Importance of the Innate Immune Response in Skeletal Muscle to Sepsis-Induced Alterations in Protein Balance

Abstract

There is growing appreciation that skeletal muscle is a fully functional component of the body’s innate immune system with the potential to actively participate in the host response to invading bacteria as opposed to being a passive target. In this regard, skeletal muscle in general and myocytes specifically possess an afferent limb which recognizes a wide variety of host pathogens via their interaction with multiple classes of cell-membrane bound and intracellular receptors, including toll-like receptors, cytokine receptors, NOD-like receptors and the NLRP inflammasome. The efferent limb of the innate immune system in muscle is equally robust and with an increased synthesis and secretion of a variety of myocyte-derived cytokines (i.e., myokines), including TNFα, IL-1, IL-6 and NO as well as multiple chemokines in response to appropriate stimulation. Herein, the current narrative review focuses primarily on the immune response of myocytes per se as opposed to other cell types within whole muscle. Moreover, as there are important differences, this review focuses specifically on systemic infection and inflammation as opposed to the response of muscle to direct injury and various types of muscular dystrophies. To date, however, there are few definitive muscle-specific studies which are necessary to directly address the relative importance of muscle-derived immune activation as a contributor to either the systemic immune response or the local immune microenvironment within muscle during sepsis and the resultant down-stream metabolic disturbances.

INTRODUCTION

The liver, spleen, lymph nodes and bone marrow are readily recognized as tissues with important immunological functions which clear the majority of invading pathogens. However, there is an increasing appreciation of the role played by striated muscle in the overall immune response to infection – especially regarding activation of the innate immune system (1, 2). That is, muscle is not merely a target for soluble immune modulators secreted by traditional immune tissues and cells, but an active participant in the whole-body immune response. This immune response can have autocrine/paracrine effects on muscle metabolism, for example, by decreasing muscle protein synthesis, increasing protein degradation, and producing anabolic resistance (3, 4). When sustained, these protein metabolic changes culminate in the loss of muscle mass (e.g., cachexia) and the impairment of contractile function (5, 6) which are associated with a marked increase in morbidity and mortality in sepsis (7, 8). Even in animals that do not succumb to the initial septic insult, there can be a sustained reduction in muscle mass, sarcoplasmic protein synthesis as well an increase in proteasome activity. Furthermore, intramuscular expression of tumor necrosis factor (TNF)-α, interleukin (IL)-1β and IL-6 remains elevated leading to a continued local immune activation and contractile defects during the recovery phase of sepsis that can delay convalescence (9, 10). Such chronic changes in muscle would be consistent with the persistent inflammation, immunosuppression and catabolism syndrome (PICS) that is observed in intensive care patients after trauma and sepsis (11). In addition, the soluble factors released from muscle have the potential to function in an endocrine-like manner thereby potentially modulating the function of distance organs and tissues. Although circulating levels of these proinflammatory cytokines vary widely with different clinical conditions (12), they are believed to be at least in part responsible for much of the organ pathology characteristically produced by bacterial infection. As muscle mass typically represents approximately 60–70% of total body weight, muscle and muscle cells appear to constitute a potentially important element of the immune system, which is relatively underappreciated.

This narrative review highlights current knowledge and recent advances that have increased our understanding related to the importance of skeletal muscle cells per se as a component of the innate immune system and their potential importance in the response to bacterial infection, and salient concepts illustrated in Figure 1. As discussed below in detail, skeletal muscle contains a number of pattern recognition receptors (PRRs) that sense a diverse array of microbial products (e.g., microbial cell wall components, foreign RNA or DNA) termed pathogen-associated molecular patterns (PAMPs), with these PAMPS originating from invading pathogens or the gut microbiota (13). Moreover, PRRs can also bind endogenous ligands released from damaged or stressed cells [damage-associated molecular patterns (DAMPs)] thereby exacerbating cellular and organ injury (14). More recently, this latter class has been recognized to include a subset of danger molecules that has been termed metabolism-associated molecular patterns (MAMPs) which are molecules derived from excess nutrients and their metabolites that are capable of activating PRRs (e.g., free fatty acids, advanced glycation end products, mitochondrial DNA, etc) (15). Herein, we focus our discussion primarily on the immune response of myocytes per se as opposed to other cell types within whole muscle (e.g., endothelial cells, satellite cells, or the small number of resident immune cells seen under basal conditions). However, in response to various types of muscle injury, there is a rapid infiltration and activation of various types of immune cells which aids in the repair and regeneration of this tissue (16). As this response of muscle to direct injury, which is seen in various types of muscular dystrophies and regeneration, differs in important ways from what is seen during systemic infection, it will not be included in the current review. Finally, this review will not focus on sepsis-induced changes in the adaptive immune system, which is also present in muscle, and has been previously reviewed (2).

