Functions and effects of creatine in the central nervous system

Abstract

Creatine kinase catalyses the reversible transphosphorylation of creatine by ATP. In the cell, creatine kinase isoenzymes are specifically localized at strategic sites of ATP consumption to efficiently regenerate ATP in situ via phosphocreatine or at sites of ATP generation to build-up a phosphocreatine pool. Accordingly, the creatine kinase/phosphocreatine system plays a key role in cellular energy buffering and energy transport, particularly in cells with high and fluctuating energy requirements like neurons. Creatine kinases are expressed in the adult and developing human brain and spinal cord, suggesting that the creatine kinase/phosphocreatine system plays a significant role in the central nervous system. Functional impairment of this system leads to a deterioration in energy metabolism, which is phenotypic for many neurodegenerative and age-related diseases. Exogenous creatine supplementation has been shown to reduce neuronal cell loss in experimental paradigms of acute and chronic neurological diseases. In line with these findings, first clinical trials have shown beneficial effects of therapeutic creatine supplementation. Furthermore, creatine was reported to promote differentiation of neuronal precursor cells that might be of importance for improving neuronal cell replacement strategies. Based on these observations there is growing interest on the effects and functions of this compound in the central nervous system. This review gives a short excursion into the basics of the creatine kinase/phosphocreatine system and aims at summarizing findings and concepts on the role of creatine kinase and creatine in the central nervous system with special emphasis on pathological conditions and the positive effects of creatine supplementation.

Introduction

Cellular energy demand and supply are balanced and tightly regulated for economy and efficiency of energy use. Cells with high and fluctuating energy requirements, such as neurons, may increase the rate of ATP hydrolysis within seconds by several orders of magnitude, but intracellular ATP levels remain amazingly constant. This stability paradox [89], [90] can be explained by the action of immediately available, fast and efficiently working energy supporting and back-up systems that connect sites of energy consumption with those of energy production via phosphoryl transfer networks [65], [64], [148], [199]. In this respect, creatine (Cr) and the creatine kinase/phosphocreatine (CK/PCr) system have recently received increasing attention. A growing number of reports now provide evidence for the eminent importance of the CK/PCr-system and Cr metabolism for normal function of the brain, as well as under neuropathological conditions. Hence the present review aims at summarizing the function and role of the CK/PCr-system in the brain and spinal cord. We tried as much as possible to incorporate the most recent work in the field. For a more extensive coverage of the literature on Cr and the CK/PCr-system, the reader is referred to the following review articles by Brosnan and Brosnan [40], Schlattner et al. [159], Wallimann et al. [194], [196], and Wyss and Kaddurah-Daouk [205].

Creatine (N-aminoiminomethyl-N-methylglycine) is a guanidino compound synthesized from the amino acids arginine, glycine and methionine. Cr is taken up in diets containing fresh meat or fish. In addition, Cr can be endogenously synthesized by the liver, kidney, pancreas, and to some extent in the brain (see Section 1.6). CK, catalyzing the reversible transfer of the N-phosphoryl group from PCr to ADP to regenerate ATP, is a major enzyme of higher eukaryotes that deal with high and fluctuating energy demands to maintain cellular energy homeostasis and to guarantee stable, locally buffered ATP/ADP ratios [24], [129], [150], [151], [190], [193], [199], [207]. The interplay between cytosolic and mitochondrial CK isoenzymes (see Section 1.3) accomplishes multiple roles in cellular energy homeostasis [97], [148], [161], [159], [160], [197], [199]. Both isoenzymes contribute to the build-up of a large intracellular pool of PCr that represents an efficient temporal energy buffer and prevents a rapid fall in global ATP concentrations upon cell activation or sudden stress conditions [129], when the cytosolic CK equilibrates the cytosolic overall ATP/ADP ratio. Due to the specific localization of mitochondrial and cytosolic CK isoenzymes, the much faster diffusion rate of PCr as compared to ATP [163], [189], and the significantly higher diffusion rate of Cr compared to ADP [100], the CK/PCr-system make available for a spatial “energy shuttle” or “energy circuit”, bridging sites of ATP generation with sites of ATP consumption (Fig. 1).

For the understanding of the functioning of the CK/PCr-circuit, the presence of subcellular CK compartments are of importance. For example, a significant fraction of cytosolic CK is structurally and functionally associated or co-localized with different, structurally bound ATPases. These ATPases include, (i) different ion pumps in the plasma membrane, (ii) the actin-activated myosin ATPase of the contractile apparatus in muscle, where CK is located at the sarcomeric M-band and I-band of the myofibrils, and (iii) the calcium pump of the muscular sarcoplasmic reticulum. In all these cases, PCr is used for local in situ regeneration of ATP, which is directly channeled from CK to the consuming ATPase. At the ATP-generating side, a part of cytosolic CK is associated with glycolytic enzymes, and even more importantly the mitochondrial proteolipid complexes containing ubiquitous mitochondrial CK (uMt-CK) are coupled to oxidative ATP production (Fig. 1).

