1. Introduction
Sepsis is a complex syndrome characterized by an overwhelming infection that results in a severe systemic inflammatory response. Sepsis causes diverse vascular, metabolic and endocrine abnormalities that lead to multiple organ failure, and often result in death [
1]. Amongst the very deleterious effects of sepsis is severe weakness, which involves both respiratory and limb skeletal muscles [
2,
3,
4,
5]. In the short term, sepsis-induced respiratory muscle weakness leads to difficulty removing patients from mechanical ventilation, increases the risk of the recurrence of respiratory failure, prolonged hospitalization and increased mortality [
6]. In sepsis survivors discharged from the intensive care unit, the long-term ramifications of sepsis-induced limb muscle weakness included functional impairment, limited physical activity and poor quality of life [
7].
There is currently a lack of effective therapies to either prevent or treat sepsis-induced skeletal muscle weakness, due in large part to the fact that its molecular and cellular bases are poorly understood. However, one clue lies at the ultrastructural level, where significant accumulations of damaged and dysfunctional mitochondria are characteristic of sepsis-induced muscle dysfunction [
8,
9]. Indeed, Bready et al. showed that, in human skeletal muscle, sepsis results in decreased complex I activity (a key enzyme of the mitochondrial electron transfer system) and declined the ATP/ADP ratio in skeletal muscles [
10]. These defects in muscle bioenergetics were also observed in a rat model of sepsis [
11]. By studying biopsies obtained from septic patients, Fredriksson K et al. described a 30% decrease in complex IV activity in limb skeletal muscles [
12]. Several studies on experimental animals also reported that sepsis results in decreased mitochondrial respiration [
13,
14,
15,
16] and an increase in the mitochondrial production of reactive oxygen species (ROS) in skeletal muscle [
17,
18]. Sepsis has also been shown to increase the levels of morphologically abnormal mitochondria, such as those with disorganized cristae, translucent vacuoles and even myelin-like structures [
13,
19,
20,
21]. Recently, Owen et al. showed that persistent muscle weakness in mice that have survived sepsis is associated with abnormal mitochondrial ultrastructure, decreased respiration, decreased activity of complexes of the mitochondrial electron transfer system and persistent oxidative damage to muscle proteins [
21].
In healthy muscles, damaged or dysfunctional mitochondria are selectively recycled in a process, known as mitophagy (selective autophagy of mitochondria), which is primarily regulated through the PINK1-Parkin pathway. Parkin, an E3 ubiquitin ligase encoded by the
Park2 gene, is a 465 amino acid protein that translocates to depolarized mitochondria to initiate mitophagy. Parkin-dependent mitophagy is regulated by PTEN-induced kinase 1 (PINK1), which acts upstream from Parkin. In healthy mitochondria, PINK1 is imported into the inner mitochondrial membrane and cleaved by PARL [
22]. Cleaved PINK1 is then released into the cytosol where it is degraded by the proteasome system. In depolarized mitochondria, the importation of PINK1 into the inner mitochondrial membrane is blocked. PINK1 is no longer degraded and becomes phosphorylated and stabilized on the outer mitochondrial membrane [
23,
24,
25,
26]. Phosphorylated PINK1 triggers the recruitment of Parkin to the mitochondria. Parkin then ubiquitinates outer mitochondrial membrane proteins, including the fusion proteins MFN1, MFN2, MIRO and TOMM20 [
27]. The degradation of MFN1 and MFN2 triggers mitochondrial fission and fragmentation, both of which are important to the recycling of mitochondria by the mitophagy pathway [
28]. The functional importance of the PINK1-Parkin mitophagy pathway in regulating skeletal muscle mitochondrial function and quality in sepsis remains unknown. Recently, we reported that the genetic deletion of Parkin leads to the poor recovery of cardiac function in septic mice and increased sepsis-induced mitochondrial dysfunction in the heart [
29]. We also demonstrated that autophagy is significantly induced in the skeletal muscles of septic mice and that the induction of autophagy is associated with increased muscle Parkin levels, suggesting that mitophagy was induced [
20,
30]. However, several morphologically and functionally abnormal mitochondria were observed in the electron micrographs of septic muscles, indicating that the mitophagy that was induced was likely insufficient to the task of completely recycling defective mitochondria [
20,
30]. Based on this reasoning, we hypothesized that enhancing mitophagy through Parkin overexpression would attenuate the impact of sepsis on skeletal muscles and their mitochondria. To test this hypothesis, Parkin was overexpressed for four weeks in the skeletal muscles of young mice using intramuscular injections of adeno-associated viruses (AAVs). The cecal ligation and perforation (CLP) procedure, a widely used model of sepsis [
31], was used to induce sepsis. Sham-operated animals served as controls. We found that Parkin overexpression prevents sepsis-induced mitochondrial morphological injury and reverses the decline in mitochondrial protein content. We also found that Parkin overexpression protects against sepsis-induced myofiber atrophy. These findings indicate that defective mitophagy in sepsis can be therapeutically manipulated as a means of counteracting sepsis-induced muscle dysfunction.
