Calcium-dependent phospholipase A₂ modulates infection-induced diaphragm dysfunction

Abstract

Calpain activation contributes to the development of infection-induced diaphragm weakness, but the mechanisms by which infections activate calpain are poorly understood. We postulated that skeletal muscle calcium-dependent phospholipase A₂ (cPLA₂) is activated by cytokines and has downstream effects that induce calpain activation and muscle weakness. We determined whether cPLA₂ activation mediates cytokine-induced calpain activation in isolated skeletal muscle (C2C12) cells and infection-induced diaphragm weakness in mice. C2C12 cells were treated with the following: 1) vehicle; 2) cytomix (TNF-α 20 ng/ml, IL-1β 50 U/ml, IFN-γ 100 U/ml, LPS 10 μg/ml); 3) cytomix + AACOCF3, a cPLA₂ inhibitor (10 μM); or 4) AACOCF3 alone. At 24 h, we assessed cell cPLA₂ activity, mitochondrial superoxide generation, calpain activity, and calpastatin activity. We also determined if SS31 (10 μg/ml), a mitochondrial superoxide scavenger, reduced cytomix-mediated calpain activation. Finally, we determined if CDIBA (10 μM), a cPLA₂ inhibitor, reduced diaphragm dysfunction due to cecal ligation puncture in mice. Cytomix increased C2C12 cell cPLA₂ activity (P < 0.001) and superoxide generation; AACOCF3 and SS31 blocked increases in superoxide generation (P < 0.001). Cytomix also activated calpain (P < 0.001) and inactivated calpastatin (P < 0.01); both AACOCF3 and SS31 prevented these changes. Cecal ligation puncture reduced diaphragm force in mice, and CDIBA prevented this reduction (P < 0.001). cPLA₂ modulates cytokine-induced calpain activation in cells and infection-induced diaphragm weakness in animals. We speculate that therapies that inhibit cPLA₂ may prevent diaphragm weakness in infected, critically ill patients.

recent studies indicate that the diaphragm becomes profoundly weak in critically ill, mechanically ventilated patients, with the force-generating capacity of the diaphragm falling, on average, to ∼20% of the normal value (8, 14, 15, 31, 40). Diaphragm weakness, in turn, is thought to greatly contribute to patient morbidity, potentiating the development of prolonged respiratory failure, as well as increasing the likelihood of death (8, 31). Several factors are thought to contribute to the genesis of diaphragm weakness in these individuals, including ventilator-induced inactivity, use of systemic corticosteroid therapies, and the presence of infection. The latter is an especially important risk factor (8, 31), since the majority of patients in critical care units are infected, with infection either precipitating critical illness (e.g., sepsis) or complicating it (nosocomial line infections, hospital acquired pneumonia, etc.).

Currently, however, there is no specific treatment to prevent or reverse infection-induced diaphragm weakness, and the mechanisms by which infections reduce diaphragm force-generating capacity are incompletely understood. Recent work, however, provides important clues regarding the potential cellular processes responsible for the induction of infection-induced skeletal muscle weakness. Some work suggests that cytokines play an important role in the development of infection-induced skeletal muscle weakness (30). It is also known that infection-induced skeletal muscle weakness is greatly attenuated in animals selectively overexpressing calpastatin, the endogenous calpain inhibitor, in skeletal muscle (37). It is, therefore, reasonable to postulate that infection-induced muscle weakness is due, at least in part, to the sequential effects of increases in circulating cytokine levels, cytokine-induced activation of muscle calpain activity, and the downstream effects of activated calpain to cleave important cytoskeletal and contractile proteins, thereby impairing skeletal muscle force generation. While this line of reasoning makes a good argument that cytokine-induced skeletal muscle calpain activation may be a key process in the induction of infection-induced skeletal muscle weakness, almost nothing is known about the mechanism(s) by which cytokines induce skeletal muscle calpain activation.

The purpose of the present study was to examine this issue and to test the hypothesis that cytokine-induced calcium-dependent phospholipase A₂ (cPLA₂) activation is a critical intermediary step required for skeletal muscle calpain activation. Specifically, we hypothesized that cytokines induce skeletal muscle cPLA₂ activation, that cPLA₂ activation elicits generation of mitochondrial superoxide, and that mitochondrial superoxide, in turn, triggers calpain activation. These studies were carried out using a skeletal muscle cell line (C2C12 cells) and intact mice. To simulate the effects of infection, skeletal muscle cell lines were exposed to a mixture of cytokines (cytomix, containing TNF-α, IL-1β, IFN-γ, and LPS) (5), and sepsis was induced in intact animals by cecal ligation puncture (CLP)-induced peritonitis (38). The role of cPLA₂ in mediating activation of calpain and enhancement of mitochondrial free radical generation was assessed by inhibiting cPLA₂ using the cPLA₂ inhibitors arachidonyltrifluoromethane (AACOCF3) and 4-{2-[5-chloro-1-(diphenylmethyl)-2-methyl-1H-indol-3-yl]-ethoxy}benzoic acid (CDIBA), respectively, in cells and animals.

METHODS

Experimental protocols: cell studies.

