Short-Term Cuprizone Feeding Induces Selective Amino Acid Deprivation with Concomitant Activation of an Integrated Stress Response in Oligodendrocytes

Abstract

Cuprizone [bis(cyclohexylidenehydrazide)]-induced toxic demyelination is an experimental approach frequently used to study de- and re-myelination in the central nervous system. In this model, mice are fed with the copper chelator cuprizone which leads to oligodendrocyte apoptosis and subsequent microgliosis, astrocytosis, and demyelination. The underlying mechanisms of cuprizone-induced oligodendrocyte death are still unknown. We analysed differences in amino acid levels after short-term cuprizone exposure (i.e., 4 days). Furthermore, an amino acid response (AAR) pathway activated in oligodendrocytes after cuprizone intoxication was evaluated. Short-term cuprizone exposure resulted in a selective decrease of alanine, glycine, and proline plasma levels, which was paralleled by an increase of apoptotic cells in the liver and a decrease of alanine aminotransferase in the serum. These parameters were paralleled by oligodendrocyte apoptosis and the induction of an AAR with increased expression of the transcription factors ATF-3 and ATF-4 (activating transcription factor-3 and -4). Immunohistochemistry revealed that ATF-3 is exclusively expressed by oligodendrocytes and localized to the nuclear compartment. Our results suggest that cuprizone-induced liver dysfunction results in amino acid starvation and in consequence to the activation of an AAR. We propose that this stress response modulates oligodendrocyte viability in the cuprizone animal model.

Introduction

Multiple sclerosis (MS), a chronic inflammatory and demyelinating disease, was first identified as a distinct neurological disorder by the French neurologist Jean-Martin Charcot, and is the most common inflammatory condition of the central nervous system (CNS) in young adults (Sellner et al. 2011; Kipp et al. 2012). It is widely accepted that MS lesions are autoimmune in origin, with myelin sheaths and oligodendrocytes being subject to an inflammatory attack (Fletcher et al. 2010; McFarland and Martin 2007). Consequently, animal models with T cell-driven inflammation and, to a certain extent, demyelination, were applied to study MS pathogenesis and new therapeutic options. The relevance of these animal models, e.g., experimental autoimmune encephalomyelitis, for distinct processes in MS pathology was critically reviewed in recent years (Marik et al. 2007; Zou et al. 1999; Kipp et al. 2009). Another commonly applied MS animal model is the toxin-induced cuprizone model (Skripuletz et al. 2011; Acs and Komoly 2012; Acs and Kalman 2012; Kipp et al. 2009). In this model, primary oligodendrocyte apoptosis is closely followed by microglia activation, astrogliosis, and finally demyelination. Besides demyelination, acute axonal damage can be observed in this model. The underlying biochemical and cellular mechanisms of these alterations are still poorly understood.

Chemically, cuprizone [oxalic acid bis(cyclohexylidene hydrazide)] is a well-known copper chelating agent, discovered and described in the early 1950s. Due to the highly chromogenic reaction of cuprizone with copper(II) ions, cuprizone was largely exploited for the quantitative determination of copper (Messori et al. 2007). Starting from the early 1970s, cuprizone attracted a lot of attention, mainly within the neuroscience scientific community, as this compound was reported to possess unique neurotoxic properties making it a valuable pharmacological tool for CNS demyelination and remyelination in various laboratory animals (Carlton 1966; Komoly et al. 1987; Ludwin 1978; Blakemore 1973; Kipp et al. 2009). However, despite many efforts, the underlying mechanisms of cuprizone-induced oligodendrocyte cell death are still not understood in detail. The most widely accepted hypothesis is that the toxic properties of cuprizone are due to a disturbance of cellular respiration, a key function of mitochondria. Feeding of cuprizone results in copper deficiency (Benetti et al. 2010) and it is believed that this induces a malfunction of the copper requiring electron carrier, cytochrome oxidase, a crucial component of the oxidative phosphorylation (Matsushima and Morell 2001; Hoppel and Tandler 1973). In line with this assumption, the activities of monoamine oxidase, cytochrome oxidase and succinyl dehydrogenase, all enzymes found in mitochondria, were found to be inhibited in the brain and liver of cuprizone-exposed animals (Venturini 1973; Petronilli and Zoratti 1990). Changes in enzyme activity were not observed in copper-chelated cuprizone-treated animals suggesting that the copper chelating nature of cuprizone is essential for the observed pathological changes (Carlton 1967). The fact that sole deficiency of copper does not provoke morphological or biochemical changes similar to those induced by cuprizone militates against any mechanism exclusively based on copper chelation (Goodman et al. 2006; Hoppel and Tandler 1973; Kumar et al. 2004). Consequently, it has been speculated that a metabolite of cuprizone is responsible for its toxic effect (Hoppel and Tandler 1973). However, there is no evidence that cuprizone undergoes metabolic transformation in the animal’s body so far.