4. Discussion
The accumulation of dysfunctional and injured mitochondria in skeletal muscles is believed to play a key role in the development of muscle weakness during sepsis [
8,
9]. In the current study, we investigated whether overexpressing Parkin, a key component of the PINK1-Parkin mitophagy pathway, could attenuate the negative impact of sepsis on skeletal muscles and their mitochondria. The current study indicates that Parkin overexpression prevented sepsis-induced accumulation of enlarged and complex mitochondria in the limb muscles of mice. Parkin overexpression also attenuated the sepsis-induced decrease in the content of complexes I and IV of the mitochondrial electron transfer system and prevented the development of limb muscle atrophy in septic mice. These results expand recent studies demonstrating that Parkin exerts protective effects on skeletal muscle health. Indeed, our group has recently reported that
Park2-/- mice have decreased limb muscle contractility, depressed muscle mitochondrial respiration, increased mitochondrial uncoupling and enhanced susceptibility to the opening of mitochondrial permeability transition pore compared to wild-type (WT) mice [
42].
Park2-/- mice also exhibit the impaired recovery of cardiac contractility and depressed cardiac mitochondrial functions in sepsis [
29]. More recently, Peker et al. have reported that Parkin knockdown in C2C12 cells results in myotubular atrophy and that
Park2-/- mice have decreased muscle mitochondrial respiration and increased levels of reactive oxygen species and fiber atrophy [
43]. Parkin overexpression in the muscles of
Drosophila melanogaster increased mitochondrial content, decreased proteotoxicity and extended lifespan [
44]. Our finding that Parkin overexpression in the Sham group increased limb muscle fiber diameters is in accordance with our recent study documenting that Parkin overexpression for several months in young mice causes muscle hypertrophy, while in old mice, Parkin overexpression attenuates ageing-related loss of muscle mass and strength, increases mitochondrial content and enzymatic activities and protects from ageing-related oxidative stress, fibrosis and apoptosis [
32]. Taken together, our current findings and published studies highlight the protective role that Parkin plays in skeletal muscle health.
Our findings that sepsis elicits distinct changes in skeletal muscle mitochondria, such as decreased VDAC level (a marker of mitochondrial content [
45,
46,
47]), the downregulation of three mitochondrial biogenesis genes (
Pgc1-α,
Tfam and
Sirt3) and decreased complexes I and IV levels, are in agreement with published studies on septic humans and experimental animals [
10,
12,
13,
14,
20]. We report for the first time that Parkin overexpression in skeletal muscle prevents the inhibitory effects of sepsis on the expression of
Tfam and on the content of complexes I and IV, as well as VDAC. Based on these results, we speculate that Parkin overexpression might have improved mitochondrial functions in septic muscles. This speculation is supported by the observation that Parkin overexpression increases mitochondrial content and enzymatic activities in normal skeletal muscles [
32,
44]. We should emphasize that, in the current study, Parkin overexpression increased
Nrf2 expression in the skeletal muscles of septic animals. Considering the role that this transcription factor has in the regulation of the expression of several anti-oxidant enzymes [
48], we anticipate that increased
Nrf2 levels in muscles overexpressing Parkin might have contributed to the protection of mitochondrial morphology and contents in septic animals.
