Electrical Stimulation Attenuates Disuse Muscular Atrophy by Modulating Endoplasmic Reticulum Stress-induced Parkin-dependent Mitophagy

Background: The aim of this study was to investigate the therapeutic effect of electrical stimulation on disuse muscular atrophy in a rabbit model of knee joint contracture and explore the role of endoplasmic reticulum stress-induced Parkin-dependent mitophagy in this process. Methods: Two sub-experiments were carried out successively in our study. In the rst sub-experiment, 24 rabbits were divided into four groups on average based on the immobilization time: Ctrl 1, I-2, I-4, and I-6 groups. In the second sub-experiment, 24 rabbits were also divided into four groups on average in accordance with the process mode: Ctrl2, ES, NR, and EST groups. To test the time-dependent changes of the rectus femoris muscles after immobilization in rabbits, and to evaluate the effect of electrical stimulation on the atrophic rectus femoris muscles, the wet weights of rectus femoris muscles were assessed in this study, along with the protein levels of atrogin-1, p-PERK, Parkin and COXIV. Results: The wet weights of rectus femoris muscles, the protein levels of atrogin-1, p-PERK and Parkin increased after immobilization. It was also revealed that the protein levels of COXIV decreased after immobilization. Electrical stimulation was effective against muscle atrophy, the elevated expression of atrogin-1, p-PERK, Parkin, and the decreased expression of COXIV. Conclusions: Immobilization of unilateral lower limb could induce rectus femoris muscle atrophy, endoplasmic reticulum stress and Parkin mediated mitophagy. Endoplasmic reticulum stress-induced Parkin-dependent mitophagy may be one of the mechanisms by which electrical stimulation can play a signicant role. The wet weights of rectus femoris muscles in the four groups were shown in the gure 4. The difference of wet weights of rectus femoris muscles between the Ctrl 2 and ES groups failed to reach a statistical difference (P>0.05). The wet weights of rectus femoris muscles in the NR group were lower than those in the Ctrl 2 group (P<0.05). Meanwhile, the wet weights of rectus femoris muscles in the EST group were higher than those in the NR group and still lower than those in the Ctrl 2 and ES groups (P<0.05).