It is well-known that cuprizone induces distinct changes in the liver (Kipp et al. 2009; Petronilli and Zoratti 1990; De and Subramanian 1982), as well as neurodegeneration, demyelination and glia activation. These pathological responses have also been described in patients suffering from severe liver diseases (Min et al. 2012; Chang et al. 2012; Utku et al. 2005; Ferreira et al. 2011; Kamei et al. 1997; Harding et al. 1995). The aim of the current study was to investigate whether metabolic liver dysfunction parallels early cuprizone-induced oligodendrogliopathy and might in part account for the toxic properties of cuprizone.

Materials and Methods

Animals and Induction of Demyelination

C57BL/6 male mice (Janvier, France) were bred and maintained in a pathogen-free environment. Animals underwent routine cage maintenance and microbiological monitoring according to the Federation of European Laboratory Animal Science Associations recommendations. Food and water were available ad libitum. Research and animal care procedures were approved by the Review Board for the Care of Animal Subjects of the district government (North Rhine-Westphalia, Germany). Demyelination was induced by feeding 8 week old (19–21 g) male mice a diet containing 0.25 % cuprizone (bis-cyclohexanone oxaldihydrazone; Sigma-Aldrich Inc., USA), mixed into a ground standard rodent chow for the indicated period as published (Kipp et al. 2011a; Clarner et al. 2012; Acs et al. 2009).

Animal Tissue Preparation

Tissue preparation was performed as previously described (Buschmann et al. 2012). For histological and immunohistochemical studies, mice were transcardially perfused with 2 % paraformaldehyde (pH 7.4). After overnight post-fixation in the same fixative, brains and livers were dissected, embedded, and then sectioned into 5 μm sections. For brain samples, cell parameter quantification was performed in slices corresponding to the levels 225 (designated as region 1; opening of third ventricle) and 265 (designated as region 2; ventral hippocampus area) of the mouse brain atlas by Sidman et al. For quantification of apoptotic cells in the liver, a random section in the middle-part of the organ was processed. For gene expression analyses, mice were transcardially perfused with ice-cold PBS, brains quickly removed, and the entire corpus callosum (CC) and/or Cx dissected using a stereo-microscopic approach as published previously (Buschmann et al. 2012; Pott et al. 2009). Tissues were immediately snap-frozen in liquid nitrogen and kept at −80 °C until further use.

Immunohistochemistry (IHC) and Cell Parameter Quantification

For IHC sections were rehydrated, if necessary unmasked by Tris/EDTA-buffer (pH 9.0) or citrate (pH 6.0) via heat-induced epitope retrieval, washed in PBS, and incubated overnight with the primary antibody diluted in PBS/5 % normal serum. Anti-active caspase 3 antibodies (Abcam, ab13847, 1:1,000) were used to detect apoptotic cells, whereas anti-ATF3 antibodies (Santa Cruz, sc188, 1:200) were used to visualize an integrated stress response. Slides were incubated with biotinylated secondary antibodies for 1 h, followed by peroxidase-coupled avidin–biotin-complex (ABC kit, Vector Laboratories, UK). 3,3-Diaminobenzidin (DAKO, Germany) was used as a peroxidase substrate. For quantification of apoptotic cells in haematoxylin–eosin (HE)-stained and anti-active caspase 3-stained brain sections, the entire CC till the border of both lateral ventricles was screened for the presence or absence of positive cells (in case of HE stains we looked for cells with condensed and fragmented nuclei, see Fig. 1a). Stained slices were digitalised and the area of the region of interest determined by ImageJ software after calibration. Density of apoptotic cells was finally calculated and is expressed as cells per mm². For quantification of apoptotic cells in the liver, slices were screened for the presence or absence of active caspase-3 positive cells in a random region of interest in the section. The same slices were digitalised and the area of the region of interest determined by ImageJ software after calibration. Density of apoptotic cells was finally calculated and is expressed as cells per mm².