Figure 1.

Schematic representation of the innate immune response by skeletal muscle cells. Muscle is a heterogenous tissue that processes an afferent sensory limb consisting of, but not limited to, receptors for various immune modulators such as toll-like receptors (TLRs) agonists, D/PAMPs damage-/pathogen-associated molecular patterns), cytokines and chemokines (IL, interleukin; IFN, interferon; TNF, tumor necrosis factor) as well as growth factors and classical hormones (e.g., catecholamines). Additionally, muscle has the cellular machinery to transduce these various ligands via NF-kB and other transcription factors to increase the synthesis and secretion of specific innate immune factors into the interstitial fluid (ISF) and blood, thus constituting the efferent limb. Other abbreviations used in the figure are defined in the text.

Analogous to the peripheral nervous system, the innate immune response in muscle is composed of an afferent and efferent limb. In this model, the afferent limb is composed of a diverse array of receptors that recognize various PAMPs and DAMPs, while the efferent limb is composed of various myocyte-synthesized soluble mediators functioning either in an autocrine/paracrine and/or endocrine fashion.

Afferent Limb of Muscle Immune Response

Toll-like receptors.

The most prominent class of innate sensors constituting the recognition arm of the immune response in muscle, and other tissues, are the toll-like receptors (TLRs) (Figure 1) (17). The TLR family of PRRs consists of 12 members in mouse (TLR1–9 and TLR11–13), but only 10 members in humans (TLR1–10) (18). The function and subsequent downstream signaling pathway for these TLRs is highly dependent on their location. In general, those receptors located on the cell surface (i.e., TLR-1, −2, −4, −5, 6 and −10) usually recognize pathogenic components such as proteins or lipids. For example, TLR1, TLR2 and TLR6 typically recognize lipoproteins, whereas TLR4 and TLR5 recognize lipopolysaccharide (LPS) and flagellin, respectively. The remaining TLRs (i.e., TLR3, −7, −8, −9, −11, −12, and −13) are localized to the endosomal compartment with double-stranded (ds) RNA and single-stranded (ss) RNA binding to TLR3 and TLR7/8, respectively, and ssDNA binding to TLR9. These sentinel molecules selectively bind a diverse array of PAMPs, DAMPs and MAMPs, and thus represent an evolutionary ancient and initial line of defense against pathogen exposure or cellular damage. The response of myocytes to different signals is dictated by the repertoire of TLRs to which they bind, and culminates in the synthesis and secretion of proinflammatory cytokines, interferon (IFN)-γ and other immune modulators (described in more detail below). A more comprehensive and detailed listing of common PRRs in innate immune has recently been published (19).

In the context of gram-negative sepsis, TLR4 which binds its canonical ligand LPS (unless specifically noted, when LPS is used in this review it refers to that derived from Escherichia coli) is recognized for its importance. Numerous studies have now demonstrated the mRNA and/or protein expression of TLR4 in whole skeletal muscle (20–22) as well as in myocytes and myocyte cell lines (20, 22, 23). Furthermore, in vivo studies have reported that skeletal muscle is capable of clearing intravenously administered LPS and provides early evidence of a physiological important ligand-receptor pair in this tissue (24). In addition, the expression of TLR 1–7 mRNA has also been detected in the murine C2C12 myoblast cell line (22). While TLR8 and TLR9 were not detected in this original study, and TLR8 was also not detected in myoblasts or human rhabdomyosarcoma cells (25), a subsequent study did report the presence of both TLR8 and TLR9 mRNA in primary muscle cells isolated from murine gastrocnemius and soleus (26) as well as in whole muscle from humans (27) and mice (28). More recently, TLR 1–6 have been detected in cultured human skeletal muscle cells, whereas the mRNA expression for TLRs 7–9 was undetectable (20, 29). There appears to be very low expression of TRL10 in human muscle, compared to other TLRs in this tissue (27), and mice have been reported to have only a pseudogene (30).

It is noteworthy that the abovementioned data pertaining to TLRs in muscle and myocytes represents mRNA and/or protein expression under basal conditions. However specific TLR ligands can also on occasion modulate the expression of either cognate or heterologous PRRs in muscle. For example, TLR2 mRNA expression was increased for up to 18 hours in C2C12 myotubes incubated with the TLR4 agonist LPS, with a maximal increase occurring 3–5 hours (22, 31). Similarly, incubation with a tripalmitoylated Cys-Ser-Lys-containing peptide (Pam; TLR2 agonist) also increased TLR2, but not TLR4, expression in myotubes. Although beyond the scope of the current review, it is also important to note that various physiological and pathophysiological conditions, including aging (32), ischemia (33), obesity (34), level of activity (35), uremia (36), alcohol (37), and inflammation (25), all of which have the potential to modulate the host response and outcomes to infection, have also been reported to alter TLR expression in skeletal muscle and myocytes.

