BCATm deficiency ameliorates endotoxin-induced decrease in muscle protein synthesis and improves survival in septic mice

Animals and experimental protocol.

The animal protocols were approved by the Institutional Animal Care and Use Committee of The Pennsylvania State University College of Medicine and adhere to the National Institutes of Health guidelines for the use of experimental animals. Studies were performed on male KO and wild-type (WT) C57BL6 littermates between 12 and 14 wk of age. BCATm KO mice have elevated circulating BCAA concentrations, but they exhibit no increase in α-keto acids in either plasma or muscle (47). These KO mice, however, do possess the cytosolic isoform of BCAT present in neurons and brain (12), which can potentially lead to the excessive local production of BCAA metabolites and neurological derangements (48). To prevent the toxic accumulation of keto acids within neural tissues, both BCATm KO and WT mice had free access to both normal chow diet (model 2018,; Global 18% protein rodent diet; Harlan Teklad, Indianapolis, IN) and an isonitrogenous and isocaloric purified BCAA-free diet (model 510081; Dyets, Bethlehem, PA) (36, 47). Male WT mice consume ∼55% and 45% of the normal chow and BCAA-free diet, respectively, whereas KO mice consume 24% and 76% of the two diets, respectively. Importantly, total food consumption and energy intake between the WT and BCATm KO mice does not differ (36, 47). Food was removed from all mice at 0700 h, and experiments were performed between 0900 and 1200 h. BCATm and WT mice administered either saline or LPS were alternated to compensate for differences in the period of time mice were without food.

The in vivo rate of protein synthesis in the gastrocnemius-plantaris complex (hereafter referred to as skeletal muscle) and heart (ventricle only) was determined in WT and BCATm KO mice 4 h after either intraperitoneal injection of 0.9% saline (e.g., basal time-matched controls) or a nonlethal dose of LPS (Escherichia coli 026:B6, 25 μg/mouse; Sigma, St. Louis, MO). The LPS dose and timing were based on preliminary studies in mice demonstrating a reproducible reduction in muscle protein synthesis (data not shown). Protein synthesis was determined using the flooding dose technique, exactly as described (54). Mice were injected intraperitoneally with [³H]-l-phenylalanine (150 mM, 30 μCi/ml; 1 ml/100 g body wt), and blood was collected 15 min later for determining the plasma phenylalanine concentration and radioactivity. Thereafter, skeletal and cardiac muscle were rapidly excised, freeze-clamped, and then stored at −70°C. Tissues were processed exactly as described (33, 53). The rate of protein synthesis was calculated by dividing the amount of radioactivity incorporated into protein by the plasma phenylalanine-specific radioactivity, and the advantages and disadvantages of this method have been reviewed (13). The specific radioactivity of the plasma phenylalanine was measured by HPLC analysis of supernatant from TCA extracts of plasma. In addition, samples of fresh muscle were homogenized for Western blot analysis and immunoprecipitation of selected proteins as described below.

Western blot analysis.

Fresh tissue was homogenized in ice-cold homogenization buffer consisting of (in mmol/l): 20 HEPES (pH 7.4), 2 EGTA, 50 sodium fluoride, 100 potassium chloride, 0.2 EDTA, 50 β-glycerophosphate, 1 DTT, 0.1 phenylmethane-sulphonylfluoride, 1 benzamidine, and 0.5 sodium vanadate (30, 33, 36). Protein was determined after centrifugation and equal amounts of protein per sample were subjected to standard SDS-PAGE, using antibodies obtained from Cell Signaling (Beverly, MA), unless otherwise specified. Specifically, Western blot analysis was performed for total and phosphorylated (Thr246) PRAS40 (Biosource, Camarillo, CA), total and phosphorylated (Ser792) raptor, and total and phosphorylated 4E-BP1 (Thr 37/46; Bethyl Laboratories, Montgomery, TX). In addition, total GβL (also called mLST8), DEPTOR (Millipore, Billerica, MA), and eIF3f or eIF3b (Abcam, Cambridge, MA) were determined by Western blot analysis. Blots were developed with enhanced chemiluminescence Western blotting reagents (Supersignal Pico; Pierce, Rockford, IL). Dried blots were exposed to X-ray film to achieve a signal within the linear range, and film was then scanned (Microtek ScanMaker IV) and quantified using Scion Image 3b software (Scion, Frederick, MD).