Fig. 1.

a HE-stained sections (CC) in control animals and cuprizone fed animals (0.25 %, 4 days) are shown on the right side. Quantification of apoptotic cells within the CC is shown on the left side. The arrow highlights an apoptotic cell (fragmented and condensed cell nucleus). b IHC against active caspase 3 (CC) after 4 days cuprizone treatment is shown on the right side. Quantification of active caspase 3-expressing cells within the CC is shown on the left side. Note that several apoptotic cells are present already after 2 days cuprizone intoxication in both regions of the CC. c IF double labelling against active caspase-3 and APC indicating an apoptotic oligodendrocyte

Immunofluorescence (IF) Double Labelling and Cell Parameter Quantification

For IF double labelling, sections were rehydrated, unmasked by Tris/EDTA buffer and heating, blocked with PBS containing 2 % heat-inactivated foetal calf serum (Gibco, Germany) and 1 % bovine serum albumin, and incubated overnight at 4 °C in a wet chamber with the indicated combination of primary antibodies diluted in blocking solution. Anti-APC (mouse IgG, Calbiochem, OP-80, 1:250) was either combined with anti-active caspase 3 (rabbit IgG, Abcam, ab13847, 1:250) to detect apoptotic oligodendrocytes or with anti-ATF3 (rabbit IgG, Santa Cruz, sc-188, 1:1,000) to visualize an integrated stress response in oligodendrocytes. After washing, sections were subsequently incubated with a combination of fluorescent anti-mouse secondary antibodies (Alexa Fluor 488 goat IgG; Invitrogen, Germany; 1:500) and fluorescent anti-rabbit secondary antibodies (Alexa Fluor 546 goat IgG; Invitrogen, Germany; 1:500) diluted in blocking solution. Sections were then incubated with Hoechst 33342 (Invitrogen, Germany; 1:10,000) diluted in PBS for nuclear staining. In order to exclude unspecific binding of the secondary antibodies to the primary ones, appropriate negative controls were performed by first incubating sections with the primary antibody of murine origin (anti-APC) and subsequently incubating these sections with the fluorescent anti-rabbit secondary antibody. Sections exposed to anti-ATF3 or anti-active caspase 3 antibodies of rabbit origin were incubated with fluorescent anti-mouse secondary antibodies. Stained and processed sections were analysed under the fluorescence microscope Leica DMI6000B, Leica DFC365FX 1.4 mp (monochrome digital camera).

RT-qPCR

Gene expression was measured using the RT-qPCR technology (BioRad, Germany), 2× SensiMix Plus SYBR & Fluorescein (Quantace, Germany) and a standardized protocol as described previously (Braun et al. 2009; Acs et al. 2009). Primer sequences, abbreviations and individual annealing temperatures are shown in Table 1. Relative quantification was performed using the ΔCt method which results in ratios between target genes and housekeeping reference gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Melting curves and gel electrophoresis of the PCR products were routinely performed to determine the specificity of the PCR reaction (data not shown).

Table 1.

Sequence of primers

Gene	Protein	S/AS	Sequence
Plp1	PLP1	S	5′-TGGCGACTACAAGACCACCA-3′
		AS	5′-GACACACCCGCTCCAAAGAA-3′
Mag	MAG	S	5′-GAGTTTGCCCCCATAATCCT-3′
		AS	5′-TCTCCGTCTCATTCACAGTCA C-3′
Atf3	ATF3	S	5′-GAGGATTTTGCTAACCTGACACC-3′
		AS	5′-TTGACGGTAACTGACTCCAGC-3′
Atf4	ATF4	S	5′-CCTGAACAGCGAAGTGTTGG-3′
		AS	5′-TGGAGAACCCATGAGGTTTCAA-3′
Gapdh	GAPDH	S	5′-TGTGTCCGTCGTGGATCTGA-3′
		AS	5′-CCTGCTTCACCACCTTCTTGA-3′

