Address correspondence and reprint requests to Axel Methner, Department of Neurology, Heinrich-Heine Universität Düsseldorf, Moorenstr. 5, 40225 Düsseldorf, Germany. E-mail: email@example.com
In amyotrophic lateral sclerosis, down-regulation of the astrocyte-specific glutamate excitatory amino acid transporter 2 is hypothesized to increase extracellular glutamate, thereby leading to excitotoxic motor neuron death. The antibiotic ceftriaxone was recently reported to induce excitatory amino acid transporter 2 and to prolong the survival of mutant superoxide dismutase 1 transgenic mice. Here we show that ceftriaxone also protects fibroblasts and the hippocampal cell line HT22, which are not sensitive to excitotoxicity, against oxidative glutamate toxicity, where extracellular glutamate blocks cystine import via the glutamate/cystine-antiporter system xc−. Lack of intracellular cystine leads to glutathione depletion and cell death because of oxidative stress. Ceftriaxone increased system xc− and glutathione levels independently of its effect on excitatory amino acid transporters by induction of the transcription factor Nrf2 (nuclear factor erythroid 2-related factor 2), a known inducer of system xc−, and the specific xc− subunit xCT. No significant effect was apparent in fibroblasts deficient in Nrf2 or xCT. Similar ceftriaxone-stimulated changes in Nrf2, system xc−, and glutathione were observed in rat cortical and spinal astrocytes. In addition, ceftriaxone induced xCT mRNA expression in stem cell-derived human motor neurons. We conclude that ceftriaxone-mediated neuroprotection might relate more strongly to activation of the antioxidant defense system including Nrf2 and system xc− than to excitatory amino acid transporter induction.
Excitatory amino acid transporters (EAATs) mediate the re-uptake of glutamate released by synaptic activity in the central nervous system (Sheldon and Robinson 2007). This attenuates excitotoxicity, a process that causes neuronal death because of over-stimulation of ionotropic glutamate receptors (Choi 1992). In amyotrophic lateral sclerosis (ALS), loss of motor neurons is associated with a reduction in the astrocytic EAAT2 in the spinal cord (Rothstein et al. 1995). Down-regulation of EAAT2 has also been described in other neurological diseases including Alzheimer’s disease, epilepsy, HIV-dementia, cerebral ischemia, hepatic encephalopathy, multiple sclerosis, and Huntington’s disease and can result in excitotoxic neuronal death (Sheldon and Robinson 2007). Thus, substances that increase EAAT2 expression might be of broad therapeutic value. Ceftriaxone (Cef), a recently identified EAAT inducer (Rothstein et al. 2005), is protective in experimental models of ALS, ischemic stroke, HIV-dementia, Huntington’s disease, and multiple sclerosis (Rothstein et al. 2005; Chu et al. 2007; Lipski et al. 2007; Ouyang et al. 2007; Rumbaugh et al. 2007; Melzer et al. 2008; Miller et al. 2008) and is currently being tested in clinical trials of ALS and psychiatric disorders (see http://www.clinicaltrials.gov).
In addition to excitotoxicity, oxidative stress plays a role in all of these diseases and neuronal cell death because of oxidative stress can be studied in vitro using oxidative glutamate toxicity, a form of glutamate-induced cell death distinct from excitotoxicity (Tan et al. 2001). In this model, increased extracellular glutamate depletes cells of cystine by blocking the cystine/glutamate antiporter system xc−, which imports cystine in exchange with intracellular glutamate (Bannai 1986). System xc− consists of two subunits, xCT and a non-specific 4F2 heavy chain (Sato et al. 1999). Cysteine, the reduced form of cystine, is required for the synthesis of GSH, the most important antioxidant in the brain (Dringen 2000; Schulz et al. 2000). System xc− inhibition results in GSH depletion, oxidative stress, and ultimately cell death (Tan et al. 2001). Oxidative glutamate toxicity has been described in neuronal cell lines (Miyamoto et al. 1989; Murphy et al. 1989; Tan et al. 2001), and immature primary neurons (Murphy et al. 1990; Davis and Maher 1994; Ratan et al. 1994). Mice lacking system xc− activity show redox imbalance (Sato et al. 2005) and brain atrophy (Shih et al. 2006), which strongly suggests that system xc− is important for antioxidant defense in the brain.
