Oxygen deprivation during ischemic or hemorrhagic stroke results in ATP depletion, loss of ion homeostasis, membrane depolarization, and excitotoxicity. Pharmacologic restoration of cellular energy supply may offer a promising concept to reduce hypoxic cell injury. In this study, we investigated whether carbimazole, a thionamide used to treat hyperthyroidism, reduces neuronal cell damage in oxygen-deprived human SK-N-SH cells or primary cortical neurons. Our results revealed that carbimazole induces an inhibitory phosphorylation of eukaryotic elongation factor 2 (eEF2) that was associated with a marked inhibition of global protein synthesis. Translational inhibition resulted in significant bioenergetic savings, preserving intracellular ATP content in oxygen-deprived neuronal cells and diminishing hypoxic cellular damage. Phosphorylation of eEF2 was mediated by AMP-activated protein kinase and eEF2 kinase. Carbimazole also induced a moderate calcium influx and a transient cAMP increase. To test whether translational inhibition generally diminishes hypoxic cell damage when ATP availability is limiting, the translational repressors cycloheximide and anisomycin were used. Cycloheximide and anisomycin also preserved ATP content in hypoxic SK-N-SH cells and significantly reduced hypoxic neuronal cell damage. Taken together, these data support a causal relation between the pharmacologic inhibition of global protein synthesis and efficient protection of neurons from ischemic damage by preservation of high-energy metabolites in oxygen-deprived cells. Furthermore, our results indicate that carbimazole or other translational inhibitors may be interesting candidates for the development of new organ-protective compounds. Their chemical structure may be used for computer-assisted drug design or screening of compounds to find new agents with the potential to diminish neuronal damage under ATP-limited conditions.
Ischemic and hemorrhagic stroke remain leading causes of death and disability worldwide (Sahni and Weinberger, 2007; Barreto, 2011). Current treatment strategies are primarily focused on the control of bleeding (Sahni and Weinberger, 2007), restoration of blood flow by thrombolytic therapies (Barreto, 2011), or lowering of intracranial pressure (Sahni and Weinberger, 2007) but are limited in success as they are aimed at the acute phase of brain damage and neglect that a significant amount of neuronal tissue damage develops progressively in the penumbra after the initial insult. Crucial molecular changes in penumbral tissue potentiate secondary brain injury, especially when ATP supply is insufficient to maintain ion homeostasis (White et al., 2000; Stankowski and Gupta, 2011). Loss of high-energy phosphate compounds results in the failure of ion-motive ATPases, membrane depolarization, excitotoxic glutamate release, calcium overload, the induction of phospholipases and proteases, the formation of radicals and inflammatory mediators, and translation of damage-promoting proteins such as inducible nitric oxide synthase, cyclooxygenase-2, or matrix metalloproteinases (White et al., 2000; Won et al., 2002; Cunningham et al., 2005; Stankowski and Gupta, 2011). There is a high correlation between the extent of neurologic recovery and the volume of penumbra that escapes infarction (Furlan et al., 1996); therefore, neuroprotective drugs with the potential to inhibit deleterious effects in the penumbra should improve clinical outcome.
Depression of protein translation, a dominant energy-consuming cellular activity, results in substantial bioenergetic savings (Hand and Hardewig, 1996). Increased availability and reallocation of cellular energy to vitality-preserving mechanisms such as maintenance of ionic equilibrium or cellular repair may become critical for survival, especially under conditions of insufficient oxygen and nutrient supply (Boutilier and St Pierre, 2000; Wang et al., 2008). Inhibition of protein synthesis during ischemia may also prevent the translational induction of mediators associated with nitration and oxidation of lipids, proteins, and DNA; inhibition of mitochondrial respiration; inflammation; increased intracranial pressure; or even hemorrhage (Won et al., 2002; Cunningham et al., 2005).
The majority of energy used in translation is consumed during elongation (Browne and Proud, 2002). In mammalian cells, peptide-chain elongation requires eukaryotic elongation factor 2 (eEF2), which promotes the translocation of peptidyl-tRNA from the A site to the P site on the ribosome (Jorgensen et al., 2006). Phosphorylation of eEF2 at Thr56 by eEF2 kinase (eEF2K) impairs interaction of eEF2 with the ribosome (Carlberg et al., 1990) and is sufficient for the inhibition of mRNA translation (Ryazanov et al., 1988). Enzymatic function of eEF2K is dependent upon calcium and calmodulin (Ryazanov, 1987); however, phosphorylation of eEF2K by cAMP-dependent protein kinase (protein kinase A or PKA) (Redpath and Proud, 1993) or AMP-activated protein kinase (AMPK) (Horman et al., 2002) may lead to acquisition of calcium-independent activity. Cytoprotective properties of enzymes involved in eEF2 regulation are well described. Activation of eEF2K promotes cell survival, reduces hypoxic injury, and regulates autophagy in response to nutrient deprivation (Terai et al., 2005; Py et al., 2009). AMPK preserves energy homeostasis (Hardie, 2004) and attenuates ischemic cell damage (Zhang et al., 2010), inflammation (Salminen et al., 2011), arrhythmias (Wong et al., 2009), hypertrophy (Chan et al., 2004), structural remodeling (Du et al., 2008), and plaque formation in Alzheimer’s disease (Greco et al., 2009). Additionally, it promotes neurogenesis, angiogenesis, and vascular function (Li and McCullough, 2010).
Carbimazole is a heterocyclic thiourylene used to treat hyperthyroidism. It prevents the thyroid peroxidase–catalyzed iodination of l-tyrosine and thus reduces the production of the thyroid hormones T3 and T4. Amelioration of acute tissue damage by carbimazole in patients has not been investigated. However, in vitro studies have shown that carbimazole may inhibit inflammation (Humar et al., 2008) and neuronal apoptosis (Humar et al., 2009). Furthermore, structurally related thiourylenes have been implicated in the reduction of excitotoxicity (Zhan et al., 1998) and radical scavenging (Steen and Michenfelder, 1980) and the induction of cytoprotective proteins (Roesslein et al., 2008), and they may be used in brain-injured patients to lower intracranial pressure and to improve cerebral oxygenation (Nordby and Nesbakken, 1984).