Cell studies were performed using the C2C12 cell line obtained from ATCC (Manassas, VA). C2C12 myoblasts were grown to 70% confluency in DMEM (Dulbecco's modified Eagle's medium) containing 10% fetal bovine serum and antibiotics. To induce differentiation, media was changed to DMEM with 2% horse serum for 7 days. We first determined whether exposure to cytomix (TNF-α 20 ng/ml, IL-1β 50 U/ml, IFN-γ 100 U/ml, and LPS 10 μg/ml) increased C2C12 cell cPLA₂ activity by comparing cPLA₂ activity for four experimental groups (n = 6 for each group): 1) control cells with saline added to the media; 2) cells with cytomix added; 3) cells treated with saline and GW4869 (10 μM), an N-SMase inhibitor; and 4) cells treated with both cytomix and GW4869. Cells were harvested 24 h after exposure and assayed for cPLA₂ activity as described below (see section cPLA₂ activity). Our use of a cytokine mixture in vitro to model the effects of sepsis is based on previous reports demonstrating that CLP-induced peritonitis results in an increase in circulating levels of TNF-α, IL-1β, and IFN-γ (11, 22, 23).

We next determined whether cytomix elicited an increase in C2C12 cell mitochondrial superoxide generation by studying the following six groups of cells (n = 4–7/group): 1) control cells with saline added to the media; 2) cells with cytomix added; 3) cells treated with saline and AACOCF3 (10 μM), a cPLA₂ inhibitor; 4) cells treated with both cytomix and AACOCF3; 5) cells treated with saline and SS31 (10 μg/ml), a mitochondrially targeted superoxide scavenger (6); and 6) cells treated with cytomix and SS31. Cells were harvested at 24 h after exposures and assayed for mitochondrial superoxide generation using the MitoSox assay described below (see section MitoSox assay). We also performed a similar study in which we measured aconitase activity as an index of mitochondrial superoxide generation. For this experiment, four groups of cells were studied (n = 4/group): 1) control cells with saline added to the media; 2) cells with cytomix added; 3) cells treated with saline and AACOCF3 (10 μM); and 4) cells treated with both cytomix and AACOCF3. Cells were harvested at 24 h after exposures, and aconitase activity was measured as described below (see section Aconitase activity).

To determine the effect of cytomix on C2C12 cell calpain activity, we utilized two indexes to assess calpain activation, including formation of a calpain-specific caspase-12 proteolytic cleavage product, as indicated on Western blots, and an assay that directly measures cell calpain activity. For these experiments, we studied six groups of cells: 1) control cells with saline added to the media; 2) cells with cytomix added; 3) cells treated with saline and AACOCF3 (10 μM); 4) cells treated with both cytomix and AACOCF3; 5) cells treated with saline and SS31 (10 μg/ml); and 6) cells treated with cytomix and SS31. Cells were harvested at 24 h after exposures, and calpain-dependent caspase-12 cleavage (n = 6–8/group) and cell calpain activity (n = 4–6/group) were determined, as described below. These same six experimental groups of cells were also used to assess the effects of cytomix on C2C12 cell calpastatin activity (n = 4–5/group) and calpastatin protein levels (n = 6/group). For this last group of experiments, cells were harvested 24 h after exposures, and calpastatin activity and protein levels were determined as described below (see sections Western blotting and Calpain activity assay).

Cell viability assay.

To determine whether cytokines may have affected muscle cell line viability, we used the Trypan blue technique. For these cell studies, C2C12 cells were grown and differentiated as described in the above protocol. On the 5th day of differentiation, myotubes were treated with or without cytomix cocktail (cytomix: TNF-α 20 ng/ml, IL-1β 50 U/ml, IFN-γ 100 U/ml, and LPS 10 μg/ml) for 24 h (n = 4 plates of cells/group). After 24 h, myotubes were washed three times with PBS (phosphate-buffered saline) and placed in PBS containing 0.1 ml of the Trypan blue stock solution (0.4% Trypan blue in PBS) for 3 min. Cells were then washed three times with PBS and imaged using a Retiga camera; five photographs were taken of each plate. Viable and nonviable cells were then counted for each image for each plate using Image J software. Cells with a normal appearance not staining with Trypan blue were considered viable, and cells that stained for Trypan blue were considered nonviable.

cPLA₂ activity.

For this assay, substrate was first prepared by mixing arachidonoyl-thio-PC with assay buffer consisting of 160 mM HEPES, 300 mM NaCl, 20 mM CaCl₂, 8 mM Triton X-100, 60% glycerol, 2 mg/ml BSA, pH 7.4 (cPLA₂ Assay Kit, Cayman Chemical, Ann Arbor, MI) and ultrapure water; this mixture was then cooled on ice. A mixture of 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) in EGTA was prepared in ultrapure water; this mixture was used to stop reactions and was kept on ice in the dark until usage. Cell samples were prepared by homogenization in buffer and subsequent protein determination of homogenates. A 96-well plate was then loaded with samples (10 μl of sample plus 5 μl of a mixture of thioetheamidate-PC and bromoenol lactone to inhibit other PLA₂ isoforms), blanks (10 μl of assay buffer and 5 μl of inhibitors), and a positive control (bee venom cPLA₂, 10 μl, plus 5 μl of inhibitors). Substrate (200 μl) was then added to each well, and plates were incubated in the dark at room temperature for 60 min. At the end of this time, 10 μl of the DTNB mixture was added, plates were mixed, and, after an additional 5 min, optical density was read at 414 nm. Activity was calculated as the optical density change over 60 min corrected for sample dilution, corrected for protein content in the samples, and corrected for the optical density of the colorimetric product.

MitoSox assay.