We recently reported that EAATs might support system xc− activity, and thereby GSH synthesis, by both increasing intracellular and decreasing extracellular glutamate and thus protecting against oxidative glutamate toxicity (Lewerenz et al. 2006). Therefore, we investigated whether Cef induces EAATs in HT22 cells, which lack ionotropic glutamate receptors (Maher and Davis 1996), thereby modulating system xc− activity and protecting against oxidative glutamate toxicity. Here, we report that Cef acts as an inducer of system xc− in glial and neuronal cells.
Tissue culture dishes were from Greiner BIO-ONE and NUNC; fetal calf serum (FCS) was obtained from Hyclone, l-glutamine, penicillin/streptomycin, 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H2DCFDA), high-glucose Dulbecco’s modified Eagle medium (DMEM), Alpha-MEM, trypsin/EDTA, and TriZol reagent for RNA purification and fluorescent secondary antibodies were from Invitrogen, Carlsbad, CA, USA; acivicin, bovine serum albumin, l-cystine, l-glutamate, HEPES, glutathione reductase, sulfasalicylic acid, 5,5-dithiobis(2-nitrobenzoic acid), homocysteic acid, β-mercaptoethanol (β-ME), reduced NADPH, and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were from Sigma, St Louis, MO, USA. DL-threo-β-benzyloxyaspartic acid (TBOA) was from Tocris Cookson, Ballwin, MO, USA. Ceftriaxone-Na+ was from Ratiopharm and Sigma-Aldrich. l-[35S]-Cystine (specific activity 40–250 mCi/mmol) and l-[3H]-glutamate were obtained from Perkin Elmer NEN, Waltham, MA, USA. Oligonucleotides were synthesized by MWG Biotech AG. All other chemicals were obtained from Merck Eurolabs, Bruchsal, Germany. MTT was reconstituted in phosphate buffered saline (PBS), TBOA in dimethyl sulfoxide. For viability assays, TBOA was dissolved in DMEM and the pH re-adjusted.
HT22 cell culture and viability assays
HT22 cell culture and viability assays were essentially performed as described (Lewerenz et al. 2003, 2006). Experimental agents were added as indicated. TBOA was added immediately before the addition of glutamate. For long term Cef treatment, 1 × 106 HT22 cells were plated in 60 mm dishes with or without the indicated concentrations of Cef. After 2 days, the cells were re-plated once at the same density and Cef concentration. After another 2 days, the cells were seeded at a density of 5 × 105 cells and grown for another 3 days at the same Cef concentration. Frequent counting proved to be essential as HT22 cells grew slightly faster in Cef (data not shown).
Mouse embryonic fibroblast culture and viability assays
Mouse embryonic fibroblasts (MEFs) derived from xCT knock-out mice (xCT KO) and nuclear factor erythroid 2-related factor 2 knock-out mice (Nrf2 KO), and respective control wild-type MEFs were a kind gift of Dr. Hideyo Sato, Yamagata University, Japan (Sasaki et al. 2002; Sato et al. 2005). Cells were cultivated at 37°C in a 10% CO2 atmosphere and DMEM containing 10% FCS, 2 mM glutamine, 10 mM HEPES, pH 7.4, 100 IU/mL penicillin, and 100 μg/mL streptomycin. Medium for xCT KO MEFs was supplement with 50 μM β-ME. For oxidative glutamate toxicity, MEFs were plated at a density of 105 cells/well in six-well plates. Nrf2 KO MEFs and wt-controls were grown with or without 300 μM Cef for 7 days. xCT KO-MEFs and wt-controls were grown to confluence in the presence of 50 μM β-ME for 5 days, washed twice with PBS and then treated for 7 days with or without 300 μM Cef without β-ME. Then, cell were harvested by trypsination and plated at a density of 2.5 × 105 cells/well into 96-well plates. After 24 h, indicated concentrations of glutamate were added and cell survival was monitored by the MTT assay after another 24 h.
Rat primary astrocytes and neuronal astrocytic co-cultures
Rat cortical astrocytes were prepared from E18 cortices as previously described (Krobert et al. 1997). Spinal astrocytes were prepared similarly. These cultures are essentially free of microglia and do not contain neurons or oligodendrocytes after the first passage. Astrocytes were grown to confluence in Alpha-MEM and 10% FCS for 2 weeks and then split 1 : 5 into 10 cm, 35 mm and 24 well dishes. For co-culture experiments, 1 × 106 and 4 × 106 freshly dissociated cortical neurons were seeded on top of astrocytes plated 24 h earlier in 24 well-plates and 35 mm-dishes, respectively. Astrocytes and co-cultures were grown for 1 week and then treated with the indicated concentrations of Cef for another week.
