running speed for 10 minutes. On day 6 and 7, rats were rested.
Weight was recorded on day 1 of each week. In week 4, ILT was performed on days 1 and 3, 40%-speed running was performed on days 2 and 4. On day 5, rats ran at 15m/min on 0° decline to exhaustion and total running time was recorded. Animals were euthanized 24h after the final running test.

Deuterium oxide labelling
Animals were stable isotope labeled with deuterium oxide using previously published protocols with modifications 68,69 . Briefly, endogenous deuterium levels were raised to about 4 % by a priming intraperitoneal injection of 0.014 mL * g−1 bodyweight deuterium oxide (99.8 % + Atom D, Euriso-Top GmbH Saarbrücken) and 0.9 % NaCl. Deuterium levels were maintained via drinking water (4 % deuterium oxide). The stable isotope labeling started five days before collection and an independent group of unlabeled rats served as a control.

Muscle collection
Rats were anaesthetized via isoflurane inhalation (2.5%). The gastrocnemius (GAS), soleus (SOL) and tibialis anterior (TA) muscles were excised from the left hindlimb, blotted dry, and weighed. Subsequently, the muscles were pinned on cork at resting length and frozen in liquid nitrogen-cooled isopentane for histological and biochemical analyses. Liver, heart and serum were also collected. On completion of tissue removal, rats were euthanized via cardiac puncture.

Histology
Frozen muscles were blocked, and serial cross-sections (10 μm) were cut using a Leica CM 3050S Slides were imaged using a Leica DMi8 inverted microscope using the HC PL FLUOTAR 10x/0.32 PH1 objective (Leica Microsystems, Wetzlar, Germany) and using the LAS X software. For comparative analysis, exposure length remained fixed for all samples.
Muscle fiber properties and fiber types were analyzed using FIJI and SMASH (MATLAB) software from the whole muscle sections.

Hydroxyproline determination of collagen content
Collagen content was determined using a hydroxyproline assay 70

Proteomics Body water enrichment analysis
Rat serum and liver were distilled overnight upside down on a bead bath at 85°C to evaporate out body water. Deuterium present in the body water were exchanged onto acetone, and deuterium enrichment in the body water was measured via gas chromatography mass spectrometry (GC-MS) 72 .

Tissue preparation for LC-MS
Tissues were homogenized in homogenization buffer (100 mM PMSF, 500 mM EDTA, EDTA-free Protease Inhibitor Cocktail (Roche, catalog number 11836170001), PBS) using a 5 mm stainless steel bead at 30 hertz for 45 seconds in a TissueLyser II (Qiagen). Samples were then centrifuged at 10,000 g for 10 minutes at 4°C. The supernatant was saved and protein was quantified using a Pierce BCA protein assay kit (ThermoFisher, catalog number 23225). 100 ug of protein was used per sample. 25 µL of 100 mM ammonium bicarbonate solution, 25 µL TFE, and 2.3 µL of 200 mM DTT were added to each sample and incubated at 60°C for 1 hour. 10 µL 200 mM iodoacetamide was then added to each sample and allowed to incubate at room temperature in the dark for 1 hour. 2 µL of 200 mM DTT was added and samples were incubated for 20 minutes in the dark. Each sample was then diluted with 300 µL H2O and 100 µL 100 mM ammonium bicarbonate solution. Trypsin was added at a ratio of 1:50 trypsin to protein (trypsin from porcine pancreas, Sigma Aldrich, catalog number T6567). Samples were incubated at 37°C overnight. The next day, 2 µL of formic acid was added. Samples were centrifuged at 10,000 g for 10 minutes, collecting the supernatant. Supernatant was dried by speedvac and resuspended in 50 µL of 0.1 % formic acid/3% acetonitrile/96.9% LC-MS grade water and transferred to LC-MS vials to be analyzed via LC-MS.

LC-MS analysis
Trypsin-digested peptides were analyzed on a 6550 quadropole time of flight (Q-ToF) mass spectrometer equipped with Chip Cube nano ESI source (Agilent Technologies). High performance liquid chromatography (HPLC) separated the peptides using capillary and nano binary flow. Mobile phases were 95% acetonitrile/0.1% formic acid in LC-MS grade water. Peptides were eluted at 350 nl/minute flow rate with an 18-minute LC gradient. Each sample was analyzed once for protein/peptide identification in data-dependent MS/MS mode and once for peptide isotope analysis in MS mode.

Statistics
Appropriate sample sizes to detect group effects were estimated on the basis of previous experiments with the same strain of animals 26,27 . Two-way analysis of variance (ANOVA) with a post hoc Tukey's multiple comparisons test was used to test the null hypothesis. An alpha of P < 0.05 was deemed statistically significant, and a P value between 0.05 and 0.1 was called a trend. Biological replicates are reported as "n" in each figure legend. "Genotype" = "Gen" indicates effects of wildtype (WT), heterozygous (HET) and desmin R349P mutant (DES) genotypes on the test. "Running" = "Run" indicates effects between sedentary (SED) or exercised (RUN) group. "Interaction" = "Int" indicates the interaction effect between "genotype" and "running". Desmin aggregates (% of whole section) in Figure 2F were tested via unpaired t-test. Effect sizes for proteomics were calculated according to Hedges' g 74 . Data in the text are reported as mean ± standard deviation, and data in the figures are visually represented as scatter dot plot with error bars indicating standard deviation. All analysis was performed with GraphPad Prism Version 8 (La Jolla, CA, USA).      Figure 4 -Changes in the Protein Quality Control system. Western Blots of phosphorylated ULK1 (Ser757) showed decreased protein levels in the running group (A). Phosphorylated ULK1 (Ser555) tended to be lower in DES and went down with running in all genotypes (B) Autophagy related 7 (Atg7) decreased in the WT group with running, but increased 2.2-fold in DES animals (C). The ratio of LC3 II to LC3 I protein levels was not statistically different between the genotypes and did not change with running (D). Running had no effect on heat shock protein 90 (HSP90) (E). Heat shock protein 27 (HSP27) levels were the same for all genotypes irrespective of running (F). Free ubiquitin (G), p62 (J) and caspase 3 (K) levels increased with running in all groups. The ubiquitination of total proteins went down with running in WT by 23%, stayed the same in HETs and increased by 26% in DES compared to the respective sedentary groups (H). PTEN-induced kinase 1 (PINK1) did not differ significantly between the groups (I). NOS (pan) levels were reduced 1.6-fold in WT and 1.4-fold in HETs after running, while NOS levels in DES stayed the same after running (L). We observed a similar effect in total OxPhoslevels (M). Gen=Genotype, Run=Running, Int=Interaction; AU=arbitrary units; Sed=sedentary controls. (n for WT sed=3, WT run=6, HET sed=4, HET run=3, DES sed=4, DES run=4). Group differences were assessed via a two-way ANOVA, error bars indicate ± SD.  Figure 5A) and glycolytic proteins ( Figure 5B) were clustered according to the NIH database DAVID (using GOTERM_CC_DIRECT and KEGG_PATHWAY selections). All fractional synthesis values of proteins were normalized to WT control. Four out of 17 mitochondrial protein synthesis values were significantly altered through running in WT. No mitochondrial protein synthesis values were significantly altered in DES ( Figure 5A). Four out of 12 glycolytic protein synthesis values were significantly altered through running in WT. Two glycolytic protein synthesis values were significantly altered in DES ( Figure 5B). (n for WT sed=3, WT run=6, DES sed=4, DES run=4)* indicates a p-value of >0.05 compared to their own sedentary control (post hoc analysis).

Animals and ethical approval
Error bars indicate ± SD