Because of the cytoprotective mechanisms attributed to heterocyclic thiourylenes, it is plausible to analyze their therapeutic potential in ischemic neurons. Loss of high-energy phosphate compounds in ischemic neurons triggers a destructive biochemical cascade culminating in neuronal cell death. Therefore, the aim of the present study was to examine carbimazole-mediated effects on global protein synthesis, its impact on high-energy phosphate supply, and whether bioenergetic savings due to translational repression reduce neuronal damage following oxygen deprivation.
Materials and Methods
Cell Culture and Treatment.
The human neuronal cell line SK-N-SH was purchased from the American Type Culture Collection (Manassas, VA) and maintained in Eagle’s minimum essential medium (MEM), supplemented with 1 mM sodium pyruvate, 2 mM GlutaMAX, 1500 mg/l sodium bicarbonate, 100 IU streptomycin, 100 IU penicillin, and 10% fetal calf serum (FCS). For culture of primary neurons, cortices of C57/BL6 neonates (P0–P2) were collected in ice-cold modified Eagle’s medium supplemented with 2 mM GlutaMAX, cut with scissors to smaller pieces, trypsinized (0.25% trypsin) for 10 minutes at 37°C, and dissociated in Hanks’ balanced salt solution containing 10 mM HEPES/KOH (pH 7.9), 3 mg/ml bovine serum albumin, 12 mM MgSO4, 0.025% DNAse I, and 0.4 mg/ml soybean trypsin inhibitor (Applicem, Darmstadt, Germany) as described previously (Lin et al., 2013). The dissociated cells were collected from the supernatant after centrifugation at 100g for 5 minutes, and neurons were enriched by a second centrifugation at 800g for 5 minutes before they were seeded on poly-d-lysine-coated tissue culture plates in Dulbecco’s modified Eagle’s medium containing 10% FCS and 2 mM GlutaMAX at 37°C with 8% CO2. After 3 hours, the medium was replaced by Neurobasal-A supplemented with 2% B27 supplement and 1 mM GlutaMAX. All materials used for cell culture were from Life Technologies (Carlsbad, CA) unless indicated otherwise.
Cells were treated with carbimazole (LTK Laboratories, St. Paul, MN), thapsigargin (Calbiochem, San Diego, CA), forskolin (Calbiochem), cAMPS-Rp (Tocris Bioscience, Bristol, UK), PKI 14-22 amide (Calbiochem), rapamycin (Calbiochem), dorsomorphin dihydrochloride (Calbiochem), camstatin (Tocris Bioscience), 1,3-dihydro-1-[1-[(4-methyl-4H,6H-pyrrolo[1,2-a][4,1]benzoxazepin-4-yl)methyl]-4-piperidinyl]-2H-benzimidazol-2-one maleate (CGS 9343B; Tocris Bioscience), cycloheximide (Sigma-Aldrich, St. Louis, MO), or anisomycin (Calbiochem) as indicated. Oxygen deprivation was induced by moving cell cultures to freshly prepared growth medium that has been pre-equilibrated in a hypoxic atmosphere for 15 hours (5% CO2, 95% N2). To induce hypoxic cell death, cell culture was continued for 72 hours in a controlled humidified atmosphere of 5% CO2 and 95% N2 at 37°C.
Cells were lysed in 50 mM Tris-HCl (pH 6.8), 100 mM dithiothreitol, 2% SDS, 10% glycerol, and 0.1% bromphenol blue in tissue culture plates. Lysates were sonicated and boiled for 5 minutes, and equal amounts of protein were separated by SDS-PAGE. Proteins were transferred onto a polyvinylidene difluoride membrane; blocked in 0.2% I-Block (Life Technologies), 50 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 0.1% Tween 20; and incubated with antibodies raised against eEF2, phospho-eEF2(Thr56), AMPK, phospho-AMPK(Thr172; clone 40H9), procaspase-3, cleaved poly(ADP-ribose) polymerase (PARP), or α-tubulin according to the manufacturer’s recommendations (Cell Signaling, Danvers, MA). Specific protein bands were visualized using horseradish peroxidase–conjugated anti-rabbit IgGs and enhanced chemiluminescence reagents (GE Healthcare, Munich, Germany).
Cells were synchronized overnight in 24-well plates using FCS/MEM before treatment with translational inhibitors for 4 hours. Cultures then were washed two times for 20 minutes in methionine/MEM (MP Biomedicals, Illkirch, France) to deplete internal methionine pools before cells were pulse-labeled with 200 μCi/ml [35S]methionine (PerkinElmer Life and Analytical Sciences, Rodgau, Germany) for 2 hours and finally lysed in 50 mM Tris-HCl (pH 6.8), 100 mM dithiothreitol, 2% SDS, 10% glycerol, and 0.1% bromphenol blue. Proteins were denatured by boiling for 5 minutes and separated by 10% SDS-PAGE before gels were fixed in 25% isopropanol and 10% acetic acid for 30 minutes, stained in 10% acetic acid containing 0.006% Coomassie Brilliant Blue G-250 for 2 hours, destained in 10% acetic acid for 6 hours, and dried on a Whatman 3MM paper (Sigma-Aldrich) for 2 hours at 80°C. [35S]Methionine incorporation into proteins was visualized by X-ray autoradiography.
Measurement of Intracellular Calcium.
