Plasma from septic shock patients induces loss of muscle protein

Two separate sets of experiments were performed. The first set of experiments was conducted to establish whether plasma in the early phase of septic shock induces muscle proteolysis and atrophy. To this end, cultured skeletal muscle myotubes were incubated with either plasma from patients with septic shock obtained within 24 hours of ICU admission or plasma from healthy subjects (controls). After 24 hours of exposure to plasma, skeletal myotubes were harvested for biochemical analysis, including myosin heavy chain content and expression of E3-ligases. The activity of nuclear factor kappa b (NFκB) was measured after one hour of exposure to plasma.

The second set of experiments was conducted to investigate if the ability of plasma from septic shock patients to induce muscle atrophy changes during the course of ICU admission and if a correlation exists between inflammatory cytokines and loss of myosin. Hence, plasma was derived at different time points during an ICU stay of patients with a septic shock and the ability of each plasma sample to induce proteolysis and atrophy in differentiated muscle cells was examined. In addition, plasma levels of cytokines were determined at each time point.

Twenty-one patients with established septic shock, according to the 2001 International Sepsis Definitions Conference [10], were included in the first set of experiments. Blood was withdrawn from the indwelling arterial catheter within 24 hours of ICU admission. In 12 control subjects blood was obtained through venapuncture.

For the second set of experiments, additional blood samples were obtained from 9 of the 21 septic shock patients at Days 2, 5 and 7 of the ICU stay. The institutional review board, the medical ethics committee, approved the current study and waived the need for informed consent.

Muscle cells, C2C12 myoblasts, were cultured into myotubes according to previously described methods [11]. Pilot studies showed stable myosin content in myotubes that differentiated for seven days and that plasma from healthy subjects does not affect myosin content in these myotubes. See Additional file 1 for data and more details.

Myosin content in myotubes was determined by Western blotting according to previous described methods. See Additional file 1 for more details.

NFκB is a ubiquitous transcription factor for a variety of cytokines. The DNA binding activity of NFκB has been shown to be highest within several hours after exposure to different stimuli [12, 13], which was confirmed by pilot-experiments in our lab. Therefore, in the current study NFκB activity was measured one hour after incubation of the myotubes with plasma.

NFκB DNA binding activity was determined by electrophoretic mobility shift assay (EMSA) conform previously described methods [14]. See Additional file 1 for details.

Expression of muscle specific E3-ligases, Muscle RING Finger protein-1 (MuRF-1) and Muscle Atrophy F-box protein (MAFbx) and ubiquitination of myosin was analyzed, conforming to previously described methods [15], see Additional file 1 for details.

Plasma levels of IL-6, IFN-γ, TNF-α and IL-1β were measured by enzyme-linked immunosorbent assay (ELISA) (for IL-6, IFN-γ and TNF-α ELISA kits from Sanquin Reagents, Amsterdam, The Netherlands, for IL-1β Quantikinekit from R&D Systems, Minneapolis, MN, USA). To further examine the role of IL-6, we first studied myosin content in myotubes that were exposed to septic shock plasma containing an antibody directed against human IL-6 (R&D Systems). In addition, we studied myosin content in myotubes exposed to control plasma containing different concentrations of recombinant human IL-6 (Invitrogen, Carlsbad, CA, USA).

Data are presented as means ± standard deviation. The statistical significances of differences regarding myosin content, MuRF-1 and MAFbx expression between myotubes exposed to plasma from controls and patient plasma were analyzed by performing Student t-tests. Differences regarding NFκB activity and cytokine concentrations were statistically tested by performing a Mann Whitney U test. Sample size calculations for the follow-up study showed that nine patients should be sufficient to detect a 25% reduction of myosin content with a standard deviation of 20%, 80% probability and an alpha level of 0.05. One-way ANOVA with post-hoc Bonferroni's multiple comparison testing was used to evaluate whether myosin content, MuRF-1 and MAFbx expression and cytokine levels at each day were statistically different from control. Repeated measures one-way ANOVA with post-hoc Bonferroni's multiple comparison testing was performed to analyze differences between Day 0 and subsequent days. The relation between myosin and IL-6 levels was explored using a linear model estimated by Generalized Least Squares to allow for correlations between the measurements caused by repeatedly measuring the same subjects; an unstructured correlation matrix was used to model this dependence. Since IL-6 levels were not normally distributed, log transformation on these values was applied. P-values below 0.05 were considered significant.

