Protective Effect of the Energy Precursor Creatine Against Toxicity of Glutamate and ␤-Amyloid in Rat Hippocampal Neurons

Gregory J. Brewer and *Theo W. Wallimann

Departments of Medical Microbiology/Immunology and Neurology, Southern Illinois University School of Medicine, Springﬁeld, Illinois, U.S.A.; and *Institute for Cell Biology, Swiss Federal Institute of Technology, Zurich, Switzerland

Abstract: The loss of ATP, which is needed for ionic (Manos et al., 1991; Molloy et al., 1992) and in vivo homeostasis, is an early event in the neurotoxicity of (Hemmer et al., 1994; Kaldis et al., 1996), CK isozyme glutamate and ␤-amyloid (A␤). We hypothesize that cells BB catalyzes the reversible conversion of PCr and ADP supplemented with the precursor creatine make more to ATP and Cr to manage different aspects of high- phosphocreatine (PCr) and create larger energy reserves energy demands in the brain (Hemmer and Wallimann, with consequent neuroprotection against stressors. In serum-free cultures, glutamate at 0.5–1 mM was toxic to 1993; Chen et al., 1995). On the other hand, mitochon- embryonic hippocampal neurons. Creatine at Ͼ0.1 mM drial CK catalyzes the reversible conversion of Cr and greatly reduced glutamate toxicity. Creatine (1 mM) could ATP to ADP and PCr to generate energy reserves in the be added as late as 2 h after glutamate to achieve pro- cytoplasm (Wyss et al., 1992). PCr reserves are depleted tection at 24 h. In association with neurotoxic protection even more rapidly than ATP during ischemia (Passon- by creatine during the first 4 h, PCr levels remained neau and Lowry, 1971) owing to their conversion to ATP constant, and PCr/ATP ratios increased. Morphologi- by CK. Although depletion of ATP is an early event in cally, creatine protected against glutamate-induced den- the neurotoxicity of glutamate (Budd et al., 1997) and dritic pruning. Toxicity in embryonic neurons exposed to ␤-amyloid (A␤) (Mark et al., 1997), PCr levels have not A␤ (25–35) for 48 h was partially prevented by creatine as been reported. Thus, an understanding of PCr levels and well. During the first6hoftreatment with A␤ plus creatine, the molar ratio of PCr/ATP in neurons increased their maintenance in the presence of the precursor, Cr, from 15 to 60. Neurons from adult rats were also partially are critical to understanding cellular energetics during protected from a 24-h exposure to A␤ (25–35) by creat- normal function and stress. ine, but protection was reduced in neurons from old Given ample PCr, the CK reaction generates ATP at a animals. These results suggest that fortified energy re- rate 10 times faster than oxidative phosphorylation and serves are able to protect neurons against important 40 times faster than glycolysis (Wallimann et al., 1992). cytotoxic agents. The oral availability of creatine may Because brain ATP levels are typically 3–4 mM and PCr benefit patients with neurodegenerative diseases. levels are ϳ5–6 mM (Erecin´ska and Silver, 1989), we Key Words: Creatine—Phosphocreatine—ATP—Ener- measured the concentrations of these critical energy mol- gy—Glutamate toxicity—␤-Amyloid. ecules, rather than ADP levels, which are maintained J. Neurochem. 74, 1968–1978 (2000). around 0.02 mM by adenylate kinase. Although the rates of many enzymes are regulated by ATP/ADP levels or the energy charge, (ATP ϩ 0.5 ADP)/(ATP ϩ ADP ϩ AMP) (Ball and Atkinson, 1975; Bishop and Atkin- The phosphocreatine (PCr)/creatine (Cr) kinase (CK) son, 1984), the immediate energy available to a mam- system is thought to be physiologically important in malian cell may be better represented by the total high- tissues with high and fluctuating energy requirements, energy pool, composed primarily of ATP and PCr. To like muscle and brain (Wallimann and Hemmer, 1994). quantify the extent of decline or the effects of Cr sup- In these cells, high-energy PCr serves not only as an immediate temporal energy buffer, but also as an energy shuttle from subcellular sites of energy production (mi- Received November 23, 1999; revised manuscript received January tochondria and/or glycolysis) to sites of energy con- 4, 2000; accepted January 5, 2000. sumption (ion pumps and various other ATPases) where Address correspondence and reprint requests to Dr. G. J. Brewer at CK isoenzymes are specifically localized in a compart- Departments of Medical Microbiology/Immunology and Neurology, Southern Illinois University School of Medicine, P.O. Box 19626, mented fashion (Bessman and Geiger, 1981; Wallimann Springfield, IL 62794-9626, U.S.A. E-mail: [email protected] et al., 1992) (Fig. 1). In the cytoplasm and membrane Abbreviations used: A␤, ␤-amyloid; CK, creatine kinase; Cr, crea- compartments of neurons, as well as glial cells in culture tine; MAP2, microtubule-associated protein-2; PCr, phosphocreatine.

1968 NEUROTOXIC PROTECTION BY CREATINE 1969

under stress and Cr supplementation. As a follow-up on a preliminary report (Brewer, 1998b), we describe the protective effects of Cr on neurons in serum-free culture as a model for the excitotoxicity of glutamate and the neurodegenerative action of A␤. Thus, Cr supplementation may turn out to be beneﬁcial for several muscular and neurodegenerative diseases (Guerrero-Ontiveros and Wallimann, 1998).

MATERIALS AND METHODS Cell culture Hippocampal neurons were isolated from rat brains at embryonic day 18 (Brewer et al., 1993), middle age (9–11 months old), or old age (35–37 months old) (Brewer, 1997). The laboratory animal protocol was approved by the Laboratory Animal Medicine Review Committee. In brief, embryonic neurons were isolated by mechanical trituration and then plated at 160 cells/mm2 on glass coverslips precoated with poly-D-lysine in 2% (vol/vol) B27 in Neurobasal medium (Life Technologies, Gaithersburg, MD, U.S.A.) with 25 ␮M glutamate and 0.5 mM glutamine. Every 3–4 days, 50% of the culture medium was exchanged for fresh medium without glutamate. Cultures were