Human embryonic stem cell-derived motor neurons expressing SOD-1 and GFP
The human embryonic stem cell (hESC) line HSF-1 was obtained from WiCell (XY, 46, NIH No.UC01). Cells were transfected with the lentiviral E/Hb9 (3.6 kb) : P/hsp68 : EGFP vector that leads to motorneuron-specific expression of enhanced green fluorescent protein (EGFP) or E/Hb9 (3.6 kb) : P/hsp68 : EGFP : SOD-1 vectors that, in addition to EGFP, separately express either wild-type or mutant superoxide dismutase 1 (SOD-1) under the motorneuron specific Hb9 promoter. Cells were then differentiated using retinoic acid, sonic hedgehog and the growth factors brain-derived neurotrophic factor, glial cell line-derived neurotrophicfactor and ciliary neurotrophic factor as described (Karumbayaram et al. 2009). Thirty-five days after induction of differentiation by growths factors, transfected hESC cultures in 6-well dishes were treated with either 30 or 300 μM Cef for 7 days.
Uptake of radiolabeled amino acids
For uptake experiments, HT22 cells passaged with or without Cef for 7 days and seeded at a density of 3 × 104 cells per well in 24-well plates and grown for 1 day in the presence or absence of Cef or astrocytic cultures grown in 24-well plates for 1 week with or without Cef were used. System xc− activity was performed using l-[35S]-cystine or in case of spinal astrocytes using l-[3H]-glutamate as described previously (Lewerenz and Maher 2009).
For measurement of EAAT activity, instead of sodium-free Hank’s buffered salt solution, sodium containing, chloride-free Hank’s buffered salt solution (choline chloride substituted by sodium gluconate, CaCl2 by calcium gluconate) was used. Uptake was performed for 20 min at 37°C with l-[3H]-glutamate diluted 1000-fold with cold l-glutamate to a final concentration of 10 μM. In addition, 100 μM TBOA or dimethyl sulfoxide were added just prior to the initiation of uptake. Samples were processed similarly as for measurement of system xc− activity. EAAT activity was defined as TBOA-sensitive l-[3H]-glutamate uptake and determined by subtracting uptake in the absence of TBOA by uptake in parallel wells in the presence of TBOA.
Quantification of total GSH
About 3 × 105 HT22 cells passaged for 1 week with or without Cef were seeded in 60 mm-dishes with the same Cef concentration. After 24 h, the medium was exchanged to Cef-free medium with the indicated concentrations of glutamate for 6 h. MEFs were seeded in six-well plates as described for oxidative glutamate toxicity and astrocytes in 35 mm dishes and grown with or without Cef for 7 days. Cells were harvested, processed and total GSH was measured as described previously (Maher and Hanneken 2005). For measurement of GSH release, astrocytes were washed twice with PBS and then incubated with 750 μL cystine-free DMEM containing 10% FCS and 200 μM of the γ-glutamyl transpeptidase inhibitor acivicin to prevent the generation of mixed GSH disulfides with cystine or enzymatic breakdown of GSH by γ-glutamyl transpeptidase. After 4 h, medium was harvested, spun at 2000 × g for 1 min and 200 μL were processed similarly to cell extracts and similarly normalized to cellular protein measured by the bicinchoninic acid-based method (Micro BCA Protein Assay; Pierce, Rockford, IL, USA).
Differentiated and transfected hESC cultures were fixed in 4% paraformaldehyde for 20 min at 20°C. Fixed cells were permeabilized with 0.2% Triton in PBS and blocked with the appropriate blocking agents. Cells were incubated with primary C4F6 anti-G93A antibody (Urushitani et al. 2007) (gift from Dr. J-P Julien, Université Laval, Quebec, CA) overnight at 4°C and with secondary antibody for 1–2 h at 20°C. Cells were viewed under a confocal microscope (Zeiss, Thronwood, NY, USA).