Cytokine receptors.

Myocytes also possess other cell membrane-bound receptors that do not directly recognize LPS or TLR ligands, but instead recognize proinflammatory cytokines present in the extracellular fluid compartment that were synthesized intramuscularly or present in the systemic circulation. The bulk of the literature in this regard indicates receptors for TNFα, IL-1, IL-6 (among others) are present in skeletal muscle and various myocyte cell lines (38–43), and their activation results in a diverse array of immunologic and metabolic responses within myocytes. However, in the context of the current discussion, it is noteworthy that incubation of myocytes directly with TNFα and IL-1β can also increase the synthesis and secretion of a IL-6 (44–46), while the incubation of myotubes with IL-1β can upregulate IL-6 and NLRP3 (inflammasome) expression (47).

NOD-like receptors.

TLRs are not the only major PRRs in the host defense mechanism residing in muscle. For example, nucleotide-binding and oligomerization domain (NOD)-like receptors (NLRs) serve as intracellular pathogen sensors in classical immune tissues, and are also found in muscle (48). Pathogenic microorganisms which escape initial detection can be present intracellularly and can be recognized by specialized NLRs. The best characterized of these are NOD1 and NOD2, which detect bacterial specific peptidoglycans [γ-D-glu-meso-diaminopimelic acid (iEDAP) and muramyl dipeptide (MDP), respectively], resulting in upregulation of proinflammatory signaling pathways via activation of nuclear factor kappa light chain enhancer of activated B cells (NF-κB). Both NOD1 and NOD2 mRNA can be detected whole tibialis muscle and soleus in mice (48) and gastrocnemius in pigs (49), as well as rat L6 myotubes (48). Moreover, both NOD1 and NOD2 mRNA were upregulated in response to in vivo administered LPS (49). Importantly, in vivo administration of an NOD1 agonist in mice produced multiorgan injury and dysfunction (50). The NOD2 ligand MDP increased mitochondrial reactive oxygen species (ROS) in L6 myotubes, which was associate with time-dependent increases in TNFα, IL-6, CXCL1/KC, MCP1 and a decrease in IL-10 (48, 51).

Inflammasome.

The innate immune response can also be activated by LPS binding to cytoplasmic PRRs which are multiprotein complexes referred to as inflammasomes, which represent a subset of NLRs (52). Foremost among these is the NLRP (nucleotide-binding oligomerization domain (NOD), leucine-rich repeat and pyrin domain-containing protein)-3 inflammasome, which in contrast to the TLRs does not function by activation of NF-κB. For its activation, a complex association is required with specific proteins. Inflammasome activation leads to pyroptosis via activation of caspase-1, and the concomitant secretion of the proinflammatory cytokines IL-1β and IL-18. Furthermore, LPS and various DAMPs have been reported to increase IL-1β secretion in primary murine skeletal muscle cells (53). Sepsis has been reported to increase both NLRP3 and IL-1β expression in whole muscle ultimately leading to elevated concentrations of circulating IL-1β (47). Other classes of PRRs, based on their protein domain homology, are known and include retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs), C-type lectin receptors (CLRs), and absent in melanoma-2 (AIM2)-like receptors (ALRs) (19); however, there is little relevant data to their relative expression and functional importance in skeletal muscle.

Proteomic profiling of C2C12 myotubes incubated with a combination of LPS and IFNγ revealed an increase in more than 1500 proteins with a differential temporal response, many of which were novel and not previously reported in myotubes (42). Essentially all of these identified proteins can be grouped in to biological process Gene Ontology (GO) terms related to Responses to External Stimulus (e.g., response to cytokine, interferon-γ, bacteria) or Response to Stress (e.g., defense response to other organisms, response to bacterium, or innate immune response). This study revealed that LPS + IFNγ increased the synthesis of a family of p47 guanylate binding proteins (GBPs) and p65 immunity-related GTPases (IRGs). These proteins can be recruited to intracellular vacuoles containing pathogens. This ultimately leads to vacuole disruption as well as the subsequent rupture and lysis of the pathogen membrane thereby releasing into the cytoplasm various microbial ligands (e.g., LPS). It is now well established that LPS can be detected in the cytosol of host cells in response to infection with Gram-negative bacteria. This intracellular LPS can activate the non-canonical inflammasome pathway triggering activation of caspase 11 and cleavage of downstream effectors such as Gasdermin D leading to pyroptosis (54). A number of peptides implicated in this signaling pathway, including Gbp2, Gbp2b, Gbp4, Gbp5, Gbp7, Gbp9, and Gbp10/Gbp6, have been detected at various time points in myotubes after LPS/IFNγ stimulation. Likewise, LPS/IFNγ also increased synthesis of IRG family members, both those that are membrane-bound (i.e., IRGM1, IRGM3/IGTP) and those that are predominantly cytosolic (i.e., IIGP1 and TGTP1) in myotubes.