incubated at 37°C in 5% CO2/95% O2 for 5–8 days. For adult neurons (Brewer, 1997), hippocampi were sliced at 0.5 mm, digested with papain, and triturated in HibernateA/B27 (Brewer FIG. 1. Relationship of Cr and PCr to ATP in the neuron, drawn and Price, 1996) (Life Technologies). Cells were separated largely from information on rat heart and chick brain. Exogenous from debris on a density gradient of Nycoprep. The neuron- Cr should be readily transported into neurons by a Cr transporter enriched fraction was collected and cultured for 7 days in or synthesized if precursors are available (Dringen et al., 1998). serum-free medium, B27/NeurobasalA with 5 ng/ml fibroblast Under resting levels of ADP, ATP produced in the mitochondria growth factor-2 (Life Technologies) and 0.5 mM glutamine is converted to PCr by mitochondrial CK (CKMi) and released into without glutamate. Adult cells were plated at 320 cells/mm2.Cr the cytoplasm. Under conditions of high Cr levels, respiratory (Sigma) was freshly dissolved every 3 days at Յ200 mM in Vmax can increase by three- to fivefold (Farrell et al., 1972; Saks et al., 1975). In the cytoplasm, the equilibrium constants for brain Neurobasal medium so that dilutions of 1:100 or 1:400 were made directly into cultures. Glutamate (Sigma) was dissolved CK, CKBB (Quest et al., 1990; Boehm et al., 1996,1998), favor conversion of PCr to ATP. ATP is consumed to maintain ion at 25 mM in Hanks’ balanced salt solution, filter-sterilized, and homeostasis, e.g., via the plasma membrane ATPase and most stored at 4°C. Dilutions were made directly into 5–8-day-old importantly the sarcoplasmic–endoplasmic reticulum Ca2ϩ- cultures or into cultures with fresh medium or Locke’s salts 2ϩ ATPase Ca pump (Wallimann and Hemmer, 1994). The result- [154 mM NaCl, 5.6 mM KC1, 2.3 mM CaCl2,1mM MgCl2, 3.6 ing ADP is rapidly consumed in the regeneration of ATP from PCr mM NaHCO3,5mM HEPES (pH 7.2), and 5 mM glucose]. A␤ by CKBB. Note that PCr can be either transported from the soma (25–35) (QCB, Hopkinton, MA, U.S.A.) was dissolved at 2.5 to distant sites or generated at sites distant from the nucleus (N), mM in deionized water and stored at Ϫ70°C until the day of such as at synaptically localized mitochondria. At synapses, PCr can also be used directly as an energy source to load synaptic use. Dilutions of 1:100 were made into fresh culture medium vesicles, especially under conditions of low potassium that for cells cultured for 6–8 days (Brewer, 1998a). Killing of would accompany high levels of synaptic activity (Xu et al., neuron-like cells was determined at the indicated times by live 1996). cell fluorescence of fluorescein from the diacetate and dead cell nuclear fluorescence of propidium iodide (Brewer et al., 1993). In contrast to protease-containing cultures in serum, back- plementation on PCr levels relative to ATP levels, we ground levels of dead cells in serum-free cultures are ϳ40% define the energy reserve status as the ratio PCr/ATP. In and do not change with time (Brewer, 1999). Percent killing was calculated as 100 Ϫ (live neuron-like cells with nonsym- resting skeletal muscle and heart, PCr levels are three metric, branched processes)/(live neuron-like cells ϩ dead times higher than ATP levels (Passonneau and Lowry, cells). Reported treatment values were corrected by subtraction 1971), but in whole brain they are at most twofold higher of the percent killing in control, untreated cultures. Twelve (Gatfield et al., 1966; Goldberg et al., 1966; Ackerman consecutive fields of 0.313 mm2 were counted per condition. et al., 1980; Erecin´ska and Silver, 1989). However, PCr levels in neurons with large energy demands may be ATP and PCr assays higher than in whole brain, where they could be diluted ATP and PCr assays were based on the methods of Lust et al. (1981). Assays were performed on each of three coverslips per by other cells. One report in cultured neurons found condition from one animal. Cells on coverslips were gently respective resting levels of PCr and ATP of 38 and 27 rinsed twice with warm HibernateA or E (Brewer and Price, nmol/mg of protein (Fitzpatrick et al., 1988), a ratio of 1996) (Life Technologies). Cells were extracted with 50 ␮lof 1.4. Here, we report resting ratios of 15–20 in neurons 0.1 M NaOH for 5 min at 37°C. Extraction with NaOH was cultivated in serum-free medium, which increase to 60 more reproducible than with perchloric acid. For measurement

J. Neurochem., Vol. 74, No. 5, 2000 1970 G. J. BREWER AND T. W. WALLIMANN of PCr, duplicate 1-␮l aliquots were neutralized in 100 ␮lof50 mM imidazole buffer (pH 7.0) containing 1 mM MgCl2, 0.02% bovine serum albumin, and 0.1 ␮MP1,P5-di(adenosine-5Ј) pentaphosphate (Sigma catalogue no. D8013, to inhibit adenylate kinase), followed by addition of 125 ␮M ADP (Sigma catalogue no. A5285) and 1.5 units/ml CK (Sigma catalogue no. C3755). Although the ADP preparation contains low levels of ATP, this concentration was chosen to optimize ATP synthesis from PCr above a background level of ATP observed without added PCr. After incubation for 20 min at 37°C, samples were frozen until assayed for ATP. Similar 1-␮l aliquots from the same samples were frozen for direct assay for ATP. Ten-microliter aliquots were removed for assay by reaction with luciferase in 190 ␮l of assay mix (Sigma catalogue no. FL-ASC) and photon counting for 10 s with an EMI- Gencom (Plainview, NY, U.S.A.) photomultiplier. PCr counts were corrected for direct ATP counts in the same samples and the efficiency of the CK reaction, based on standards (usually ϳ80%). A standard curve was linear from 0.1 to 1,000 pmol of ATP and 1 to 10 pmol of PCr. In calculating the amount of ATP or PCr per live cell, we assume that cells that have lost their ability to retain fluorescein and stain with propidium iodide (dead cells) will not retain ATP or PCr. Although there were FIG. 2. Glutamate toxicity is greater in Locke’s salts (E) than in gradations in fluorescein fluorescence intensity, we did not B27/Neurobasal (F) with hormones, vitamins, and amino acids. observe simultaneous green and red fluorescence. Preliminary Neurons were cultured for 8 days followed by a full medium studies determined that ATP levels correlated better with live change to either Locke’s salts or fresh B27/Neurobasal with cells than with protein concentration, probably because dead addition of 0.5 mM glutamate for the indicated times. A: Increase cells contain protein but have lost their ATP. in neuron killing [ANOVA F(1,22) ϭ 5.4]. B: Decrease in ATP levels [t test (df ϭ 4) ϭ 18 and 25 for the 4- and 7-h time points, Immunocytology and statistics respectively]. Data are mean Ϯ SE (bars) values. Some error bars Immunocytology for the cytoskeletal proteins microtubule- are smaller than the symbols. Similar results were obtained in associated protein-2 (MAP2) and tau were performed as pre- two other experiments. viously described (Brewer, 1998a). In brief, cells in culture medium were either untreated or treated with 1 mM glutamate or glutamate plus 1 mM Cr for 7 or 24 h, fixed with 4% balanced salts than in the nutrient medium B27/Neuro- paraformaldehyde, blocked, and reacted with rabbit anti-tau basal. Correspondingly lower ATP levels were reached (1:2,000; Sigma) and mouse anti-MAP2 (1:250) for1hatroom in Locke’s solution, levels that were one-half to one-third temperature. Specific binding was detected with Cy2 and rho- of the ATP levels in the original culture medium. There- damine antibodies (Jackson ImmunoResearch, West Grove, fore, to ensure that observed effects were due to gluta- PA, U.S.A.). Images were acquired using a Nikon 60ϫ/1.4 n.a. mate rather than nutrient deprivation, all further experi- oil objective and an Ikegami Newvicon camera with Global ments with embryonic hippocampal neurons were per- Lab Image software (Data Translation, Marlboro, MA, U.S.A.) at constant gain, pedestal, and threshold or photographed with formed in the complete medium, Neurobasal/B27, which Kodak Ektachrome P800 film. Mean immunoreactive areas contains amino acids, vitamins, antioxidants, hormones, divided by the number of cells in the field were determined and and other nutrients (Brewer et al., 1993). Based on stain- normalized to control values with Excell software. Controls ing with neurofilament and glial fibrillary acidic protein, without primary antibodies were negative. these cultures are Ͼ95% neurons, so that the effects of Means and SEs are reported in all figures. These values, various stressors can be interpreted as acting directly on ANOVA, and Student’s t tests were calculated with the assis- neurons and not indirectly through glia. tance of PlotIt software (Scientific Programming Enterprises, Hazlett, MI, U.S.A.). Cr protects neurons from glutamate toxicity Based on the hypothesis that neurons die from gluta- RESULTS mate toxicity because of a loss of energy, we attempted to supplement energy reserves by pretreatment with the Neuron culture models of glutamate toxicity are well energy precursor, Cr. After 24 h of Cr treatment, neurons established for both acute and chronic application of were exposed to glutamate for 24 h, followed by mea- glutamate. Most of these approaches remove the nutrient surement of killing. Figure 3A shows that 0.5 mM glu- culture medium during exposure to glutamate. We prefer tamate kills significantly more neurons than the control not to impose nutrient and antioxidant deprivation so that treatment ( p Ͻ 0.001). Cr at 0.03 mM is not protective a single variable of excitotoxin exposure is not com- ( p Ͻ 0.001). In the presence of glutamate and Cr at pounded by metabolic stress. To test the importance of 0.125–2 mM, neuron killing is not significantly different this precept, killing and intracellular ATP levels were from the zero glutamate control. These are physiological monitored during exposure to glutamate. On exposure to levels because Cr supplementation in healthy humans glutamate (Fig. 2), we found greater killing in Locke’s can raise arterial Cr levels from 0.07 to 0.3 mM (Poort-

