Pharmacological inhibition of GSK-3 in a guinea pig model of LPS-induced pulmonary inflammation: II. Effects on skeletal muscle atrophy

Outbred, male, specified pathogen-free Dunkin Hartley guinea pigs (Harlan, Heathfield, UK) were used in this study. All protocols described in this manuscript were approved by the University of Groningen Committee for Animal Experimentation.

Thirty-six guinea pigs, 12 ± 4 wks of age were randomly assigned to four experimental groups (n = 9), namely: (1) vehicle-treated, saline-challenged; (2) SB216763-treated saline-challenged; (3) vehicle-treated, LPS-challenged, and (4) SB216763-treated, LPS-challenged. The guinea pigs were treated twice per week for 12 consecutive weeks by intranasal instillation of 100 μl SB216763 (sterile, 2 mM in 10% (v/v) DMSO in saline) or vehicle (100 μl 10% (v/v) DMSO in sterile saline). After the intranasally instilled solution was aspirated, the animals were kept in an upright position for an additional 2 min, to allow sufficient spreading of the fluid throughout the lungs. The animals were intranasally instilled with 100 μl LPS (10 mg/ml in sterile saline) or sterile saline, 30 min post SB216763 or vehicle instillation. SB216763 is a selective GSK-3 inhibitor (3-(2,4-dichlorophenyl)-4-(1-methyl-1H-indol-3-yl)-1H-pyrrole-2,5-dione) (Tocris Cookson, Bristol, UK) and the LPS was derived from Escherichia coli, serotype 055:B5 (Sigma-Aldrich, MO, USA). Twenty-four hours after the last instillation, the guinea pigs were sacrificed by experimental concussion, followed by rapid exsanguination. Next, the lungs and a series of hind limb muscles including the M. gastrocnemius, M. tibialis anterior, M. plantaris and M. extensor digitorum longus (EDL) were collected using standardized dissection methods. Independent muscle weights of a single hind limb were measured and all tissues were immediately flash-frozen in liquid nitrogen.

The EDL muscles were embedded in Tissue-Tek (Sakura Finetek, the Netherlands) and sectioned on a Leica CM3050 S cryostat at −20°C. Subsequently, serial cross-sections (5 μm) were stained with the following primary antibodies: anti-Type I MyHC (#A4840) (Developmental Studies Hybridoma Bank, Iowa City, IA, USA), and anti-laminin (#L-9393) (Sigma-Aldrich) to determine the fiber cross-sectional area (CSA) and fiber type distribution. The sections were incubated with the following secondary antibodies: goat anti-mouse IgM Alexa Fluor 555 (#A-21426) and goat anti-rabbit IgG Alexa Fluor 350 (#A-11069) (both from Invitrogen, CA, USA). Digital images of the stained sections were taken under 200X total magnification using an Eclipse E800 microscope (Nikon, Japan) connected to a digital camera (DXM, 1200 NF, Nikon, Japan). The CSA was measured after having identified five non-overlapping regions containing a total of 100–200 individual fibers per animal, which were then analyzed using Lucia Software (version 4.81).

The murine skeletal muscle cell line C₂C₁₂ (ATCC # CRL1772) was cultured in growth medium (GM), composed of low glucose (1 g/l) Dulbecco’s Modified Eagle Medium (DMEM) containing antibiotics (50 U/ml Penicillin and 50 μg/ml Streptomycin) and 9% (v/v) Fetal Bovine Serum (FBS) (all from Gibco, MD, USA). The C₂C₁₂ cells were plated overnight in GM at 10⁴/cm² on BD Matrigel coated (1:50 in DMEM low glucose) 35 mm dishes as described previously (both from BD Biosciences, MA, USA) [43]. To study effects on myogenesis, differentiation was induced by growth factor withdrawal [44], replacing GM with differentiation medium (DM) (DMEM, low glucose, with 1% heat-inactivated FBS and antibiotics). The synthetic GC dexamethasone (hereinafter referred to as Dex) (Sigma-Aldrich), TNF-α (Calbiochem, CA, USA), with or without LiCl (Sigma-Aldrich) or CHIR99021 (2 μM) (kindly provided by Dr. Cohen, MRC Protein Phosphorylation Unit, University of Dundee, UK) were directly added to the culture medium upon the induction of differentiation and again 24 h later when the cells were provided with fresh DM. The myocytes were allowed to differentiate for a total of 72 h, in absence or presence of Dex (10 μM) or TNF-α (1 ng/ml) prior to analysis of myogenesis markers.