Fluorescence-activated cell sorting
Differentiated and transfected hESC cultures were collected from the dishes by TrypLETM (Invitrogen) treatment and were washed twice with DMEM/F12 containing B27. The cells were collected in 5 mL tubes fitted with cell strainer caps (BD falcon-352235) and EGFP-expressing cell were isolated by fluorescence-activated cell sorting (FACS). Typically, cultures contained about 1% of EGFP-positive cells and approximately 10 000 cells were obtained per dish. The FACS-purified motor neurons were centrifuged at 1000 g for 5 min and the pellet was saved in RNAlater solution (Ambion, Austin, TX, USA) as prescribed.
Quantitative real-time PCR
Total RNA was extracted using the RNeasyTM kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions and reverse transcribed with oligo(dT)-primers using the SUPERSRIPTTM first strand cDNA synthesis system (Invitrogen). Real-time PCR was performed using Taqman technology with Fam/Dark-quencher probes from the universal ProbelibraryTM (Roche Molecular Biochemicals, Indianapolis, IN, USA) or individually designed Fam/Tamra probes (MWG) on an ABI Prism 7500 Sequence Detection System using the thermocycling conditions proposed by the manufacturer. Real-time PCR reactions were carried out at a final concentration of 300 nM for forward and reverse primers, and 250 nM for the Taqman probes (for sequences see Appendix S2). Each reaction contained 2 μL containing about 60 ng of cDNA, 12.5 μL of the TaqMan universal PCR Master Mix (Abi) in an individual reaction volume of 20 μL. Glyceraldehydes 3-phosphate dehydrogenase (GAPDH) and hypoxanthine-guanine phosphoribosyltransferase (HPRT) served as endogenous control-genes and showed no differential expression after incubation with Cef. Relative differences in abundance were calculated by the ΔΔCT comparator method.
Cell fractionation, sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblotting
HT22 cells grown for 1 week with or without 300 μM ceftriaxone were plated at a density of 8.8 × 105 cells per 10 cm dish and prior growth conditions were continued for 24 h. Membrane, cytosolic and nuclear fractions of HT22 cells and pure astrocytes in 10 cm dishes were obtained by the differential detergent fractionation (Appendix S1). Astrocyte-neuron co-cultures in 35-mm dishes were directly lysed in 200 μL sodium dodecyl sulfate-sample buffer and boiled for 5 min. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis, western blotting and quantification was performed as previously describend (Lewerenz and Maher 2009). The primary antibodies used were: anti-Nrf2 (#SC13032; 1/1000) from Santa Cruz Biotechnology, Santa Cruz, CA, USA, anti-xCT (1/1000, a generous gift from Sylvia Smith, Medical College of Georgia), anti-EAAT2 (1/10 000, a generous gift from J. D. Rothstein, Johns Hopkins University), anti-β-actin (1/200 000, A5441) (Sigma).
Reactive oxygen species measurement
HT22 cells were seeded onto 96-well black walled microtiter plates at a density of 5 × 103 cells per well. The next day, the cells were treated with different concentrations of ceftriaxone for 15 min, 1 h, 4 h or 6 h prior to replacing the medium with 100 μL loading medium (phenol red-free DMEM containing 25 mM HEPES, 2% dialysed FCS and 10 μM CM-H2DCFDA). As a positive control, 250 μM tertiary-butylhydroperoxide (tBOOH) was added to the cells for 15 min prior to the addition of CM-H2DCFDA to some wells. After 30 min in CM-H2DCFDA, the medium was replaced with fresh loading medium without CM-H2DCFDA and the fluorescence (λ excitation = 495 nm, λ emission = 525 nM) was determined using a Gemini fluorescent plate reader. Each treatment was done in quadruplicate. DCFDA fluorescence was normalized to control cells not exposed to ceftriaxone or tBOOH.
Determination of the Trolox equivalent activity concentration
Trolox equivalent activity concentration (TEAC) values were determined as described by (Re et al. 1999) but modified for a plate reader (Maher 2006).