J. Neurochem., Vol. 74, No. 5, 2000 NEUROTOXIC PROTECTION BY CREATINE 1971

without glutamate ( p ϭ 0.003) but nonsignificantly lower with Cr ( p ϭ 0.055) (Fig. 4B). Figure 4C shows that PCr levels remained constant during the course of glutamate treatment in the presence of Cr, but that without Cr, PCr levels were significantly lower during the first 4 h [ANOVA, F(1,4) ϭ 74, p ϭ 0.04]. PCr levels in surviving cells after7hofglutamate treatment were not significantly affected by Cr. Because PCr is a buffer for ATP, a measure of reserve energy is the ratio of PCr/ ATP. Figure 4D clearly shows that glutamate causes an early rise in the PCr/ATP ratio but that the rise is significantly greater in neurons supplemented with Cr [ANOVA, F(1,4) ϭ 44, p ϭ 0.003]. At 7 h and later, there is no difference in reserve energy in surviving neurons with or without Cr. Cr reduces dendritic pruning by glutamate Mattson and Kater (1989) demonstrated that glutamate preferentially causes destruction of dendrites compared FIG. 3. A: Cr addition before glutamate protects embryonic with axons. Neurons treated with glutamate, with and neurons against glutamate toxicity. Embryonic hippocampal without Cr, were fixed and immunostained for the cy- neurons cultured for 5 days were treated with Cr, as indicated, along with a full change to fresh medium. After 24 h, 0.5 mM glutamate was added as indicated. After another 24 h, neuron killing was determined. B: Cr addition up to 2 h after glutamate exposure is neuroprotective. Either no Cr or 1 mM Cr was added at the indicated times relative to the addition of 1 mM glutamate. Significant killing occurs with glutamate without Cr (1/no) or with Cr (1/ϩ4) added 4 h after glutamate. Similar results were obtained in a second experiment. Data are mean Ϯ SE (bars) values. n.s., not significant compared with control. mans et al., 1997) and muscle Cr from 20 to 25 mmol/L of muscle water (Greenhaff, 1997). Figure 3B shows a study designed to test how late Cr can be added to achieve protection from glutamate toxicity. Cr alone does not affect killing. Cr addition from 24 h before glutamate and up to 2 h after glutamate exposure is neuroprotective; killing was not significantly different from control levels without glutamate. Addition of Cr 4 h after glutamate was too late to preserve viability; killing was significantly greater than control treatment without glutamate ( p Ͻ 0.001). FIG. 4. Short time course of Cr protection from glutamate tox- Cr supplementation increases intracellular PCr and icity. Embryonic neurons 6 days in culture were treated with 1 mM glutamate and without (hatched columns) or with 1 mM Cr boosts the reserve energy ratio (cross-hatched columns). At the indicated times, (A) neuron To determine the effects of Cr supplementation on killing was determined, and extracts were prepared for analysis cellular energetics in the presence of glutamate, extracts of (B) ATP, (C) PCr, and (D) calculated ratio of PCr to ATP. In A, were obtained at various times from neurons treated with ANOVA F(1,22) ϭ 5.2 indicates significant protection by Cr against toxicity. In B, ATP levels declined with time of exposure glutamate and Cr followed by measurement of ATP and to glutamate during the first 4 h. ATP levels were significantly PCr levels. Figure 4A shows significant neuroprotection lower for samples treated with Cr, ANOVA F(1,4) ϭ 74. After from 2 to 24 h by Cr added at the same time as glutamate 24 h, ATP levels in surviving cells were significantly lower than at [ANOVA, F(1,22) ϭ 5.2, p ϭ 0.03]. In Fig. 4B, ATP the start, but levels in cells treated with Cr were not significantly levels declined over the first4hintheCr-treated cells higher than those in untreated cells ( p ϭ 0.08). In C, PCr levels remained high in the presence of Cr, but fell in the absence of Cr, compared with untreated cells [ANOVA, F(1,4) ϭ 74, p ANOVA F(1,4) ϭ 8.1. Cells that survived 15–24 h of treatment ϭ 0.001], likely owing to the conversion of new intra- with glutamate showed PCr levels similar to those of untreated cellular Cr to PCr with consumption of ATP. At times cells, with no long-term effect of Cr. In D, the ratio of PCr/ATP later than 4 h, the difference between addition and no increases in the first4hofglutamate treatment, with significantly higher levels obtained in the presence of Cr, ANOVA F(1,4) ϭ 44. addition of Cr is not significant, although the final levels Data are mean Ϯ SE (bars) values. In some cases, error bars are of ATP in surviving cells after 24 h are significantly too small to show. Similar results were obtained in a second lower from treatment with glutamate without Cr than experiment.

J. Neurochem., Vol. 74, No. 5, 2000 1972 G. J. BREWER AND T. W. WALLIMANN

FIG. 5. Phase-contrast (A, D, and G), tau (B, E, and H) and MAP2 (C, F, and I) immunostained neurons. After 6 days in culture, embryonic neurons were left untreated (A–C), treated with 1 mM glutamate (glu; D–F), or glu and 1 mM Cr (G–I). After 7 h, cells were fixed and pro- cessed for immunocytology. In the presence of glu, note the presence of labeled axons (arrow- heads in E) but the severe dendritic pruning (arrows in F). Note that stronger dendritic labeling re- mains in cells treated with Cr and glu (arrows in I). Bar in A ϭ 50 ␮m. J and K: Quantitative analysis of immunoreactivity for MAP2 (J) and tau (K), normalized to the control with zero addition. Proba- bilities show comparisons with no additions. Data are mean Ϯ SE (bars) values and are representative of three cell preparations with measures from 10 fields with ϳ10 cells per field. n.s., not significant compared with control.

toskeletal elements tau, which is enriched in axons, and Cr and A␤ toxicity MAP2, which is largely restricted to soma and dendrites. Neurons in culture have been used as a model for the Figure 5 shows the dendritic pruning caused by gluta- degenerative effects of A␤ in Alzheimer’s disease mate (D and F). Note the presence of labeled axons (Yankner et al., 1990; Pike et al., 1993). In these studies, (arrow in E). In the presence of Cr and glutamate, den- the fragment of residues 25–35 is responsible for the drite morphology (G) and dendritic labeling with MAP2 majority of the neurotoxicity (Pike et al., 1995). Deple- is stronger (I) than with glutamate without Cr (F). Quan- tion of ATP is an early event in A␤ toxicity (Mark et al., titative analysis of MAP2 (J) and tau (K) confirms the 1997). Therefore, supplementation with the energy pre- greater effect of glutamate to reduce MAP2 staining for cursor, Cr, might be neuroprotective. Figure 6A shows dendrites (60% decrease) than tau staining for axons (no significant toxicity to embryonic neurons of a 48-h treat- decrease, t test, p Ͻ 10Ϫ6). However, Cr treatment pro- ment with 25 ␮M A␤ (25–35) ( p Ͻ 0.002). Cr doses tects against some of this loss of MAP2 immunoreactiv- from 0.5 to 2 mM partially protect against A␤ toxicity ity ( p ϭ 0.002). It is surprising that tau staining increases [ANOVA, F(3,8) ϭ 9.5, p ϭ 0.005]. Figure 6B shows in the presence of Cr and glutamate, compared with that ATP levels in neurons surviving this 48-h treatment neurons treated with glutamate ( p ϭ 0.006). A control of are slightly elevated by 26% in the presence of Cr Cr alone also indicated increased tau immunoreactivity compared with neurons surviving A␤ treatment without compared with untreated neurons. Cr ( p ϭ 0.06). Note that Cr alone significantly raises

J. Neurochem., Vol. 74, No. 5, 2000 NEUROTOXIC PROTECTION BY CREATINE 1973

Cr and A␤ toxicity in adult and aged neurons The capability to isolate and culture neurons from adult hippocampus of any age rat (Brewer, 1997) allows comparisons of A␤ toxicity with age and the protective effects of Cr. We recently found that toxicity to A␤ is age-related, with increased levels of killing in neurons from 3-year-old rats compared with middle-aged and embryonic rats (Brewer, 1998a), yet viability without added stressors is unaffected by age (Brewer, 1997, 1999). Figure 8 gives an example of this age-related toxicity. In the presence of 2 mM Cr, added at the same time as A␤, neuron killing is signiﬁcantly reduced compared with A␤ without Cr [ANOVA, F(1,6) ϭ 13, p ϭ 0.01]. However, the relative amount of protection was least with the oldest neurons (9%) compared with 21% for middle- aged neurons and 14% for embryonic neurons in this small sample from three animals of each age.