Efferent Limb of Muscle Immune Response

For the above-mentioned PRRs to be physiologically important they must be able to bind to the appropriate ligand, activate intracellular signaling pathways, and increase the transcription and translation of specific myokines (e.g., cytokines synthesized and released by muscle) thereby modulating the immunological microenvironment of muscle (Figure 1). In this respect, myokines have the potential to modulate the immunological milieu in an autocrine/paracrine manner by directly binding to PRRs or indirectly by the regulating the action of other cell types, including resident immune cells, within muscle or by increasing the recruitment of immune cells into muscle. In addition, these soluble molecules can be released into the systemic circulation and may functionally impact non-muscle tissues via a more classical endocrine mechanism. However, what remains largely unknow is whether the secretion of these soluble immune modulators from muscle is sufficient to impact distant organ function (2).

Extensive data from independent laboratories clearly highlight TLR-mediated increases in multiple myokines which have the potential to initiate and perpetuate the inflammatory response. In this regard, both intravenous and intramuscular injection of LPS increased expression of TNFα, IL-6 and nitric oxide synthase (NOS)-2 with increases occurring relatively rapidly (e.g., with 30 min) and being relatively transient (e.g., several hours). Furthermore, in in vitro studies using intact muscles indicated a similar cytokine secretion profile for both muscle with a predominance of slow-twitch versus fast twitch fibers (55). The elevation in these immune modulators appear TLR4-dependent as they are absent or blunted in TLR4 signaling-deficient C3H/HeJ mice that are unresponsive to LPS (56, 57). Additionally, muscle also expresses mRNA for the late-phase cytokine high-mobility-group protein-1 (HMGB1), a nuclear chromatin protein organizing DNA and regulating gene expression. The expression of HMGB1 in muscle is increased during sepsis (57) and appears TLR4-dependent (56). However, the relative contribution of muscle as a source of the increase in HMGB1 – as well as the other cytokines elaborated during sepsis and inflammation –remains poorly defined. One limitation of these types of in vivo studies is they cannot exclude the possibility that the synthesis and secretion of various soluble mediators is driven by the influx of immunocompetent cells into muscle or resident non-myocytes. However, what is evident is that cultured myocytes are responsive to LPS and other TLR ligands, and the consensus from the literature is that muscle cells per se can synthesize and secrete IL-1α, IL-1β, IL-6, IL-8, IL-15, TNFα, CXCL1 (KC), NO and MCP-1 (CCL2), in addition to other potentially inflammatory cytokines and chemokines (31, 55, 58). Hence, muscle fibers are a likely cellular origin for these immunomodulators during infection.

IL-6 was the first and, arguable, the best characterized of the many reported myokines with studies clearly demonstrating the importance of muscle-derived IL-6 as a contributor to circulating levels of the cytokine (59). Both myotubes and myoblasts responded with an increase in IL-6 in a dose- and time-dependent manner to peptidoglycan, LPS from Porphyromonas gingivalis and PAM, all of which specifically activate TLR2 (22). IL-6 mRNA and protein was also increased in a dose- and time-dependent manner in LPS-treated C2C12 myoblasts (56). Furthermore, IL-6 mRNA and IL-6 protein secretion was upregulated in human myotubes when incubated with specific ligands for TLRs 2–6 (20). This latter study also reported an LPS-induced increase in TNFα, IL-12 and IL-1 receptor antagonist (IL-1ra), but no change in IL-1α in myotubes. Additionally, the TLR3 ligand dsRNA mimetic polyinosinic-polycytidylic acid [Poly (I:C)] and TLR7 ligands were capable of increasing IL-6, albeit to a lesser extent than LPS (22, 26). In contrast, IL-6 expression was unaltered in C2C12 myoblasts in response to a bacterial DNA mimetic containing cytosine-phospho-guanine dinucleotide (CpG) motifs, a strong TLR9 ligand (22), consistent with some reports that muscle lacks TLR9 (23). Direct intramuscular injection of either a TLR4 or TLR2 ligand was also capable of increasing IL-6 protein in vivo (22). Importantly, the intramuscular injection of LPS into C3H/HeJ mice, which carry a mutation in the receptor’s TIR (toll/interleukin-1 receptor) domain, failed to elicit an increase in IL-6 (22). Mechanistically, inhibitors of the c-Jun N-terminal kinase (JNK) and p38 pathways blocked LPS-induced increases in IL-6, whereas inhibition of the ERK signaling pathway was without effect, thus suggesting that PAMPs exhibit signaling specificity (46). LPS-incubated myocytes also increased the amount of IL-6 protein secreted into the conditioned media via activation of the NF-kB pathway (22). Finally, although not often studied, there is a potential for the interaction of various TLR ligands and stress hormones which are elevated during sepsis. For example, the catecholamine epinephrine alone increased IL-6 expression in C2C12 myocytes, mediated via the β2-adrenergic receptor, whereas the combination of LPS and epinephrine produced a synergistic increase in IL-6 synthesis in muscle (60).