Background
Joint contracture is a common disease in rehabilitation medicine, which can have a great impact on patients' quality of life [1]. After the onset of joint contracture, muscle atrophy occurs, intra-articular tissue adhesion happens, and bone changes take place around the joint [2,3]. Furthermore, the function of impaired joints may be limited, which can affect the patients' abilities in daily living. Prompt and effective treatment is important for the prognosis of joint contracture. however, the timing of rehabilitation treatment and the choice of rehabilitation intervention methods are still controversial.
Joint contracture is often accompanied by morphological changes of various tissues around the joint.
According to the anatomical factors limiting joint range of motion, it can be divided into myogenic factors (muscle, tendon, fascia, etc.) and arthrogenic factors (including bone, cartilage, joint capsule and ligament, etc.). The limiting factors of joint motion are mainly muscular factors in the early stage of joint contracture, while articular factors plays the leading role in the middle and late stages [4,5]. To our knowledge, the treatment of joint contracture is very di cult after the advanced stage. Therefore, the suppression of myogenic factors in the early stage of joint contracture is of great importance to the prognosis of these patients.
Myogenic contracture is primarily manifested as skeletal muscle atrophy, which is characterized by the reduction of muscle mass, cross-sectional area and muscle length in muscles around the affected joint [5,6]. After the formation of joint contracture caused by joint immobilization, the atrophied muscle tissue around the joint showed a light cytoplasm staining and reticular pattern, with a relative increase in stroma and nucleus, as well as innervation and aggregation of the nucleus [7,8]. This phenomenon indicates that the catabolism of muscle protein in skeletal muscle tissue was enhanced. Our previous studies suggest that the occurrence of disuse muscular atrophy is related to the happening of joint contracture [5,9]. The decrease in muscle wet weights is one of the most direct re ections of muscle atrophy [10].
Previous studies have suggested that atrogin-1 is one of the important muscle atrophy-related genes [11], so the elevated expression of atrogin-1 can be used to re ect muscle atrophy. However, the mechanism of periarticular muscle disuse atrophy following joint immobilization is still not thoroughly understood.
Disuse muscular atrophy is regulated by a variety of mechanisms, of which autophagy -lysosomal pathway may be one of the most important signaling pathways [12]. Autophagy is divided into selective autophagy and non-selective autophagy. Most of the previous studies focused on non-selective autophagy, with only a few on selective autophagy. Parkin-mediated mitophagy signaling pathway is one of the classical signaling pathways that regulate mitophagy at present [13]. Under cellular stress conditions, the reduction of circumstances of mitochondrial membrane potential may cause the mutation of mitochondrial DNA mutation and increase in unfolded proteins [14]. Intracellular protein kinase Pink1 aggregates in the outer membrane of the damaged mitochondria and is self-phosphorylated. Pink1 is phosphorylated by the ubiquitin ligase Parkin and is recruited to mitochondria to activate Parkin signal.
Concurrently, the process also promotes the phosphorylation of ubiquitin on the outer membrane of mitochondria, which further activates Parkin. Ubiquitin binds to LC3-II through autophagy receptor proteins (p62, etc.) to form autophagosomes. The autophagosome then fuses with the lysosome to form the autophagosome, and the mitochondria and related proteins are nally degraded by lysosomal hydroxylase [15]. Previous studies have suggested that Parkin-mediated mitophagy may be activated after atrophic muscle changes happened following joint immobilization [16]. Recently, scientists have recognized that mitophagy signaling pathways may be related to skeletal muscle atrophy. However, the exact mechanism that may regulate skeletal muscle mitophagy is still unclear.
The endoplasmic reticulum (ER) is an intracellular organelle in which proteins are modi ed. Previous studies have shown that mitophagy can be regulated by endoplasmic reticulum stress [17]. The endoplasmic reticulum stress responds to the load of unfolded proteins by activating intracellular signal transduction pathways, which is known as unfolded protein response (UPR) [18]. Previous study has suggested that UPR pathways play pivotal roles in muscle stem cell homeostasis, myogenic differentiation and regeneration of injured skeletal muscle [19]. In general, at least three branches of mechanisms can regulate the expressions of a large number of genes that can maintain homeostasis or induction in the endoplasmic reticulum apoptosis. These are protein kinase RNA-like ER kinase (PERK), activated transcription factor-6 (ATF6) and inositol-demanding enzyme-1 (IRE1) [20]. Among these three pathways, PERK signaling pathway can participate in the formation of myotubules and have a regulatory effect on the synthesis of myocytes [21,22]. PERK is responsible for reducing the overload of misfolded proteins, thereby alleviating ER stress. Kang C et al. showed that mitophagy mediated by the Parkin signaling pathway is elevated in the atrophic muscle after immobilization of unilateral lower limb [23]. The study of Deval C et al. also suggested that mitophagy levels in gastrocnemius and tibialis anterior muscles were increased after lower limb immobilization [24]. These results indicated that the level of mitophagy may increase when skeletal muscle atrophy occurs. Previous studies have suggested that Parkin-mediated mitophagy can be regulated by the PERK signal in placental tissues [25]. However, few studies have concentrated on the regulatory role of PERK on Parkin in atrophic skeletal muscles.
Low frequency electrical stimulation (LFES) is a safe and effective physical agents therapy method in rehabilitative medicine, which can increase contractile function of muscle ber which can then be used for the treatment of skeletal muscle atrophy [26]. Rectus femoris muscle is a commonly used treatment site of electrical stimulation for patients with quadriceps muscle atrophy in clinical practice. The current study involves an investigation of the effects of electrical stimulation on disuse muscular atrophy in a rabbit model of knee joint contracture and an exploration of the molecular mechanisms that underlie the initiation and progression of this pathology. We hypothesized that PERK-regulated mitophagy may play an important role in the pathology of joint contracture and that electrical stimulation may inhibit disuse muscular atrophy through PERK-regulated mitophagy.