For this assay, a 5 mM stock solution of MitoSox (Invitrogen, San Diego, CA) was prepared in DMSO. MitoSox was then diluted with cell differentiation media to generate a 5 μM working solution. C2C12 cells grown on coverslips and incubated with various media solutions were then washed with media and overlain with the MitoSox working solution for 20 min in darkness. Coverslips were then washed three times with buffer, mounted on a laser Doppler microscope (University of Kentucky Imaging Core), and imaged using an excitation frequency of 396 nm and an emission frequency of 560 nm. Images were captured, stored on a flash drive, and analyzed using SigmaScan imaging software.

Aconitase activity.

A commercially available Aconitase Assay Kit (Abcam, Cambridge, MA) was used for these assays. In brief, samples (harvested cells, diaphragm muscle) were prepared in kit sample preparation buffer, and samples were then assayed for protein content. Samples containing 100 μg of protein were then mixed with kit assay buffer to a final volume of 50 μl. Aconitase substrate was prepared by mixing 3 ml of kit buffer with 200 μl of an isocitrate standard and 50 μl of a manganese standard. Samples and blanks were then placed in a 96-well plate, aconitase substrate was added (200 μl) to each well, and optical density at 240 nm was measured (incubation temperature 37°C) over 30 min. The rate of rise of the 240-nm signal was taken as an index of aconitase activity.

Western blotting.

Western blot techniques were employed to measure protein levels in C2C12 cells. For these determinations, cells were first homogenized in buffer (10 mM β-glycerophosphate, 50 mM sodium fluoride, 1 mM sodium orthovanadate, 20 mM HEPES, 2 mM EDTA, 250 mM sodium chloride, 2 μg/ml leupeptin, 2 μg/ml aprotinin, 1 mM PMSF, 0.5 μg/ml benzamidine, 1 mM DTT, pH 7.4), and homogenate protein levels were determined using the Bradford assay (BioRad, Hercules, CA). Equal amounts of proteins from samples were then diluted with an equal volume of loading buffer (126 mM Tris·HCl, 20% glycerol, 4% SDS, 1.0% 2-mercaptoethanol, 0.005% bromophenol blue, pH 6.8) and loaded onto Tris glycine polyacrylamide gels. Protein mixtures were separated by electrophoresis (Novex Minicell II, Carlsbad, CA). Proteins were then transferred to polyvinylidene fluoride membranes and incubated over night at 4°C with primary antibodies to targeted proteins. Membranes were then incubated with horseradish peroxidase-conjugated secondary antibodies, and antibody binding was detected on film using enhanced chemiluminescence (Western Lightning, Perkin Elmer, Boston, MA). Densitometry of filmed gels was performed using a Microtek scanner (Carson, CA) and UN-SCAN-IT software (Silk Scientific, Orem, UT). Membranes were reprobed with antibodies to GAPDH (Santa Cruz Biotechnology) to verify equal lane loading. We chose GAPDH because this protein is not altered in skeletal muscle by sepsis. We examined the following proteins using the designated primary antibodies: anti-caspase-12 (cs2202, Cell Signaling Technology, Beverly, MA), anti-GAPDH (ab9484, Abcam, Cambridge, MA), and anti-calpastatin (sc-20779; Santa Cruz Biotechnology, Santa Cruz, CA).

Calpain activity assay.

Calpain activity was assessed using a commercially available kit (Abcam, Cambridge, MA), as previously described (27, 37). This kit includes a proprietary set of buffers to allow determination of calpain activity in tissues before homogenization. In brief, samples were prepared in the proprietary extraction buffer according to the manufacturer's specifications. Sample protein levels were measured, and 100 μg of protein from each sample were diluted to a final volume of 85 μl with extraction buffer. An additional aliquot of each sample (100 μg of protein) was mixed with 1 μl of calpain inhibitor and sample volume brought to 85 μl with extraction buffer. Extraction buffer alone was used as a blank. All mixtures were placed in a 96-well plate, and then reaction buffer (10 μl) and calpain substrate (5 μl) were added to wells. An initial reading was taken, plates were incubated in the dark at 37°C for 1 h, and a final fluorescent reading was made (excitation 400 nm, emission 505 nm). The difference in increases in fluorescent activity over time for a given sample in the presence and absence of the specific calpain inhibitor was taken as an index of calpain enzyme activity.

Calpastatin activity.

Calpastatin activity was measured by determining the inhibitory effect of muscle extracts on calpain-mediated degradation of BODIPY-FL-casein, as described by Thompson et al. (39) with some modification. Muscle extracts were first heated at 100°C for 5 min to destroy calpain activity. Purified μ-calpain (800 ng) from porcine erythrocytes (Calbiochem-Novabiochem, San Diego, CA) and aliquots (50 μg protein) of muscle extract were added to microtiter plate wells, and the volume was adjusted to 100 μl by adding buffer. The reaction was then started by adding 100 μl BODIPY-FL-casein (16 μg casein/ml), and calpastatin activity (expressed as %inhibition of calpain activity) was calculated from the difference in calpain activity measured in the absence or presence of muscle extract.

Experimental protocols: animal studies.

To provide an assessment of the effect of inhibition of cPLA₂ activity on diaphragm force generation in an animal model of infection, we studied mice undergoing infection-induced by CLP (details of this procedure are described below). All animal studies were approved by the University of Kentucky Institutional Animal Care and Use Committee, and all procedures were carried out in keeping with the University of Kentucky animal use regulations.