As a morphological parameter of myogenesis, the myogenic index was determined to quantitate myoblast fusion. The C₂C₁₂ cells were induced to differentiate for 72 h either in the presence or absence of Dex or TNF-α. After 72 h of differentiation the cells were washed twice in 1× PBS (RT), subsequently fixed in methanol and stained in May-Grünwald Giemsa (Sigma-Aldrich) according to the manufacturer’s instructions. Pictures were taken at 40× and 100× magnifications using an inverted light microscope (Eclipse E800, Nikon, Japan) connected to a digital camera (DXM, 1200 NF, Nikon, Japan). The 100× magnified images were taken in series of four with a fixed overlap. The total number of nuclei in 4 or more fields was counted, and nuclei were assigned to one of three classes: (1) single nucleated myoblasts; (2) dividing or fusing bi-nucleated myoblasts, and (3) multi-nucleated (> 2 nuclei) myotubes. Per condition, 1900 or more nuclei were counted and assigned to either of the above-mentioned classes.

Measurements of Troponin I (TnI) promoter activity during differentiation were performed by creating a stable C₂C₁₂ cell line carrying a genomic TnI promoter-luciferase reporter gene as described previously [45]. To determine the luciferase activity, the cells were washed twice in ice-cold 1× PBS, lysed in 1× reporter lysis buffer (Promega, Madison, WI, USA) and stored at −80°C. The lysates were spun at 14000 rpm (4°C) prior to analysis, and the soluble fraction was used to measure the luciferase activity according to the manufacturer’s instructions (Promega). The total protein concentration was assessed using a Bio-Rad protein assay kit (Bio-Rad, CA, USA) according to the manufacturer’s instructions. The data was corrected for total protein content.

Myogenic differentiation was assessed biochemically by measuring muscle creatine kinase (MCK) activity. After the induction of differentiation, the C₂C₁₂ cells were washed twice in ice-cold 1× PBS, subsequently lysed in 0.5% Triton X-100, and scraped from the dish with a cell scraper (rubber policeman). The lysates were centrifuged for 2 min at 14000 rpm (4°C), and the supernatant was aliquoted and stored at −80°C to determine the protein content or MCK activity in the presence of 1.25% BSA. The MCK-activity was measured spectrophotometrically (Stanbio Laboratory, TX, USA) [46]. The specific activity was calculated after correction for total protein content [47].

The muscle tissue was homogenized in ice-cold 1X whole cell lysate buffer (WCL) (50 mM Tris–HCl, pH 7.4; 150 mM NaCl; 1 mM EDTA; 1 mM Na₃VO₄; 5 mM NaF; 10% glycerol; 0.5% Nonidet P-40; 1 mM DTT; 1 mM PMSF; 10 μg/ml leupeptin; 1% aprotinin; 10 mM β-glycerophosphate and 1 mM Na-pyro-PO₄) using a rotating blade tissue homogenizer (Polytron PT 1600E, Kinematica, Switzerland). The C₂C₁₂ cells were washed twice in ice-cold 1× PBS after which they were lysed in 1× reporter lysis buffer and scraped of the dish using cell scrapers. The total protein concentration was assessed by the Thermo Scientific Pierce BCA Protein Assay kit (Pierce Biotechnology, IL, USA) according to the manufacturer’s instructions. The protein lysates were boiled for 5 min at 95°C after addition of 4× Laemmli sample buffer (0.25 M Tris–HCl pH 6.8; 8% (w/v) SDS; 40% (v/v) glycerol; 0.4 M DTT and 0.04% (w/v) Bromophenol Blue). For SDS-PAGE 1–25 μg of protein was loaded per lane and separated on a Criterion XT Precast 4 - 12% Bis-Tris gel (Bio-Rad), followed by transfer to a 0.45 μm Whatman Protran Nitrocellulose Transfer membrane (Whatman GmbH, Germany) by electroblotting (Bio-Rad Criterion Blotter). The nitrocellulose blots were incubated overnight (4°C) with specific antibodies directed against: myosin light chains 1 (MyLC-1) and −3 (MyLC-3) (#F310) (Developmental Studies Hybridoma Bank, Iowa City, IA, USA), myosin heavy chain fast (MyHC-f) (#M4276) (Sigma-Aldrich), p-eIF2Bϵ (Ser⁵³⁹) (#44-530G) (Invitrogen), p-mTOR (Ser²⁴⁴⁸) (#2971), mTOR (#2983), p-Akt (Ser⁴⁷³) (#9271), Akt (#9272), p-GSK-3β (Ser⁹) (#9336), GSK-3β (#9332), p-p70S6K (Thr³⁸⁹) (#9206), p70S6K (#2708), p-4E-BP1 (Thr^37/46) (#9459), 4E-BP1 (#9452), p-S6 (Ser^235/236) (#2211), p-FoXO1 (Ser²⁵⁶) (#9461), FoXO1 (#2880), p-FoXO3a (Ser²⁵³) (#9466), FoXO3a (#2497) and GAPDH (#2118) (all from Cell Signaling Technology), diluted in 1× TBS/0.1% Tween-20. The blots were probed with a peroxidase conjugated secondary antibody (#PI-1000) (Vector Laboratories, CA, USA), and visualized using Supersignal WestPico Chemiluminescent Substrate (Pierce Biotechnology) according to the manufacturer’s instructions and exposed to Super RX film (FUJIFILM, Japan). The Western blot films were digitalized using a Bio-Rad GS-800 Densitometer and subsequent quantification was done using Quantity One software (version 4.5.0) (both from Bio-Rad).