Long-term ceftriaxone treatment protects HT22 cells against oxidative glutamate toxicity and induces EAAT expression and activity
In HT22 cells, 7 days of pre-treatment with Cef concentration-dependently protected against oxidative glutamate toxicity. Whereas all untreated cells succumbed to 10 mM glutamate, 7% of cells treated with 100 μM and 19% of cells treated with 300 μM Cef survived (Fig. 1a). Within the first 4 days, Cef-mediated protection was only minimal but continued to increase over time (Fig. 1b). As reported for glutamate-resistant HT22 cells (Lewerenz et al. 2006), wild type HT22 cells express EAAT1–3 but not EAAT4 or EAAT5 mRNA. EAAT1 was about 1000-fold more abundant than EAAT3 and EAAT2 about 32-fold more abundant than EAAT3 assuming equal PCR efficiency (Fig. 1c left panel). Cef increased mRNA expression of EAAT1, EAAT2 and EAAT3 2.3-, 1.5-, and 1.9-fold, respectively (Fig. 1c right panel). In line with these results, TBOA-sensitive EAAT activity was increased by 37% upon 300 μM Cef treatment for 7 days (Fig. 1d).
Ceftriaxone-induced protection in HT22 cells is associated with increased expression of xCT, increased activity of system xc−, and induction of Nrf2
HT22 cells do not express ionotropic glutamate receptors (Maher and Davis 1996). Therefore, the protective effect of Cef cannot be explained by a reduction in excitotoxicity. Instead, we hypothesized that increased glutamate uptake would lead to an increase in cystine import through system xc− and a subsequent increase in GSH by both increasing intracellular glutamate, the driving force for system xc−-mediated cystine uptake, and decreasing inhibitory extracellular glutamate concentrations as recently described (Lewerenz et al. 2006). System xc− exchanges intracellular glutamate with extracellular cystine, the major building block of the important antioxidant GSH (Bannai 1986). Indeed, in the absence of glutamate, GSH in Cef-treated HT22 cells was ∼42% higher compared to control cells and the absolute difference persisted following a 6 h-treatment with increasing concentrations of glutamate (Fig. 2a).
If the protection was caused exclusively by an EAAT-driven increase in system xc−-mediated cystine uptake, inhibition of glutamate uptake with the specific non-transportable EAAT inhibitor, TBOA, should abolish Cef-mediated protection. However, 1 mM TBOA only attenuated but did not abolish cell protection by Cef (Fig. 2b), suggesting additional, EAAT-independent effects of Cef. Because of its important role in oxidative glutamate toxicity and GSH synthesis, we quantified xCT mRNA expression by real-time PCR and measured system xc− function directly as sodium-insensitive, glutamate-sensitive 35S-cystine uptake. Both methods showed a prominent induction of xCT mRNA expression (three-fold) and system xc− activity (41%) in response to Cef treatment. As the sodium-free conditions used to measure system xc− activity inactivate EAATs (Fig 2c), the increase in system xc− activity upon Cef treatment can only be explained by up-regulation of system xc− itself. Unfortunately, xCT protein, the specific subunit of system xc−, is expressed at levels too low to be detected by western blotting in HT22 cells (not shown). We conclude that Cef not only induces EAATs but also xCT and system xc− in HT22 cells.
xCT belongs to a family of phase II response genes involved in defense against oxidative stress (Sasaki et al. 2002) that contain an antioxidant-response element (ARE) in their promoter regions (Ishii et al. 2000). The nuclear factor erythroid 2-related factor 2 (Nrf2) binds to the ARE and activates transcription (Nguyen et al. 2003). Thus, we hypothesized that ceftriaxone increases nuclear Nrf2 protein levels. Western blot analysis of nuclear HT22 cell extracts showed that Cef significantly increased Nrf2 levels 1.8-fold compared to control cells (Fig 2d).
Ceftriaxone-mediated increased glutathione content is independent of the induction of EAAT activity, but correlates with an increase in system xc− activity
To verify the suspected EAAT-independent effects of Cef on system xc− in primary cultures, we exploited the fact that astrocytes only express EAAT2 when co-cultured with neurons (Swanson et al. 1997). We first compared the effect of Cef on EAAT protein induction and function in rat cortical astrocytes co-cultured with primary cortical neurons with pure cortical astrocytes. As expected only astrocytes co-cultured with neurons exhibited the typical stellate phenotype and increased EAAT2 expression and activity measured as TBOA-sensitive 3H-glutamate uptake in response to 300 μM Cef (Fig. 3a). Cef had no effect on EAAT function in pure cortical astrocytes (Fig. 3b), although EAAT function could be induced by the known EAAT2-inducer dibutyryl cAMP (Fig. S1) as previously reported (Swanson et al. 1997). Moreover, EAAT activity in there cultures was not modified by the absence or presence of antibiotics in the culture medium (Fig. S1).