We studied four groups of animals (n = 4/group), including the following: 1) sham-operated, vehicle-treated controls; 2) CLP-operated animals treated with vehicle; 3) sham-operated animals given CDIBA (4-{2-[5-chloro-1-(diphenylmethyl)-2-methyl-1H-indol-3-yl]-ethoxy}benzoic acid), administered as a single intraperitoneal injection immediately after surgery in a dose of 5 mg/kg in a volume of 100 μl DMSO + saline, a cPLA₂ inhibitor; and 4) CLP operated animals given CDIBA (16). At 24 h after surgery, animals were euthanized, and diaphragms removed. Muscle strips were dissected from a portion of the diaphragm and used for determination of the diaphragm force-frequency relationship in vitro. The remaining diaphragm was frozen, stored in a −80°C freezer, and subsequently used to determine diaphragm protein level calpain activity and cPLA₂ activity (these procedures are described in paragraphs above this section). Details of the CLP procedure and measurement of diaphragm force are described below.

CLP-induced sepsis.

Sham surgery or CLP was performed as previously published (25, 35). Briefly, mice were anesthetized with isoflurane (2–4%). Once a steady plane of anesthesia was achieved, the abdomen was sterilely prepped, an incision made (∼1.5 cm in length), the cecum was identified, and a portion ligated (∼0.5 cm). A 21-gauge sterile needle was used to puncture the ligated cecum through and through. Thereafter, the abdominal musculature was approximated and sutured, and the skin was closed with surgical staples. For sham surgery, the abdomen was opened and closed without cecal ligation or puncture. Mice were resuscitated with 60 ml/kg of saline administered subcutaneously following the surgical procedure and received buprenorphine (0.05 mg/kg) subcutaneously immediately and then every 12 h following surgery. Animals were euthanized at 24 h after surgery using an intraperitoneal injection of pentobarbital (150 mg/kg). Diaphragms were then removed and used for determination of force generation, to assess muscle protein concentration, to determine diaphragm calpain activity, and to assess diaphragm cPLA₂ activity.

Diaphragm force-frequency curves and muscle mass.

Diaphragm force-frequency relationships were assessed as previously reported (26–28, 32, 35, 37). Muscle strips were dissected from the left midcostal portion of the diaphragm and mounted in water-jacketed glass organ baths filled with Krebs-Henseleit solution (25°C, curare 50 mg/l, pH 7.40, NaCl 135 mM, KCl 5 mM, dextrose 11.1 mM, CaCl₂ 2.5 mM, MgSO₄ 1 mM, NaHCO₃ 14.9 mM, NaHPO₄ 1 mM, insulin 50 U/l, 95% O₂/5% CO₂). One end of the strip was tied to the base of the organ bath, and the other to a force transducer (Scientific Instruments, Heidelberg, Germany), followed by delivery of supramaximal currents from a biphasic constant-current amplifier connected to a Grass S48 stimulator (Grass, West Warwick, RI) using platinum mesh field electrodes. After a 15-min equilibration period, muscle length was adjusted to optimum length, and strips were sequentially stimulated with trains of 1-, 10-, 20-, 50-, 100-, and 150-Hz stimuli (train duration 800 ms, 30 s between adjacent trains) with force recorded with a Kipp-Zonen chart recorder (Bohemia, NY). Muscle cross-sectional area was then calculated as muscle strip weight divided by muscle density (1.06), and muscle length and muscle-specific force were calculated as raw force divided by cross-sectional area. Total weight of the costal diaphragm was calculated by adding the weight of the strips used for the force-frequency curve to the weight of the remaining muscle. Diaphragm muscle tissue that was not used for force-frequency determinations was frozen, stored at −80°C, and subsequently used to assess protein concentration using the Bradford assay kit (BioRad Laboratories, Hercules, CA), as well as the levels of calpain and cPLA₂ activities.

Statistical analysis.

ANOVA was used to compare variables (e.g., force) across groups of cells and animals treated with different agents, with post hoc testing (Tukey's) to determine differences between groups. A P value of <0.05 was taken as indicating statistical significance for experiments.

RESULTS

Effect of cytokine administration on C2C12 myotube cPLA₂ activity and mitochondrial superoxide generation.

We found that addition of cytomix to the C2C12 myocytes evoked a threefold increase in cell cPLA₂ activity, as shown in Fig. 1 (P < 0.001 for comparison of cPLA₂ activity for saline-treated control cells compared with cells treated with cytokines). Since our recent work indicates that cytokine effects on skeletal muscle may be mediated, in large part, via activation of muscle neutral sphingomyelinase (27), we also determined if administration of a neutral sphingomyelinase inhibitor, GW4869, would alter the effect of cytomix to augment C2C12 cell cPLA₂ activity. We found that GW4869 administration completely blocked the effect of cytomix to increase C2C12 cell cPLA₂ activity (P < 0.001 for comparison of cPLA₂ activity in cells treated with cytomix vs. cells treated with both cytomix and GW4869; Fig. 1).

Fig. 1.

Mean cPLA₂ activity for saline-treated control C2C12 cells (open bars), cytomix-treated cells (solid bars), GW4869 plus cytomix-treated cells (shaded bars), and cells treated with GW4869 alone (cross-hatched bars). Cytomix elicited a significant increase in cPLA₂ activity level (*P < 0.001; asterisks indicate statistical significance in this and all other figures). Error bars indicate 1 SE. GW4869 blocked the cytokine-induced increase in cPLA₂ activity.

To rule out the possibility that our findings were not artifactually influenced by an effect of cytomix to kill C2C12 cells, we also determined the viability of cells exposed to control media and media containing cytomix using the Trypan blue assay. We found that 98.7 ± 0.3% of harvested cells that were exposed to control media were viable by the Trypan blue assay after 24 h of incubation, and an equal fraction of harvested cells exposed to cytomix for 24 h (i.e., 98.6 ± 0.2%) were viable.