Metabolite Quantification and Determination of Liver Parameters

All plasma samples were measured in one batch using the AbsoluteIDQ™ kit p180 (BIOCRATES Life Sciences AG, Innsbruck, Austria). Out of 10 μL plasma, amino acid levels were quantified simultaneously by electrospray ionization (ESI), flow injection analysis, tandem mass spectrometry (FIA–MS/MS), and ESI liquid chromatography tandem mass spectrometry (LC–MS/MS). Details of the procedures of the AbsoluteIDQ™ p180 kit and abbreviations used are published elsewhere (Illig et al. 2010; Jourdan et al. 2012; Suhre et al. 2011). Sample handling was performed by a Hamilton Micro Lab Star robot (Hamilton Bonaduz AG, Bonaduz, Switzerland) and an Ultravap nitrogen evaporator (Porvair Sciences, Leatherhead, U.K.). Mass spectrometric (MS) analyses were carried out using an API 4000 mass spectrometer (AB Sciex Deutschland GmbH, Darmstadt, Germany) coupled to a 1200 Series HPLC (Agilent Technologies Deutschland GmbH, Böblingen, Germany) and a HTC PAL auto sampler (CTC Analytics, Zwingen, Switzerland) controlled by the software Analyst 1.5.1. Measurements were performed as described in the manufacturer manual UM-P180. Analytical specifications for LOD and evaluated quantification ranges, further LOD for semi-quantitative measurements, identities of quantitative and semi-quantitative metabolites, specificity, potential interferences, linearity, precision and accuracy, reproducibility and stability were described in Biocrates manual AS-P180. The LODs were set to three times the values of the zero samples (PBS). The LLOQ and ULOQ were determined experimentally by Biocrates. Data evaluation for quantification of metabolite concentrations and quality assessment was performed with the MetIDQ™ software package which is part of the AbsoluteIDQ™ kit p180. Concentrations were calculated with reference to appropriate internal standards as detailed by the manufacturer. The methods of the AbsoluteIDQ™ p180 kit have been proven to be in conformance with the FDA Guideline “Guidance for Industry—Bioanalytical Method Validation, May 2001” which implies proof of reproducibility within a given error range. Concentrations of the analysed metabolites are given in μmol/L. Metabolite measurements were performed at the Genome Analysis Centre of the Helmholtz Zentrum München, Germany.

Quantification of liver parameters was performed in serum samples using routine methods established in the Institute for Laboratory Animal Science and Experimental Surgery, RWTH-Aachen University.

Statistical Analysis

At least six mice per experimental group were included for gene expression analyses by RT-qPCR and amino acid quantification, whereas at least five animals were included for determination of immunohistochemical parameters in the brain. Group size for metabolic parameters was N = 5 whereas group size for quantification of apoptotic cells in the liver was N = 4. Differences between groups were analysed by Mann–Whitney test using GraphPad Prism (GraphPad Software Inc, USA). All data are given as arithmetic mean ± SEM unless stated otherwise. P values ≤0.05 were considered to be statistically different. If not stated otherwise * indicates P values ≤0.05, ** indicates P values ≤0.01, and *** indicates P values ≤0.001, respectively.

Results

We have recently demonstrated that oligodendrocyte death starts early after initiation of the cuprizone diet (Buschmann et al. 2012). To verify early cuprizone-induced oligodendrocyte loss, mice were fed cuprizone (0.25 %) for 2 and 4 days, and brain sections were immunohistochemically stained for active caspase-3 or histochemically stained with HE to visualize the activation of apoptotic pathways. Furthermore, mRNA levels of the myelin marker proteins proteolipid protein (PLP) and myelin-associated glycoprotein (MAG) were determined in the isolated CC by means of RT-PCR.