LPS also increased TNFα in isolated myocytes (22, 26), but not the anti-inflammatory cytokine IL-10 (26). In addition, the expression of the chemokine IL-8 (CXCL8) was increased after exposure of human myocytes to TLR3–6 ligands (20). Furthermore, stimulation with TLR2 or TLR4 ligands increased CCL2 (aka chemokine monocyte chemoattractant protein (MCP)-1) and CXCL1 (keratinocytes-derived chemokine; KC, in mice) (23), which are representative members of the two largest chemokine families. Human skeletal muscle cells also responded to TLR2 agonists by increasing the synthesis of CCL20 and CXCL6 (29). As a result of this increased chemokine expression, such findings suggest the potential to enhance the recruitment of specific leukocyte populations into muscle and further modulation of the immune environment. In these studies, the ability of TLR2 ligands to increase MCP-1 and KC was largely via the calcineurin pathway, while TLR4 ligands increased the expression of these two chemokines via the NF-κB pathway (23).

While secretome analysis has demonstrated human skeletal muscle cells have the potential produce more than 600 myokines (61), there is sparse information pertaining to the effects of TLR ligands on the synthesis of most myokines. For example, there are no reports on LPS- or TLR ligand-induced changes in irisin, myonectin, decorin, fibroblast growth factor (FGF) 21 and secreted protein acidic and rich in cysteine (SPARC) in skeletal muscle, or their biological significance in muscle during sepsis. The exception being myostatin, a known muscle-specific negative regulator of muscle mass, which is increased by LPS in C2C12 myoblasts, an effect antagonized with a specific TLR4 inhibitor (62). Moreover, myostatin deficiency also prevented the sepsis-induced increase in plasma HMGB1, macrophage inhibitory cytokine (MIC)-1 as well as the reduction in muscle mass while increasing survival (8).

Effect of Immune Activation on Muscle Protein Balance

An increase in whole-body nitrogen excretion and reduction in lean body mass (LBM) are some of the earliest reported hallmarks of prolonged sepsis and systemic inflammation (63). Over the short term, the net protein breakdown and release of amino acids is seen as beneficial as it supports the various anabolic functions (e.g., gluconeogenesis and acute phase protein synthesis) of the liver and the maintenance of gut function (64). However, when sustained, the loss of muscle mass impairs muscle contractile function (4) and increases morbidity and mortality (7, 8). This reduction in LBM results from an imbalance between the rate of protein synthesis which is generally observed to be decreased and the rate of protein degradation (e.g., proteasome activity, calpain activity, and autophagy) which is often elevated (3, 65). While sepsis and inflammatory cytokines are capable of altering basal rates of protein synthesis and degradation, they also antagonize nutrient-, hormone- and contraction-induced changes in both sides of the protein balance equation resulting in a condition referred to as anabolic resistance (3, 4, 66). As described in more detail below, this dual effect often complicates efforts to elucidate whether the immune modulator is directly impacting protein balance or the metabolic disturbance was secondary to the development of growth factor resistance (17).

Potential role of TNFα.