We next examined the effect of cytomix administration on C2C12 cell superoxide generation. We employed two different techniques to assess C2C12 cell free radical generation, including the MitoSox assay and aconitase activity. MitoSox is a fluorescent compound that has an increase in fluorescence in response to reaction with superoxide and several other reactive oxygen species. As shown in Fig. 2, cells treated with both MitoSox and cytomix manifested a large increase in fluorescence compared with control cells treated with saline and MitoSox (P < 0.001 for comparison of emission light intensity for the two groups). We also found that concomitant administration of SS31, a mitochondrially targeted superoxide scavenger (6), blocked the cytomix-induced increase in MitoSox fluorescence of C2C12 myotubes, arguing that the cytomix-induced increase in MitoSox fluorescence represented a response to mitochondrial superoxide generation and not a nonspecific signal related to other reactive oxygen species. More importantly, we also found that administration of AACOCF3, a cPLA₂ inhibitor (18), blocked cytomix-induced increases in MitoSox activity, indicating that cytomix-mediated increases in C2C12 cell superoxide generation are downstream of cPLA₂ activation (P < 0.001 for comparison of MitoSox activity for cells treated with cytomix compared with cells treated with both cytomix and AACOCF3).

Fig. 2.

Left: representative images of C1C12 cells incubated with the MitoSox fluorescent probe (pseudocolored green); labels above each individual image indicate the conditions under which cells were incubated, including saline-treated control cells (top left), cytomix-treated cells (top right), AACOCF3-treated cells (middle left), cells treated with AACOCF3 + cytomix (middle right), cells treated with SS31 (bottom left), and cells treated with SS31 + cytomix (bottom right). Right: mean MitoSox fluorescence intensity for the six experimental groups. Cytomix elicited a significant increase in MitoSox fluorescence (*P < 0.001), and this increase was blocked in cytomix-treated cells by administration of either AACOCF3 or SS31.

We also employed the aconitase assay to assess C2C12 mitochondrial superoxide generation (12). Aconitase is a Kreb's cycle enzyme that is only found in mitochondria in muscle cells and is specifically inactivated by reaction with superoxide ions (12). As shown in Fig. 3, C2C12 cell aconitase activity significantly decreased following addition of cytomix to cell media (P < 0.001 for comparison of aconitase activity in cells treated with saline and cells treated with cytomix), consistent with an effect of cytomix to elicit an increase in generation of superoxide ions in C2C12 mitochondria. Administration of AACOCF3 to the media blocked the effect of cytomix to decrease C2C12 cell aconitase activity, consistent with our hypothesis that cytomix increases CC12 cell mitochondrial superoxide generation by a cPLA₂-dependent mechanism.

Fig. 3.

Mean aconitase activity for saline-treated control C2C12 cells (open bars), cytomix-treated cells (solid bars), AACOCF3 plus cytomix-treated cells (shaded bars), and cells treated with AACOCF3 alone (cross-hatched bars). Cytomix elicited a significant decrease in aconitase activity (*P < 0.001). AACOCF3 blocked the cytokine-induced reduction in aconitase activity.

Cytomix-induced C2C12 cell calpain activation and its relationship to cPLA₂ activation and mitochondrial superoxide generation.

We employed two indexes to assess C2C12 cell calpain activation. Previous work has shown that caspase-12 cleavage by calpain results in production of a smaller, 38-kDa caspase-12 fragment (1, 19, 20). We, therefore, assessed this caspase-12 molecular fragment by Western blotting and used the level of this protein as an index of cell calpain activation. We also directly measured C2C12 cell calpain activity using a proprietary assay that utilizes a solution that prevents calpain activation during cell lysis, thereby providing a measure of cell calpain activity that is not artifactually altered by the isolation technique. We found that calpain-specific caspase-12 degradation product (CSCDP) levels in C2C12 cells increased significantly following cytomix administration (Fig. 4, P < 0.001 for comparison of CSCDP levels for control and cytomix-treated cells), consistent with an effect of cytokines to increase C2C12 cell calpain activity. We also found that concomitant administration of either AACOCF3, a cPLA₂ inhibitor, or SS31, a mitochondrially targeted superoxide scavenger, blocked cytokine-induced increases in C2C12 CSCDP levels. These findings are consistent with the possibility that cytomix-mediated calpain activation is downstream of cytomix-induced increases in cPLA₂ activity and mitochondrial superoxide generation. In addition, we also found that cytomix administration elicited an increase in directly measured C2C12 cell calpain activity (Fig. 4, P < 0.001 for comparison of calpain activity for control and cytomix-treated cells), and that concomitant administration of either AACOCF3 or SS31 also blocked cytomix-induced increases in this later index of C2C12 cell calpain activation.

Fig. 4.