DISCUSSION In these studies of cultured hippocampal neurons, the FIG. 6. Cr partially protects embryonic neurons against A␤ toxicity after 48 h. A: Cr doses of 0.5–2 mM partially protect against deleterious effects of two toxic agents, glutamate and A␤ toxicity, ANOVA F(3,8) ϭ 9.5. B: ATP levels per live cell A␤, are signiﬁcantly reduced by Cr supplementation. For increase slightly with Cr compared with A␤ without Cr in surviv- glutamate, Cr can prevent killing of neurons up to 2 h ing cells. C: PCr levels per live cell are elevated in the presence of Cr. D: PCr/ATP levels are elevated in the presence of Cr. Extracts were obtained, and toxicity was measured after 48 h. Cr was added at the same time as A␤. Data are mean Ϯ SE (bars) values of cultures from three independent cell preparations (A) or one representative of three experiments (B–D). n.s., not signiﬁ- cant compared with control.

ATP levels in these cells ( p ϭ 0.002). In Fig. 6C, A␤ treatment without Cr causes a small but nonsignificant decline in PCr levels. However, in the presence of A␤ and Cr (0.5–2 mM), there is a highly significant rise in PCr levels ( p ϭ 0.005). Cr alone raises intracellular PCr levels ( p ϭ 0.04). In Fig. 6D, the PCr/ATP ratio is elevated by 61% (from 18 to 29; p ϭ 0.02) for surviving cells in the presence of A␤ and Cr. From two other experiments, PCr/ATP values increased from 28 to 46 (64%) and from 36 to 43 (19%) without and with Cr, respectively. Although the elevation of PCr level is significant in neurons surviving treatment with A␤ in the presence of Cr, the initial response of neurons may be more dramatic. FIG. 7. Short time course of Cr protection from A␤ toxicity. Marginal killing by A␤ is detected as early as6h(p Embryonic neurons in culture for 8 days were treated with 25 ␮M ϭ 0.03), which is protected by Cr (Fig. 7A). In Fig. 7B, A␤ and1mM Cr as indicated. Extracts were prepared at the indicated times. A: Neuron killing by A␤ was significantly higher ATP levels do not change significantly in 6 h except a after 6 h without Cr but was not significantly changed after 6 h paradoxical 20% increase in the absence of Cr ( p with Cr. B: ATP levels remain unchanged in the presence of A␤ ϭ 0.007). In Fig. 7C, PCr levels are seen to increase and Cr, whereas they increase significantly after6hinthe significantly from 1 to 6 h [ANOVA, F(4,10) ϭ 24, p presence of A␤ without Cr, compared with Cr. C: PCr levels increase in the presence of A␤ and Cr, ANOVA F(4,10) ϭ 24. PCr Ͻ 0.0001]. PCr levels were also 30% higher in the levels were significantly lower after6hinthepresence of A␤ absence of Cr ( p ϭ 0.03). In Fig. 7D, the reserve energy without Cr compared with Cr. In D, the ratio of PCr to ATP rises ratio of PCr/ATP rises from 15 to 60 in 6 h, an increase sharply for cells treated with A␤ and Cr, ANOVA F(4,10) ϭ 3.7. of fourfold [ANOVA, F(4,10) ϭ 3.7, p ϭ 0.04]. The The PCr/ATP level is significantly lower after6hinthepresence of A␤ without Cr compared with Cr. Data are mean Ϯ SE (bars) reserve energy ratio for A␤ treatment without Cr was values and are representative of three experiments. n.s., not slightly elevated ( p ϭ 0.04). significant compared with control.

J. Neurochem., Vol. 74, No. 5, 2000 1974 G. J. BREWER AND T. W. WALLIMANN

Depending on the ATP levels, either apoptosis or necrosis will result (Nicotera et al., 1998). Mitochondrial CK octamer was recently shown to be part of the permeability transition pore complex involving mitochondrial outer membrane porin and inner membrane adenine nucleotide translocase (Brdiczka et al., 1998). Cr and cyclocreatine were demonstrated to protect permeability transition pore opening in mitochondria containing the CK octamer (O’Gorman et al., 1997). Thus, some of the effects seen with Cr in cultured neurons are likely to be due to stabilization and protection of the permeability transition pore and mitochondrial integrity (Brdiczka FIG. 8. Cr partially protects against 25 ␮M A␤ toxicity in neurons from old rats (solid columns) and young rats (cross-hatched et al., 1998). columns) as well as embryonic rats (open columns). A␤ treat- Neuroprotection may also involve induction of tran- ment produces significantly more neuron killing compared with scription of CK. Transcription of CK isozyme BB can be the control without A␤. In the presence of 2 mM Cr, added at the up-regulated by activation of cyclic AMP-dependent same time as A␤, neuron killing is significantly reduced com- protein kinase A (Kuzhikandathil and Molloy, 1994). pared with A␤ without Cr, ANOVA F(1,6) ϭ 13, p ϭ 0.01. Data are mean Ϯ SE (bars) values from neurons prepared from three Across a wide range of tissues, levels of Cr correlate with animals of each age. levels of CK (Iyengar, 1984). In addition, phosphorylation of CK lowers its Km for Cr by 50% (Quest et al., 1990). As studied in isolated mitochondria, Cr can be after the chronic addition of glutamate. Cr also reduces rate-limiting because Cr can increase respiration by the loss of MAP2 immunoreactivity in dendrites associ- three- to fivefold (Farrell et al., 1972; Saks et al., 1975). ated with glutamate toxicity while increasing tau immu- An additional mechanism of action of Cr supplementa- noreactivity. An NMDA channel blocker has also been tion may be the recently discovered interplay between reported to reduce the loss of MAP2 immunoreactivity AMP-activated protein kinase and CK. Not only is CK (Laferrie`re et al., 1999). These may be the first studies of activity regulated by phosphorylation via AMP-activated PCr levels in hippocampal neurons in serum-free culture protein kinase, but more interesting is that AMP-acti- in which Cr levels are controlled. Part of the mechanism vated protein kinase is directly regulated by the ATP/ of Cr protection appears to be a preservation of the levels AMP and PCr/Cr ratios (Ponticos et al., 1998). Thus, of PCr, which decline in the presence of glutamate with- increasing the cellular energy charge by elevating these out Cr. In the case of A␤, ATP did not decline, but ratios via Cr supplementation would reduce AMP-acti- actually increased in the absence of Cr. In the presence of vated protein kinase activity, acting as a ratiometric Cr and A␤, PCr levels significantly increased with time. energy sensor, not only reducing CK activity but other These results suggest that there are differences in the targets such as lipid metabolism (Kemp et al., 1999). energy demands of neurons challenged with glutamate or Together, these studies indicate multiple mechanisms for A␤, that glutamate stresses neuron energy levels more increasing cellular energy reserves through CK, if suffi- than A␤ at the concentrations studied. cient substrate Cr is available. It is surprising that resting levels of the PCr/ATP ratio, Additional facets of a central mechanism of neuropro- which start at 5–15, can rise as high as 60 in the presence tection based on energy supplies come from the work of of Cr. Possible changes in cell volume may have addi- Aksenov et al. (1998). In cultured embryonic hippocam- tional effects on the concentrations of these compounds, pal neurons exposed to A␤, CK isozyme BB activity is but the ratio is a measure independent of cell volume. rapidly reduced, but in those neurons that survive, CK is This rise in PCr/ATP ratio may serve to power ion elevated. Our short-time kinetics in which ATP declined pumps, especially the energetically demanding sarco- but PCr remained elevated could be explained by re- plasmic–endoplasmic reticulum Ca2ϩ-ATPase pump duced levels of CK isozyme BB in the cytoplasm, (Wallimann and Hemmer, 1994), whose activities are whereas mitochondrial CK remained active. required to maintain ion homeostasis, excitability, and The critical function of CK for the energetics of Ca2ϩ Ca2ϩ signaling. These ion pumps are coupled to CK and homeostasis has been lucidly demonstrated by transgenic require a high local ATP/ADP ratio for efficient function. mice lacking the muscle CK isozyme (Steeghs et al., Excitotoxicity is accompanied by a Ca2ϩ surge 1997). These mice develop severe problems with muscle (Khodorov et al., 1996; Stout et al., 1998), which leads to relaxation and Ca2ϩ handling. Thus, part of the neuro- a vicious cycle of chronic Ca2ϩ overload and further protective effects of the energy precursor, Cr, could be deterioration of the cellular energy status, if energy re- explained by the improvement of cellular energetics un- serves are insufficient for Ca2ϩ pumping. der stressful conditions. The fact that the PCr/ATP ratio Chronic Ca2ϩ overload, elicited by continuous expo- can rise in neurons much higher than expected from the sure of neurons to neurotoxins, will induce Ca2ϩ uptake CK equilibrium reaction is indicative of subcellular com- by mitochondria (Kruman and Mattson, 1999) and open- partments of CK or its substrates, especially PCr. Our ing of the mitochondrial permeability transition pore. finding of differential effects of Cr on axonal and den-