Intracellular cytoplasmic calcium measurements were performed by fura-2 labeling of SK-N-SH cells. Cultures were synchronized overnight by starvation, treated with carbimazole or thapsigargin in calcium-free or calcium-containing RPMI 1640 medium (Genaxxon Bioscience, Ulm, Germany) for 1 hour, and loaded with 2 μM fura-2 acetoxymethyl ester (AM) (Life Technologies) for the last 30 minutes of the experiment at 37°C in the dark. Fluorescence was excited at 340 nm and emission was recorded at 510 nm by a spectrofluorophotometer (SpectraMax GeminiXS; Molecular Devices, Sunnyvale, CA).
Measurement of Intracellular cAMP.
Cells were lysed with moderate shaking in 0.1 M HCl for 15 minutes at room temperature before cellular debris was removed by centrifugation at 600g. The cAMP content in cellular lysates was determined by a competitive enzyme immunoassay as directed by the manufacturer’s instructions [Cyclic AMP (Direct) Enzyme Immunometric Assay (EIA), non-acetylated format; Assay Designs, Ann Arbor, MI]. The absorption of each sample was quantified at 405 nm by a SpectraMax Plus384 plate reader (Molecular Devices). Each value was normalized to internal controls and a freshly prepared standard curve ranging from 0.2–200 pmol/ml cAMP.
Cytotoxicity was determined by the colorimetric Cytotoxicity Detection Kit (Roche Applied Science, Penzberg, Germany), which measures the activity of lactate dehydrogenase (LDH) released from the cytosol of damaged cells. In brief, cell culture supernatants were taken at different time points after induction of hypoxia, added to LDH substrate, and incubated for 30 minutes at 25°C. The absorbance at 490 nm was measured by a SpectraMax Plus384 spectrophotometer with a reference wavelength of 690 nm. LDH values corresponding to 100% neuronal death were established by addition of 1% Triton-X 100 to untreated control cells, which induced total cell lysis and maximal LDH release. Determined LDH values were displayed as a percentage of neuronal death compared with total cell lysis.
Caspase Activity Assay.
SK-N-SH cells were lysed in 10 mM HEPES/KOH (pH 7.9), 350 mM NaCl, 1% Nonidet P-40, 1 mM MgCl2, 0.5 mM EDTA, 0.1 mM EGTA, 5 mM dithiothreitol, 2.5 mM phenylmethylsulfonyl fluoride, and 20 μg/ml aprotinin. Extracts were diluted 1:10 in 100 mM HEPES/KOH (pH 7.5), containing 2 mM dithiothreitol and 60 μM fluorogenic caspase-3 substrate N-acetyl-(aspartic acid–glutamic acid–valine–aspartic acid) 7-amino-4-methylcoumarin (Ac-DEVD-AMC; Enzo Life Sciences, Loerrach, Germany). Caspase-3-like activity was determined by a Gemini XS plate reader at 340/460 nm (Molecular Devices) for 30 minutes at 27°C. Values were normalized to protein content and expressed as increasing relative light units (RLU) per second.
Measurement of Intracellular ATP.
Relative intracellular ATP content was determined by the bioluminescence-based ATP Assay Kit (Calbiochem, Merck KGaA, Darmstadt, Germany), which quantifies the ATP-dependent oxidation of luciferin by luciferase. Briefly, cells were lysed in 250 μl of ice-cold ATP-releasing buffer containing 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% Nonidet P-40, 100 μg/ml phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, and 70 μg/ml aprotinin by repeated freeze/thaw cycles. The cell lysates were diluted 1:5 by addition of 40 μl of nucleotide-releasing buffer, and luciferase activity was determined for 10 seconds per sample upon automated injection of 50 μl of a luciferin/luciferase mixture using a luminometer (Microluminat Plus LB 96P; Berthold Technologies, Bad Wildbad, Germany). Values were expressed as RLU and normalized to protein content, which was determined by the bicinchoninic acid (BCA) Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL).
Results are expressed as mean ± S.D. for the indicated number of separate experiments. Statistical differences between experimental groups were determined by performing one-way analysis of variance (ANOVA) followed by the Dunnett’s multiple-comparisons test, unpaired t test, or by two-way ANOVA, followed by the Bonferroni post hoc test as indicated. Differences between groups were considered to be significant at P < 0.05. Statistical analyses were carried out using the Prism software package (GraphPad Software Inc., La Jolla, CA).
Carbimazole Inhibits Global Protein Synthesis by Induction of eEF2 Phosphorylation.
Translational elongation of the polypeptide chain is regulated by eEF2, a translocase that is responsible for the movement of the mRNA along the ribosome (Jorgensen et al., 2006). The activity of eEF2 is tightly controlled by posttranslational modification (Carlberg et al., 1990), and phosphorylation of eEF2 at Thr56 is associated with translational repression (Ryazanov et al., 1988). To determine whether carbimazole affects translational elongation, human neuronal SK-N-SH cells were treated with increasing doses of carbimazole, and eEF2 phosphorylation was analyzed by immunoblotting. Our results clearly indicate that phosphorylation of eEF2 at Thr56 is induced by carbimazole in a dose-dependent manner (Fig. 1A; Supplemental Fig. 1A). The inhibitory modification of eEF2 was already detectable at 50 μM and was markedly pronounced in the presence of ≥0.5 mM carbimazole. Phosphorylation of eEF2 was induced within 10–60 minutes and persisted for at least 24 hours (Fig. 1B; Supplemental Fig. 1B). Under these conditions, carbimazole mediated no cytotoxic effects, because a significant increase in intracellular LDH activity was not detected in cell culture supernatants of SK-N-SH cells that were cultured in the presence or absence of carbimazole for 72 hours (Figs. 3D and 5B). The total amount of eEF2 was unchanged by exposure to carbimazole, indicating that increases in phospho-eEF2 are due to enzymatic modification but not increased protein synthesis (Fig. 1, A and B, lower blots).