Tissue- and compartment-specific isoenzymes of CK do exist which is crucial to their functions in cellular energy metabolism [68]. Most vertebrate tissues express two CK isoenzyme combinations, either dimeric, cytosolic, muscle-type MM-CK together with mostly octameric sarcomeric mitochondrial sMt-CK, or alternatively, brain-type BB-CK, together with uMt-CK [199]. The CK isoenzyme combination, MM-CK with sMt-CK, is expressed in differentiated sarcomeric muscle, cardiac [69] and skeletal [203]. On the other hand, the combination BB-CK with uMt-CK is prominently expressed in brain [99], neuronal cells [39], retina photoreceptor cells [198], [202], hair cell bundles of the inner ear [168], smooth muscle [94], kidney [81], endothelial cells [53], spermatozoa [100] and skin [158]. CK isoforms were shown to be present throughout the central and peripheral nervous system in the fetal rat brain [48], [91], [99]. Using Western blot analysis, we were able to show high levels of BB-CK and uMt-CK expression in human fetal spinal cord [58], rhombencephalon, ventral mesencephalon, ganglionic eminence and cerebral cortex (Fig. 2). Hybrid cytosolic MB-CK, on the other hand, is expressed only transiently during muscle differentiation but persists at low levels in adult cardiac muscle (for reviews see [199], [195], [207]).

Octameric uMt-CK is localized in the cristae, as well as in the intermembrane space of mitochondria, preferentially at the contact sites between inner and outer mitochondrial membrane [95], [156], [159], [160], [202], [207] (Fig. 1). Mitochondrial and cytosolic CK have diverged million years ago [66], suggesting that compartmentalized CK isoenzymes have evolved very early during evolution in the context of functional coupling between uMt-CK and oxidative phosphorylation [101], [149], [190] and metabolite channeling [157], [159], [160].

The importance of the CK system for brain function has been highlighted by experiments using either CK knockout mice or by depletion of brain Cr by pharmacological intervention. Mice with a gene knockout of cytosolic BB-CK showed diminished open-field habituation, a slower learning curve in the water maze, and demonstrated a loss in hippocampal mossy fiber connections [96]. Undetectable PCr and 30% reduced Cr levels in the brain of double knockout transgenic mice, lacking both BB-CK and uMt-CK, have been reported [92]. These CK double knockout mice showed a much more severe phenotype, compared to the single CK isoenzyme knockout mice, since obviously, the lack of cytosolic BB-CK can somehow to some extent be compensated by the presence of mitochondrial uMt-CK and vice versa. The CK double knockout mice showed significantly reduced body and brain weights as compared to wild-type controls and were also behaviorally affected with severely impaired spatial learning, a lower nest-building activity and a diminished acoustic startle reflex [176]. Feeding of normal mice with the Cr analog, beta-guanidino propionic acid (GPA), a competitive inhibitor of the creatine transporter (CRT), resulted in a significant decrease of total Cr pools in muscle and brain, resulting in a muscle and behavioral phenotype [128]. In humans, new creatine-deficiency syndromes, affecting either endogenous Cr synthesis or Cr transport, have been discovered recently (for review see [156]). Patients suffering from this syndrome do have an almost complete lack of Cr in the brain and present with severe neurological symptoms, such as developmental and speech delay, epileptic seizures, autism and severe mental retardation (for details see below). Hence, either ablating the CK isoenzymes or inducing a marked reduction of their substrate in the brain, lead to similar and rather severe phenotypes. These observations provide strong evidence for the important physiological significance of the CK/PCr-system for normal brain function and indicate a need for a better understanding of Cr metabolism in the human body and particularly in the brain (for review see [205]).

Importantly to note, the brain, which constitutes only about 2% of the body mass, may spend up to 20% of the body's energy consumption [169]. A very high turnover of ATP is therefore necessary to maintain electrical membrane potentials, as well as signaling activities of the central and peripheral nervous system. Hence, energy production via oxidative phosphorylation and thus the production of ATP and PCr are critical to cerebral function. During physiological function of neurons, rapid changes in ATP demands are occurring while cellular energy reserves are small. An effective coupling of ATP-generating and ATP-consuming processes is needed to maintain a sufficiently high-energy transfer since cellular processes are widely distributed and sites of high-energy consumption are often localized at remote locations from the neuronal cell body, i.e., synapses [6]. For this reason, the CK/PCr-system is assumed to play a critical function in neuronal ATP metabolism [71], [86]. In line with this notion, several reports have demonstrated that the CK/PCr-circuit plays a key role in the energy metabolism of the brain and spinal cord [39], [45], [86], [195], [206]. Consequently, Cr depletion in brain is associated with disruption of neuronal functions, e.g., loss of hippocampal mossy fiber connection [92], and changes in mitochondrial structure, showing intramitochondrial uMt-CK-rich inclusion bodies [208] that are typical for several clinical pathological conditions, such as encephalomyopathies and mitochondrial myopathies (for review see [201]). As mentioned above, patients with Cr-deficiency syndrome show mental retardation, speech delay, autism and even brain atrophy [175].