J. Neurochem., Vol. 74, No. 5, 2000 NEUROTOXIC PROTECTION BY CREATINE 1975 dritic compartments in neurons supports a nonhomoge- plus Cr proved most effective in attenuating the hypoxia neous distribution of Cr transport or CK or PCr. In vivo current (Chung et al., 1998). Recently, neuroprotective findings support a sequestered pool of PCr or Cr to which effects not only of Cr, but also of its analogues, guanidi- CK does not have access or which is not in equilibrium nopropionic acid and cyclocreatine, have been reported with other adenylnucleotides (Hochachka and Mossey, in vivo. Cyclocreatine has a sparing action on brain 1998). Expected reaction rates are further complicated by energy reserves (Woznicki and Walker, 1980), and gua- intracellular acidosis with glutamate treatment (Wang nidinopropionic acid enhances animal survival during et al., 1994), tending to shift the kinase toward ATP hypoxia (Holtzman et al., 1997, 1998a). Cr suppresses synthesis and away from PCr synthesis (Furter et al., seizures in the hypoxic immature rat (Holtzman et al., 1993). These equilibrium arguments are supported by 1998b). Mitochondrial poisons were used to induce char- our observed fall in PCr levels in the presence of gluta- acteristics of Huntington’s disease in rats (Matthews mate, without extracellular Cr, but do not explain in a et al., 1998). For these rats that were given Cr supple- simple single reaction the observed fall in ATP levels, ments in their food, lesion sizes were reduced two- to exacerbated by Cr. Energy consumption (lower ATP) fivefold over controls. The same laboratory has also could easily exceed the ability of the cell to regenerate found large benefits of Cr with a transgenic mouse model high-energy phosphates. of amyotrophic lateral sclerosis (Klivenyi et al., 1999). The fact that an inborn error of Cr metabolism involv- Cr is either synthesized by certain cells or actively ing the last step of Cr biosynthesis, guanidinoacetate- transported into tissues with high energy demands by a methyl transferase, leads to a severe extrapyramidal neu- sodium-dependent transporter [CreaT (Sora et al., rological disorder (Stockler et al., 1994) indicates that Cr 1994)]. Areas of high expression of the transporter in the synthesis in the brain is physiologically important. Be- adult rat brain include the hippocampus, white matter cause the severity of symptoms was greatly alleviated by tracts, and endothelial cells of the choroid plexus oral Cr supplementation, it seems that Cr can be taken up (Saltarelli et al., 1996) and correspond with areas of high by brain cells. The Cr transport system has been reported expression of CK isozyme BB (Kaldis et al., 1996). In in glial cells (Mo¨ller and Hamprecht, 1989) and matches addition, Cr can be synthesized in rat brain (Defalco and the high activity of CK, especially of cytosolic CK Davies, 1961). In humans, de novo Cr synthesis occurs isozyme BB, also seen in white matter by in vivo 31P- mostly in the kidney, pancreas, and liver (Greenhaff, NMR CK flux measurements (Holtzman et al., 1993). 1997). Cr is also available in meat in the diet at a level of Thus, neurons were assumed to synthesize their own Cr, 11% of total nitrogen, ϳ0.4% of beef mass (Khan and but this has not yet been demonstrated experimentally. Cowen, 1977; Dvorak, 1981). Cr loading by adults at a Astroglial cells in culture were shown to synthesize Cr rate of 5 g/day for 5 days increases arterial Cr levels from from glycine and to secrete guanidinoacetate, the precur- 74 to 275 (Ϯ 213) ␮M (Poortmans et al., 1997). The sor of Cr (Dringen et al., 1998). Therefore, it was sug- large SD indicates significant variation among subjects. gested that neurons take up guanidinoacetate and synthe- In muscle, PCr levels can rise an average of 20% accom- size their own Cr. It is shown here, however, that neurons panied by a 20% mean increase in peak work perfor- are obviously able to take up Cr from the medium mance (Greenhaff, 1997) and a loss of fat and increase of because on Cr supplementation, the PCr levels and the lean body mass (Vandenberghe et al., 1997). Based on PCr/ATP ratio both increased markedly. It is interesting these results, Cr supplementation is now widely used by that the Cr effects were seen at rather low concentrations athletes to boost work performance and speed up recov- and were surprisingly still effective after delayed addi- ery from exhaustive exercise (Greenhaff et al., 1994). In tion of up to 2 h. If neurons have the same transporter as human heart disease, Cr reduces postventricular arrhyth- astroglial cells, then the Km of 20 ␮M for transport of Cr mias in patients with myocardial infarcts (Ruda et al., (Dai et al., 1999) can explain the effectiveness of Cr at as 1988). Although Cr supplementation failed to increase low as 0.1 mM for neurons exposed to the faster deple- heart ejection fraction in patients with chronic heart tion of energy by the glutamate stressor. Thus, the choice failure, these patients showed significant increases in leg between synthesis, uptake, or both by different brain strength (Gordon et al., 1995). Cr also produced in- cells may depend on the metabolic condition of these creased handgrip strength in patients after congestive cells and the extracellular availability of Cr. heart failure (Andrews et al., 1998). In congenital human A neuroprotective action of Cr has been reported with mitochondrial cytopathies, Cr increases handgrip guinea pig and rat hippocampi, where preincubation of strength by 19% and leg strength by 11% (Tarnopolsky brain slices with this natural compound was shown to et al., 1997). Cr markedly improved muscle cell survival prolong synaptic transmission (Whittingham and Lipton, and alleviated the severely impaired Ca2ϩ homeostasis 1981) and to prevent anoxic damage (Carter et al., 1995). of dystrophic mdx myotubes under stress in culture (Pu- Similar protective effects were noted with ex vivo brain- lido et al., 1998). Finally, Cr protects the livers of trans- stem slices from Cr-fed animals where electrical activity genic mice made to express CK brain-type isozyme BB of the central respiratory network was protected from in the liver, which is normally absent, against ATP anoxia in animals treated with Cr but not in controls depletion and a drop in pH during fructose overload (Wilken et al., 1998). In addition, in hippocampal CA1 (Brosnan et al., 1991), hypoxia, and ischemia (Miller neurons of rat brain slices, intracellularly applied ATP et al., 1993), as well as against liver toxins (Kanazawa