Posttranslational modification of translational regulators is a common mechanism to control de novo protein synthesis (Browne and Proud, 2002). To determine whether carbimazole-dependent phosphorylation of eEF2 at Thr56 corresponds to translational inhibition, SK-N-SH cells were metabolically labeled with [35S]methionine for 2 hours, and the amount of de novo synthesized polypeptides was visualized by SDS-PAGE and autoradiography. Preliminary experiments ensured that incorporation of [35S]methionine into the growing polypeptide chains was directly proportional to the length of labeling time within 20 minutes to 6 hours (data not shown). As shown in Fig. 2, A and B, exposure of SK-N-SH cells to carbimazole resulted in a dose-dependent reduction of the basal rate of protein synthesis. Treatment of cells for 4 hours with cycloheximide, a known inhibitor of translation, also resulted in a marked decrease in protein synthesis (Fig. 2, A and B). In contrast, total protein levels visualized by Coomassie Brilliant Blue G-250 staining of gels before autoradiography were similar, demonstrating the comparison of equivalent samples and excluding cytotoxic effects of the tested compounds (Fig. 2C). These findings indicate that carbimazole does not selectively downregulate protein expression but instead exerts a global inhibitory effect on protein translation. The ability of carbimazole to decrease global protein synthesis is consistent with the observed increase in eEF2 phosphorylation.
Carbimazole Moderately Elevates Cytoplasmatic Calcium Levels.
Next, the mechanism involved in translational regulation by carbimazole was investigated. eEF2 phosphorylation at Thr56 is mediated by eEF2K (Nairn and Palfrey, 1987). To prove that eEF2K is directly involved in carbimazole-mediated phosphorylation of eEF2, SK-N-SH cells were incubated with 1-hexadecyl-2-methyl-3-(phenylmethyl)-1H-imidazolium iodide (NH 125), a specific eEF2K inhibitor. As shown in Fig. 3A (densitometry in Supplemental Fig. 2A), NH125 abrogated eEF2 phosphorylation in the presence of carbimazole, confirming that this highly specific kinase is crucial for carbimazole-mediated translational regulation.
Because eEF2K is dependent on calcium and calmodulin for activity (Ryazanov, 1987), we next analyzed the effect of carbimazole on cytoplasmatic calcium content. Free intracellular calcium imaging was performed by labeling SK-N-SH cells with the membrane-permeable fluorometric calcium indicator fura-2 AM. As illustrated in Fig. 3B, carbimazole increased the fura-2 fluorescence by 4- to 6-fold. This increase was dependent on the presence of extracellular calcium, indicating the opening of plasma membrane-gated calcium channels. In contrast, thapsigargin elevated cytoplasmic calcium levels also in the absence of extracellular calcium, thus demonstrating the depletion of intracellular calcium stores.
Intracellular calcium overload is a critical event in excitotoxicity (Won et al., 2002). Therefore, caspase-3 activity assays were performed to analyze whether elevated calcium levels due to carbimazole or thapsigargin treatment induce apoptosis. As shown in Fig. 3C, thapsigargin induced strong caspase-3-like activity, whereas carbimazole mediated no significant cleavage of the caspase-3-specific substrate Ac-DEVD-AMC. Similarly, LDH leakage into cell culture supernatants, used as a nonspecific marker for cell injury, was only observed in the presence of thapsigargin (Fig. 3D). This indicates that the moderate increase in intracellular calcium by carbimazole treatment does not facilitate cytotoxicity but may be more likely involved in calcium-dependent intracellular signaling.
Protein Kinases in the Regulation of eEF2 Phosphorylation.
Calcium in resting cells is about 0.1 μM, and the observed 4- to 6-fold increase might be close to optimum, described for the induction of calcium/calmodulin-dependent proteins (Diggle et al., 1998). To further characterize the biologic consequences of a carbimazole-dependent intracellular calcium increase, the calcium/calmodulin antagonists CGS 9343B and camstatin were used. Surprisingly, neither CGS 9343B nor camstatin influenced carbimazole-dependent eEF2 phosphorylation, arguing for a calcium/calmodulin-independent eEF2 modulation by carbimazole (Fig. 4A; Supplemental Fig. 2, B and C).
Several protein kinases regulate eEF2K activity (Browne and Proud, 2002). Phosphorylation of eEF2K by PKA results in acquisition of calcium-independent eEF2K activity and phosphorylation of eEF2 (Redpath and Proud, 1993). In fact, accumulation and release of cAMP as a response to thyroid-stimulating hormone and thyroid-stimulating antibodies have been described, but direct effects of carbimazole on adenylate cyclase and cAMP levels remained unidentified (Bidey et al., 1981). Therefore, we examined whether carbimazole affects cAMP signaling and determined the amount of cAMP in carbimazole-treated SK-N-SH cells by a competitive enzyme immunoassay. We observed a transient increase in cytosolic cAMP in neuronal SK-N-SH cells 60 minutes after carbimazole treatment (Fig. 4B). However, the biologic significance of this marginal increase is unclear. For comparison, 100 μM of the cAMP analog 8-(4-chlorophenylthio)-cAMP is used to induce biologic responses (Hovland et al., 1999), and when we used forskolin, a labdane diterpene that is commonly applied to induce adenylate cyclase–dependent cAMP synthesis, cAMP levels were increased 100–200 times higher than with carbimazole (data not shown). To determine the relevance of the observed carbimazole-mediated cAMP increase, we applied the cell-permeable cAMP analog cAMPS-Rp or the myristoylated peptide inhibitor PKI 14-22 amide as specific inhibitors of PKA and subsequently measured PKA-independent eEF2 phosphorylation in the presence of carbimazole. As demonstrated in Fig. 4C (densitometry in Supplemental Fig. 2, D and E), treatment of SK-N-SH cells with cAMPS-Rp or PKI 14-22 amide had no effect on carbimazole-dependent eEF2 phosphorylation, indicating that the cAMP/PKA pathway is not directly involved in translational regulation by carbimazole.