As demonstrated in Fig. 1, oligodendrocyte death starts early after initiation of the cuprizone diet. Numerous apoptotic oligodendrocytes (i.e., condensed and/or fragmented nuclei of cells in a chain-like formation in HE-stained sections, arrow in Fig. 1a) were seen after 2 and 4 days in the CC of cuprizone fed animals. In region 1, mean numbers of apoptotic cells were 38.1 ± 6.5 after 2 days and 36.9 ± 5.6 after 4 days of cuprizone intoxication, respectively. In region 2, mean numbers of apoptotic cells were 22.22 ± 2.0 after 2 days and 31.6 ± 4.7 after 4 days of cuprizone intoxication, respectively (Fig. 1a). The apoptotic nature of these cells was further verified by anti-active caspase 3 staining (arrow in Fig. 1b). In region 1, mean numbers of caspase-3⁺ cells were 71.4 ± 16.0 after 2 days and 76.9 ± 6.3 after 4 days of cuprizone intoxication, respectively. In region 2, mean numbers of caspase-3⁺ cells were 54.7 ± 9.7 after 2 days and 62.3 ± 12.6 after 4 days of cuprizone intoxication, respectively (compare to bars in Fig. 1b). Control animals did not show apoptotic cells. Additional flourescent double-labelling experiments revealed that active caspase-3 positive cells express the mature oligodendrocyte marker protein APC (Fig. 1c). We further investigated the effect of cuprizone treatment on PLP (myelin) 1 and MAG gene expression in the isolated CC after a 4 day treatment period. As shown in Fig. 2, mRNA levels of both proteins were significantly downregulated in the investigated white matter tract. Expression levels declined to 7.6 ± 0.8 % for PLP and 7.9 ± 0.6 % for MAG compared to control levels (100 %), respectively. These data convincingly demonstrate that cuprizone intoxication induces a fast deterioration of oligodendrocyte physiology and function.

Fig. 2.

Real-time RT-sqPCR expression analysis of PLP and MAG within the entire CC is presented. Each bar represents the averaged fold induction compared to untreated control mice. Values were normalized against a housekeeping gene (GAPDH) and expressed as percentage of controls

The liver plays an important role in amino acid metabolism (Sookoian and Pirola 2012; Montejo Gonzalez et al. 2011; Yoshizawa 2012). Disturbances in amino acid levels can induce a cellular stress response called integrated stress response (Sikalidis et al. 2011; Sayers et al. 2013; Kilberg et al. 2009, 2012). Thus, amino acid levels in the blood of control and 4 days cuprizone-exposed animals were determined. None of the nine essential amino acids, namely valine, methionine, tryptophane, isoleucine, leucine, threonine, lysine, phenylalanine, and histidine, were changed in the plasma of cuprizone-fed animals compared to controls (Fig. 3). In contrast, the three non-essential amino acids, alanine (Ala), glycine (Gly), and proline (Pro), displayed significant lower levels in the plasma of cuprizone fed animals compared to controls (Fig. 4). Ala levels declined from 369.7 ± 66.4 μM in control animals to 246.7 ± 11.8 μM in cuprizone fed animals, Gly from 226.5 ± 19.3 μM in control animals to 163.8 ± 7.9 μM in cuprizone fed animals, and Pro from 104.6 ± 21.4 μM in control animals to 65.7 ± 2.3 μM in cuprizone fed animals. Levels of all other non-essential amino acids did not significantly differ between control and cuprizone fed animals, namely serine (Ser), asparagine (Asn), ornithine (Orn), tyrosine (Tyr), arginine (Arg), aspartic acid (Asp), glutamic acid (Glu), and glutamine (Gln) (Fig. 5).

Fig. 3.

Levels of essential amino acids in the plasma of control and cuprizone fed animals (0.25 %, 4 days) are shown. Note that none of the essential amino acid levels differ between control and cuprizone intoxicated groups. For abbreviations see text

Fig. 4.

Levels of non-essential amino acids in the plasma of control and cuprizone fed animals (0.25 %, 4 days) are shown. Note that three of the non-essential amino acids (compare also Fig. 5) Ala, Gly and Pro display reduced plasma levels in the blood of cuprizone-intoxicated animals

Fig. 5.

Levels of non-regulated, non-essential amino acids in the plasma of control and cuprizone fed animals (0.25 %, 4 days) are shown. For abbreviations see text

To further investigate whether liver integrity is substantially disturbed after short-term cuprizone exposure, we performed a set of liver function tests in the serum of cuprizone-fed animals. Levels of alkaline phosphatase, aspartate aminotransferase, gamma glutamyl transpeptidase as well as total protein and total billirubin did not differ between control and cuprizone-exposed animals (data not shown). Interestingly, levels of alanine aminotransferase (ALT) were found to be significantly lower after 1 week cuprizone exposure (Fig. 6a). Furthermore, we deemed it mandatory to screen for apoptotic cells in the liver. As shown in Fig. 6b, we observed higher numbers of active caspase-3⁺ cells after short-term cuprizone exposure. These results confirm that cuprizone induces stress in liver cells. However, this cell stress is not paralleled by increased levels of liver enzymes in the serum as usually observed under sever pathological liver conditions.