We then compared the effect of Cef on GSH content and system xc− activity measured as glutamate-sensitive 35S-cystine uptake in these cultures. In both culture systems, Cef increased GSH content and system xc− activity in a similar pattern. In mixed cultures, GSH content increased from ∼80% at 30 μM to ∼200% at 300 μM, which was paralleled by an increase in system xc− activity from 16% at 30 μM to 84% at 300 μM Cef. In pure astrocytic cultures GSH content rose by 27% at 30 μM and 86% at 300 μM Cef, whereas system xc− activity increased from 24% at 30 μM to 73% at 300 μM as compared to cultures not treated with Cef (Fig. 3c). In absolute values, the xc− but not the EAAT transport activity was much enhanced in co-cultures as compared to pure astrocytic cultures (Fig. 3d). However, as these experiments were not conducted in parallel, the results should be interpreted with caution. To prove that the increase in GSH is because of increased synthesis and not decreased export of GSH, we measured GSH export in pure astrocyte cultures treated with Cef for 7 days. Treatment with 30 μM and 300 μM Cef actually increased GSH export by 32% and 94%, respectively (Fig. 3e). We also noticed that the effect on EAAT activity in astrocytes co-cultured with neurons was already saturated with 30 μM Cef consistent with the reported EC50 of 3.5 μM for EAAT2 induction in organotypic spinal cord slice cultures (Rothstein et al. 2005), whereas the effect on GSH content and system xc− activity could be further increased by higher concentrations of Cef. The increase in GSH content at higher Cef concentrations therefore more closely resembles the pattern of system xc− than EAAT activity induction. We conclude that in astrocytes the effect of Cef on GSH levels and system xc− activity parallel each other and can occur in the absence of a functionally significant induction of EAAT function.
Ceftriaxone also increases system xc− activity and intracellular and released GSH in spinal cord astrocytes
In human ALS and the animal model of ALS, mutant SOD-1 transgenic mice, motor neuron loss is most prominent in the spinal cord (Gurney et al. 1994; Bruijn et al. 2004). Non-neuronal cells that surround spinal cord motor neurons modulate neurodegeneration in SOD-1 transgenic mice (Clement et al. 2003). Pathways regulating astrocytic EAAT expression depend on the region of the central nervous system from which they are derived (Schluter et al. 2002). Therefore, we investigated the effect of Cef on system xc− activity in primary spinal astrocyte monocultures. Again, Cef treatment for 7 days did not increase EAAT function, but did induce a significant increase in system xc− activity, intracellular GSH, and GSH release even more robustly than in cortical cultures (Fig 3f).
Ceftriaxone induces xCT and Nrf2 protein and this induction mediates its activity in mouse embryonic fibroblasts
We next measured Nrf2 and xCT protein expression by immunoblotting in our two culture systems: cortical astrocytes differentiated by co-culture with neurons and pure cortical astrocytes. In both culture systems, 300 μM Cef prominently increased the expression of xCT and Nrf2 matching the results obtained functionally with GSH content and system xc− activity. Cef increased xCT protein 9.3-fold and Nrf2 6.3-fold in mixed cultures and 2.9-fold (xCT) and 5.0-fold (Nrf2) in pure astrocytic cultures (Fig. 4a). To prove that these two proteins indeed participate in the protective function of Cef, we treated Nrf2 KO and control wild-type (WT) MEFs with or without 300 μM Cef for 7 days before monitoring GSH content and their sensitivity against oxidative glutamate toxicity. Cef significantly increased GSH content only in WT, but not in Nrf2 or xCT KO MEFs (Fig. 4b). Also, Cef significantly protected WT MEFs, but did not change the sensitivity of Nrf2 KO MEFs against oxidative glutamate toxicity (Fig. 4c). We therefore conclude that Cef protection against oxidative glutamate toxicity is probably caused by an up-regulated antioxidant defense mediated by Nrf2 and subsequent xCT expression, which then increases cystine import and glutathione synthesis and release.