It is well accepted that sepsis and other inflammatory states are generally associated with elevated TNFα in both blood and often in peripheral tissues, including skeletal muscle. Several laboratories have reported the systemic administration of TNFα reduced basal rates of muscle protein synthesis (3, 67), and this reduction was seen in both the synthetic rate for sarcoplasmic and myofibrillar proteins (67). Moreover, the TNF-induced reduction in skeletal muscle protein synthesis was associated with a temporally-associated suppressed activity of the mechanistic target of rapamycin complex (mTORC)-1 as evidenced by the decreased phosphorylation of eukaryotic initiation factor (eIF)-4G and 4E-binding protein (BP)-1. These changes were associated with a concomitant reduction in the binding of eIF4G with eIF4E and reduced rate of translation initiation (68, 69). One possible mechanism for this inhibition is the ability of TNFα impair amino acid uptake by skeletal muscle (70). Chronic TNFα infusion over several days also impaired both hepatic and muscle insulin action (71), the latter of which would be anticipated to limit protein synthesis muscle. On the other hand, pretreatment of rats with TNF binding protein (TNFBP) partially or completely prevented the sepsis-induced decrease in skeletal muscle mass and protein content (69, 72). This response was mediated at least in part by the ability of TNFBP to ameliorate the sepsis-induced reduction in muscle protein synthesis by correcting the reduction in translational efficiency. Pretreatment of septic rats with TNFBP largely ameliorated the altered distribution of eIF4E as well as the reduced phosphorylation of 4E-BP1, ribosomal protein S6, and mTOR (69). Finally, TNFBP partially prevented the ability of leucine to stimulate muscle protein synthesis in septic rats and this effect appears mediated by the reversal of the sepsis-induced inhibition of mTORC1 activation (73, 74).

TNFα can also directly impact both sides of the muscle protein balance equation (75). In vivo administration of TNFα has also been reported to impact the other side of the protein balance equation – protein breakdown. Early reports indicated TNFα increased muscle protein degradation (76). Moreover, when TNFα was administered in vivo, the rate of protein degradation remained elevated even when muscle was excised and examined in vitro (77). Intravenous infusion of TNFα, but not IL-1, increased both free and conjugated ubiquitination in muscle in a time-dependent manner suggesting a role for non-lysosomal protein degradation (78). The incubation of myotubes with TNFα reduced total protein content and was in part due to the concomitant reduction in total protein synthesis (79) as well as the synthesis of myosin heavy chain (80). In contrast, the TNFα produced a concomitant upregulation of the muscle-specific E3 ligases Muscle RING (MuRF1) and atrogin-1/ MAFbx (81, 82). Collectively, these changes in protein balance were sufficient to decreased muscle fiber diameter in myotubes cultured with TNFα in a time- and concentration-dependent manner (81, 82). The ability of TNFα to dysregulate protein metabolism could be direct effect of the cytokine on various metabolic pathways or mediated secondary to the coordinate reduction in intracellular concentration of the anabolic hormone insulin-like growth factor (IGF)-I (83).

A causative role for TNFα for the sepsis-induced muscle anabolic resistance has also been postulated. For example, incubation of C2C12 myotubes with TNFα produced an insulin resistant state mediated by the inhibition of IRS/PI3K activity (84) and completely inhibited the increased protein synthesis typically observed in response to IGF-I and serum (79). This resistance of protein synthesis toward various anabolic stimuli appears to be largely mediated via the ability of TNFα to antagonize growth factor-induced stimulation of the mTORC1 signal transduction pathway (85, 86). In addition, incubation of myotubes with septic serum, high in TNFα, induced autophagy and was inhibited by activation of the AKT/mTORC1 (87).

Potential role of IL-6.

IL-6 is another prominent proinflammatory cytokine modulating the innate immune system. Although elevated circulating IL-6 concentrations appear to be a relatively good biomarker for sepsis severity (88), such elevations have also been reported to have limited importance for determining overall mortality (89). Increased concentrations of IL-6, either systemic or intramuscular, can produce muscle atrophy and decrease myofibrillar protein content (90, 91). Importantly, short-term infusion of IL-6 into humans also decreased muscle protein synthesis (92) which may be mediated in part by an increase in AMP-activated protein kinase (AMPK) activity (93). Alternatively, or in addition to, IL-6 also activates STAT (signal transducer and activator of transcription)-3 signaling thereby down-regulating p90RSK/eEF2 and mTOR/S6K1 axes (90, 94). Data pertaining to the effect of IL-6 on muscle protein breakdown, however, are less consistent. In vivo administration of IL-6 has been reported to either increase myofibrillar breakdown when assessed under in vitro conditions (95) or have no effect (96). Likewise, no change in proteolysis was detected in mice injected with IL-6 or in myotubes incubated with the cytokine (97). Finally, infusion of IL-6 in humans has been reported to decrease muscle protein breakdown (92). Despite the lack of a consistent effect on proteolysis, IL-6 produced atrophy in C2C12 myotubes as evidenced by a decreased fiber diameter. As noted above, some of the catabolic effects of these proinflammatory cytokines may be attributed to the concomitant development of growth factor resistance. For example, IL-6 also impaired insulin signaling and this defect was not detected in myocytes isolated from IL-6 deficient mice (98). Conversely, IL-6 transgenic mice had a decreased muscle mass that is prevented by an anti-IL-6 receptor antibody (99).