Top: examination of the effect of AACOCF3 on cytomix-induced, calpain-mediated caspase-12 cleavage and include a representative caspase-12 Western blot (top left) and mean densities for the calpain-dependent caspase-12 cleavage band (designated CSCDP in the representative Western blot). The representative blot includes lanes (from left to right) for cells incubated with no cytomix or AACOCF3, cells incubated with cytomix, cells incubated with cytomix plus AACOCF3, and cells incubated with AACOCF3 alone. A Western blot for GAPDH is included as a loading control. Cells incubated with cytomix manifested a large increase in the intensity of the CSCDP band (*P < 0.001 for comparison of control to cytomix-treated cell groups). Administration of AACOCF3 attenuated the cytomix effect on CSCDP (*P < 0.001 for comparison of CSCDP levels between cytomix and cytomix + AACOCF3 groups). Middle: examination of the effect of SS31 on cytomix-induced, calpain-dependent caspase-12 cleavage, with a representative Western blot on the left and mean group data on the right. Cytomix treatment increased CSCDP band formation, and SS31 blocked this effect (*P < 0.001 for comparison of CSCDP levels between cytomix and cytomix + SS31 groups). Bottom: mean cell calpain activity levels, with cytomix-treated cells (red bar) manifesting higher calpain activity compared with saline-treated control cells (black bar; *P < 0.001). Administration of either AACOCF3 or SS31 blocked the cytomix-induced increase in cell calpain activity (*P < 0.001 for comparison of AACOCF3 + cytomix and SS31 + cytomix groups to the cytomix-treated group).

There are several potential cellular mechanisms by which cellular calpain activity can increase. One such mechanism is depletion or inactivation of cell calpastatin, the normal endogenous inhibitor of calpain (13). For this reason, we assessed the effects of cytomix administration on C2C12 cell calpastatin activity and calpastatin protein content. As shown in Fig. 5, cytomix administration elicited a dramatic reduction in C2C12 calpastatin activity (P < 0.01 for comparison of calpastatin activity for control and cytomix-treated C2C12 cells), and concomitant administration of either AACOCF3 or SS31 prevented the cytomix-induced reduction in calpastatin function (P < 0.01 for comparison of calpastatin activity between cytomix and cytomix plus AACOCF3-treated cells, and P < 0.01 for comparison of calpastatin activity between cytomix and cytomix plus SS31-treated cells). Cytomix did not, however, reduce calpastatin protein levels (Fig. 5, bottom). There was a trend for AACOCF3 to increase calpastatin protein levels in both saline and cytomix-treated cells, but this response did not achieve statistical significance.

Fig. 5.

Top: mean calpastatin activity (%inhibition of BODIPY-FL-casein cleavage by calpain) for saline-treated control C2C12 cells (black bar), cytomix-treated cells (red bar), AACOCF3 + cytomix-treated cells (green bar), cells treated with AACOCF3 alone (yellow bar), SS31 + cytomix-treated cells (blue bar), and cells treated with SS31 alone (pink bar). Cytomix elicited a large reduction in calpastatin activity (*P < 0.01). Both AACOCF3 and SS31 blocked the cytokine-induced reduction in calpastatin activity. Bottom: data for calpastatin protein levels, with the left bottom panel displaying a representative Western blot for calpastatin protein and the right bottom panel displaying mean densitometry of calpastatin protein levels for the four experimental groups. There was no significant effect of cytomix on calpastatin protein level.

Effect of administration of a cPLA₂ inhibitor on diaphragm function and calpain activity in the CLP model of sepsis.

CLP-induced peritonitis elicited a large reduction in the force-generating capacity of the diaphragm, producing dramatic reductions in muscle-specific force generation of the diaphragm (i.e., force per unit cross-sectional area) in response to stimulation frequencies ranging from 1 to 150 Hz (see Fig. 6, P < 0.001 for comparison of diaphragms from control and CLP groups at all stimulation frequencies). Administration of CDIBA, a cPLA₂ inhibitor, to CLP animals markedly attenuated the effects of CLP on diaphragm strength, with force generation at all stimulation frequencies 1–150 Hz significantly higher for the CDIBA + CLP group compared with CLP animals (P < 0.001 for all stimulation frequencies). CLP did not, however, induce reductions in either diaphragm mass or total protein content (Fig. 6), findings consistent with previous reports from our group examining diaphragm function in animal models of infection (36). CLP did, however, elicit large increases in diaphragm calpain activity (P < 0.001), as shown in Fig. 6. Administration of CDIBA to CLP animals blocked the CLP-mediated increase in diaphragm calpain activity, paralleling the effect of CDIBA to preserve diaphragm-specific force generation following CLP.

Fig. 6.

Top: mean force-frequency relationships for sham-operated control animals (■), cecal ligation perforation (CLP) animals (●), CLP animals treated with CDIBA (▼), and sham-operated animals treated with CDIBA (▲). CLP induced large reductions in diaphragm force generation, reducing the force generated at all stimulation frequencies (*P < 0.001 for all comparisons). CDIBA significantly attenuated the effects of CLP on diaphragm force generation (*P < 0.001 for comparison of force between CLP and CDIBA + CLP at all stimulation frequencies). Middle: diaphragm mass (normalized to animal weight), diaphragm total protein content (normalized to animal weight), and diaphragm protein/unit diaphragm mass for the four experimental animal groups. Neither CLP nor CDIBA altered theses indexes of diaphragm mass or protein content. Bottom: calpain activity for diaphragm protein samples from the four animal groups (open bar represents the control group data, solid bar presents CLP group data, shaded bar indicates CDIBA group data, and the cross-hatched bar indicates CDIBA + CLP group data). Diaphragm calpain activity was significantly increased in the CLP group, and administration of CDIBA blocked this CLP-induced effect (*P < 0.001 for comparison of control to CLP groups, and *P < 0.001 for comparison of CLP to CDIBA + CLP groups).

We also found that cPLA₂ activity levels for diaphragm samples were higher for the CLP animals than for the control group, and CDIBA administration to CLP animals reduced diaphragm cPLA₂ activity to control levels. Specifically, diaphragm cPLA₂ activity averaged 362 ± 46, 817 ± 56, 345 ± 3 3, and 298 ± 40 pmol·min⁻¹·mg⁻¹, respectively, for control, CLP, CDIBA + CLP, and CDIBA groups (P < 0.001 for comparison of CLP to the other three groups).