Ceftriaxone increases xCT mRNA in human stem cell-derived motor neurons and the induction is modified by mutant G93A SOD-1
To study the effect of Cef on motor neurons, we used human motor neurons derived from human embryonic stem cells (hESCs), which can be identified by expression of EGFP driven by the motor neuron-specific enhancer Hb9 as described previously (Karumbayaram et al. 2009) (Fig. 5a). We treated differentiated hESCs for 7 days with 30 or 300 μM Cef, purified EGFP-expressing cells by FACS, and quantified xCT mRNA expression by real-time RT-PCR. Treatment with Cef led to a concentration-dependent increase in xCT mRNA expression in hESC-derived motor neurons (Fig. 5b). We then used cells over-expressing mutant G93A SOD-1, which causes familial ALS or wild-type human SOD-1 as a control (Karumbayaram et al. 2009). The expression of G93A SOD-1 was confirmed using a G93A mutant-specific antibody (Urushitani et al. 2007) (Fig. 5c). Basal xCT expression did not differ significantly between neurons expressing mutant G93A or wild type SOD-1 (Fig. 5d). As in hESC-derived motor neurons expressing EGFP only, 7 days of Cef significantly induced xCT mRNA in hESC-derived motor neurons expressing mutant G93A or wild-type SOD-1. Nevertheless, the magnitude of induction was significantly lower in mutant G93A SOD-1 than in wild-type SOD-1 expressing neurons (Fig. 5e).
Ceftriaxone is a slow inducer of Nrf2 devoid of any pro- or antioxidant activity
Based on the above data, we hypothesized that Cef is a novel, slow, and therefore atypical inducer of Nrf2. Whereas a 4 h pre-treatment with 25 μM tert-butyl hydroquinone (tBHQ), a classical Nrf2 inducer, protected HT22 cells against oxidative glutamate toxicity, no effect was seen for a similar treatment with 300 μM Cef (Fig. 6a). Similarly, while tBHQ strongly induced nuclear Nrf2 protein after a 4 h treatment, Cef-induced induction of Nrf2 was not significantly different from control at this time point, but became apparent with a 5-fold induction after 24 h (Fig. 6b). The development of a robust protection by Cef took even longer (Fig 1b).
Many inducers of Nrf2 have either oxidant or antioxidant properties (Kensler et al. 2007). Reactive oxygen species (ROS) production in living neuronal cells can be measured by intracellular oxidation of the non-fluorescent dye H2DCFDA to the fluorescent compound DCFDA (Oyama et al. 1994). If Cef was a pro- or an antioxidant, then a change in reactive oxygen species detection measured by DCFDA fluorescence should be present after exposure of cells to Cef. After treatment of HT22 cells with 30 or 300 μM Cef, DCFDA fluorescence was slightly and transiently diminished after 15 and 60 min but not different from control values after longer exposure. The pro-oxidant tBOOH served as a positive control (Fig. 6c). Thus, we tested whether Cef is a direct antioxidant using the TEAC assay. In contrast to rapid and strong inducers of Nrf2 such as fisetin (Hanneken et al. 2006) and tBHQ (Wang and Jaiswal 2006), which show significant antioxidant activity, Cef showed no direct antioxidant activity in the TEAC assay (Fig. 6d).
Ceftriaxone has been reported to increase EAAT2 expression and induce neuroprotection in animal models of diverse neurological diseases (Rothstein et al. 2005; Chu et al. 2007; Ouyang et al. 2007; Miller et al. 2008). It was generally assumed that the up-regulation of EAAT2 is responsible for Cef-mediated neuroprotection by virtue of its ability to reduce extracellular glutamate levels and subsequent excitotoxicity (Sheldon and Robinson 2007). Here, we demonstrate that Cef robustly induces the expression of the glutamate/cystine exchanger, system xc−, in four different neuronal or glial cell culture models: hippocampal HT22 cells, cortical astrocytes differentiated by co-culture with cortical neurons, cortical and spinal cord astrocyte monocultures, as well as increased mRNA expression of the specific system xc− subunit, xCT, in human stem cell-derived motor neurons. Induction of system xc− activity by Cef strongly correlated with increased intracellular GSH levels, the most important endogenous small molecule antioxidant in the brain (Dringen 2000; Schulz et al. 2000). In addition, Cef increased GSH release from cortical and spinal astrocytes, supporting the assumption that the increase in GSH is because of increased synthesis and not decreased export. Experiments with cortical and spinal astrocytic monocultures showed that the effect of Cef on GSH metabolism is independent of EAAT induction since Cef-induced increases in GSH were present in pure astrocytic cultures without increase in functional EAAT expression. In vivo astrocytes contain ∼4-fold higher GSH levels compared to neurons (Sun et al. 2006) and astrocytic GSH release and subsequent extracellular degradation is thought to be an important source of cysteine for neuronal GSH synthesis (Dringen and Hirrlinger 2003). Thus, induction of xCT in vivo might have a dual protective activity on motor neurons by, first, increasing their ability to directly import cystine and, second, by providing cysteine as a break-down product of astrocyte-derived GSH for GSH synthesis.