In contrast to PKA, the mTOR pathway induces eEF2-dependent translational elongation by inhibitory phosphorylation of eEF2K (Browne and Proud, 2002). To determine, whether inactivation of eEF2K by mTOR is reversed by carbimazole, carbimazole-mediated eEF2 phosphorylation was analyzed in the presence of the specific mTOR inhibitor rapamycin. As shown in Fig. 4D and Supplemental Fig. 2F, treatment of cells with rapamycin had no effect on carbimazole-dependent eEF2 phosphorylation, indicating that the mTOR pathway is also not participating in translational regulation by carbimazole.
Carbimazole Induces AMPK.
Various reports indicate that heterocyclic thiourylenes attenuate mitochondrial respiration (Aldridge and Parker, 1960; Chance et al., 1963). Energy starvation may then induce AMPK, an enzyme that is critically involved in the regulation of energy homeostasis (Oakhill et al., 2012). Because carbimazole is a heterocyclic thiourylene derivate, we speculated that an interaction of this compound with oxidative phosphorylation might lead to a shift in cellular AMP/ATP ratios and consequently in AMPK/eEF2K-dependent phosphorylation of eEF2. To test this hypothesis, AMPK activity was determined in the presence of carbimazole by visualizing AMPK autophosphorylation at Thr172 in the activation loop, which is required for activity (Oakhill et al., 2012). As shown in Fig. 4E (densitometry in Supplemental Fig. 1C), exposure of neuronal SK-N-SH cells to carbimazole revealed a marked increase in the levels of Thr172-phosphorylated AMPK 1 hour after onset of treatment, which persisted for at least 6 hours. The total amount of AMPK was unchanged by exposure to carbimazole, indicating that increases in phospho-AMPK are due to enzymatic modification and not enhanced protein synthesis (Fig. 4E, lower blot).
To investigate the significance of AMPK in eEF2 inhibition, we compared the phosphorylation pattern of eEF2 at Thr56 in the presence or absence of dorsomorphin dihydrochloride, a specific AMPK inhibitor. As shown in Fig. 4F (densitometry in Supplemental Fig. 2G), dorsomorphin hydrochloride completely blocked carbimazole-induced eEF2 phosphorylation, confirming AMPK/eEF2K-dependent eEF2 inactivation.
Carbimazole Protects Neuronal SK-N-SH Cells from Hypoxic Cell Damage.
Secondary brain damage following hypoxic cell injury evolves progressively and thus offers a therapeutic window for pharmacologic intervention (Heiss and Graf, 1994). To analyze whether carbimazole protects neurons from hypoxia-related cell damage, SK-N-SH cells were exposed to oxygen deprivation before intracellular LDH release into cell culture medium was measured as an index of neuronal injury. As shown in Fig. 5A, 48 hours of hypoxia induced significant cell injury as indicated by marked LDH leakage. Surprisingly, we observed that cell damage due to oxygen deprivation was most pronounced in serum-activated neuronal SK-N-SH cells, whereas serum-starved cells remained largely unaffected. The difference in susceptibility implies that metabolically active cells are particularly vulnerable to hypoxic cell death in contrast to resting cells, which may be resistant to injury by oxygen deprivation. Therefore, we suspected that the observed translational arrest by carbimazole may contribute to metabolic repression and thus affects hypoxic cell damage in otherwise nonrestricted cells. To confirm our assumption, cell damage was accessed in carbimazole-treated, oxygen-deprived SK-N-SH cells. As demonstrated in Fig. 5B, hypoxia-induced cell injury was almost completely abolished by carbimazole 48 hours after onset of deoxygenation. At this time point, carbimazole reduced the extent of hypoxia-mediated LDH leakage to levels comparable with the spontaneous release of untreated, normoxic control cells. However, carbimazole-mediated protection decreased after 72 hours, probably due to a limited half-life of carbimazole in serum-conditioned medium. The morphologic studies confirmed the results of LDH leakage; i.e., hypoxia induced detachment of SK-N-SH cells and thus decreased viable cellular density, whereas carbimazole prevented detachment of neurons and cellular density was comparable to untreated normoxic control cells (data not shown).
Inhibition of Global Protein Synthesis by Carbimazole Preserves Availability of ATP.
ATP depletion has deleterious effects on ion homeostasis and energy metabolism, resulting in glycolysis-dependent acidosis, excitotoxicity, and neuronal cell death (Won et al., 2002). Protein synthesis is a dominant energy-consuming cellular process, and suppression of translation might result in substantial bioenergetic savings (Hand and Hardewig, 1996). Therefore, we investigated whether carbimazole-mediated translational repression affected intracellular ATP content in normoxic or oxygen-deprived SK-N-SH cells, using an ATP-dependent bioluminescence assay (Fig. 6). Carbimazole did not significantly reduce availability of ATP in normoxic cells as demonstrated by comparable luciferase activity in cell lysates of nontreated versus carbimazole-treated SK-N-SH cells. This is surprising, as AMPK, a most sensitive detector for increased AMP/ATP ratios, was activated in the presence of carbimazole (Fig. 4E; Supplemental Fig. 1C). However, in oxygen-deprived cells, maximal ATP depletion was already detectable after 12 hours, with no further significant reduction for the next 60 hours in hypoxic cell culture medium. Carbimazole prevented hypoxia-dependent ATP deprivation in an early phase for up to 48 hours. The ATP-dependent bioluminescence was only insignificantly reduced in lysates of carbimazole-treated, hypoxic SK-N-SH cells compared with normoxic control conditions. This indicates that translational repression by carbimazole is linked to the conservation of high-energy phosphates in the form of ATP, thus opposing hypoxic cell injury. At 72 hours, ATP availability in carbimazole-treated hypoxic cells was still considerably increased compared with oxygen-deprived control cells but noticeably reduced in comparison with normoxic cells. The observed reduction in ATP availability in carbimazole-treated hypoxic cells may be due to a time-dependent loss of carbimazole activity and correlates with the reduced protection from hypoxic damage (Fig. 5B). In conclusion, conservation of cellular energy by inhibition of global protein synthesis and its reallocation to essential ATP-consuming processes may be the key for cell survival under conditions of decreased oxygen levels as ATP supply is limited.