Patient characteristics are shown in Table 1. All patients met septic shock criteria. Thirteen of the 21 patients received one bolus of 100 mg hydrocortisone before blood withdrawal. Additional characteristics of patients included in the second set of experiments are shown in Additional file 1, Table S1. None of the control subjects reported any significant past medical history or current use of prescribed medication.

Because the current data demonstrate that plasma from patients with septic shock induces atrophy at ICU admission, we performed additional experiments to follow this atrophic response in the course of ICU stay.

The use of cultured C2C12 skeletal myotubes in the current study is highly appropriate to specifically study the contribution of plasma ligands in muscle wasting, since the use of non-diseased muscle tissue virtually excludes any contribution of intrinsic muscle abnormalities. Other in vitro models have been used previously by others, including dissected muscle bundles [8, 18]. A disadvantage of this latter model is the absence of muscular microcirculation. This induces oxidative stress and limits the supply of ligands to the muscle bundle [19, 20]. This may explain why incubation with non-septic human plasma also increased protein degradation in these models [8]. Therefore, we considered that model not suitable for the current research questions.

C2C12 skeletal myotubes reach a considerable degree of differentiation, as indicated by expression of fast twitch skeletal muscle troponin T, alpha-actin and tropomyosin [21]. Permeabilized myotubes generate force when perfused with calcium solutions, with a similar calcium sensitivity of force generation compared to mature skeletal muscle fibers [21]. Nevertheless, data from our own lab [11] and others [22, 23] demonstrated differences in some physiological processes compared to mature skeletal muscle, such as intracellular calcium handling. Accordingly, C2C12 cultured myotubes are suitable for addressing specific research questions, but limitations should be recognized.

Sepsis affects skeletal muscle physiology at different stages of excitation-contraction coupling, including contractile protein function, muscle protein content (atrophy), membrane excitability and mitochondrial function [24]. We found that myosin content is reduced by approximately 25% after exposure of muscle to plasma from patients with septic shock. These data are in line with previous studies on rodents, showing a rapid loss of muscle mass upon induction of sepsis [6, 25, 26]. The triggers that activate muscle proteolysis in early septic shock are largely unknown. Observational studies have proposed several risk factors, including cytokines, corticosteroids, hyperglycemia and immobilization [1]. Yet, as their effects intermingle in vivo, it is very difficult to establish which of these risk factors do play a role in skeletal muscle wasting in critically ill patients. By specifically studying the potency of plasma to induce muscle wasting, the current study demonstrates that plasma ligands play a prominent role in inducing muscle proteolysis in patients with septic shock. A previous study found that serum of critically ill patients affect membrane excitability and the excitation-contraction coupling process of isolated muscle fibers [27]. Altogether these findings indicate that circulating factors contribute to the development of muscle weakness in critically ill patients.

We demonstrated that the atrophic response to plasma obtained beyond the day of ICU admission was less prominent, but still present. These findings underscore the necessity of early interventions in the prevention of muscle atrophy in septic shock patients. Since the set-up in the current study specifically addresses the effect of plasma ligands, we do not exclude that other factors, such as immobilization [28] and production of inflammatory mediators by the muscle itself [29, 30] also contribute to muscle wasting in septic patients, in particular during prolonged ICU stay.