Fig. 6.

a ALT levels in the serum after 1 week cuprizone exposure. b A representative IHC of active caspase-3 on the right side whereas the quantification of positive cells is shown on the right side. Note that more apoptotic cells can be found in the liver of cuprizone-exposed animals

As mentioned above, amino acid deprivation can induce an integrated stress response. The two major transcription factors which are activated in response to amino acid deprivation are ATF3 and ATF4 (Fu and Kilberg 2013). In the next step, we have therefore tested the hypothesis that selective cuprizone-induced amino acid deprivation induces ATF3 and ATF4 expression in the brain after cuprizone-induced demyelination. As demonstrated in Fig. 7a (left side), ATF3 expression levels were significantly increased after short-term cuprizone exposure by ~100-fold compared to control levels. ATF4 mRNA levels were slightly increased by ~50 % compared to controls. Since oligodendrocytes are particularly vulnerable in this animal model (Buschmann et al. 2012; Kipp et al. 2011b), we speculated that oligodendrocytes are the target cell population up-regulating ATF3 expression in the brain of cuprizone-exposed animals. As shown in Fig. 7a (right side), control animals displayed very weak immunoreactivity for ATF3 in the CC. The staining was mainly located in the perinuclear space (arrows in Fig. 7a) indicating the inactivity of this transcription factor. In sharp contrast, strong immunoreactivity for ATF3 protein was observed after 4 days cuprizone exposure. Most of the anti-ATF3 signal was observed in the nuclear compartment indicating the activity of ATF3. Many of the ATF3-expressing cells were found to be inter-fascicular oligodendrocytes (arrowhead in Fig. 7a). To investigate whether ATF3 expression is restricted to oligodendrocytes, we performed double-labelling experiments. As demonstrated in Fig. 7b, most if not all ATF3 expressing cells express the mature oligodendrocyte marker protein APC.

Fig. 7.

a Real-time RT-sqPCR expression analyses of the transcription factors ATF3 and ATF4 within the entire CC and cortex is presented on the left side. Each bar represents the averaged fold induction over untreated control mice. Values were normalized against a housekeeping gene (GAPDH) and expressed as % of controls. IHC against ATF3 (CC) in control animals and after 4 days cuprizone treatment is shown on the right side. Arrows indicate perinuclear ATF3 in control animals, the arrowhead shows nuclear localization of ATF3 after 4 days cuprizone exposure in inter-fascicular oligodendrocytes. b IF double labelling against ATF3 and APC is shown on the right side indicating nuclear localized ATF3 in oligodendrocytes after 4 days of cuprizone intoxication. Number of ATF3 expressing cells quantified in fluorescence-labelled sections is shown on the left side. Area within the column indicates the percentage of ATF3-expressing cells which express as well APC (~90 %)

Discussion

An animal model to study the correlation of disturbed oligodendrocyte function and demyelination is the toxic demyelination model “cuprizone”. In this model, primary oligodendrocyte apoptosis is closely followed by microglia activation, astrogliosis, and finally active demyelination (Kipp et al. 2009; Acs and Komoly 2012; van der Star et al. 2012; Blakemore 1972). Here, we confirm that oligodendrocyte apoptosis starts few days after initiation of the cuprizone diet. Although most studies focus on later time points in this model (Acs et al. 2009; Bruck et al. 2012; Clarner et al. 2012; Hussain et al. 2013), early cuprizone-induced oligodendrocyte dysfunction has already been described in the early 1970s (Kesterson and Carlton 1971). More recently, Hesse and colleagues demonstrated that oligodendroglial cell death and downregulation of myelin genes starts days after initiation of the cuprizone diet and weeks before demyelination is apparent (Hesse et al. 2010). These findings were confirmed by other research groups including our (Kang et al. 2012; Buschmann et al. 2012). Gudi and colleagues demonstrated a strong upregulation of the expression of distinct growth factors such as neuregulin-1 and glial cell line-derived neurotrophic factor after 1 week of cuprizone exposure (Gudi et al. 2011).