DISCUSSION

Recent studies indicate that mechanically ventilated critically ill patients develop severe respiratory and limb muscle weakness, with skeletal muscle strength for this group of patients averaging only a fraction of that found in normal, healthy adults (8, 14, 15, 31, 40). Specifically, three recent studies (8, 14, 31) found that the transdiaphragmatic pressure generated in response to maximal bilateral phrenic nerve stimulation averages ∼9 cmH₂O in mechanically ventilated intensive care unit patients, a value 20–25% of that reported for normal subjects using this same technique. This profound skeletal muscle weakness is thought to have important consequences, with weaker patients requiring a much longer time to wean from mechanical ventilation and having a far greater mortality than stronger patients (8, 31). Infection appears to be an especially important risk factor for the development of diaphragm weakness in intensive care unit patients, with two recent reports indicating that the diaphragm strength of infected mechanically ventilated patients is one-half that of noninfected mechanically ventilated patients (8, 31). Studies using animal models of infection have reported similar findings, with severe diaphragm weakness reported in response to endotoxin administration, CLP-induced peritonitis, pneumonia, and other animal models of infection (9, 10, 27, 29, 33, 34, 37).

While the precise mechanism(s) by which infections induce skeletal muscle dysfunction is incompletely understood, previous studies do provide important clues regarding the etiology of this process. Some work suggests cytokines play an important role in modulating the induction of diaphragm weakness in animal models of infection (30). Other reports indicate that infections induce activation of several proteolytic enzyme systems in skeletal muscle, including the calpain, caspase, and proteasome protein degradation pathways (4, 24, 29, 34). In support of the role of calpain in the development of infection-induced skeletal muscle weakness, multiple markers of calpain activation (increased talin and spectrin cleavage, autolytic cleavage of calpain-1 and calpain-2, increased muscle levels of calpain activity) increase in skeletal muscles in animal models of infection (27, 29, 37). In addition, administration of chemical inhibitors of active calpain has been shown to preserve skeletal muscle strength when given to infected animals (29). Furthermore, a recent study also found that muscle-specific overexpression of calpastatin, the endogenous calpain inhibitor, markedly attenuates the diaphragm weakness elicited in response to CLP-induced peritonitis (37).

To date, however, research has not identified the specific upstream mechanisms by which infections activate calpain in skeletal muscle. The purpose of the present study, therefore, was to investigate this issue. We also thought that a better understanding of the cellular processes which mediate infection-induced calpain activation could ultimately lead to the development of translational therapies to prevent or reverse skeletal muscle weakness in infected patients. The results of the present work support the hypothesis that cPLA₂ modulates infection-induced diaphragm dysfunction. Specifically, we found that cytokines activate cPLA₂ and calpain in the C2C12 muscle cell line, and that cPLA₂ inhibition blocks calpain activation in cytokine-treated cells. In addition, we also found that administration of CDIBA, a cPLA₂ inhibitor, preserves diaphragm force-generating capacity and blocks diaphragm calpain activation in the CLP animal model of sepsis.

Our data also suggest that cPLA₂ produces its downstream effects on muscle calpain activation and contractile function by increasing mitochondrial free radical generation. In support of this argument, we found that cytokine administration to the C2C12 muscle cell line increased two indexes of muscle mitochondrial superoxide generation, and that inhibition of cPLA₂ blocked increases in both of these indexes. The first technique we employed to assess mitochondrial free radical formation involved measurement of fluorescence following incubation of C2C12 cell with MitoSox, a mitochondrially targeted derivative of hydroethidine. This molecule localizes to mitochondria and reacts with superoxide to form a product with a characteristic fluorescent signature (17). We found that addition of cytokines to C2C12 cells resulted in an increase in the MitoSox fluorescent signal. Addition of a selective mitochondrial superoxide scavenger (6, 21), SS31, blocked this cytokine-mediated increase in fluorescence, indicating that augmented superoxide generation was responsible for the cytokine-mediated increase in the MitoSox signal. A second index we employed to assess mitochondrial free radical generation was measurement of cell aconitase activity. Muscle aconitase is only present in mitochondria, and this enzyme is specifically inactivated by mitochondrially generated superoxide (12). We found that C2C12 cell aconitase activity decreased following cytokine exposure, and that this reduction was prevented by addition of SS31, a specific mitochondrially targeted superoxide scavenger. As a result, our observations that AACOCF3, a cPLA₂ inhibitor, prevented cytokine-induced increases in muscle cell MitoSox activity and cytokine-induced reductions in aconitase activity suggest that cytokine-induced muscle mitochondrial superoxide generation is secondary to cPLA₂ activation. This finding is consistent with previous work showing that phospholipid breakdown products generated by cPLA₂ can interact with mitochondria to induce superoxide formation (2).