An important question is which plasma factors in septic shock patients initiate muscle atrophy. First, inflammatory cytokines such as IL-6, TNF-alpha, IL-1-beta and IFN-γ are often implicated in muscle wasting diseases. In our study, the latter three cytokines were below the detection limit in all healthy subjects and the majority of the septic shock patients, suggesting that these were not a major factor in the development of atrophy in our model. Plasma IL-6 levels were elevated in patients with septic shock and plasma levels significantly correlated with the severity of myosin loss. Moreover, blocking IL-6 in plasma from septic shock patients diminished the atrophic response in skeletal myotubes. These data indicate a prominent role of IL-6 in inducing muscle atrophy during septic shock. Noteworthy, the addition of IL-6 to plasma from controls did not induce atrophy of skeletal myotubes, even when a supra-physiological IL-6 concentration of 50 ng/ml was applied. Thus, while IL-6 in plasma from septic shock patients is important to induce severe muscle atrophy, other plasma factors seem to be needed as well. Second, hyperglycemia has been associated with muscle wasting in critically ill patients [31] and hyperglycemia induces protein degradation in cultured muscle [32]. Yet, patients in the current study were normoglycemic (average glucose level 6.8 ± 0.9 mM), as strict glucose control is part of our routine clinical care. Accordingly, it is unlikely that hyperglycemia did directly contribute to muscle wasting in our model. Third, neuromuscular blocking agents have been associated with the development of muscle atrophy [33], although this has been challenged in recent clinical studies [34]. Nevertheless, the last bolus of rocuronium was administered more than four hours before blood withdrawal, ruling out an effect of rocuronium in muscle wasting in our study. Fourth, high doses of corticosteroids have been associated with skeletal muscle wasting [35]. In the current study, 13 out of 21 patients received low dose (maximal one bolus of 100 mg i.v.) hydrocortisone prior to blood withdrawal. No significant difference in atrophy response at Day 1 was observed between patients that had received hydrocortisone and steroid naive patients (Figure 3C). Moreover, at the time more patients had received hydrocortisone (Day 2 to Day 7) the atrophic response was lower than on Day 1. Plasma cortisol concentrations in five of the studied septic shock patients were above normal levels (150 to 700 nM). Yet, these plasma samples provoked similar reductions of myosin in myotubes as plasma samples with normal cortisol levels (data not shown). Finally, metabolic acidosis is known to induce muscle proteolysis by a glucocorticoid-dependent mechanism [36]. Although plasma pH in most patients was acidic, there was no significant relation between plasma pH and myosin concentration (data not shown).

The main focus of this study was to investigate whether myosin loss is triggered by plasma from patients with septic shock, but we also studied activation of proteolysis in these skeletal myotubes. The ubiquitin-proteasome pathway is the main proteolytic system in eukaryotic cells and controls both protein quality and quantity [37]. During the course of this pathway proteins are linked to a chain of ubiquitin molecules under regulation of E3-ligases such as MuRF-1 and MAFBx and, subsequently, recognized and degraded by the proteasome. Recent evidence from our lab indicates that myosin degradation follows this pathway, as proteasome inhibition restores myosin content and muscle function in animal models for respiratory muscle weakness [15]. Moreover, components of the ubiquitin-proteasome pathway are up-regulated in skeletal muscle of septic shock patients [38, 39]. The current study adds to these earlier observations as we show that plasma from septic shock patients increases ubiquitinated myosin levels and activates MuRF-1 and MAFBx in skeletal myotubes. MuRF-1 has been shown to specifically ubiquitinate myosin, thereby promoting myosin degradation [40, 41]. MuRF-1 expression is under control of the transcription factor NFκB [42]. Indeed, exposure to patient plasma increases the activity of NFκB in myotubes within one hour. MAFbx expression occurs independent from NFκB activity, but is also associated with loss of muscle proteins [43]. Noticeably, in mice overexpression of circulating IL-6 enhances MAFbx mRNA and induces loss of muscle mass [44]. In line with that study, we found that MAFbx expression diminished after Day 2 on the ICU and followed a similar trend as IL-6, which in turn is inversely related to myosin content. Therefore, these data further support the notion that IL-6 is involved in the initiation of skeletal muscle atrophy in septic shock patients.