xCT is transcriptionally induced by the nuclear factor Nrf2, which binds to AREs in the promoter region of the xCT gene (Sasaki et al. 2002). Our observation that increased xCT expression and system xc− activity was paralleled by increased Nrf2 protein levels in response to Cef treatment strongly suggests that Nrf2 induces xCT. This hypothesis is substantiated by the observations that Cef did not protect MEFs deficient in Nrf2 against oxidative glutamate toxicity, and that Nrf2 and xCT deficiency strongly and independently diminished the effect of Cef on GSH levels. Besides xCT, Nrf2 also induces enzymes involved in GSH biosynthesis (glutamate cysteine ligase and glutathione synthase), use (glutathione S-transferase and glutathione reductase), and export (multidrug resistance protein 1) (Shih et al. 2003). Nevertheless, the diminished Cef-mediated GSH increase in xCT KO MEFs indicates the critical role of xCT induction.
Whereas the classical Nrf2 inducer tBHQ protects against oxidative glutamate toxicity within 4 h, the Cef-induced protection of HT22 cells took days to develop. In addition, we show that Cef neither acts as a direct antioxidant nor induces oxidative stress. Thus, Cef represents a novel, slow, and rather atypical inducer of Nrf2. It remains to be shown whether Cef induces Nrf2 in a phosphatidylinositol 3-kinase-dependent pathway as recently described for dietary flavonoids (Bahia et al. 2008).
The EAAT-inhibitor TBOA reduced Cef-mediated protection in the HT22 cells, where we observed simultaneous induction of EAAT activity and Nrf2 by Cef. EAATs were suggested to support system xc− function by supplying intracellular glutamate to drive cystine import and to reduce competitive inhibition by extracellular glutamate (Rimaniol et al. 2001; Lewerenz et al. 2006). We think that Cef-mediated EAAT induction in HT22 cells is responsible for part of the observed protective effect via this mechanism. Moreover, system xc− activity inevitably leads to glutamate release (Bannai 1986) and, in a cell culture model of ischemic stroke, a combination of post-hypoxic induction of system xc− and concomitant reduction of EAAT activity has been found to exacerbate glutamate-mediated neuronal cell death (Fogal et al. 2007). Thus, in diseases with both oxidative stress and a deficiency in EAAT function, compounds like Cef that simultaneously induce system xc− and EAAT activity might provide a substantial neuroprotective benefit. Our observation that the Cef-mediated xCT induction is preserved in the presence of mutant SOD-1 in human stem cell-derived motor neurons points to the potential therapeutic benefits of ceftriaxone in ALS. Of note, mutant G93A slightly attenuated the effect of Cef on xCT expression compared to wild type SOD-1. As we did not control for equal expression of both proteins, the results should be interpreted with care. Nevertheless, our results are in line with observations that suggest an impaired Nrf2-mediated gene transcription in the presence of mutant SOD-1 in motor neuron-like cells and in familial ALS carrying SOD-1 mutations as described previously (Kirby et al. 2005).
In summary, we demonstrate that the neuroprotective β-lactam antibiotic Cef not only increases the expression of EAATs but, additionally, is an inducer of Nrf2, system xc− and GSH in vitro. Thus, Cef has a dual neuroprotective effect with the induction of EAATs theoretically supporting the action of system xc− and neutralizing the increase in glutamate release resulting from enhanced system xc− activity, which is a possible negative side effect. Further studies are needed to reveal whether both actions play a role in the observed Cef-mediated neuroprotection in vivo.
This work was funded by the American ALS association (ALSA starter grant 1266 to Axel Methner and Dave Schubert) and Forschungskommission der Heinrich-Heine Universität Düsseldorf grant 9772353 to Philipp Albrecht.