J. Neurochem., Vol. 74, No. 5, 2000 1976 G. J. BREWER AND T. W. WALLIMANN et al., 1998). This again stresses the energy sparing and Acknowledgment: We thank John Torricelli for excellent cell protective effects of the PCr/CK system. Together, technical assistance. This work was supported in part by grant these results indicate no reported adverse effects, very RO1 AG13435 from the National Institute on Aging, by Life good oral availability, and clear benefit in tasks requiring Technologies, by Swiss NF grant 3100-50824.97, and by the brief maximal muscle work output. Swiss Society for Muscle Diseases. In support of the relevance of these findings to Alz- heimer’s disease, Aksenov et al., (1997) found cytoplas- REFERENCES mic CK isozyme BB was reduced by 80% in homoge- nates prepared from Alzheimer’s disease brains com- Ackerman J. J. H., Grove T. H., Wong G. G., Gadian D. G., and Radda pared with age-matched controls. The loss of CK is also G. K. (1980) Mapping of metabolites in whole animals by P NMR using surface coils. Nature 283, 167–170. significant because of its particular sensitivity to inacti- Aksenov M. Y., Aksenova M. V., Payne R. M., Smith C. D., Markes- vation by reactive oxygen species (Aksenov et al., 1997; bery W. R., and Carney J. M. (1997) The expression of creatine David et al., 1998), especially peroxynitrite (Kuzhikan- kinase isoenzymes in neocortex of patients with neurodegenera- dathil and Molloy, 1994; Stachowiak et al., 1998). Cr, tive disorders: Alzheimer’s and Pick’s disease. Exp. Neurol. 146, together with Mg2ϩ-nucleotide, is protective against per- 458–465. Aksenov M. Y., Aksenova M. V., Markesbery W. R., and Butterfield oxynitrite inactivation of all CK isoenzymes, by protect- D. A. (1998) Amyloid ␤-peptide (1–40)-mediated oxidative stress ing the enzyme from being modified in the active site, in cultured hippocampal neurons—protein carbonyl formation, most likely involving the critical Cys278 residue (Sta- CK BB expression, the level of Cu, Zn, Mn SOD mRNA. J. Mol. chowiak et al., 1998). Nitric oxide donors often elicit Neurosci. 10, 181–192. Andrews R., Greenhaff P., Curtis S., Perry A., and Cowley A. J. (1998) apoptosis (Bonfoco et al., 1996). At least some of the The effect of dietary creatine supplementation of skeletal muscle toxic effects of A␤ appear to be mediated by nitric oxide metabolism in congestive heart failure. Eur. Heart J. 19, 617–622. and related events (Hu and el-Fakahany, 1993). Support Ball W. J. and Atkinson D. E. (1975) Adenylate energy charge in for reduced energy levels in Alzheimer’s disease brain Saccharomyces cerevisiae during starvation. J. Bacteriol. 121, comes from magnetic resonance studies. Pettegrew et al. 975–982. Bessman S. P. and Geiger P. J. (1981) Transport of energy in muscle: (1994) found decreases of 10% in PCr levels in the brains the phosphorylcreatine shuttle. Science 211, 448–452. of mildly demented patients. However, brains from mod- Bishop P. D. and Atkinson D. E. (1984) Adenine nucleotide control of erately severe cases did not show the decline, and Bot- the rate of oxygen uptake by rat heart mitochondria over a 15- to tomley et al. (1992) did not detect significant changes. 20-fold range. Arch. Biochem. Biophys. 230, 335–344. These negative findings are not surprising for two rea- Boehm E., Veksler V., Mateo P., Lenoble C., Wieringa B., and Ven- tura-Clapier R. (1998) Maintained coupling of oxidative phos- sons. First, Alzheimer’s disease does not affect the whole phorylation to creatine kinase activity in sarcomeric mitochondrial brain; visible pathology affects Ͻ7% of the brain (Lass- creatine kinase-deficient mice. J. Mol. Cell. Cardiol. 30, 901–912. mann et al., 1995). Therefore, detecting even large Boehm E. A., Radda G. K., Tomlin H., and Clark J. F. (1996) The changes in a small fraction of the neurons is theoretically utilization of creatine and its analogues by cytosolic and mitochondrial creatine kinase. Biochim. Biophys. Acta 1274, 119–128. difficult if not impossible. Second, the above magnetic Bonfoco E., Leist M., Zhivotovsky B., Orrenius S., Lipton S. A., and resonance measurements were made under resting con- Nicotera P. (1996) Cytoskeletal breakdown and apoptosis elicited ditions. Considering our data and biochemical principles, by NO donors in cerebellar granule cells require NMDA receptor PCr may remain at modest levels until neurons are activation. J. Neurochem. 67, 2484–2493. stressed, for example, by high levels of activity that Bottomley P. A., Cousins J. P., Pendrey D. L., Wagle W. A., Hardy C. J., Eames F. A., McCaffrey R. J., and Thompson D. A. (1992) consume PCr faster than it can be restored by oxidative Alzheimer dementia: quantification of energy metabolism and phosphorylation or glycolysis. Here we show that iso- mobile phosphoesters with P-31 NMR spectroscopy. Radiology lated neurons from young and old adult hippocampus are 183, 695–699. also protected from A␤ toxicity by Cr, although protec- Brdiczka D., Beutner G., Ry¨ck A., Dolder M., and Wallimann T. (1998) The molecular structure of mitochondrial contact sites. tion by Cr in older neurons was less than in young Their role in regulation of energy metabolism and premeability neurons. If sufficient Cr is available, healthy neurons transition. Biofactors 8, 235–242. may be able to maintain PCr levels that would rapidly Brewer G. J. (1997) Isolation and culture of adult rat hippocampal decline without exogenous Cr. neurons. J. Neurosci. Methods 71, 143–155. Clinical trials of Cr in Alzheimer’s disease and stroke Brewer G. J. (1998a) Age-related toxicity to lactate, glutamate and beta-amyloid in cultured adult neurons. Neurobiol. Aging 19, patients may be justified, in view of the protective effects 571–568. of Cr reported here for neurons in culture and the dra- Brewer G. J. (1998b) Protective effect of the energy precursor creatine matic protection of Cr in animal models of hypoxia against toxicity of glutamate and beta-amyloid in rat hippocampal (Holtzman et al., 1998b; Wilken et al., 1998), Hunting- neurons. Soc. Neurosci. Abstr. 24, 1456. ton’s disease (Matthews et al., 1998), and amyotrophic Brewer G. J. (1999) Regeneration and proliferation of embryonic and adult rat hippocampal neurons in culture. Exp. Neurol. 159, 237– lateral sclerosis (Klivenyi et al., 1999). The oral avail- 247. ability of Cr, its ability to increase brain PCr levels Brewer G. J. and Price P. J. (1996) Viable cultured neurons in ambient (Matthews et al., 1998), and its apparent lack of side carbon dioxide and hibernation storage for a month. Neuroreport effects provide further reason for optimism to establish 7, 1509–1512. Brewer G. J., Torricelli J. R., Evege E. K., and Price P. J. (1993) Cr supplementation as an adjuvant or preventive thera- Optimized survival of hippocampal neurons in B27-supplemented peutic for several neurodegenerative and age-related dis- Neurobasal, a new serum-free combination. J. Neurosci. Res. 35, eases. 567–576.