The raw data were entered into SPSS (version 20.0) for statistical analysis. All values are represented as means and error bars indicate the standard error of the mean (SEM). Comparisons of mean values were tested parametrically, using a one-way ANOVA followed by a post hoc Fischer’s LSD test. The changes in body weight were tested using a mix-model design ANOVA. Mean value comparisons of in vitro data were tested non-parametrically, using the Mann–Whitney U-test. A two-tailed probability value (p < 0.05) between groups was considered statistically significant.

Throughout the experimental procedures, neither LPS nor the concomitant administration of LPS and SB216763 significantly affected the increase in body weight of the guinea pigs (Figure 1A). However, from week 4 onwards the increase in body mass of the SB216763-treated saline-challenged group was significantly lower compared with the vehicle-treated, saline-challenged group (p < 0.05) (Figure 1A). Repeated LPS administration consistently appeared to decrease muscle wet weights (M. plantaris: -2%, M. gastrocnemius: -8%, M. tibialis: -5%, M. EDL: -7%), although this did not reach statistical significance (Figure 1B). Intriguingly, SB216763-treatment significantly prevented the LPS-induced reduction in these skeletal muscle weights (except for M. EDL). To verify the effects on muscle mass, the myofiber CSA of the EDL muscle was determined. The glycolytic EDL muscle predominantly consisted of Type II fibers (96.4%, data not shown), and immunohistochemical staining revealed that chronic LPS administration significantly decreased the mean Type II fiber CSA compared with vehicle control-treated muscle (Figure 1C). The decline in Type II fiber CSA following LPS was further substantiated by examining the fiber size distribution curves, which revealed a leftward shift (smaller fiber size) compared with the fiber distribution of vehicle-treated control animals (Figure 1D). Strikingly, pharmacological GSK-3 inhibition abrogated the reduction of mean Type II fiber CSA in response to LPS (Figure 1C and Figure 1D). Unexpectedly, GSK-3 enzyme inhibition caused a significant decrease in mean Type II fiber CSA in EDL muscle of vehicle-treated animals (Figure 1C). Nevertheless, collectively these data indicate that muscle atrophy induced by chronic LPS challenge is prevented by GSK-3 inhibition despite sustained pulmonary inflammation.

To address the potential contribution of altered protein synthesis signaling to the muscle atrophy phenotype, the protein levels and the phosphorylation state of mTOR and its downstream effectors p70S6K and 4E-BP1 as well as Akt, the upstream activator of mTOR were assessed. The phosphorylated (p)-Akt to Akt ratio in LPS control muscle was unchanged following a 12 week treatment regimen with intranasally instilled LPS. Likewise, the p-Akt levels in muscle exposed to SB216763 alone or in combination with LPS remained unaltered, comparable to vehicle/saline-treated controls (Figure 2A). Similarly, the phosphorylation state and abundance of GSK-3β, a direct downstream substrate of Akt, was unaffected in any of the conditions. Chronic pharmacological GSK-3 inhibition by SB216763 in the lung did not result in detectable alterations in the phosphorylation state of the GSK-3β substrate eIF2Bϵ (Figure 2A).