One of the more definitive studies in this context has been recently published by Yang et al (100). This study reported the sepsis-induced decrease in muscle mass, cross sectional area and contractile function were all prevented in IL-6 knockout mice; however, survival was not assessed. Thus, as sepsis-induced increases in muscle proteolysis are still detected in IL-6 knockout mice (97), IL-6 appears to have a more pronounced effect at reducing muscle protein synthesis than impairing proteolysis. A recently published study was performed using conditional muscle IL-6 knockdown (101). The loss of IL-6 production by muscle per se resulted in a generalized decrease in essentially all cytokines and chemokines assessed (e.g., TNFα, IL-6, IL-10, IFNγ, IL-1α, IL-12, −13 and −15, CM-CSF, CXCL-1, −2, and −10, and CCL-4 and −5) at 6 h in female mice with many of these immune modulators remaining decreased at a 12-h time point. Although muscle mass and indices of protein balance were not assessed in this study, it is noteworthy that despite this generalized suppression of sepsis-induced cytokine response there is no difference in mortality between wild-type and muscle IL-6 knockdown mice. Moreover, the reduction in cytokines in septic mice was much more variable in male mice with the conditional muscle IL-6 knockdown, again emphasizing the importance of studying sepsis-induced change in immune function in both sexes.

Potential role of IL-1.

In comparison to TNFα and IL-6, there are fewer studies that have investigated the potential role of IL-1 on sepsis-induced metabolic derangements in muscle. Early studies indicated the acute in vivo administration of IL-1β had the ability to decrease protein synthesis and increase proteolysis in skeletal muscle (76, 102). Subsequent studies where IL-1β was continuously infused for 6 days also demonstrated a suppression of muscle protein synthesis. This inhibition resulted from a decrease in mRNA translational efficiency that was associated with a reduction in eukaryotic initiation factor 2B (eIF2B) epsilon (103). Moreover, IL-1 has also been reported to increase total and myofibrillar protein breakdown (104). As for the aforementioned cytokines, some of these IL-1-induced changes may be cause by the development of insulin resistance observed in response to the infusion of IL-1 (105). Importantly, neutralizing the in vivo effects of IL-1β by administration of a IL-1 receptor antagonist (IL-1ra) prevented the sepsis-induced inhibition of muscle protein synthesis and loss of muscle mass (106). IL-1ra also prevented the sepsis-induced decrease in the IGF-I concentration in blood and muscle, and it was noted there was strong linear correlation between the IGF-I protein content and muscle protein synthesis (107). Treatment of C2C12 myotubes with IL-1β decreased myotube diameter in a dose-dependent manner which was associated with elevations in the atrogenes MuRF1 and atrogin-1 (47, 108). Additionally, the sepsis-induced increase in muscle IL-1β (and IL-6) as well as and plasma IL-1β was reduced in NLRP3 KO mice. These KO mice also exhibited a smaller sepsis-induced reduction in body weight and muscle cross-sectional diameter as well as improvement in survival. NLRP3 deletion has also been reported to protect against insulin resistance (109), a defining characteristic of sepsis (110), which may be mediated by an increase in GLUT4 translocation to the plasma membrane (111). However, because the concentration of IL-1β was reduced in both muscle and blood of NLRP3 KO mice, it was not possible to determine whether the amelioration of sepsis-induced atrophy was mediated by an endocrine or autocrine/paracrine mechanism.

One of the most rigorous studies to date to directly address the importance of muscle-derived cytokines in sepsis has been published by Laitano et al (112). This study used a conditional knockout of skeletal muscle-specific myeloid differentiation factor 88 (Myd88) in mice, a molecule that is central to canonical TLR and IL-1β signal transduction. Median circulating concentrations of pro- and anti-inflammatory cytokines, including TNFα, IL-1β, IL-1α, IL-4, IL-10, IL-12p40, IL12-P70, IL-13, IL-15, IL-17, IP10, MIP-1α and −1β, and MIP-2, were decreased in knockout mice compared to wild-type controls at the 6-h time point after induction of sepsis. A subset of these cytokines remained reduced at a later 12-h time point. Moreover, skeletal muscle deletion of Myd88 also altered trafficking of immune cells populations to the site of infection (i.e., peritoneal cavity) and decreased short-term survival in female, but not male, mice. Overall, these data are some of the most definitive and directly address the functional importance of skeletal muscle as authentic component of the innate immune system.

Potential role of NO.