J. Neurochem., Vol. 74, No. 5, 2000 NEUROTOXIC PROTECTION BY CREATINE 1977

Brosnan J. M., Chen L., Wheeler C. L., Van Dyke T. A., and Koretsky Bergmann glial cells and muscle-specific creatine kinase in Pur- A. P. (1991) Phosphocreatine protects ATP from a fructose load in kinje neurons. Eur. J. Neurosci. 6, 538–549. transgenic mouse liver expressing creatine kinase. Am. J. Physiol. Hochachka P. W. and Mossey M. K. (1998) Does muscle creatine 260, C1191–C1200. phosphocreatine have access to the total pool of phosphocreatine Budd S. L., Castilho R. F., and Nicholls D. G. (1997) Mitochondrial plus creatine? Am. J. Physiol. 274, R868–R872. membrane potential and hydroethidine-monitored superoxide gen- Holtzman D., Tsuji M., Wallimann T., and Hemmer W. (1993) Func- eration in cultured cerebellar granule cells. FEBS Lett. 415, 21–24. tional maturation of creatine kinase in rat brain. Dev. Neurosci. 15, Carter A. J., Mu¨ller E., Pschorn U., and Stransky W. (1995) Preincu- 261–270. bation with creatine enhances levels of creatine phosphate and Holtzman D., Meyers R., O’Gorman E., Khait I., Wallimann T., Allred prevents anoxic damage in rat hippocampal slices. J. Neurochem. E., and Jensen F. (1997) In vivo brain phosphocreatine and ATP 64, 2691–2699. regulation in mice fed a creatine analog. Am. J. Physiol. 272, Chen L., Roberts R., and Friedman D. (1995) Expression of brain-type C1567–C1577. creatine kinase and ubiquitous mitochondrial creatine kinase in the Holtzman D., Brown M., O’Gorman E., Allred E., and Wallimann T. fetal rat brain: evidence for a nuclear energy shuttle. J. Comp. (1998a) Brain ATP metabolism in hypoxia resistant mice fed Neurol. 363, 389–401. guanidinopropionic acid. Dev. Neurosci. 20, 469–477. Chung I., Zhang Y., Eubanks J. H., and Zhang L. (1998) Attenuation of Holtzman D., Togliatti A., Khait I., and Jensen F. (1998b) Creatine hypoxic current by intracellular application of ATP regenerating increases survival and suppresses seizures in the hypoxic imma- agents in hippocampal CA1 neurons of rat brain slices. Neuro- ture rat. Pediatr. Res. 44, 410–414. science 86, 1102–1107. Hu J. and el-Fakahany E. E. (1993) Beta-amyloid 25–35 activates nitric Dai W. X., Vinnakota S., Qian X. J., Kunze D. L., and Sarkar H. K. oxide synthase in a neuronal clone. Neuroreport 4, 760–762. (1999) Molecular characterization of the human CRT-1 creatine Iyengar M. R. (1984) Creatine kinase as an intracellular regulator. transporter expressed in Xenopus oocytes. Arch. Biochem. Bio- J. Muscle Res. Cell. Motil. 5, 527–534. phys. 361, 75–84. Kaldis P., Hemmer W., Zanolla E., Holtzman D., and Wallimann T. David S., Shoemaker M., and Haley B. E. (1998) Abnormal properties (1996) Hot spots of creatine kinase localization in brain: cerebel- of creatine kinase in Alzheimer’s disease brain: correlation of lum, hippocampus and choroid plexus. Dev. Neurosci. 18, 542– reduced enzyme activity and active site photolabeling with aber- 554. rant cytosol-membrane partitioning. Mol. Brain Res. 54, 276–287. Kanazawa A., Tanaka A., Iwata S., Satoh S., Hatano E., Shinohara H., Defalco A. J. and Davies R. K. (1961) The synthesis of creatine by the Kitai T., Tsunekawa S., Ikai I., Yamamoto M., Takahashi R., brain of the intact rat. J. Neurochem. 7, 308–312. Chance B., and Yamaoka Y. (1998) The beneficial effect of Dringen R., Verleysdonk S., Hamprecht B., Willker W., Leibfritz D., phosphocreatine accumulation in the creatine kinase transgenic and Brand A. (1998) Metabolism of glycine in primary astroglial mouse liver in endotoxin-induced hepatic cell death. J. Surg. Res. cells: synthesis of creatine, serine, and glutathione. J. Neurochem. 80, 229–235. 70, 835–840. Kemp B. E., Mitchelhill K. I., Stapleton D., Michell B. J., Chen Z. P., Dvorak Z. (1981) Creatine as an indicator of net muscle proteins. J. Sci. and Witters L. A. (1999) Healing with energy demand: the AMP- Good Agric. 32, 1033–1036. activated protein kinase. Trends Biol. Sci. 24, 22–25. Erecin´ska M. and Silver I. A. (1989) ATP and brain function. J. Cereb. Khan A. W. and Cowen D. C. (1977) Rapid estimation of muscle Blood Flow Metab. 9, 2–19. proteins in beef–vegetable protein mixtures. J. Agric. Food Chem. Farrell E. C., Baba N., Brierley G. P., and Grumer H. D. (1972) On the 25, 236–238. creatine phosphokinase of heart muscle mitochondria. Lab. Invest. Khodorov B., Pinelis V., Vergun O., Storozhevykh T., and Vinskaya N. 27, 209–213. (1996) Mitochondrial deenergization underlies neuronal calcium Fitzpatrick S. M., Cooper A. J. L., and Hertz L. (1988) Effects of overload following a prolonged glutamate challenge. FEBS Lett. ammonia and ␤-methylene-DL-aspartate on the oxidation of glucose and pyruvate by neurons and astrocytes in primary culture. 397, 230–234. J. Neurochem. 51, 1197–1203. Klivenyi P., Ferrante R. J., Matthews R. T., Bogdanov M. B., Klein Furter R., Furter-Graves E. M., and Wallimann T. (1993) Creatine A. M., Andreassen O. A., Mueller G., Wermer M., Kaddurah- kinase: the reactive cysteine is required for synergism but is Daouk R., and Beal M. F. (1999) Neuroprotective effects of nonessential for catalysis. Biochemistry 32, 7022–7029. creatine in a transgenic animal model of amyotrophic lateral Gatfield P. D., Lowry O. H., Schulz D. W., and Passonneau J. V. (1966) sclerosis. Nat. Med. 5, 347–350. Regional energy reserves in mouse brain and changes with isch- Kruman I. I. and Mattson M. P. (1999) Pivotal role of mitochondrial aemia and anaesthesia. J. Neurochem. 13, 185–195. calcium uptake in neural cell apoptosis and necrosis. J. Neuro- Goldberg N. D., Passonneau J. V., and Lowry O. H. (1966) Effects of chem. 72, 529–549. changes in brain metabolism on the levels of citric acid cycle Kuzhikandathil E. V. and Molloy G. R. (1994) Transcription of the intermediates. J. Biol. Chem. 241, 3997–4003. brain creatine kinase gene in glial cells is modulated by cyclic Gordon A., Hultman E., Kaijser L., Kristjansson S., Rolf C. J., Nyquist AMP-dependent protein kinase. J. Neurosci. Res. 39, 70–82. O., and Sylven C. (1995) Creatine supplementation in chronic Laferrie`re N. B., Tremblay R., Murray C. L., Monette R., Brown D. L., heart failure increases skeletal muscle creatine phosphate and Durkin J. P., and Morley P. (1999) AR-R15896AR blocks the muscle performance. Cardiovasc. Res. 30, 413–418. early NMDA-induced loss of MAP2 in primary cortical cultures. Greenhaff P. L. (1997) The nutritional biochemistry of creatine. J. Neurol. Res. 21, 524–528. Nutr. Biochem. 8, 610–618. Lassmann H., Bancher C., Breitschopf H., Wegiel J., Bobinski M., Greenhaff P. L., Bodin K., Soderlund K., and Hultman E. (1994) Effect Jellinger K., and Wisniewski H. M. (1995) Cell death in Alzhei- of oral creatine supplementation on skeletal muscle phosphocre- mer’s disease evaluated by DNA fragmentation in situ. Acta atine resynthesis. Am. J. Physiol. 266, E725–E730. Neuropathol. (Berl.) 89, 35–41. Guerrero-Ontiveros M.-L. and Wallimann T. (1998) Creatine supple- Lust W. D., Feussner G. K., Barbehenn E. K., and Passonneau J. V. mentation in health and disease. Effects of chronic creatine inges- (1981) The enzymatic measurement of adenine nucleotides and tion in vivo: down-regulation of the expression of creatine trans- P-creatine in picomole amounts. Anal. Biochem. 110, 258–266. porter isoforms in skeletal muscle. Mol. Cell. Biochem. 184, Manos P., Bryan G. K., and Edmond J. (1991) Creatine kinase activity 427–437. in postnatal rat brain development and in cultured neurons, astro- Hemmer W. and Wallimann T. (1993) Functional aspects of creatine cytes, and oligodendrocytes. J. Neurochem. 56, 2101–2107. kinase in brain. Dev. Neurosci. 15, 249–260. Mark R. J., Pang Z., Geddes J. W., Uchida K., and Mattson M. P. Hemmer W., Zanolla E., Furter-Graves E., Eppenberger H. M., and (1997) Amyloid ␤-peptide impairs glucose transport in hippocam- Wallimann T. (1994) Creatine kinase isoenzymes in chicken cer- pal and cortical neurons: involvement of membrane lipid peroxi- ebellum: specific localization of brain-type creatine kinase in dation. J. Neurosci. 17, 1046–1054.