Animals and experimental materials
Animal care and experimental procedures were performed in accordance with the Guidelines for Animal Experimentation of Anhui Medical University and were approved by the Animal Ethics Committee of Anhui Medical University (LLSC20190761). 48 male New Zealand white rabbits were purchased from the Experimental Animal Center of Anhui Medical University.
The method of establishment of knee contracture model in this study was the same as those in our previous studies [2,3,5]. To make the animal model, the rabbits underwent unilateral immobilization of the knee joint at full extension using a plaster cast from the groin to the proximal toes. The rabbits were divided into 2 parts in our experiments. In the rst part, in order to explore the effects of immobilization on disuse muscular atrophy and the expression levels of p-PERK and Parkin mediated mitophagy, 24 rabbits were randomly divided into 4 groups based on the immobilization time: Ctrl1 group (free movement for 6 weeks, followed by no immobilization), I-2 group (free movement for 4 weeks, followed by immobilization for 2 weeks), I-4 group (free movement for 2 weeks, followed by immobilization for 4 weeks) and I-6 group (immobilization for 6 weeks), with 6 rabbits in each time cohort. 18 rabbits underwent unilateral immobilization of the left knee joint at full extension using a plaster cast from the groin to the proximal toes. The rabbits in the I-2 group, I-4 group, and I-6 group underwent 2, 4 and 6 weeks of immobilization respectively. In the second part, in order to investigate the therapeutic effect of low-frequency electrical stimulation on disuse muscular atrophy and explore the possible mechanism concerning PERK mediated mitophagy, 24 rabbits were randomly divided into 4 subgroups corresponding to subsequent intervention measures: 1) a Ctrl2 group, in which no immobilization or low-frequency electrical stimulation was used and the rabbits can move freely for 7 weeks; 2) a ES group, in which the rabbits freely moved for 4 weeks, followed by 3 weeks of 10 Hz low-frequency electrical stimulation treatment; 3) a NR group, where the rabbits' left knee joints were xed for 4 weeks, followed by natural recovery for 3 weeks; 4) a EST group, where the rabbits' left knee joints were immobilized for 4 weeks, followed by 3 weeks of 10 Hz lowfrequency electrical stimulation treatment (Table 1). Low frequency electrical stimulation treatment Rabbits in ES and EST groups received 10Hz low-frequency electrical stimulation with an Electronic Acupuncture Treatment Instrument (SDZ-), and the time was set at 20 minutes, once a day for 3 weeks. The intervention site of electrical stimulation was the quadriceps femoris of the left hind limb of the rabbits. After the hair of the left hind leg was shaved off, two 3 × 3 cm² non-woven silica gel electrode sheets were attached to the skin of the left hind leg, with a distance of 0.5 cm between the two sheets.
The output current of electronic acupuncture instrument was less than 10mA. We adjusted the size of the output current to cause the quadriceps muscle contraction, but not cause the rabbits' excessive struggling. The current was set to 5 mA within the rabbits' tolerance range.

Tissue preparation
At the end point, an overdose of sodium pentobarbital was used to kill the rabbits. The immobilized hind limbs of rabbits were disarticulated at the hip joints and the rectus femoris muscles were separated.
Measurement of wet weights of rectus femoris muscles After being separated from the rabbits, the wet weights of rectus femoris muscles were measured using an electronic scale and then recorded. Subsequently, the 1/3 of rectus femoris muscles were stored in a refrigerator and the remaining 2/3 were used for other experiments in our study.

Immunohistochemistry
Atrogin-1 protein expression was measured by IHC staining in paraformaldehyde-xed, para nembedded tissue slides from the rectus femoris specimens. Firstly, the tissue slides were depara ned in xylene and rehydrated in different concentration gradients of ethyl alcohol. After being washed with phosphate-buffered saline (PBS) three times, the slides underwent antigen retrieval in citrate buffer (pH 6.0) in the microwave oven. A solution of 3% hydrogen peroxidase (H2O2) was used to quench the endogenous peroxidase activity of the slides for 5 min. Next, the slides were incubated with the primary antibody of anti-atrogin-1 (1 : 300, Bioss) at 37℃ for 60 minutes. The slides were incubated at 37℃ for

Statistical analysis
All data in our study are presented as Mean ± SD and were analyzed using an analysis of variance with a post-hoc LSD test for comparison between individual groups in the two experiments. Results for wet weights, the expressions of atrogin-1, p-PERK, Parkin and COXIV were compared among all groups with the analysis of variance test. A P-value of less than 0.05 was chosen as the signi cance threshold.

Results
Immobilization induced the reduction of muscle mass in rabbits Figure 1 illustrates the wet weights of femoris muscles in the four groups which have underwent different period of immobilization. After 2 weeks of immobilization, the wet weights of femoris muscles were decreased and compared with the Ctrl 1 group (P<0.05). Moreover, there was a signi cant difference between the wet weights of the femoris muscles in the I-4 group compared with those in the I-2 group (P<0.05). The wet weights of femoris muscles in the I-6 group were lower than those in the I-4 group, although failed to reach a statistical signi cance (P>0.05).

Immobilization induced the increased protein expression of atrogin-1
The protein levels of atrogin-1 examined with immunohistochemistry were shown in gure 2. The protein levels of atrogin-1 increased with the prolonging of immobilization time. After 2 weeks of immobilization, the protein levels of atrogin-1 in the I-2 group increased compared with those in the Ctrl 1 group (P<0.05). Moreover, the protein levels of atrogin-1 in the I-4 group were higher than those in the I-2 group (P<0.05).
The protein levels of atrogin-1 in the I-6 group were also higher than those in the I-4 group although not statistically signi cant (P>0.05).