The 4E-BP1·eIF4E and eIF4G·eIF4E complexes were quantified as described previously (30, 36). Briefly, eIF4E was immunoprecipitated from aliquots of supernatants by using an anti-eIF4E monoclonal antibody (Drs. Jefferson and Kimball; Hershey, PA). Antibody-antigen complexes were collected by using magnetic beads subjected to SDS-PAGE and quantified as above.

To maintain the protein-protein interactions within mTORC1, a portion of muscle was homogenized in 3[(3-cholamidopropyl)dimethylammonio]-propanesulfonic acid (CHAPS) buffer consisting of (in mmol/l): 40 HEPES (pH 7.5), 120 NaCl, 1 EDTA, 10 pyrophosphate, 10 β-glycerophosphate, 50 sodium fluoride, 1.5 sodium vanadate, 0.3% CHAPS, and 1 protease inhibitor cocktail tablet (14). The homogenate was mixed on a platform rocker and clarified by centrifugation. An aliquot of the resulting supernatant was combined with anti-raptor, anti-mTOR or anti-eIF3b antibody, and immune complexes were isolated with goat anti-rabbit BioMag IgG beads (PerSeptive Diagnostics, Cambridge, MA). Beads were collected, washed with CHAPS buffer, precipitated by centrifugation, and subjected to SDS-PAGE and analyzed as above.

Ribonuclease protection assays.

Total RNA was extracted from tissues in a mixture of phenol and guanidine thiocyanate (TRI Reagent; Molecular Research Center, Cincinnati, OH) by using the manufacturer's protocol. Riboprobes were synthesized from a multiprobe mouse cytokine template set (rCK-1; Pharmingen, San Diego, CA) by using an in vitro transcription kit (Pharmingen). Primer sequences for TNF-α, IL-6, nitric oxide synthase (NOS)-2, and inhibitor of NF-κB-δ (IκBζ) have been reported (41). The labeled riboprobe was hybridized with RNA overnight by using a ribonuclease protection assay. Protected RNAs were separated by using a 5% acrylamide gel, and dried gels were exposed to a phosphorimager screen (Molecular Dynamics, Sunnyvale, CA). The resulting data were quantified by using ImageQuant and normalized to L32.

Determination of plasma metabolites and hormone concentrations.

The plasma concentrations of insulin and IGF-I were determined by enzyme-linked immunosorbent assays (Linco Research, St. Charles, MO and Immunodiagnostic Systems, Fountain Hills, AZ, respectively). BCAAs were determined using reverse-phase HPLC after precolumn derivatization of amino acids with phenylisothiocyanate.

Sepsis survival study.

In a second study, the effect of BCATm deletion and elevated BCAA concentrations on survival was determined in septic mice. For this study, polymicrobial peritonitis was produced by cecal ligation and puncture (CLP), as described previously (41). Briefly, mice were anesthetized using isoflorane (Abbott Laboratories, North Chicago, IL). The abdomen was shaved, the skin was prepared with povidone iodine, and a midline incision (1.5 cm) was made below the diaphragm. The cecum was isolated, ligated, punctured twice with a 22-gauge needle, and a small amount of cecal material extruded to ensure patency. The cecum was returned to the abdomen, the muscle incision closed with 6-0 surgical suture (Ethicon, Somerville, NJ), and metal wound clips were used to close the skin incision. Before suturing the skin, two to three drops of lidocaine (Abbott Laboratories) were administered to the wound for analgesia. All mice were housed individually after surgery, and each received 1 ml of warmed (37°C) 0.9% sterile saline containing 0.05 mg/kg of buprenorphine (Reckitt Benckiser Pharmaceuticals, Richmond, VA) administered subcutaneously every 12 h for the remainder of the experimental protocol. Sham controls were subjected to the same surgical laparotomy and cecal isolation, but the cecum was neither ligated nor punctured. The survival study was performed in a manner whereby the investigators were blinded to the animal genotype and/or sepsis treatment until completion of the study. Survival was assessed every 12 h for 7 days. Although the primary endpoint of this second experiment was survival, any mouse determined to be moribund was anesthetized and killed instead of waiting for spontaneous death (6, 41). Before and after surgery, animals had unrestricted access to food and water, and food consumption was measure daily. At the end of the observation period, whole body lean mass was measured using an ¹H-NMR system (model LF90; Bruker Minispec, Woodlands, TX) as previously reported (47). Mice were then anesthetized, the gastrocnemius and heart were removed, and their wet weight was determined.