Although it is in our days well appreciated that cuprizone intoxication rapidly induces oligodendrocyte stress and subsequent apoptosis, the underlying mechanisms are just partly understood. Already in 1969, it was observed that cuprizone consistently induced the formation of enlarged mitochondria in hepatocytes and oligodendrocytes (Suzuki and Kikkawa 1969; Blakemore 1972; Suzuki 1969). Since cuprizone is chemically a copper chelator and copper is an important co-factor for enzymes of the mitochondrial respiratory chain, such as the cytochrome c oxidase (complex IV) (Zeng et al. 2007; Russanov and Ljutakova 1980), it has been postulated that cuprizone-induced copper deficiency impairs mitochondrial activity, and in consequence homoeostasis in oligodendrocytes. The particular vulnerability of oligodendrocytes to this intoxication was linked to their vast amount of myelin synthesis, although this hypothesis lacks any experimental evidence. However, systemic copper deficiency does not result in similar histopathological changes in the brain as observed in the cuprizone model (Kumar et al. 2011; Bertfield et al. 2008). It was, therefore, suggested that other effects might in addition account for copper-induced cytotoxicity in oligodendrocytes. Since cuprizone affects liver function, we investigated whether cuprizone alters plasma amino acid levels which critically depend on hepatocyte function (Sookoian and Pirola 2012; Montejo Gonzalez et al. 2011; Yoshizawa 2012). As demonstrated in this work, three of the non-essential amino acids (Ala, Gly, and Pro) displayed lower levels in the plasma of cuprizone-fed animals compared to controls. This deficiency might impact oligodendrocyte physiology. For example, myristoylated alanine-rich C-kinase substrate (MARCKS), a major substrate of activated protein kinase C, has been linked to oligodendrocyte development. Alanine deficiency affects the activity of MARCKS (Bhat et al. 1995; Siskova et al. 2006) and therefore maybe oligodendrocyte function as well. Gly triggers calcium influx in oligodendrocytes (Belachew et al. 2000) which might modulate their physiology. Finally, a Pro rich protein, namely Prmp, has been shown to be important for linking the extracellular matrix to the actin cytoskeleton in oligodendrocytes (Nielsen et al. 2006). Future studies, however, have to address whether a deficiency in these amino acids indeed influences oligodendrocyte development and/or their function.

It is well-known that the depletion of individual amino acids at the cellular level can trigger a cellular stress response which is called “the amino acid response pathway (AAR)” (Kilberg et al. 2009). The general control nonderepressible 2 kinase acts as a sensor of amino acid levels by binding uncharged transfer tRNA which leads to phosphorylation of the alpha subunit of the translation initiation factor eIF2 and increased translation of the activating transcription factor-4 (ATF4). Accumulation of ATF4 protein activates a broad spectrum of downstream target genes, including ATF3 (Pan et al. 2007). We assume that the deprivation of Ala, Gly, and Pro after cuprizone exposure triggers AAR in oligodendrocytes which seem to be specifically vulnerable to reduced amino acid levels. Gene expression studies and IHC convincingly showed that ATF3 is activated especially in oligodendrocytes after short-term cuprizone administration. Results of different studies suggest that the ATF3 signalling cascade is involved in pro-apoptotic effects in various cell types and experimental models (Liu et al. 2012; Song et al. 2011; Yoon et al. 2011). However, pro-survival effects of ATF3 have been reported as well (Lv et al. 2011; Zhang et al. 2011; Kiryu-Seo et al. 2010). Interestingly, ATF3 expression was also detected in one patient with Marburg’s variant of MS (Lovas et al. 2010), in an animal model for Pelizaeus–Merzbacher Disease (Southwood et al. 2002), and after lysolecithin-induced demyelination (Kiryu-Seo et al. 2010). Future studies have to show the relevance of ATF3 induction for oligodendrocyte death in the cuprizone model, as well as in MS patients.

In summary, we demonstrated that early cuprizone-induced oligodendrocyte apoptosis is paralleled by a selective reduction in amino acid levels and that this event is temporally linked to an AAR in oligodendrocytes as evidenced by the nuclear expression of the transcription factor ATF3.