Our data also suggest that muscle mitochondrial superoxide production may inactivate calpastatin, the normal endogenous inhibitor of cellular calpain activity (13). In support of this possibility, we found that calpastatin activity was markedly reduced in C2C12 muscle cells following exposure of cells to cytokines. This is shown in Fig. 5, top, which demonstrates diaphragm calpastatin activity is dramatically reduced following exposure of muscle cells to cytokines, and that administration of either a mitochondrially targeted superoxide scavenger (SS31) or a cPLA₂ inhibitor (AACOCF3) prevents this cytokine-induced effect. The fact that calpastatin protein levels were not diminished by cytokine exposure, but calpastatin activity was almost completely abolished, suggested that this protein was inactivated but not destroyed following cytokine exposure. As a result, our findings are consistent with the possibility that cytokines elicit increases in skeletal muscle cPLA₂ activity, that cPLA₂ activation induces increased mitochondrial production of reactive oxygen species, that reactive oxygen species inactivate calpastatin, and that reduction in calpastatin functional capacity is linked to calpain activation.

Many of the findings in the present study are consistent with a previous report examining the role of cPLA₂ as a mediator of muscle dysfunction following denervation. In this previous work, by Bhattacharya et al. (2), muscle denervation was found to increase muscle cPLA₂ activity. Heightened cPLA₂ activity was found to induce muscle atrophy by a calcium-dependent mechanism, an observation consistent with the possibility that cPLA₂ activation stimulated calpain activity. In another study, investigators found that muscle force loss in the hindlimb suspension animal model of inactivity was due, in part, to activation of muscle calpain activity (25). This latter work found that muscle weakness in response to muscle inactivity was markedly reduced in animals with skeletal overexpression of calpastatin, the endogenous calpain inhibitor. Taken together, these studies of inactivity-induced muscle dysfunction argue that heightened cPLA₂ activity and activation of muscle cell calpain may be responsible for loss of muscle strength and size. As a result, both inactivity (the stimulus examined in these previous reports) and cytokine mediated-induced muscle dysfunction (the present report) may induce muscle weakness by activating similar cellular pathways (i.e., cPLA₂ and calpain activation).

The focus of the present project was to determine whether there was a potential role of cPLA₂ in modulating calpain activation and force loss in the diaphragm following the induction of CLP. The purpose of this study was not to determine which specific calpain isoforms were contributors to the sepsis-induced increases in calpain activity levels. Our laboratory's previous work, however, suggests that two calpain isoforms, calpain-1 and calpain-2, are activated in the diaphragm in response to sepsis and undergo autocatalytic cleavage (29). Our finding, in the present study, that calpastatin becomes inactivated following CLP, would be consistent with our previous results. Calpastatin is the normal endogenous inhibitor of both calpain-1 and calpain-2 isoforms. As a result, sepsis-induced calpastatin inactivation would be expected to augment the effective activity of both calpain-1 and calpain-2.

There are, however, some technical limitations to the methods employed in the present report. We used chemical inhibitors to block cPLA₂ activity and, while the agents employed are arguably the most specific cPLA₂ inhibitors currently available, all chemical inhibitors have the potential for nonspecific off-target effects. In past studies, we utilized dominant-negative genetic constructs to block activation of various genes in both cell and animal studies (32, 35), but no cPLA₂ dominant-negative constructs are currently available. We have also successfully used small interfering RNA to block gene activity in C2C12 cells in the past, but we have been unable to successfully transfect C2C12 cells with small interfering RNA to cPLA₂, despite several attempts. On the other hand, chemical inhibitors of cPLA₂ have been used in clinical trials (7). Thus our finding that a chemical inhibitor of cPLA₂ can block sepsis-induced diaphragm dysfunction raises the possibility that a similar agent could prove to be useful as a therapy to prevent sepsis-induced muscle dysfunction.

One additional potential criticism of this work is that we examined the response of C2C12 cells to the combined effects of several cytokines rather than dissecting the responses to individual cytokines. We did so, however, because the septic response to infections invariably results in simultaneous increases in multiple cytokines, as previously reported for the CLP model (11, 22, 23). As a result, utilization of a cytokine mixture is a more realistic model of the milieu muscles will face in vivo.

In conclusion, these data indicate that skeletal muscle cPLA₂ is activated by cytokines, is linked to mitochondrial superoxide generation, and that cPLA₂-induced superoxide generation contributes to calpain activation in skeletal muscle cells. In addition, we found that administration of a cPLA₂ inhibitor, CDIBA, to intact animals markedly attenuates infection-induced diaphragm dysfunction. This latter finding is important because there are currently no treatments to prevent or reverse infection-induced skeletal muscle dysfunction. Several potential treatments have been proposed, including testosterone analogs and inhibitors of the proteasome proteolytic system, but recent studies indicate that both testosterone analogs and proteasomal inhibitors have deleterious effects that prevent their use as treatments in critically ill patients (3, 36). Based on the present findings, we speculate that therapies that inhibit cPLA₂ and/or scavenge mitochondrial superoxide radicals can prevent infection-induced calpain activation in the diaphragm, attenuating the development of diaphragm weakness in infected, critically ill patients.

GRANTS

This work was supported by Veterans Administration Merit Award 1I01BX002132 (G. S. Supinski) and National Heart, Lung, and Blood Institute Grants R01HL113494 (G. S. Supinski), R01HL080429 (G. S. Supinski), R01HL081525 (G. S. Supinski), and R01HL112085 (L. A. Callahan).

DISCLOSURES

No conflicts of interest, financial or otherwise are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: G.S.S. and L.A.C. conception and design of research; G.S.S., A.P.A., L.W., and X.-H.S. performed experiments; G.S.S. and L.A.C. analyzed data; G.S.S. and L.A.C. interpreted results of experiments; G.S.S. and L.A.C. prepared figures; G.S.S. drafted manuscript; G.S.S. and L.A.C. edited and revised manuscript; G.S.S., A.P.A., L.W., X.-H.S., and L.A.C. approved final version of manuscript.