Mitochondria form a dynamic network constantly undergoing fusion and fission events that tightly regulate the shape (i.e., morphology), size and number of mitochondria [
41,
49]. In the present study, we show that sepsis significantly alters mitochondrial morphology by increasing the proportion of enlarged and more complex IMF mitochondria. These results extend previous observations, showing that sepsis causes major alterations of the mitochondrial ultrastructure in skeletal muscle [
13,
19,
20,
21]. This increase in mitochondrial size and complexity in septic muscles might have been caused by decreased DRP1 contents and activation [
41], which are expected to alter the fusion/fission balance towards increased mitochondrial fusion. Since mitochondrial fission is required for mitochondrial degradation through mitophagy [
50], it is possible that decreased DRP1 content and activation may play a role in the accumulation of damaged and dysfunctional mitochondria in septic muscles by impairing muscle capacity to recycle dysfunctional mitochondria through the mitophagy pathway. It should also be noted that a decrease in DRP1 content per se might have also played a role in myofiber atrophy. Indeed, a recent study showed that muscle-specific DRP1 deletion results in severe muscle dysfunction, characterized by atrophy, weakness, fiber degeneration and mitochondrial dysfunction [
51]. Importantly, we found that Parkin overexpression attenuated sepsis-induced changes in mitochondrial morphology and rendered muscle mitochondria to be smaller, more circular and simpler, relative to muscles expressing GFP. These findings are in agreement with previous reports, indicating that Parkin overexpression in skeletal muscles and neurons stimulates mitochondrial fragmentation [
44,
52]. We speculate that the decrease in mitochondrial size and complexity in septic muscles overexpressing Parkin might have facilitated the recycling of damaged/dysfunctional mitochondria.
A puzzling result of the present study is the increase in the proportion of enlarged and more complex mitochondria in the Parkin overexpressing muscles of sham-operated mice. The mechanisms behind the differences in the effects of Parkin on mitochondrial morphology in the sham and CLP groups remain unclear. We should point out, however, that although Parkin overexpression altered the mitochondrial morphology in skeletal muscles, none of the parameters related to mitochondrial dynamics that we investigated were affected by Parkin overexpression. Indeed, no significant impact of Parkin overexpression on the expression and protein content of MFN2, OPA1 and DRP1 was evident. Furthermore, Parkin overexpression had no effect on DRP1 phosphorylation on Ser616, suggesting that there was no change in DRP1 activity. Further studies are therefore required to identify the mechanisms underlying the differential impact of Parkin overexpression on skeletal muscle mitochondrial morphology in healthy and septic animals.
In the current study, we report data indicating that autophagy was induced in septic skeletal muscles, as evidenced by the increased expression of several autophagy-related genes, including
Lc3b,
Gabarapl1, and
Sqstm1, and by the increase in the LC3B-II/LC3B-I ratio in the muscles of CLP mice. These findings are in agreement with previous reports, which documented increased muscle autophagy in various models of sepsis [
20,
30,
53]. An interesting finding in our study is that Parkin overexpression had no effects on the expression of autophagy-related genes and the LC3B-II/LC3B-I ratios in the sham and CLP groups. Given the key role that Parkin plays in the recycling of dysfunctional mitochondria by autophagosomes [
23,
50], our results indicate that basal and activated autophagy levels in normal and septic muscles, respectively, were sufficient to deal with increased mitophagy in muscles overexpressing Parkin. We also observed that BNIP3 mRNA and protein levels increased significantly in septic skeletal muscles, and that this induction was not influenced by Parkin overexpression. The BNIP3 protein localizes to the mitochondria and promotes PINK1-Parkin-independent mitophagy by interacting with the LC3 protein, resulting in the recruitment of autophagosomes to damaged mitochondria [
54]. The lack of changes in BNIP3 levels in response to Parkin overexpression suggests that BNIP3-mediated mitophagy functions in an independent fashion to that of the PINK1-Parkin pathway. It is worth mentioning that the present study suffers from several limitations. First, we did not directly assess whether Parkin overexpression actually translates into increased mitophagic flux. Although it was recently reported that Parkin overexpression is sufficient to trigger higher mitochondrial clearance in cardiomyocytes [
55], further studies should investigate whether Parkin overexpression is sufficient to increase mitophagy in healthy and septic skeletal muscles. Another important limitation arises from the fact that muscle contractility was not assessed in the present study. Further studies are therefore required to define whether Parkin overexpression can attenuate sepsis-induced skeletal muscle weakness.