Immobilization induced the activation of PERK-mediated mitophagy
The protein levels of p-PERK, Parkin and COXIV were examined in the rectus femoris of rabbits which had been immobilized for different timing and the results were shown in Figure 3. Immobilization increased the protein levels of p-PERK, Parkin and decreased the protein levels of COXIV. The protein levels of p-PERK in the I-2 group were higher than those in the Ctrl1 group (P<0.05). The protein levels of p-PERK in the I-4 group were higher than those in the C group (P<0.05) and the I-2 group (P<0.05). The protein levels of p-PERK in the I-6 group were lower than those in the I-2 and I-4 groups (P<0.05). The protein levels of p-PERK in the I-6 group were higher than those in the Ctrl1 group but failed to reach a statistical difference (P>0.05). The protein levels of Parkin in the I-2 group were higher than those in the Ctrl1 group (P<0.05). The protein levels of Parkin in the I-4 group were higher than those in the Ctrl1 group (P<0.05) and lower than those in the I-2 group (P<0.05). The protein levels of Parkin in the I-6 group were higher than those in the Ctrl1 group but failed to reach a statistical difference (P>0.05). The protein levels of Parkin in the I-6 group were lower than those in the I-2 and I-4 groups (P<0.05). The protein levels of COXIV in the I-2 group were lower than those in the Ctrl1 group (P<0.05). The protein levels of COXIV in the I-4 group were lower than those in the Ctrl1 group, but failed to reach a statistical difference (P>0.05). The protein levels of COXIV in the I-4 group were higher than those in the I-2 group (P<0.05). The protein levels of COXIV in the I-6 group were lower than those in the Ctrl1 and I-4 groups (P<0.05) and higher than those in the I-2 group (P<0.05).

Low-frequency electrical stimulation improved skeletal muscle atrophy
The wet weights of rectus femoris muscles in the four groups were shown in the gure 4. The difference of wet weights of rectus femoris muscles between the Ctrl 2 and ES groups failed to reach a statistical difference (P>0.05). The wet weights of rectus femoris muscles in the NR group were lower than those in the Ctrl 2 group (P<0.05). Meanwhile, the wet weights of rectus femoris muscles in the EST group were higher than those in the NR group and still lower than those in the Ctrl 2 and ES groups (P<0.05).
Low-frequency electrical stimulation corrected the abnormal elevation of atrogin-1 expression The results of immunohistochemistry concerning atrogin-1 were shown in gure 5. There was no statistical difference in the protein expression of atrogin-1 between the Ctrl 2 and ES groups (P>0.05). In the NR group, the protein levels of atrogin-1 were higher than those in the Ctrl 2 group (P<0.05). Moreover, the protein levels of atrogin-1 in the EST group were lower than those in the NR group (P<0.05), but still higher than those in the Ctrl 2 group (P<0.05).

Low-frequency electrical stimulation alleviated PERKmediated mitophagy
The protein levels of p-PERK, Parkin and COXIV were examined in the rectus femoris of rabbits which had underwent different treatments ( gure 6). The difference of protein levels of p-PERK in the Ctrl2 and the ES groups failed to reach a statistical signi cance (P> 0.05). The protein levels of p-PERK in the NR group were higher than those in the Ctrl2 and ES groups (P<0.05). The protein levels of p-PERK in the EST group were lower than those in the NR group (P<0.05). The difference of protein levels of Parkin in the Ctrl2 and the ES groups failed to reach a statistical threshold (P> 0.05). The protein levels of Parkin in the NR group were higher than those in the Ctrl2 and the ES groups (P<0.05), and the protein levels of Parkin in the EST group were lower than those in the NR group (P<0.05). The difference of protein levels of COXIV in the Ctrl2 and the ES groups is within the statistical signi cance (P> 0.05). The protein levels of COXIV in the NR group were lower than those in the Ctrl2 and the ES groups (P<0.05). The protein levels of COXIV in the EST group were higher than those in the NR group (P<0.05) and lower than those in the Ctrl 2 group (P<0.05).