Statistical analysis.

Data for each condition are summarized as means ± SE where the number of mice per treatment group is indicated in the legend of the figure or table. Statistical evaluation of the data was performed using two-way ANOVA with post hoc Student-Neuman-Keuls test when the interaction was significant. Differences in survival were analyzed by a Kaplan-Meier survival plot and the log-rank statistic. Differences were considered significant when P < 0.05.

Plasma BCCA and hormone concentrations.

Under basal postabsorptive conditions, BCATm KO mice had elevated plasma concentrations for each BCAA [leucine (∼10-fold), isoleucine (∼6-fold), and valine (∼5-fold)] compared with time-matched pair-fed WT mice (Table 1). LPS did not significantly alter the plasma concentration of leucine, isoleucine, or valine in either WT or BCATm KO mice. There was no statistically significant interaction between gene deletion and LPS, so the BCAA concentrations did not differ between BCAT KO and WT mice injected with LPS.

The plasma insulin concentration did not differ between WT and BCATm KO mice under basal conditions (Table 1). In contrast, hyperinsulinemia was detected in both WT and BCATm KO mice after LPS, and the increment in plasma insulin was comparable in both groups. Neither BCATm deletion nor LPS administration significantly altered the plasma IGF-I concentration.

Muscle protein synthesis.

In vivo-determined protein synthesis in skeletal muscle from BCATm KO mice was increased 25%, compared with WT control values under basal (e.g., no LPS) conditions (Fig. 1, top). LPS decreased muscle protein synthesis 43% in WT mice and 33% in BCATm KO mice, compared with their respective control values. As a result, the protein synthetic rate was not different in muscle from WT mice under basal conditions and BCATm KO mice after LPS. In contrast to skeletal muscle, protein synthesis in heart was not altered by either the deletion of the BCATm gene and/or LPS administration (Fig. 1, bottom).

4E-BP1 phosphorylation and eIF4F complex formation.

We quantified the phosphorylation of 4E-BP1 as a marker of mTOR kinase activity, because this protein is a known mTOR substrate (7). The extent of 4E-BP1 phosphorylation was increased 60% in skeletal muscle of BCATm KO mice, compared with time-matched control values (Fig. 2A). Conversely, LPS decreased 4E-BP1 phosphorylation in both groups, but the extent of 4E-BP1 dephosphorylation was greater in WT than BCATm KO mice. These changes were independent of a change in total 4E-BP1 content, as assessed by quantitation of the α-, β-, and γ-isoforms of the protein in muscle homogenates among the four groups [WT-control = 5,460 ± 445 arbitrary units (AU), BCATm KO-control = 5,520 ± 566 AU, WT-LPS = 5,402 ± 298 AU, and BCATm KO-LPS = 5,606 ± 522 AU; P = not significant (NS)]. The hyperphosphorylation of 4E-BP1 leads to the redistribution of eIF4E from the inactive eIF4E·4E-BP1 to the active eIF4E·eIF4G complex (15). However, under control conditions, the amount of eIF4E·4E-BP1 complex in skeletal muscle did not differ between WT and BCATm KO mice (Fig. 2B). LPS appeared to selectively increase eIF4E·4E-BP1 content in WT, but not in KO mice. As anticipated, the binding of eIF4E to eIF4G was increased in BCATm KO mice under control conditions compared with WT control values (Fig. 2C). Conversely, the amount of eIF4E·eIF4G was reduced in both groups after LPS, but the absolute amount of the complex was lower in WT than KO mice. The above-mentioned changes in eIF4E distribution were independent of a change in total eIF4E in skeletal muscle among the four experimental groups (data not shown).