Furthermore, the ratio of p-mTOR over total mTOR was unaffected in any of the conditions. The phosphorylation state of p70S6K, a downstream substrate of mTOR, was unaffected by LPS instillation or GSK-3 inhibition (Figure 2B). In contrast, phosphorylation of S6, a substrate of p70S6K, tended to be reduced upon LPS instillation, but these findings did not reach statistical significance (Figure 2B). Finally, repeated LPS administration or GSK-3 inhibition did not affect p-4E-BP1 or total 4E-BP1 protein abundance, as another downstream substrate of mTOR (Figure 2B). Both phosphorylated levels of FoXO1 as well as total FoXO1 protein abundance remained unaltered following either LPS -or SB216763-treatment (Figure 2C). In contrast, the p-FoXO3a to FoXO3a ratio was reduced in response to concomitant LPS and SB216763-treatment, which is indicative of increased FoXO3a activity (Figure 2C). Altogether these data imply that gross alterations in skeletal muscle protein turnover signaling could not account for the muscle atrophy observed in response to chronic pulmonary inflammation, nor the prevention thereof by pharmacological GSK-3 inhibition.

In addition to alterations in protein turnover, impaired myogenesis may lie at the basis of sustained muscle wasting [48, 49]. Moreover, systemic inflammation resulting from pulmonary inflammation can trigger muscle atrophy [26], and inflammatory cytokines have been shown to contribute to muscle wasting through the inhibition of myogenic differentiation [43]. To investigate whether pharmacological GSK-3 inhibition prevents impaired myogenesis, differentiating C₂C₁₂ myoblasts were cultured in the presence or absence of LiCl and/or TNF-α. LiCl is a direct and indirect inhibitor of GSK-3 and has been widely used to investigate the role of GSK-3 [50, 51]. TNF-α supplementation resulted in diminished myogenesis of C₂C₁₂ myocytes (Figure 3A). Subsequent quantification of myotube formation, by determining the myogenic index, clearly demonstrated that TNF-α reduced myoblast fusion (Figure 3B). Conversely, LiCl increased myotube formation, and importantly, markedly attenuated the TNF-α-induced decrease in myotube formation (Figure 3B). TNF-α significantly decreased the myofibrillar protein abundance, i.e. MyHC-f, MyLC-1 and MyLC-3, whereas LiCl stimulated their expression (Figure 3C). Notably, LiCl significantly abrogated the reduction in contractile protein content in response to TNF-α (Figure 3C). In addition to reduced expression of sarcomeric/contractile proteins, TNF-α supplementation markedly decreased MCK activity. Conversely, enzymatic GSK-3 inhibition increased basal MCK activity and prevented the TNF-α-induced decline in MCK activity (Figure 3D). The differentiation-induced transcriptional activation of the TnI promoter was diminished in response to TNF-α, and increased following GSK-3 inhibition (Figure 3E). In line with the other markers of myogenesis, LiCl-treatment significantly reversed the reduction in TnI promoter transactivation in response to TNF-α.

Systemic inflammation increases circulating levels of cortisol; a potent trigger of muscle atrophy [52, 53]. Repeated intranasal LPS instillation in guinea pigs resulted in an increase in plasma cortisol levels (229%, ± 48.4%), which was unaffected by SB213763-treatment (172%, ±SEM 51.2%). Previously it was demonstrated that the synthetic GCs prednisolone as well as Dex strongly impair myogenesis [54]. The addition of Dex to the culture medium during differentiation resulted in impaired C₂C₁₂ myotube formation (Figure 4A). Similar to the results obtained with TNF-α, pharmacological GSK-3 significantly prevented impairment of myoblast fusion in the presence of Dex (Figure 4B). Furthermore, Dex significantly decreased the muscle-specific protein expression of MyHC-f, MyLC-1 and MyLC-3, while LiCl supplementation completely prevented this effect (Figure 4C). Moreover, Dex markedly reduced MCK activity (Figure 4D) and TnI promoter transactivation (Figure 4E), which was prevented in the presence of LiCl (Figure 4D and Figure 4E, respectively). To ascribe the preventive effects of LiCl on impaired myogenic differentiation by TNF-alpha or Dex to inhibition of GSK-3 enzymatic activity, the structurally unrelated GSK-3 inhibitor CHIR99021 was deployed. Incubation of differentiating myoblasts with CHIR99021 prevented or attenuated TNF-alpha-induced blockade of myogenic fusion or MyLC accumulation (Figure 5A, B), similar as observed with LiCl. Likewise, pharmacological GSK-3 inhibition using CHIR99021 reversed the Dex-induced impairment of myogenesis (Figure 5A, C).