Discussion
Joint contracture, especially advanced joint contracture, can bring great inconvenience to the activities of daily living in patients and make the treatment challenging [27]. The disease brings many adverse effects to the quality of life in patients, social medical resources and the health economy. Hence, early and timely rehabilitation treatment is of great importance to the prognosis of the disease. Previous studies by others as well as our own suggest that myogenic contracture was the main cause in the early stage of joint contracture, and muscle atrophy was one of the most important aspects of myogenic contracture [28]. Consequently, the treatment of disuse muscle atrophy is of great importance to the early prevention and treatment of joint contracture. Electrical stimulation is a well-known physical agent therapy for the treatment of muscle atrophy. However, few studies have concerned about the effect of low frequency electrical stimulation on muscle atrophy in a knee joint contracture model, and the exact mechanism underlying the therapeutic effect of electrical stimulation still remains unclear. In order to investigate the effect of electrical stimulation on muscle atrophy in a rabbit model of knee joint contracture and explore the underlying mechanism, we designed the experiments in this study.
Loss of wet weight is one of the most intuitive representations of muscle atrophy [29,30]. Biological ndings in this study showed that the wet weight of rectus femoris muscles decreased with the prolonging of immobilization time. The results indicated that skeletal muscle atrophy gradually increased with the progression of joint contracture. The degree of muscular atrophy in the natural recovery group was lower than that of the normal group in our study. The results illustrated that short-term natural recovery alone cannot completely restore muscular atrophy, and prompt rehabilitation intervention is necessary. The difference of wet weights in the Ctrl2 and the ES groups failed to reach a statistical signi cance but the wet weight in the EST group was statistically lower than that in the NR group. The results hinted that simple electrical stimulation had no obvious negative effect on normal rabbit rectus femoris muscle, but electrical stimulation can reduce the degree of muscle atrophy in a rabbit model of knee joint contracture.
A reduction in skeletal muscle mass is considered to be signi cantly associated with increased expression of E3 ubiquitin ligases [31]. Atrogin-1 is one of E3 ubiquitin ligases expressed in skeletal muscle, which can control both the processes associated with the breakdown of cytoskeletal proteins as well as processes associated with protein synthesis [32,33]. Hence, atrogin-1 is up-regulated in skeletal muscle atrophy and cachectic syndromes [34,35]. Our experiments showed signi cantly increased atrogin-1 protein levels in rectus femoris after immobilization, suggesting that muscle atrophy occurred following joint immobilization. Importantly, electrical stimulation can partly reverse the elevated protein levels of atrogin-1. The results indicated that electrical stimulation may rectify muscle atrophy via the suppression of atrogin-1.
Skeletal muscle atrophy following prolonged immobilization (IM) is known to be a catabolic state characterized by increased proteolysis. Selective autophagy of mitochondria (which is also called mitophagy) can be used to degrade redundant mitochondria speci cally those that have previously been segregated on account of mitochondrial ssion [36]. Previous research has indicated that overactivated mitophagy plays a pivotal role in the event of muscle atrophy [37]. Parkin mediated mitophagy is one of the most important mitophagy pathways, and COXIV is a well-known mitophagy marker protein [38]. In our study, the activation of mitophagy on the muscles after immobilization could be detected by examining the upregulation of Parkin and the downregulation of COXIV. The treatment with electrical stimulation resulted in the downregulation of Parkin and the upregulation of COXIV, which indicated that overactivated mitophagy was inhibited in the study. These results indicated that the inhibition of abnormal activation of mitophagy may be one of the mechanisms by which electrical stimulation plays an antiatrophic role on the disuse atrophy in a rabbit model of knee joint contracture.
Studies to date have shown that endoplasmic reticulum stress can regulate mitophagy in other areas [39,40,41]. Gao K et al. [39] reported that Pb-induced ER stress may play a regulatory role in the upstream of mitophagy. Guo J et al. [40] reported that ER stress can induce miR-346 and modulated autophagic ux in Hela cells. Zhang X et al. [41] demonstrated that the ER stress induced by TM and TG can protect against the transient ischemic brain injury, and the PARK2-mediated mitophagy may be underlying the protection of ER stress. In our study, immobilization can induce the elevated protein levels of p-PERK, which indicated that ER stress may be activated. Along with the activation of ER stress, the mitophagy levels were also elevated in our study. Out ndings show that PERK mediated ER stress may play a regulatory role in the elevated Parkin mediated mitophagy in the athophic muscles in this animal model. Moreover, electrical stimulation can partly rectify the abnormally activated ER stress and mitophagy. These ndings may provide a new strategy to rescue muscle atrophy by inhibiting mitophagy through ER stress suppression.

Conclusions
We have shown for the rst time that endoplasmic reticulum stress-induced Parkin-dependent mitophagy may play an important role in the development of muscle atrophy in a rabbit model of knee joint contracture. Electrical stimulation can alleviate disuse muscular atrophy in the animal model by modulating PERK-regulated mitophagy.