To confirm the above-mentioned data for myocardial protein synthesis, we also determined 4E-BP1 phosphorylation. Our data indicate that Thr37/46 phosphorylation of 4E-BP1 was not statistically altered by either BCATm gene deletion and/or LPS administration in heart (data not shown), and additional analysis of mTORC1 in heart was not pursued because of this lack of change.

mTORC-1.

BCATm deletion and/or LPS were without effect on the total amount of raptor, PRAS40, GβL, and DEPTOR (data not shown), four proteins that together with mTOR comprise mTORC1 (14, 24). Although PRAS40 phosphorylation did not differ between WT and BCATm KO mice under control conditions, LPS decreased PRAS40 phosphorylation by 40% in both groups (Fig. 3, top). Furthermore, raptor phosphorylation increased twofold in both WT and KO mice in response to LPS (Fig. 3, bottom).

Neither BCATm gene deletion nor LPS altered the amount of mTOR or PRAS40 bound to raptor (Fig. 4, A and B, respectively). In contrast, the ability of 4E-BP1 to bind raptor tended to be increased (P < 0.08) under control conditions in KO mice, compared with WT controls (Fig. 4C). In response to LPS, the amount of the 4E-BP1·raptor complex was reduced in muscle from both WT and KO mice, but the decrease was greater in LPS-treated WT mice. This change in 4E-BP1·raptor binding was independent of the amount of raptor in the immunoprecipitate (data not shown). Finally, BCATm KO mice demonstrated an increased binding of raptor with eIF3 under control conditions (Fig. 4D). The amount of the raptor·eIF3 complex in muscle was decreased by LPS in WT mice, but not in BCATm KO mice.

Muscle cytokines.

As part of the innate immune system, whole muscle and cultured myocytes secrete a number of inflammatory cytokines that can impair protein synthesis (8). Therefore, TNF-α IL-6, NOS2, and IkBζ mRNA expression were determined. The muscle mRNA content for these immunomodulators did not differ between WT and BCATm KO mice under control conditions (Fig. 5). As anticipated, LPS increased muscle mRNA expression of each modulator in WT mice. However, this LPS-induced increase was smaller in magnitude for each of the assessed inflammatory mediators.

Sepsis survival study.

To determine whether the elevation in leucine and other BCAA, as well as the reduced inflammatory response, would translate to a survival advantage in BCATm KO mice, 7-day survival was determined by using a standard CLP model of peritonitis. This model is generally recognized to mimic the clinical scenario with greater fidelity than the injection of LPS, which is more appropriate in studying the acute phase of bacterial infection (3). The survival of both WT and BCATm KO mice under control conditions (e.g., following sham surgery) was 100% (Fig. 6A). WT mice exhibited only 20% survival at 7 days after CLP, and this was significantly (P < 0.05) lower than the 65% survival rate observed in BCATm KO mice after CLP.

Both groups of septic mice showed a 60–80% decrease in food consumption the first day after CLP, compared with time-matched control values (Fig. 6B). Food consumption gradually increased and was not different from time-matched control values on days 5–7. Hence, a difference in food consumption cannot explain the survival advantage seen in the BCATm KO mice, compared with WT mice, in response to CLP.

The body weight of the BCATm KO mice immediately prior to the induction of sepsis was 9% lower than WT control values (28.4 ± 0.8 vs. 25.5 ± 0.4 g; P < 0.05), consistent with previous reports (36, 47). In the sham control mice, a comparable difference in body weight was detected at the end of the 7-day observation period (Fig. 7A). Sepsis induced by CLP decreased body weight 10% on day 7 in WT mice, but no decrement in body weight was detected in BCATm KO mice. Whole body lean body mass determined by ¹H-NMR did not differ between WT and BCATm KO mice prior to CLP (22.1 ± 0.1 and 21.9 ± 0.1 g, respectively; P = NS). However, similar to the change in body weight, the sepsis-induced decrease in lean body mass of 11% in WT mice was not seen in BCATm KO mice (Fig. 7B). Likewise, the directly determined wet weight of the gastrocnemius did not differ between WT and BCATm KO mice under nonseptic control conditions, but muscle weight was selectively decreased in WT and intermediate in the BCATm KO mice in response to CLP (Fig. 7C). In contrast, the wet weight of the heart did not differ among any of the four experimental groups (data not shown).