J. Neurochem., Vol. 74, No. 5, 2000 1978 G. J. BREWER AND T. W. WALLIMANN

Matthews R. T., Yang L. C., Jenkins B. G., Ferrante R. J., Rosen B. R., porter in situ and in transfected HeLa cells. Dev. Neurosci. 18, Kaddurah-Daouk R., and Beal M. F. (1998) Neuroprotective ef- 524–534. fects of creatine and cyclocreatine in animal models of Hunting- Sora I., Richman J., Santoro G., Wei H., Wang Y., Vanderah T., ton’s disease. J. Neurosci. 18, 156–163. Horvath R., Nguyen M., Waite S., Roeske W. R., and Yamamura Mattson M. P. and Kater S. B. (1989) Excitatory and inhibitory neu- H. I. (1994) The cloning and expression of a human creatine rotransmitters in the generation and degeneration of hippocampal transporter. Biochem. Biophys. Res. Commun. 204, 419–427. neuroarchitecture. Brain Res. 478, 337–348. Stachowiak O., Dolder M., Wallimann T., and Richter C. (1998) Miller K., Halow J., and Koretsky A. P. (1993) Phosphocreatine Mitochondrial creatine kinase is a prime target of peroxynitrite- protects transgenic mouse liver expressing creatine kinase from induced modification and inactivation. J. Biol. Chem. 273, 16694– hypoxia and ischemia. Am. J. Physiol. 265, C1544–C1551. 16699. Molloy G. R., Wilson C. D., Benfield P., de Vellis J., and Kumar S. Steeghs K., Benders A., Oerlemans F., de Haan A., Heerschap A., (1992) Rat brain creatine kinase messenger RNA levels are high in Ruitenbeek W., Jost C., van Deursen J., Perryman B., Pette D., primary cultures of brain astrocytes and oligodendrocytes and low Bruckwilder M., Koudijs J., Jap P., Veerkamp J., and Wieringa B. in neurons. J. Neurochem. 59, 1925–1932. (1997) Altered Ca2ϩ responses in muscles with combined mito- Mo¨ller A. and Hamprecht B. (1989) Creatine transport in cultured cells chondrial and cytosolic creatine kinase deficiencies. Cell 89, 93– of rat and mouse brain. J. Neurochem. 52, 544–550. 103. Nicotera P., Leist M., and Ferrando-May E. (1998) Intracellular ATP, Stockler S., Holzbach U., Hanefeld F., Marquardt I., Helms G., Requart a switch in the decision between apoptosis and necrosis. Toxicol. M. H., Sˇnicke W., and Frahm I. (1994) Creatine deficiency in the Lett. 103, 139–142. brain: a new treatable inborn error of metabolism. Pediatr. Res. O’Gorman E., Beutner G., Dolder M., Koretsky A. P., Brdiczka D., and 36, 409–414. Wallimann T. (1997) The role of creatine kinase in inhibition of Stout A. K., Raphael H. M., Kanterewicz B. I., Klann E., and Reynolds mitochondrial permeability transition. FEBS Lett. 414, 253–257. I. J. (1998) Glutamate-induced neuron death requires mitochon- Passonneau J. V. and Lowry O. H. (1971) Metabolite flux in single drial calcium uptake. Nat. Neurosci. 1, 366–373. neurons during ischemia and anesthesia. Curr. Probl. Clin. Bio- Tarnopolsky M. A., Roy B. D., and MacDonald J. R. (1997) A chem. 3, 198–212. randomized, controlled trial of creatine monohydrate in patients Pettegrew J. W., Panchalingam K., Klunk W. E., McClure R. J., and with mitochondrial cytopathies. Muscle Nerve 20, 1502–1509. Muenz L. R. (1994) Alterations of cerebral metabolism in prob- Vandenberghe K., Gillis N., Van-Leemputte M., Van-Hecke P., Van- able Alzheimer’s disease—a preliminary study. Neurobiol. Aging stapel F., and Hespel P. (1997) Long term creatine intake is 15, 117–132. beneficial to muscle performance during resistance training. Pike C. J., Walencewicz A., Glabe C. G., and Cotman C. W. (1993) J. Appl. Physiol. 80, 452–457. Neurodegeneration induced by beta-amyloid peptides in vitro: the Wallimann T. and Hemmer W. (1994) Creatine kinase in non-muscle role of peptide assembly state. J. Neurosci. 13, 1676–1687. tissues and cells. Mol. Cell. Biochem. 133, 193–220. Pike C. J., Walencewicz-Wasserman A. J., Kosmoski J., Cribbs D. H., Wallimann T., Wyss M., Brdiczka D., Nicolay K., and Eppenberger Glabe C. G., and Cotman C. W. (1995) Structure–activity analyses of ␤-amyloid peptides: contributions of the ␤25–35 region to H. M. (1992) Intracellular compartmentation, structure and func- aggregation and neurotoxicity. J. Neurochem. 64, 253–265. tion of creatine kinase isoenzymes in tissues with high and fluc- Ponticos M., Lu Q. L., Morgan J. E., Hardie D. G., Partridge T. A., and tuating energy demands: the “phosphocreatine circuit” for cellular Carling D. (1998) Dual regulation of the AMP-activated protein energy homeostasis. Biochem. J. 281, 21–40. kinase provides a novel mechanism for the control of creatine Wang G. J., Randall R. D., and Thayer S. A. (1994) Glutamate-induced kinase in skeletal muscle. EMBO J. 17, 1688–1699. intracellular acidification of cultured hippocampal neurons dem- 2ϩ Poortmans J. R., Auquier H., Renaut V., Durussel A., Saugy M., and onstrates altered energy metabolism resulting from Ca loads. Brisson G. R. (1997) Effect of short-term creatine supplementa- J. Neurophysiol. 72, 2563–2569. tion on renal responses in men. Eur. J. Appl. Physiol. 76, 566– Whittingham T. S. and Lipton P. (1981) Cerebral synaptic transmission 567. during anoxia is protected by creatine. J. Neurochem. 37, 1618– Pulido S., Passaquin A. C., Leijendekker W. J., Challet C., Wallimann 1621. T. R., and Y¨ egg U. T. (1998) Creatine supplementation improves Wilken B., Ramirez J. M., Probst I., Richter D. W., and Hanefeld F. intracellular Ca2ϩ handling and survival in mdx skeletal muscle (1998) Creatine protects the central respiratory network of mam- cells. FEBS Lett. 439, 357–362. mals under anoxic conditions. Pediatr. Res. 43, 8–14. Quest A. F. G., Soldati T., Hemmer W., Perriard J. C., Eppenberger Woznicki D. T. and Walker J. B. (1980) Utilization of cyclocreatine H. M., and Wallimann T. (1990) Phosphorylation of chicken phosphate, an analogue of creatine phosphate, by mouse brain brain-type creatine kinase affects a physiologically important ki- during ischemia and its sparing action on brain energy reserves. netic parameter and gives rise to protein microheterogeneity in J. Neurochem. 34, 1247–1253. vivo. FEBS Lett. 269, 457–464. Wyss M., Smeitink J., Wevers R. A., and Wallimann T. (1992) Mito- Ruda M. Y., Samarenko M. B., Afonskaya N. I., and Saks V. A. (1988) chondrial creatine kinase: a key enzyme of aerobic energy metab- Reduction of ventricular arrhythmias by phosphocreatine olism. Biochim. Biophys. Acta 1102, 119–166. (Neoton) in patients with acute myocardial infarction. Am. Xu C. J., Klunk W. E., Kanfer J. N., Xiong Q., Miller G., and Pettegrew Heart J. 116, 393–397. J. W. (1996) Phosphocreatine-dependent glutamate uptake by Saks V. A., Chernousova G. B., Gukovsky D. E., Smirnov V. N., and synaptic vesicles. A comparison with ATP-dependent glutamate Chazov E. I. (1975) Studies of energy transport in heart cells. Eur. uptake. J. Biol. Chem. 271, 13435–13440. J. Biochem. 57, 273–290. Yankner B. A., Duffy L. K., and Kirschner D. A. (1990) Neurotrophic Saltarelli M. D., Bauman A. L., Moore K. R., Bradley C. C., and and neurotoxic effects of amyloid beta protein. Science 250, Blakely R. D. (1996) Expression of the rat brain creatine trans- 279–282.

J. Neurochem., Vol. 74, No. 5, 2000