Hypoxia Induces Late and Transient Phosphorylation of eEF2 and AMPK.
Stroke itself induces marked reductions in protein synthesis, which might limit the therapeutic benefit of an additional protein synthesis inhibitor. Therefore we tested whether hypoxia resulted in eEF2 and AMPK phosphorylation. Our observations depicted in Fig. 7 show that oxygen deprivation induced phosphorylation of both eEF2 and AMPK (densitometry in Supplemental Fig. 3, A and B). However, these posttranslational modifications were a late and transient event. At that time hypoxia-induced ATP depletion was already maximal (Fig. 6), which indicates that translational arrest by hypoxia is induced as a consequence of energy deprivation and therefore cells are unable to maintain energy homeostasis. Indeed, at the time of hypoxia-induced eEF2 and AMPK phosphorylation, neurotoxicity due to oxygen deprivation was highly pronounced and could not be further prevented (Fig. 5). These results indicate that adaptive preservation of energy is necessary before ATP depletion by hypoxia is complete.
Translational Inhibitors Inhibit Hypoxic Cell Injury and Preserve Intracellular ATP Content.
Since there is evidence that carbimazole may inhibit neuronal apoptosis independently of translational repression (Humar et al., 2009), we asked whether other translational inhibitors similarly reduce hypoxic cell damage. As shown in Fig. 8A, the protein synthesis inhibitor cycloheximide completely abrogated hypoxic injury for 72 hours. In the presence of cycloheximide, the amount of LDH was comparable in supernatants of normoxic and oxygen-deprived cells. Coincubation with anisomycin also prevented neuronal hypoxic damage but displayed biphasic protection (Fig. 8B). In an early phase of hypoxia (≤48 hours), anisomycin (1 μg/ml) was most efficient in preventing hypoxic damage, whereas after 72 hours lower doses of anisomycin (10 ng/ml) were necessary to maintain protection, probably due to a toxic effect of persistent anisomycin treatment. In fact, high-dose anisomycin treatment has been associated with exacerbated neuronal injury (Hong et al., 2007). Accordingly, we observed that LDH leakage in the presence of high anisomycin doses was comparable in normoxic and hypoxic SK-N-SH cells (Fig. 8B). In conclusion, these results demonstrate the therapeutic potential of translational inhibitors in neuroprotection during hypoxia but indicate that their therapeutic dose must be carefully titrated according to the temporal progression of the injury process.
Next we investigated whether neuroprotection by cycloheximide or anisomycin also correlated with a stabilization of ATP content in hypoxic SK-N-SH cells. As shown in Fig. 8, C and D, we observed that hypoxic neuronal SK-N-SH cells displayed a massive loss in availability of intracellular ATP, which corresponds to the experiments shown in Fig. 6. However, cycloheximide completely prevented ATP depletion following oxygen deprivation. In fact, ATP levels in cycloheximide-treated cells were considerably increased when compared with untreated normoxic control extracts. Anisomycin also significantly ameliorated ATP depletion following oxygen deprivation. However, maintenance of cellular energy supply by anisomycin was time- and concentration-dependent and correlated with the biphasic protection against hypoxic injury, as observed in Fig. 8B. High-dose anisomycin treatment was especially effective in the beginning of hypoxic injury; however, a long-lasting ATP supply was only sustained when low therapeutic doses of anisomycin were used, indicating cytotoxic side effects of this compound.
The preservation of intracellular ATP levels in cycloheximide- or anisomycin-treated hypoxic cells generally correlated with protection from hypoxic cell damage. Based on these observations, we postulate that conservation of cellular energy by inhibition of global protein synthesis and its reallocation to essential ATP-consuming processes provides a promising therapeutic approach to rescue potentially vulnerable hypoxic cells when ATP generation is limited.
Translational Control by Carbimazole in Primary Cortical Neurons.
The rate of translation in primary cortical neurons and SK-N-SH cells might significantly vary. Furthermore, considering that neurons are highly differentiated cells, with translation occurring in both the somata and dendrites, protein inhibition in these two compartments may have differential effects in cell death or cell survival that will probably not become visible in SK-N-SH cells. Therefore, primary cortical cell cultures were established to analyze the effects of carbimazole on regulatory posttranslational modifications of eEF2 and AMPK in terminally differentiated neurons. As observed in Fig. 9A, carbimazole induced a robust phosphorylation of eEF2 (densitometry in Supplemental Fig. 3C). Phosphorylation of AMPK followed a similar kinetic, indicating that the posttranslational modifications of eEF2 and AMPK are both regulated by a common pathway (Supplemental Fig. 3, C and D).
Because eEF2 phosphorylation is associated with translational repression, we next analyzed whether carbimazole also inhibits protein synthesis in primary differentiated neurons. Metabolic labeling of cortical neurons with [35S]methionine indicated that global protein synthesis is inhibited by carbimazole (Fig. 9B; Supplemental Fig. 4), which correlated with eEF2 and AMPK phosphorylation (Fig. 9A; Supplemental Fig. 3, C and D).
In SK-N-SH cells, the inactivating modification of eEF2 and translational repression preserved intracellular energy content and ameliorated hypoxic damage (Figs. 5 and 6). To test whether pharmacologic inhibition of protein synthesis also reduces hypoxic injury of differentiated primary neurons, cortical cell cultures were exposed to hypoxia in the presence or absence of carbimazole and cell damage was assessed by an LDH release assay. In analogy to SK-N-SH cells, oxygen deprivation resulted in significant neurotoxicity (Fig. 9C). Cell death was independent of caspase-3, whereas pharmacologic induction of calcium overload by thapsigargin induced procaspase-3 cleavage, PARP cleavage, and eEF2 phosphorylation (Fig. 9D; Supplemental Fig. 3, E–G). This indicates that loss of ion homeostasis, as observed during excitotoxicity, might also induce translational arrest. However, translational arrest induced by hypoxia or calcium overload will be inefficient to salvage neurons because at this time cell injury is already extensive (Fig. 9C) and the progression of programmed cell death following caspase activation is thought to be irreversible.