Alimentary Cr, present in fresh fish and meat, is taken up by an intestinal CRT [138] and transported into the blood stream, where it mixes with endogenously synthesized Cr. Approximately 50% of daily Cr requirement in humans (totaling 3–4 g of Cr/day) is endogenously synthesized by a two-step synthesis involving the enzymes arginine:glycine amidino transferase (AGAT), producing guanidino acetate (GAA) as an intermediate, and S-adenosyl-l-methionine:N-guanidinoacetate methyltransferase (GAMT). The liver is the main organ of endogenous Cr synthesis. To get into the brain, Cr has to pass the blood–brain barrier (BBB), where CRTs are expressed and localized on the luminal and basal side of microcapillary endothelial cells, but not in the astrocytes sitting on these microcapillaries [36], [37]. Since the latter are lined with astrocytic feet, which apparently do not express CRT, the restricted transport of Cr from the blood through the BBB into the brain might only be possible through the limited surface of microcapillaries that are not covered by astrocytic endings (for discussion see [35]). This may be an explanation why Cr uptake into the brain and saturation of the endogenous Cr pool takes much longer, as compared to muscle [93]. After passaging through the BBB, Cr is then actively taken up from the extracellular fluid of the brain, by those cells that express the CRT, i.e., neurons and oligodendrocytes, but not astrocytes, which as known up to date are lacking the CRT [36], [37], [134] (see Fig. 3). In neural cells, Cr is charged-up by CK to high-energy PCr (for review see [205]). AGAT and GAMT can be detected in the embryonic [37] as well as in the adult [36] brain. Hence, there seems to be a potential for endogenous Cr synthesis in the brain [36]. Notably, AGAT and GAMT are not present only in astrocytes, but also in neurons and oligodendrocytes, giving the potential of Cr synthesis to all main cell types of the brain. This does not seem true for CRT, that is not expressed in astrocytes. Trafficking of Cr synthesized by astrocytes between astrocytes and neurons or oligodendrocytes has been suggested [36], [184] (Fig. 3). Furthermore, a recent finding demonstrated that Cr is not only synthesized and taken up by neurons, but also released in an action-potential dependent, excitotic manner, providing strong evidence for its role as a neuromodulator in the brain [5]. Notably, a number of important questions, concerning details of Cr metabolism, like regulation of trans-cellular Cr transport, uptake of Cr into the brain and intracellular trafficking and excitotic release of Cr after neuronal stimulation inside the brain, have still to be clarified in more detail.

The view that Cr exerts its functions exclusively via effects in cellular energy metabolism [205] and by enhancing the cellular energy status [82] cannot explain a number of recently reported findings (see below). Hence, Cr is assumed to have additional functions in the CNS. For example, a direct anti-apoptotic effect of elevated cellular Cr levels has been reported. In combination with the action of uMt-CK inside mitochondria, Cr prevented or delayed mitochondrial permeability transition pore opening, an early event in apoptosis [56], [133]. Moreover, Cr supplementation was demonstrated to have antioxidant properties via a mechanism involving a direct scavenging of reactive oxygen species [164] or alternatively, reducing the production of mitochondrially generated reactive oxygen species. The latter is facilitated by the stimulatory effects of Cr on mitochondrial respiration [101] that allows for efficient recycling of ADP inside mitochondria by uMt-CK, leading to tight coupling of mitochondrial respiration with ATP synthesis and suppression of reactive oxygen species formation [128]. Notably, protective effects of Cr against oxidant and UV stress has been detected in keratinocytes and on human skin [114]. Furthermore, Cr was reported to normalize mutagenesis of mitochondrial DNA and its functional consequences caused by UV irradiation of skin cells [23]. These findings point to effects of Cr for suppression of the generation of reactive oxygen species that lead to cell damage and inactivation of CK. Another recent study provided evidence that Cr-mediated neuroprotection can occur independent of changes in the bioenergetic status but rather by effects on cerebral vasculature leading to improved circulation in the brain [139]. Finally, a recent study demonstrated that Cr is able to protect cultured cells from hyper-osmotic shock by means of a significant increase of Cr uptake into cells, indicating that Cr can act as a compensatory osmolyte [4]. Indeed, Cr has been suggested as one of the main brain cell osmolytes based on experiments using hypo-osmotic perfusion of cortical brain tissue [33], [34].