Skeletal muscle is now recognized as an important organ in redox-related signaling in general and NO function in particular (113). All three nitric oxide synthase (NOS) isoforms are present in whole muscle, including the inducible form NOS2 (i.e., iNOS), which is transcriptionally upregulated by DAMPs and cytokines under both in vivo and in vitro conditions (31, 45). Activation of NOS2 increases NO synthesis within myocytes and skeletal muscle fibers which can alter cellular metabolism. For example, the effect of LPS and/or IFNγ on mTOR signaling in cultured C2C12 skeletal muscle cells has been examined in detail (31, 114). The combination of LPS and IFNγ decreased autophosphorylation of mTOR and its authentic downstream substrates 4EBP-1, ribosomal protein S6 kinase (S6K)-1, and S6. However, LPS alone was insufficient to decrease mTORC1 activity or protein synthesis in myocytes, but it did upregulate the expression of IL-6 and NOS2. The reliance on IFNγ for this mTORC1 suppression appears to result from its ability to enhance the magnitude and duration of LPS-stimulated NOS2 expression by stabilizing NOS2 mRNA. Hence, the protracted period over which NO levels are elevated causes greater damage to regulatory proteins in growth signaling pathways via nitrosylation. Consistent with this observation, NOS inhibitors prevented the LPS/IFNγ-induced decrease in protein synthesis as well as defects in mTORC1 signaling, and prevented the decrease in myocyte atrophy. Similarly, incubation of C2C12 myotubes with TNFα or IL-1 in combination with IFNγ also markedly increases NO production (45, 115).

Potential role of IGF-I.

Whether the role of the various above-mentioned cytokines on protein balance is direct or mediated secondary to the concomitant reduction in the intramuscular synthesis and action of IGF-I has not been fully explored. However, there is an abundance of circumstantial evidence supporting the relative importance of intracellular IGF-I as a key regulator of protein balance in muscle (116, 117). Systemic infusion of IGF-I (118) as well as the local administration of the growth factor (119) increases muscle protein synthesis via activation of mTORC1 (120, 121). Furthermore, sepsis, LPS as well as TNFα and IL-1 have all been reported to decrease IGF-I in muscle (83, 119, 122–124) and are associated with a reduced rate of mTOR-dependent protein synthesis (125). The cytokine and LPS-induced decrease in IGF-I occurred at both the mRNA and protein level in muscle, and at least for LPS this effect was mediated via TLR4 (75). As IGF-I can also be anti-inflammatory, decreases in muscle IGF-I are often associated with increased expression various cytokines and activation of the NF-κB pathway (126). It is likely that TNFα decreases IGF-I gene transcription as it does not alter mRNA half-life (83). Neutralization of the sepsis-induced increase in either TNFα or IL-1 prevented the typical reduction in muscle IGF-I and protein synthesis (106, 127), and supplementation of septic rats with IGF-I either systemically or locally prevented the decrease in muscle IGF-I as well as the reduction in protein synthesis and muscle atrophy (119, 128). TNFα enhanced c-Jun phosphorylation and specific inhibition of the JNK signaling pathway antagonized the TNFα-induced drop in IGF-I mRNA in myocytes. Finally, TNFα has also been reported to blunt the growth hormone-induced increase in IGF-I within muscle and myocytes (83, 129). Collectively, these data suggest the increase in autocrine-derived cytokines in response to LPS and possibly sepsis may contribute to the suppression of muscle IGF-I and protein synthesis. The sepsis-induced reduction in intramuscular IGF-I may also be in part responsible for the increase in the E3-ligase atrogin-1 and protein ubiquitination in muscle driving proteolysis (130).

Conclusion

The data in the literature clearly demonstrate all elements of an intact innate immune system are present within skeletal muscle and are immunological active (Figure 1). Additional studies will undoubtedly expand on the mechanistic aspects of the various individual components and pathways described. What is unclear currently and represents an unmet need in this area is a more in-depth understanding of the physiological importance of the skeletal muscle immune system to the metabolic and contractile properties of this tissue as well as its clinical significance in modulating systemic immunity. Answers for these questions will not be found in future studies which are correlative or descriptive. Indeed, studies will only be definitive when they include the knockout or knockdown of individual elements of the innate immune system in a muscle-specific manner. In addition, it is well recognized that muscle is a heterogenous tissue and that the basal and/or activation of myocytes per se undoubtedly has differences when compared to other cell types within whole muscle (e.g., epithelial cells, and resident and/or recruited classical immune cells). Finally, many studies have examined only a limited array of immune modulators and future studies will need to examine both pro- and anti-inflammatory mediators and pathways using more integrative approaches to assess the relevance of skeletal muscle to overall immunity in sepsis both acutely and during convalescence.