In analogy to SK-N-SH cells, damage of cortical neurons by oxygen withdrawal was significantly ameliorated by carbimazole (Fig. 9C). To test whether cytoprotection by carbimazole is mediated by the maintenance of energy balance in hypoxic cortical neurons, cell lysates were analyzed for their ATP content (Fig. 9E). Hypoxia resulted in a significant decline of ATP levels in cortical neurons, but the depletion of ATP in the absence of oxygen was significantly diminished by carbimazole. In conclusion, these results confirm that conservation of cellular energy by inhibition of global protein synthesis by carbimazole is neuroprotective in both SK-N-SH cells and terminally differentiated cortical neurons, especially when oxygen supply is limited.
In our study, we have shown that carbimazole diminishes hypoxic neuronal damage by inhibiting global protein synthesis. Cycloheximide and anisomycin also reduced cellular damage of oxygen-deprived neurons, indicating that repressors of translation are potential candidates for treatment of cerebral hypoxia and ischemia.
Protein synthesis is a major consumer of cellular energy (Rolfe and Brown, 1997). Under conditions of limited oxygen availability, repression of translation may result in energy conservation, which can instead be used to maintain cellular mechanisms necessary for survival. We observed that translational repressors prevented or significantly delayed the depletion of ATP stores in oxygen-deprived neurons. In addition, independent studies demonstrate that molecules that maintain mitochondrial respiratory activity during hypoxia or extrinsic application of high-energy metabolites also ameliorate ischemic brain damage (Bouaziz et al., 2002; Ying et al., 2007). Our observations may also explain neuroprotective effects mediated by therapeutic hypothermia. Therapeutic hypothermia has been associated with the attenuation of deleterious cascades, many of them dependent on ATP availability or translation, including the regulation of cerebral metabolism, excitotoxicity, inflammation, the integrity of the blood-brain barrier, and acidosis (Moore et al., 2011).
Translational inhibition by cycloheximide and anisomycin are well characterized, whereas the clinical use of carbimazole is primarily based upon thyroid peroxidase and dehalogenase inhibition. Here we demonstrate that carbimazole additionally interferes with translational elongation by inducing a specific modification of eEF2. Phosphorylation of eEF2 at Thr56 prevents translocation of the nascent protein chain from the A site to the P site of the ribosome (Ryazanov et al., 1988; Carlberg et al., 1990) and offers a plausible explanation for the observed translational repression and conservation of intracellular energy. Furthermore, we could establish the molecular mechanism responsible for carbimazole-mediated eEF2 phosphorylation, which includes AMPK-dependent activation of eEF2K. Interestingly, both kinases are associated with cytoprotective molecular mechanisms, including reduced hypoxic injury, stress resistance, autophagy, inhibition of plaque formation in Alzheimer’s disease, regulation of energy homeostasis, and endothelial nitric oxide synthesis that improves blood flow in ischemic brain tissue (Hardie, 2004; Terai et al., 2005; Greco et al., 2009; Py et al., 2009; Li and McCullough, 2010; Zhang et al., 2010). However, whether carbimazole exerts all of these protective mechanisms remains to be established.
Carbimazole-mediated effects on translation were observed beginning at concentrations 5 times higher than serum levels of thionamides typically found during treatment of hyperthyroidism (Skellern et al., 1974). However, patients obtain long-term antithyroid treatment, which may last for years, and the turnover of thionamides is slow in tissue compared with a short plasma half-life (Jansson et al., 1983). This might result in drug accumulation as demonstrated for related agents (Turcant et al., 1985).
Structural analogs of carbimazole inhibit respiration and uncouple oxidative phosphorylation (Aldridge and Parker, 1960). We did not directly analyze the effects of carbimazole on cellular respiration; however, we observed activation of AMPK, an enzyme that monitors intracellular AMP/ATP ratios, which are naturally increased when generation of energy is impaired. However, ATP levels were not noticeably reduced by carbimazole, which corresponds to other studies, demonstrating that inhibition of respiration by thiourylenes is incomplete and does not appreciably depress ATP levels under basal metabolic conditions (Aldridge and Parker, 1960; Chance et al., 1963).
Anisomycin also efficiently reduced hypoxic cell damage, but whether protection is based on translational arrest or modulation of p38 mitogenMAPK activity (Hong et al., 2007) is unclear. We observed that the p38 inhibitor SB202190 reduced neuronal cell damage during oxygen deprivation or calcium overload (M. Humar, unpublished observation), and we previously reported that carbimazole inhibits p38 activation (Humar et al., 2007). However, protection by SB202190 was less pronounced and restricted to a short time frame, indicating that in addition to inhibition of p38, anisomycin or carbimazole must activate further protective mechanisms, e.g., translational repression and reduction of cellular energy demand.
Stroke itself induces marked reductions in protein synthesis, which could limit the therapeutic benefit of an additional protein synthesis inhibitor. However, our data indicate that translational repression by hypoxia occurs as a late and transient event. Consequently, eEF2 phosphorylation due to hypoxia is inefficient as a protective mechanism, because ATP pools are already irreversibly depleted and the residual intracellular energy supply is not sufficient to salvage oxygen-deprived neurons. Furthermore, we observed that intracellular calcium overload, typical for excitotoxicity, favored eEF2 phosphorylation and apoptosis, confirming that translational arrest during stroke is induced when neuronal damage is already irreversible. However, pharmacologic preconditioning of neurons by translational inhibitors reduces cellular metabolism and ATP consumption before ATP pools are critically depleted and before neurons are severely damaged due to hypoxia. This may limit progressive tissue damage in the penumbra that evolves over hours or even days after stroke. Moreover, it was shown that carbimazole inhibits calcium-dependent caspase-3 activation (Humar et al., 2009).
Our study demonstrated the pharmacologic prevention of hypoxic damage, which involves the induction of AMPK. However, AMPK is also critical for nutrient sensing, and during stroke nutrient restriction contributes to increased AMP/ATP ratios and neuronal tissue damage. Recently it has been described that nutrient restriction may result in translational repression. In addition, this pathway is exploited by tumor cells in adapting to metabolic stress (Leprivier et al., 2013). Induction of AMPK by carbimazole may provide a survival benefit before it is too late to restore energy balance and thus may prevent the progressive expansion of penumbral tissue damage during stroke in close analogy to hypoxia-induced injury.
Hypoxic brain damage depends on several independent pathways. Selective pharmacologic inhibition of a single pathway will most likely be insufficient in preventing the deleterious effects of hypoxic brain damage (White et al., 2000; Stankowski and Gupta, 2011). Clinical trials with brain-injured patients also showed poor clinical efficacy when pharmacologic intervention was applied at a single target (Labiche and Grotta, 2004). Carbimazole and other structurally related compounds exert a wide array of cytoprotective effects. They ameliorate the consequences of calcium overload by preventing the activation of calcium-dependent enzymes and protein phosphatase 2B–dependent neuronal apoptosis (Humar et al., 2009). Furthermore, they inhibit inflammation and induce a cytoprotective heat shock response (Humar et al., 2008; Roesslein et al., 2008). Drugs that downregulate many independent deleterious pathways should be more efficient in limiting secondary neuronal tissue damage than drugs that inhibit only single targets.
Bioenergetic conservation might not be the only cytoprotective mechanism attributed to translational arrest. Hypoxic cell injury is also determined by translation of inflammatory mediators, inducible nitric oxide synthase, cyclooxygenase-2, or matrix metalloproteinases (White et al., 2000; Won et al., 2002; Cunningham et al., 2005). Inhibition of global protein synthesis could prevent their translation, further reducing neuronal damage.
The pathophysiology of several diseases is based on translation and overexpression of damage-inducing proteins, such as in hypertrophy, inflammation, malignant transformation, or structural remodeling. Inhibition of protein synthesis may present an intriguing therapeutic option to alleviate characteristic symptoms of these diseases. Previously, we have demonstrated that translational inhibition results in the inhibition of cell cycle progression (Schwer et al., 2013), which may slow down hyperproliferative diseases like cancer, and carbimazole treatment has been associated with deceleration of tumor progression in some patients (http//www.docstoc.com/docs/73386197). On the other hand, general inhibition of protein synthesis might prevent cellular regeneration and tissue repair. Interestingly, many cytoprotective proteins are preferentially synthesized during stress conditions associated with translational arrest (Yueh and Schneider, 2000; Hernandez et al., 2004). We observed that some carbimazole analogs induce HSP70, which was associated with cytoprotection (Roesslein et al., 2008). Because tissue repair is a late event following organ damage, we suggest first to minimize tissue damage by active pharmaceutical intervention, because the therapeutic restoration of neuronal tissue is unsolved. However, at a later time point, translation needs to be restored to allow tissue repair based on the synthesis of vital proteins. Further studies will be necessary to establish a balanced therapeutic approach to this dynamic process.
Thiopental, a structural analog of carbimazole, also promotes neuroprotection (Nordby and Nesbakken, 1984). Thiopental induces phosphorylation of eEF2, translational repression, and protection from hypoxic cell death, indicating that modulation of this cellular pathway is a characteristic feature of heterocyclic thiourylenes. Oxyderivates of thiourylenes display a significantly reduced ability to perform translational repression, demonstrating that the modulation of intracellular signal transduction significantly depends on the sulfur side chain of the thiourea group (M. Humar, unpublished observations). Derivates of heterocyclic thiourylenes may reveal a pharmacologically relevant scaffold for the development of novel organ-protective compounds.
In summary, the presented findings suggest that carbimazole and other inhibitors of translation protect neurons from hypoxic cell injury by preserving the ATP content during conditions of limited oxygen availability. Carbimazole-mediated translational repression and energy conservation was based on the activation of AMPK, resulting in eEF2 phosphorylation and inactivation. Thus, heterocyclic thiourylenes may represent interesting candidates for the development of new organ-protective compounds. Furthermore, our results suggest the evaluation of other general translational inhibitors, agents that maintain mitochondrial respiratory activity during hypoxia, and the therapeutic application of high-energy molecules as potential strategies to ameliorate hypoxic tissue damage.
Participated in research design: Humar.
Conducted experiments: Lehane, Guelzow, Zenker, Humar.
Performed data analysis: Lehane, Guelzow, Schwer, Erxleben, Humar.
Wrote or contributed to the writing of the manuscript: Lehane, Schwer, Heimrich, Buerkle, Humar.
- N-acetyl-(aspartic acid–glutamic acid–valine–aspartic acid) 7-amino-4-methylcoumarin
- AMP-activated protein kinase
- acetoxymethyl ester
- analysis of variance
- CGS 9343B
- 1,3-dihydro-1-[1-[(4-methyl-4H,6H-pyrrolo[1,2-a][4,1]benzoxazepin-4-yl)methyl]-4-piperidinyl]-2H-benzimidazol-2-one maleate
- eukaryotic elongation factor 2
- eukaryotic elongation factor 2 kinase
- fetal calf serum
- lactate dehydrogenase
- minimum essential medium
- mammalian target of rapamycin
- NH 125
- 1-hexadecyl-2-methyl-3-(phenylmethyl)-1H-imidazolium iodide
- poly(ADP-ribose) polymerase
- cAMP-dependent protein kinase or protein kinase A
- relative light units
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics