AGC1-malate aspartate shuttle activity is critical for dopamine handling in the nigrostriatal pathway

Authors

  • Irene Llorente-Folch,

    1. Departamento de Biología Molecular, Centro de Biología Molecular Severo Ochoa UAM-CSIC, and CIBER de Enfermedades Raras (CIBERER), Universidad Autónoma de Madrid, Madrid, Spain
    Search for more papers by this author
  • Ignasi Sahún,

    1. Programme of Genes and Disease, Center for Genomic Regulation, and CIBER de Enfermedades Raras (CIBERER), Barcelona, Spain
    Search for more papers by this author
  • Laura Contreras,

    1. Departamento de Biología Molecular, Centro de Biología Molecular Severo Ochoa UAM-CSIC, and CIBER de Enfermedades Raras (CIBERER), Universidad Autónoma de Madrid, Madrid, Spain
    Search for more papers by this author
  • María José Casarejos,

    1. Department of Neurobiology and CIBERNED, Hospital Ramón y Cajal, Carretera de Colmenar km 9,1, Madrid, Spain
    Search for more papers by this author
  • Josep María Grau,

    1. Department of Internal Medicine and Pathology, Hospital Clinic, University of Barcelona Medical School, Barcelona, Spain
    Search for more papers by this author
  • Takeyori Saheki,

    1. Institute of Resource Development and Analysis, Kumamoto University 2-2-1 Honjo, Kumamoto, Japan
    Search for more papers by this author
  • María Angeles Mena,

    1. Department of Neurobiology and CIBERNED, Hospital Ramón y Cajal, Carretera de Colmenar km 9,1, Madrid, Spain
    Search for more papers by this author
  • Jorgina Satrústegui,

    1. Departamento de Biología Molecular, Centro de Biología Molecular Severo Ochoa UAM-CSIC, and CIBER de Enfermedades Raras (CIBERER), Universidad Autónoma de Madrid, Madrid, Spain
    Search for more papers by this author
  • Mara Dierssen,

    1. Programme of Genes and Disease, Center for Genomic Regulation, and CIBER de Enfermedades Raras (CIBERER), Barcelona, Spain
    Search for more papers by this author
  • Beatriz Pardo

    Corresponding author
    • Departamento de Biología Molecular, Centro de Biología Molecular Severo Ochoa UAM-CSIC, and CIBER de Enfermedades Raras (CIBERER), Universidad Autónoma de Madrid, Madrid, Spain
    Search for more papers by this author

Address correspondence and reprint requests to Beatriz Pardo, Centro de Biología Molecular Severo Ochoa UAM-CSIC, and CIBER de Enfermedades Raras (CIBERER), Universidad Autónoma de Madrid, 28049-Madrid, Spain. E-mail: bpardo@cbm.uam.es

Abstract

The mitochondrial transporter of aspartate-glutamate Aralar/AGC1 is a regulatory component of the malate-aspartate shuttle. Aralar deficiency in mouse and human causes a shutdown of brain shuttle activity and global cerebral hypomyelination. A lack of neurofilament-labeled processes is detected in the cerebral cortex, but whether different types of neurons are differentially affected by Aralar deficiency is still unknown. We have now found that Aralar-knockout (Aralar-KO) post-natal mice show hyperactivity, anxiety-like behavior, and hyperreactivity with a decrease of dopamine (DA) in terminal-rich regions. The striatum is the brain region most affected in terms of size, amino acid and monoamine content. We find a decline in vesicular monoamine transporter-2 (VMAT2) levels associated with increased DA metabolism through MAO activity (DOPAC/DA ratio) in Aralar-KO striatum. However, no decrease in DA or in the number of nigral tyrosine hydroxylase-positive cells was detected in Aralar-KO brainstem. Adult Aralar-hemizygous mice presented also increased DOPAC/DA ratio in striatum and enhanced sensitivity to amphetamine. Our results suggest that Aralar deficiency causes a fall in GSH/GSSG ratio and VMAT2 in striatum that might be related to a failure to produce mitochondrial NADH and to an increase of reactive oxygen species (ROS) in the cytosol. The results indicate that the nigrostriatal dopaminergic system is a target of Aralar deficiency.

Abbreviations used
3-MT

3-methoxy-tyramine

5-HT

serotonin

AGC1

aspartate-glutamate carrier

COMT

catechol-ortho-methyl-transferase

DA

dopamine

DARPP32

dopamine and cAMP regulated phosphoprotein of 32 KDa

DAT

dopamine transporter

DOPAC

3,4 dihydroxy-phenyl acetic acid

GSH

reduced glutathione

GSSG

oxidized glutathione

HVA

homovanillic acid

MAO

monoamine oxidase

MAS

malate-aspartate shuttle

NAA

N-acetylaspartate

NA

noradrenaline

PD

Parkinson′s disease

PND

post-natal day

ROS

reactive oxygen species

SERT

serotonin transporter

TH

tyrosine hydroxylase

Tyr

tyrosine

VMAT2

vesicular monoamine transporter 2

Aralar is the brain isoform of the mitochondrial transporter of aspartate/glutamate mainly expressed in neurons (Del Arco and Satrústegui 1998; Ramos et al. 2003; Pardo et al. 2006, 2011) and its expression is increased during maturation in parallel to malate-aspartate shuttle (MAS) activity (Ramos et al. 2003). Aralar deficiency leads to a loss of respiration on malate plus glutamate, a shutdown of MAS activity, and a drop in brain and neuronal aspartate levels (Jalil et al. 2005). Neurons from Aralar-KO mice have a clear metabolic impairment in glucose oxidation because of the lack of a functional MAS, which results in lower pyruvate levels and increased lactate production (Pardo et al. 2011). In intact cultured neurons, the maximal respiratory capacity of Aralar-KO neurons is clearly reduced as compared with control (Gómez-Galán et al. 2011), reflecting the limitation in pyruvate supply to mitochondria in the absence of a functional MAS. Other NADH shuttle systems, as the glycerophosphate shuttle (GPS) may partially compensate for the lack of Aralar. Indeed, although GPS was thought to be non-functional in neurons and astrocytes which lack the major isoform of cytosolic glycerophosphate dehydrogenase (Nguyen et al. 2003), the mRNA of a second cytosolic glycerol-P-dehydrogenase (GPD1-like), functionally identified in heart (Liu et al. 2010), is expressed in neurons and astrocytes (Cahoy et al. 2008). However, even if GPS exerts some compensation, it is unable to prevent the limitation in pyruvate supply caused by Aralar deficiency.

Aralar-KO mice show a drop of N-acetylaspartate (NAA), hypomyelination, and a progressive failure to synthesize glutamine in brain astrocytes, suggesting that glutamatergic neurotransmission may be compromised in the older animals (Jalil et al. 2005; Pardo et al. 2011). These mice present motor problems, tremor, seizures, and premature death (Jalil et al. 2005). Impaired development or degeneration of neuronal processes unrelated to myelin deficits has been observed in Aralar-KO mouse brain (Ramos et al. 2011). Recently, a patient with an homozygous loss of function mutation in SLC25A12 was reported to show a loss of mitochondrial respiration on malate plus glutamate, low NAA levels, hypomyelination, arrested psychomotor development, hypotonia, and seizures (Wibom et al. 2009). These findings are relevant as the main features of Aralar deficiency (reduced NAA levels and hypomyelination) are common in mouse and human, and support the importance of the Aralar-KO mouse for the study of the global cerebral hypomyelination (OMIM ID #612949). To gain further insight into the behavioral deficits of the Aralar-KO mouse, in comparison with human AGC-1 deficiency, we have studied in more detail motor abilities and general behavior in Aralar-KO mice from post-natal day 13 (PND13) to PND20. In addition, we have analyzed neurochemical changes in specific brain areas of the Aralar-KO mouse that could be responsible for the behavioral failures observed. Further studies have been performed on healthy and adult mice expressing only half-a-dose Aralar.

Our data indicate that Aralar-KO mice show deficits in motor coordination, ataxic gait, increased reactivity, hyperactivity, and anxiety-like behavior, together with a pronounced decline in the striatal levels of DA. Besides, both PND20 Aralar-KO and 18-month-old Aralar-hemizygous mice display DA mishandling in striatum. The present results reveal a high susceptibility of DAergic neurons, specifically those DAergic groups of the nigrostriatal system, to Aralar-MAS dysfunction. A large body of evidence supports that nigrostriatal DAergic neurons are highly vulnerable to oxidative stress (Mena et al. 1993, 1997, 2002; Pardo et al. 1995; Canals et al. 2001), and to impairments of energy metabolism (Zeevalk et al. 1997; Watabe and Nakaki 2008; Pickrell et al. 2011; Sterky et al. 2012). Our results evidence that the striatum is a preferential target of Aralar deficiency which leads to a reduction and/or mishandling of DA and suggest a role of oxidative stress caused by Aralar deficiency as the origin of DA mishandling in striatum.

Materials and methods

Methods including histological and immunohistochemical studies in brain and histomorphological studies of muscle are described in Supplementary material.

Animals

Male SVJ129 x C57BL6 mice carrying a deficiency for ARALAR expression (Aralar wild-type, WT; Aralar heterozygous, HT; and Aralar Knock-out, KO) were obtained from Lexicon Pharmaceuticals, Inc. (The Woodlands, TX, USA) (1). The mice were housed in a humidity- and temperature-controlled room on a 12-h light/dark cycle, receiving water and food ad libitum. Genotype was determined by PCR using genomic DNA obtained from tail or embryonic tissue samples (Nucleospin tissue kit; Macherey-Nagel, Düren, Germany) as described previously (Jalil et al. 2005). All the experimental protocols used in this study were approved by the local Ethics Committees at the Center of Molecular Biology ‘Severo Ochoa’, Autónoma University (UAM), Madrid, and at the Center for Genomic Regulation, Barcelona.

General procedure for post-natal observations

All the pregnant females were allowed to deliver spontaneously. Each pup was checked for gross abnormalities and the day after delivery was designated as PND1 of age for neonates (estimation error on time of birth ± 8 h). The pups were individually marked with ink and were nursed by their natural dams until weaning. During the testing protocol, whole litters were separated from the dam less that 10 min and maintained in a warm environment. Males and females were pooled to perform the neurodevelopmental screening. All the measures were performed between 5:00 a.m. and 8:00 a.m. All experiments procedures were approved by the Animal Care Committee of the Centre for Genomic Regulation. For the developmental screening 42 animals, males and females, were employed from five different litters.

Neurobehavioral development

Reflex and sensory function

Visual placing

After the opening of the eyes, the pup was suspended by the tail and lowered toward the tip of a pencil without the vibrissae touching it, on PND17. The response was considered positive when the paws were extended to touch it.

Blast response

Exaggerated jumping or running behavior in response to a gentle puff of air, on PND11.

Tactile orientation

The test assessed the head turning (orienting) response triggered by the application to one side of the perioral area a cotton Q-tip, beginning on PND11.

Vibrissae orientation

The pups were suspended by the tail and lowered toward the tip of a cotton Q-tip. At contact of the cotton with the vibrissae, the pup raised its head and performed a placing response, beginning on PND8.

Preyer reflex/startle response

The response of the pups to a moderate sound burst consisting of a moderately brink flick of the pinna or startle response was recorded, beginning on PND11.

Toe pinch

The test assessed the presence or absence of withdrawal answer against a mild painful stimulus, exerting pressure in a hind paw with the fingers, on PND17.

Reaching response

The animal is held by the tail above a flat surface and it is noted if the forepaws are stretched out to make contact with the surface, on PND17.

Touch escape

Response of the animal to a finger stroke from above was recorded and scored, beginning on PND13, as follow: 0 = no response, 1 = moderate (rapid response to light stroke); and 2 = vigorous (escape response to approach).

Homing test

On PND19, individual pups were transferred to a cage containing new sawdust with a bite portion of sawdust of the home litter ‘goal arena’. The pups were located at the opposite side of the goal arena, near to the wall. The time taken to reach the home litter sawdust was recorded (cut-off time 60 s).

Paw print test

To examine the step patterns of the hind limbs during forward locomotion, mice were required to traverse a straight, narrow tunnel. The experiments to evaluate the walking pattern of the mice were adapted from previous study by Martínez de Lagrán et al. (2004) and performed on PND18. The hind paws of the pups were coated with blue non-toxic waterproof ink. Animals were then placed at one end of a long narrow tunnel (10 × 10 × 30 cm), and in the opposite end there were placed part of their nest. They spontaneously enter and partially or totally traverse the tunnel. A clean sheet of white paper was placed on the floor of the tunnel to record the paw prints. The pattern of three consecutive steps (the first four steps were excluded from the analysis) was analyzed and the following parameters assessed, averaged over consecutive steps: (i) stride length, the averaged distance between each stride; (ii) hindpaw base width, measured as the average distance between left and right hind footprint overlap. These values were determined by measuring the perpendicular distance of a given step to a line connecting its opposite preceding and proceeding.

Beam balance test

On PND19, individual pups were located in a trip of 40 × 2 cm elevated 25 cm from the surface. They began the task in the middle of the wooden trip and they should travel to reach the end of the trip with a cut-off time of 60 s. Latency to the first movement, arrival latency and latency to fall were recorded. There was a first training session where the animals performed the task and learned about the mechanism. Mice were guided along the trip holding them by the tail to avoid any fall if they were not able to do it themselves. In a period of an hour the test was repeated, in this case without any help and counting the times.

Open field test

The open field was a white melamine box (70 × 70 × 50 cm high) divided into 25 equal squares and under high intensity light levels (lighted, 500 Lux) or in darkness (with weak red light). Mice tend to avoid brightly illuminated, novel, open spaces, so the open field environment acts as an anxiogenic stimulus and allows for measurement of anxiety-induced locomotor activity and exploratory behavior s. Thus, two zones, center (1764 cm2) and periphery (3136 cm2) were delineated, being the center more anxiogenic. At the beginning of the test session, mice at PND15–16 were left in the periphery of the apparatus and during 5 min we measured and analyzed the latency to cross from the periphery to the center, total distance traveled, average speed and time spent in several sectors of the field (i.e., the border areas vs. the open, central area). Observation was made in an actimeter (Panlab, Barcelona, Spain) by computerized analysis of movements.

To test the actively behavioral effect of amphetamine in 18-month-old Aralar mice (Aralar- WT and Aralar-HT, n = 7), we performed an acute intraperitoneal administration protocol using three different drug concentrations: 1.5, 3.0, and 5.0 mg/kg, and saline as a control (sham). Aging animals performed the task immediately after the amphetamine treatment and developed the task once every other day increasing the drug dose concentration. Locomotor activity was measured in actimetry boxes (45 × 45 cm; Harvard Apparatus, Panlab) and movements on the ground were monitored via a grid of infrared beams and used as an index of locomotor activity (counts). All data were collected with Acti-Track software (Harvard Apparatus).

Statistical analysis

For behavioral and motor tests, variance homogeneity and normality of data were tested by means of Levene and Shapiro–Wilk tests, respectively. Simple comparisons between genotypes mice were performed using the two-tailed unpaired Student′s t-test with Whitney's correction to account for the different variances in the populations being studied. If the data did not meet specifications required for parametric analysis, non-parametric analysis of variance was used (Kruskal–Wallis) followed by comparisons between groups (Mann–Whitney U-test). Data were expressed as mean F SEM. In all tests, a difference was considered significant if the obtained probability value was p < 0.05. These statistical analyses were performed with the commercial software package Statistica 7.0 (StatSoft, Tulsa, OK, USA).

Brain regions and tissue preparation for amino acid analysis

Mice at PND19 were anesthetized, the whole brain was immediately removed from the skull and the brain regions were dissected according to Carlsson and Lindqvist (1973) and Itier et al. (2003) into the DA-rich limbic portion, the corpora striata (striatum), diencephalon, brainstem, cerebellum, and cerebral cortex. Regions were sonicated in 3% perchloric acid (PCA), neutralized, and centrifuged at 10 000 g for 15 min. Supernatants were lyophilized and dissolved in 0.2 M lithium citrate loading buffer pH 2.2 for quantification with an automatic amino acid analyzer Biochrom 20 (Pharmacia, Uppsala, Sweden) using a precolumn derivatization with ninhydrin and a cationic exchange column.

Measurements of monoamines in selected brain regions

The levels of DA and its metabolites, 3-methoxy-tyramine (3-MT), 3,4 dihydroxy-phenyl acetic acid (DOPAC) and homovanillic acid (HVA), noradrenaline (NA) and its metabolite, 4-hydroxy-3-methoxy-phenyl-glycol (MHPG), serotonin (5-HT) and its metabolite, 5-hydroxy-indole-acetic acid (5-HIAA) were measured by HPLC with an ESA coulochem detector, according to Mena et al. (1995). Briefly, samples from the same brain regions indicated above were sonicated in 8 volumes (w/v) of 0.4 N perchloric acid (PCA) with 0.5 mM Na2S2O5 and 2% EDTA and then centrifuged for 10 min. Monoamine levels were determined from 20 uL of the supernatant. The chromatographic conditions were as follows: a column Nucleosil 5C18; the mobile phase, a 0.1 M citrate/acetate buffer, pH 3.9 with 10% methanol, 1 mM EDTA and 1.2 mM heptane sulfonic acid; and the detector voltage conditions: D1 (+0.05), D2 (−0.39), and the guard cell (+0.4).

Measurements of glutathione in brain regions

Total glutathione (Gsx) levels were measured by the method of Tietze (1969). A sample (40 μL) of the sonicated brain region supernatant in 0.4 N PCA was neutralized with four volumes of phosphate buffer (0.2 M NaH2PO4, 0.2 M Na2HPO4, 0.5 M EDTA, and pH 7.5). Fifty microliters of this preparation were mixed with 5,5′-dithiobis-(2-nitrobenzoic acid) (0.6 mM), NADPH (0.2 mM), and glutathione reductase (1 unit) and the reaction was monitored in a P96 automatic microtiter reader at 412 nm for 6 min. Oxidized glutathione (GSSG) was measured as described by Griffith (1980). After the neutralization with the phosphate buffer, the sample remaining was mixed with 2-vinylpyridine (1.2 μL) at 25°C for 1 h and the reaction was carried out as described earlier. Reduced glutathione (GSH) was obtained by subtracting GSSG levels from Gsx levels.

Western blot in brain tissue

Aliquots (20 μg of protein) of brain lysates (in 20 mM Tris-HCl, 10 mM AcK, 1 mM dithiothreitol , 1 mM EDTA, protease inhibitor cocktail tablet (complete Mini, EDTA-free; Roche Diagnostics, Mannheim, Germany) 0.25% NP-40, pH 7.4) were centrifuged (12 000 g, 30 min at 4°C) and electrophoresed in an 8% sodium dodecyl sulfate acrylamide gel. Proteins were transferred electrophoretically to nitrocellulose membranes, which were blocked in 5% (w/v) dry skimmed milk (Sveltesse, Nestle) in Tris-buffered saline (10 mM Tris-HCl pH 7.5, 150 mM NaCl plus 0.05% (v/v) Tween-20) for 2 h, and further incubated with antibodies against Aralar (Del Arco and Satrústegui 1998) (polyclonal antibody, 1 : 1000), dopamine markers (tyrosine hydroxylase, TH polyclonal antibody Millipore (Billerica, MA, USA), 1 : 5000; dopamine transporter, DAT (SLC6A3) monoclonal antibody Millipore, 1 : 2000; vesicular monoamine transporter, VMAT2 (SLC18A2) polyclonal antibody Millipore 1 : 1000; dopamine and adenosine 3′, 5′-monophosphate-regulated phosphoprotein (32 kDa), DARPP-32 polyclonal antibody Millipore 1 : 10 000), GABAergic marker [glutamic acid decarboxylase (GAD)65, Chemicon (Caramillo, CA, USA), polyclonal antibody, 1 : 1000], glial markers [GFAP polyclonal antibody Dakopatts (Glostrup, Germany) 1 : 500, for astrocytes; and OX-6 monoclonal antibody Serotec (Kidlington, UK) 1 : 500, for microglia], and β-actin (monoclonal antibody Sigma (St Louis, MO, USA), 1 : 10 000), for 1 h at 25°C. Signal detection was performed with and enhanced chemiluminescence substrate (Western lighting-ECL; PerkinElmer, Baesweiler, Germany).

Statistical analysis

For biochemical data, the statistical significance of the differences was assessed by One-way analysis of variance (anova) followed by a post hoc Student's, Duncan/Tukey or Student-Newman-Keuls t-test method, as indicated. The results are expressed as mean ± standard error of the mean (SEM).

Results

Hyper-reactivity, demotivation, and motor discoordination are observed in Aralar-KO mice

A battery of behavioral and motor tests was carried out to elucidate the specific problems related to AGC1-MAS deficiency. When performing the toe pinch test, Aralar-KO mice presented hyper-reactivity as compared to WT mice (Fig. 1b), with no alteration in other sensory functions (i.e., visual placing, blast response, tactile and vibrissae orientation; not shown). Gross postural alterations became evident in Aralar/AGC1 KO mice analyzing the reaching response capacity (Fig. 1a). Aralar-KO animals showed a limb clasping phenotype, instead of showing a normal escape posture. Hyper-reactivity could also explain the results obtained in the touch escape test (Fig. 1d), where KO mice showed an exacerbated response to a finger stroke from above with no aversive effect in Aralar-WT or Aralar-hemizygous (HT) littermates.

Figure 1.

Hyperreactivity, hyperactivity, and social impairments in Aralar-knock-out (KO) mice. Reaching response (a) and toe pinch at post-natal day (PND)17 (b), homing test (c, denoting exploratory, social, and motor behavior) and touch scape (d) are illustrated in wild-type (open circles or bars), Aralar-hemizygote (gray circles or bars) and Aralar-KO mice (filled circles or bars). (e–g) Open field studies show high levels of hyperactivity and anxiety-like behavior in Aralar-KO mice. Parameters as total traveled distance (e), mean speed of walk (f) and resting time between displacements (g) in the center and periphery of the arena and at darkness (basal)-light (aversive) conditions are valutated in wild-type (WT), Aralar-hemizygote, and Aralar-KO mice. In (b and d) non-parametric analysis of variance was used (Kruskal–Wallis) followed by comparisons between groups (Mann–Whitney U-test). Simple comparisons between genotypes mice were performed in (c) using the two-tailed unpaired Student′s t-test with Whitney's correction. One-way anova with Bonferroni test for post hoc analyses was performed in open field test. Data are expressed as the mean ± SEM (= 8–13 mice per group). * 0.05, ** 0.01, *** 0.001.

To evaluate the psychomotor state, social behavior and motivations, we performed the homing test (Fig. 1c). The latency of Aralar wild-type (WT) mice to reach the opposite side of the cage stimulated by the presence of the nest was 7.2 ± 1.3 s, showing high motivation to reach the goal. In contrast, Aralar-KO mice did not show any displacement toward the target during the test. As the motor activity, measured in the open field test (Fig. 1e–g), seemed to be intact in Aralar-KO mice, the phenotype observed in the homing test may reflect demotivation and/or inability to initiate a voluntary motor response, one of the signs of basal ganglia dysfunction, particularly involving the caudate-putamen (Hauber 1998; Palmiter 2008).

The results of the pawprint analysis showed a gait disturbance in Aralar-KO mice (Fig. 2a and b). KO mice exhibited a significantly shorter stride length and hindpaw base width than their WT littermates (Fig. 2a), according to their smaller size, additionally presenting an erratic direction of path and in-coordination (Fig. 2b). Beam balance test clearly demonstrated a motor impairment of Aralar-KO mice (Fig. 2c) who tended to lock in a fixed spastic posture while on the beam and almost all Aralar-KO mice fell off the bar. The delayed onset for the first movement in KO mice on the bar is near the maximum latency estimated in the task, 60 s, while WT mice improved the performance in the second trial starting the movement in 26.5 ± 7.8 s and reaching the end of the beam in 29.3 ± 7.5 s. These motor deficits could be attributable to striatal dysfunction (Menéndez et al. 2006; Taylor et al. 2011) or increased anxiety/fear-related responses.

Figure 2.

Aralar-knock-out (KO) mice have a dramatic motor discoordination. In the pawprinting test (a, b) it is noticeable that Aralar-KO mice have a shorter stride length and a shorter hindpaw length than the wild-type (WT) siblings (a), and an erratic direction of walk (b). Asymmetric gait at the right leg is shown in the footprint left by the hind limbs of Aralar-KO mice walking on the paper (b). The beam balance test (c) was performed in wild-type (open circles), Aralar-hemizygote (gray circles), and Aralar-KO mice (filled circles). 1 and 2 indicates the number of assay. Statistical analysis was performed by two-tailed unpaired Student′s t-test with Whitney's correction. Data are expressed as the mean ± SEM (= 10–13 mice per group). * 0.05, ** 0.01, *** 0.001.

Thus, KO mice showed a lack of motor coordination in the hindlimbs with no muscle affectation (Figure S1), indicating a failure in midbrain and/or forebrain structures.

Enhanced locomotor, exploratory activity, and emotionality in Aralar-KO mice

The open field test, used to study spontaneous locomotor activity and anxiety-related behavior in mouse (Carola et al. 2002), was assessed in two different environments, in darkness less aversive conditions (with red light) and in lighted aversive conditions (Fig. 1e–g). Both WT and Aralar-KO were able to perform properly the task at PND15–16. Total distance traveled in darkness was statistically higher in Aralar-KO, with strong thigmotaxic behavior (preference to the periphery), compared with WT mice (Fig. 1e), indicating a more anxious-like behavior. Aralar-KO mice were as fast as WT (mean speed, 8.7 ± 0.9 cm/s; Fig. 1f), but with lower resting time (Fig. 1g) indicating hyperactivity. In the aversive (lighted) condition, Aralar-KO mice traveled significantly less distance in periphery than WT and than they did in darkness conditions (Fig. 1e). Moreover, they showed a striking burst of speed in the center of the arena (Fig. 1f). Grooming and fecal boli in the lighted versus. darkness condition were equally increased in all mice, however, the vertical activity (rearing), which involves hindlimb strength, tended to be higher in Aralar-KO mice in darkness condition than in WT mice (not shown), and this is also consistent with hyperactivity. These results indicate a rise in anxiety-, emotionality-related behavior s and reactivity in Aralar-KO animals.

Striatum is the main target for AGC1 deficiency

Although expression of Aralar-AGC1 has been extensively reported to be mainly restricted to and highly expressed in brain neurons (Ramos et al. 2003; Berkich et al. 2007; Xu et al. 2007; Cahoy et al. 2008; Pardo et al. 2011), no neuronal cell death in brain from Aralar-KO mice was detected (not shown), but evident hypomyelination and loss of neurofilaments was reported occurring in specific brain regions (Jalil et al. 2005; Sakurai et al. 2010; Ramos et al. 2011). Aralar-KO mice have brain abnormalities consisting of a marked enlargement in brain lateral ventricles (Fig. 3a) (Jalil et al. 2005), also reported in human AGC-1 deficiency (Wibom et al. 2009). Figure 3 shows that the enlargement of lateral ventricles is related to a reduction in the size of striatum. Thus, the striatum/brain ratio size was significantly reduced to 80% in Aralar-KO mice versus WT (13.33 ± 0.04% and 10.76 ± 0.03% in WT and Aralar-KO, respectively; p = 0.0286), while the hippocampus/brain ratio size was unchanged (11.77 ± 0.02 and 12.74 ± 0.02% in WT and Aralar-KO, respectively; Fig. 3b). Interestingly, the AGC1-deficient patient has also a smaller caudate-putamen than expected (Wibom et al. 2009).

Figure 3.

Aralar-knock-out (KO) mice show no morphological abnormalities in brain, but an enlargement in lateral ventricule and a reduction in size of striatum. View of coronal sections are shown stained with cresyl violet in wild-type (wt) and Aralar-KO animals at post-natal day (PND)20 (scale bar, 500 μm) (a). (b) Quantitation of the area for striatum and hippocampus versus that of whole brain is represented, as percentage, in Aralar-WT (open bars) and Aralar-KO mice (filled bars; = 4). (c) View of coronal sections of the brain from WT and Aralar-KO mice at PND20 as observed at the high power (scale bar, 500 μm), (= 6 mice). (d) The number of positive-neurons for tyrosine hydroxylase (TH) in the substantia nigra (SN) and midbrain was found to be equal in Aralar-KO as compared with control mice. Statistical analysis was performed by two-tailed unpaired Student′s t-test with Whitney's correction. Data are expressed as the mean ± SEM * 0.05.

Consistent with a preferential effect of Aralar deficiency in striatum, we found that the drop in whole brain Glutamine content previously reported (Pardo et al. 2011) is more prominent in the striatum (Table 1) than in all the other brain regions analyzed, reaching 29% of WT levels in Aralar-KO mice. However, in cerebellum and particularly in brainstem Glutamine levels hardly drop (in cerebellum to 82% of WT levels, p < 0.05; and no change in brainstem) perhaps because of the presence of low levels of citrin, a component of malate-aspartate shuttle homologous to Aralar, in these brain areas (Contreras et al. 2010). Striatal GABA was also significantly reduced to 60% in Aralar-KO (Table 1), with no changes in all the other brain regions.

Table 1. Amino acid content in brain extracts from Aralar-WT mice and Aralar-KO mice at 20 days (striatum, diencephalon, hipocampus, brainstem, cerebral cortex, and cerebellum)
Aminoacids (nmol/g tissue) StriatumDiencephalonLimbic systemBrain stemCer. CortexCerebellum
  1. Results are expressed in nmol per g of tissue. Values are the mean ± SEM (n = 3–6). Statistical analysis was performed by one-way analysis of variance followed by Student-Newman-Keuls t-test.

  2. a

    p < 0.05 Aralar-KO versus Aralar-WT mice.

  3. b

    p < 0.01 Aralar-KO versus Aralar-WT mice.

  4. c

    p < 0.001 Aralar-KO versus Aralar-WT mice.

AspartateAralarWT2264 ± 1752466 ± 1221894 ± 2322409 ± 1562356 ± 912379 ± 62
Aralar KO454 ± 41c429 ± 21c399 ± 48b562 ± 34c430 ± 54c706 ± 61c
(% vs WT)20%17%21%23%18%29%
SerineAralarWT1101 ± 90790 ± 741014 ± 159514 ± 191000 ± 33709 ± 48
Aralar KO160 ± 16c164 ± 13b138 ± 9b168 ± 26c196 ± 24c215 ± 18b
(% vs WT)14%21%13.6%33%20%30%
AlanineAralarWT783 ± 97573 ± 39870 ± 128525 ± 18805 ± 22497 ± 19
Aralar KO297 ± 63b180 ± 18c211 ± 15b238 ± 36c261 ± 23c201 ± 19c
(% vs WT)38%31%24%45%32%40%
GlutamateAralarWT7512 ± 4877615 ± 3218166 ± 10765840 ± 2588326 ± 2528071 ± 237
Aralar KO4030 ± 238b3935 ± 230c4446 ± 577b4141 ± 235b5797 ± 575b5202 ± 418c
(% vs WT)54%52%54.4%71%70%64%
GlutamineAralarWT3019 ± 2582880 ± 1432548 ± 3012653 ± 2052449 ± 2313808 ± 143
Aralar KO895 ± 166c1387 ± 233b1133±268b2569 ± 1851471 ± 264b3119 ± 352a
(% vs WT)29%48%44.4%60%82%
GABAAralarWT2523 ± 2602563 ± 2371568 ± 2211543 ± 1331558 ± 681442 ± 103
Aralar KO1512 ± 141b2682 ± 1161223 ± 1591979 ± 194a1863 ± 88b1062 ± 93
(% vs WT)60%128%119%

On the other hand, Table 1 shows that none of the brain regions analyzed differed with respect to the drop in whole brain aspartate, serine and alanine levels (Pardo et al. 2011). Thus, the striatum appeared to be the brain region most affected by aralar deficiency.

Monoamine metabolism is impaired in brain from Aralar-deficient mice

Given the marked inability of Aralar-KO mice to perform motor tasks and the hyper-reactivity, hyper-activity, and anxiety observed in the tests performed, and the prominent effect of Aralar deficiency in the striatum, we decided to investigate the metabolism of monoamines in several brain regions, and particularly in striatum, as it is closely involved in the former functions (see Stein et al. 2006 for references).

NA content was similar in WT and Aralar-KO mice in any of the regions studied (Table 2).

Table 2. Monoamine metabolism in striatum, diencephalon, limbic system, and brainstem of Aralar-WT mice and Aralar-KO mice
Monoamines (ng/g tissue) StriatumDiencephalonLimbic systemBrain stem
  1. Results are expresssed in ng per g of tissue. Values are the mean ± SEM (n = 6). Statistical analysis was performed by one-way analysis of variance followed by Student-Newman-Keuls t-test.

  2. a

    p < 0.05 Aralar-KO mice versus Aralar-WT mice.

  3. b

    p < 0.01 Aralar-KO mice versus Aralar-WT mice.

  4. c

    p < 0.001 Aralar-KO mice versus Aralar-WT mice.

DAAralarWT5690 ± 525341 ± 921412 ± 22097.7 ± 18
Aralar KO3661 ± 175b431 ± 88810 ± 82a164 ± 51
(% vs. WT)64%57%
3-MTAralarWT330 ± 39n.d ≤ 659.7 ± 6.2n.d ≤ 6
Aralar KO142 ± 29bn.d ≤ 6n.d ≤ 6n.d ≤ 6
(% vs. WT)43%
DOPACAralar WT387 ± 1084 ± 10141 ± 1443.8 ± 2.7
Aralar KO362 ± 38127 ± 10a71 ± 17b73.7 ± 8b
(% vs. WT)151%50%168%
HVAAralar WT681 ± 35219 ± 21136.5 ± 12.161.6 ± 6.4
Aralar KO468 ± 64a279 ± 18115.2 ± 6.685.2 ± 7.5a
(% vs. WT)68%138 ± 12.2%
NAAralar WT251 ± 20846 ± 22336 ± 201223 ± 60
Aralar KO265 ± 301008 ± 98320 ± 28.81359 ± 36.5
(% vs. WT)
5-HTAralar WT333 ± 671501 ± 242585 ± 551988 ± 39
Aralar KO250 ± 89768 ± 88a460 ± 581445 ± 172a
(% vs. WT)51%72%
5-HIAAAralar WT256 ± 102859 ± 95337 ± 40484 ± 16
Aralar KO99 ± 28375 ± 42c218 ± 29.5a285 ± 25c
(% vs. WT)43%64%59%

Regarding the serotoninergic system, Aralar deficiency resulted in a significant decrease both in 5-HT levels and specially those of its intracellular degradation product 5-HIAA in diencephalon (to 51% and 43%, respectively, vs. control) and brainstem (to 72% and 59%, respectively) (Table 2), with similar changes (not always significant) in striatum and limbic system (Table 2), which have much lower 5HT content than the former areas. Of note, mRNA for serotonin transporter (SERT/Slc6a4), the plasma membrane transporter of serotonin which terminates the action of serotonin, is greatly increased (5.5-fold of WT value) in Aralar-KO brain (Table S3), suggesting an increased reuptake of this neurotransmitter in brain from Aralar-KO mice.

In the brainstem, where the DAergic neuronal somata of the nigrostriatal pathway are located, no obvious change in the number of TH-positive cells was apparent (Fig. 3d). Neither the substantia nigra nor the midbrain from Aralar-KO mice showed changes in TH-immunolabeling (Fig. 3c) or expression of TH (not shown) as compared with WT. Moreover, DA and its metabolites were increased or unchanged in brainstem and diencephalon, the major regions enriched in DAergic somata, from Aralar-KO mice (Table 2).

In contrast, Aralar deficiency resulted in changes in DA and its metabolites in the regions enriched in DAergic projections, striatum, and limbic system. Aralar-KO striatum showed a substantial reduction in DA (to 64%) and its metabolites, 3-MT (to 43%) and HVA (to 68% of controls) (Table 2). However, the content of DOPAC was not changed, resulting in a significant increase in DOPAC/DA ratio (138% vs. control, Fig. 4a). The very low levels of 3-MT (Wood and Altar 1988; Brown et al. 1991) in striatum from Aralar-deficient mice and the increase in the ratio of the MAO-derived DA metabolite, DOPAC, over the catechol-ortho-methyl-transferase (COMT)-derived DA metabolites, 3-MT and HVA, in Aralar-KO striatum (DOPAC/HVA, 1.38-fold of WT (Fig. 4a); and DOPAC/3-MT, 2.35-fold of WT (data not shown)) suggest an impairment in DA release in these animals. As DOPAC formation requires the intraneuronal MAO–aldehyde dehydrogenase pathway, these results also suggests an increase in intraneuronal metabolism of DA in Aralar-KO mice (see Fig. 4c).

Figure 4.

Striatum appeared to be the most affected region in Aralar-knock-out (KO) brain. (a) Differential effects of Aralar deficiency on dopamine (DA) metabolism in brain regions enriched in DA terminals (striatum and limbic system). DA and its metabolites, except 3,4 dihydroxy-phenyl acetic acid (DOPAC), are significantly decreased in Aralar-KO striatum. DA turnover is increased in Aralar-KO striatum. (b) The enhanced striatal DA turnover provoked an increase in cellular oxidative stress as measured by GSH/GSSG ratios. (c) Scheme of tyrosine metabolism into the presynaptic and post-synaptic neurons in striatum, representing metabolites, enzymes, and proteins involved. (d–f) Expression of dopamine markers in Aralar-KO striatum and limbic system. Representative western blot of tyrosine hydroxylase (TH), dopamine transporter (DAT) in striatum (d); and vesicular monoamine transporter 2 (VMAT2), dopamine and cAMP regulated phosphoprotein of 32 KDa (DARPP32) proteins in striatum (e) and limbic system (f) with their respective densitometric histograms. β-actin was used as charge control. Results are expressed as the mean ± SEM (= 6 mice per group). Statistical analysis was performed by one-way anova followed by Newman–Keuls test.* 0.05, ** 0.01, *** 0.001.

Aralar-KO limbic system shows an important reduction in DA (57%) and DOPAC (50%); but only a slight non-significant decrease in HVA (Table 2). Given that DOPAC/DA ratio was unchanged, the drop in DA content in the limbic system of Aralar-KO mice does not appear to be because of increased intracellular metabolism of DA (Fig 4a).

In conclusion, the Aralar-KO mouse brain shows no changes in NA but a significant decrease in 5-HT only in diencephalon and brainstem, brain regions rich in 5-HT neurons. As for the DA system, although DA and metabolites did not change in brain regions rich in DA somata, that is, diencephalon and brainstem, marked decreases in DA were found in areas enriched in DAergic nerve terminals. These results (Fig. 4a) suggest a clear increase in intraneuronal DA metabolism in striatum, but no obvious changes in the limbic system.

DA markers of presynaptic and post-synaptic terminals in striatum and limbic system from Aralar-KO mice

A further analysis for DA markers in striatum and limbic system is shown in Fig. 4d–f. Remarkably, there is a significant reduction of the DAergic markers VMAT2, the presynaptic vesicular transporter of monoamines, and DARPP32, the dopamine and cAMP regulated phosphoprotein of 32 kDa present in striatal post-synaptic medium spiny neurons (Fienberg et al. 1998) in striatum (Fig. 4e), but not in the limbic system (Fig. 4f) of Aralar-KO mice. However, no changes in TH and the dopamine transporter DAT were found in striatum from Aralar-KO mice indicating no gross modifications in the density of presynaptic DAergic terminals (Fig. 4d). The content of GFAP and OX6 as gliosis markers was unchanged in Aralar-deficient striatum compared with controls (not shown).

DARPP32 is mainly expressed in mature medium spiny neurons which are GABAergic (Fienberg et al. 1998). We have previously noted a 40% decrease in striatal GABA levels in Aralar-KO mice (Table 1). However, the remarkable lack of post-synaptic DARPP32 in striatum from Aralar-deficient mice (Fig. 4e) was not accompanied by any change in the GABA synthesis enzyme GAD65, a more immature marker for GABA-containing neurons (not shown). These results suggest a deficiency in maturation of medium spiny GABA neurons in Aralar-deficient striatum.

Regarding presynaptic DA terminals, the fall in VMAT2 (of about 42%) suggests that the vesicular storage of DA is also impaired in Aralar-deficient striatum.

Increased oxidative stress in Aralar-KO striatum

The increase in the DOPAC/DA ratio in Aralar-KO striatum (Fig. 4a) indicates an increased intracellular oxidation of DA which would lead to an increase of H2O2 formation via mitochondrial MAO activity (Fig. 5) which could result in a selective oxidative stress in DAergic neurons (Spina and Cohen 1989). To verify this hypothesis, we measured the content of reduced (GSH) and oxidized glutathione (GSSG) in striatum and other brain regions (limbic system and brainstem) as readout of the cellular redox state (White et al. 1986; Spina and Cohen 1989).

No changes in GSH levels were found in striatum but the content of GSSG was more than two-fold higher in Aralar-KO compared to WT mice (Fig. 4b). GSH and GSSG content in Aralar-KO mice was unchanged in the other brain regions analyzed (limbic system and brainstem; Figure S2). The decreased GSH/GSSG ratio in Aralar-deficient striatum suggests that this brain region is subjected to high oxidative stress (Figs 4b and 5).

Aralar-hemizygous adult mice also show increased intraneuronal dopamine metabolism

The affectation observed in striatum from Aralar-KO mice (PND20) prompted us to analyze monoamine metabolism in striatum of adult and healthy mice expressing only half-a-dose of Aralar (Fig. 6). Mice expressing 50% normal Aralar/AGC1 (hemizygous) are indistinguishable from wild-type siblings and thrive and survive normally (not shown). In striatum, DA and 5-HT content was not altered by half-a-dose of Aralar expression versus control (Fig. 6a), but DOPAC content was significantly increased. Hemizygous-Aralar (HT) adult mice display an increased metabolism of DA by MAO (DOPAC/DA ratio was 135% higher in striatum of Aralar-HT vs. Aralar-WT mice at 18 months of age), (p = 0.0055, n = 7; Fig. 6a).

Figure 5.

Mechanism for toxicity induced by the lack of Aralar in DAergic nigrostriatal nerve terminals. Lack of Aralar-MAS activity causes a decrease in mitochondrial NADH because both lack of redox transfer by the shuttle and limited pyruvate supply to mitochondria; and also decreased mitochondrial NAD(P)H content. As glutathione system (and thioredoxin) for ROS detoxification is reduced by the mitochondrial NADPH pool, increased ROS is expected to happen in Aralar-deficient mitochondria. Mitochondrial ROS (O2- and H2O2) diffuse to the cytosol, presumably provoking increased levels of cytosolic alpha-synuclein aggregates and synuclein-VMAT2 complexes in the presynaptic DA nerve terminals of Aralar-knock-out (KO) mice, with loss of VMAT2. Consequently, an increase in non-vascularized cytosolic DA produces an enhancement in DA autoxidation and in enzymatic oxidation via MAO activity (with higher DOPAC/DA ratio). Both pathways involve an overproduction of ROS, as reflected by increased GSSG, in Aralar-KO striatum. This further potentiates VMAT2 decline and possibly loss of function in the DA terminal with the subsequent mishandling of DA. AGC, aspartate-glutamate carrier; Asp, aspartate; AAT, aspartate aminotransferase; DA, dopamine; DOPAC, 3,4-dihydroxy-phenylacetic acid; G6P, glucose 6 phosphate; Glu, glutamate; Gluc, glucose; GA3P, glyceraldehyde 3-phosphate; GPx, glutathione peroxidase; GR, glutathione reductase; GSH, reduced glutathione; GSSG, oxidized glutathione; α-KG; α-ketoglutarate; Lac, lactate; Mal, malate; MAO, monoamine oxidase; MDH, malate dehydrogenase; NNT, NADH-NADP-transhydrogenase; OAA, oxalacetic acid; OGC, α-ketoglutarate-malate carrier; Pyr, pyruvate; VMAT2, vesicular monoamine transporter 2.

Figure 6.

Alteration of striatal dopamine handling in adult and healthy Aralar-hemizygous mice at 18-months old. (a) Cytosolic dopamine (DA) metabolism is significantly increased in Aralar-HM striatum as compared with wild-type mice, without changes in the content of monoamines between genotypes. Values (ng/g tissue) in control striatum are as follows: DA, 14583 ± 258; 3,4 dihydroxy-phenyl acetic acid (DOPAC), 782 ± 21; homovanillic acid (HVA), 923 ± 24; serotonin (5-HT), 1252 ± 36; 5-HIAA, 291 ± 13. (= 7). (b) Representative western blot of vesicular monoamine transporter 2 (VMAT2), dopamine and cAMP regulated phosphoprotein of 32 KDa (DARPP32) proteins in striatum with their respective densitometric histograms. β-actin was used as charge control. (= 6). (c–f) Increased sensitivity to the motor effects of systemic amphetamine in Aralar-HM mice. Open field parameters as average speed (c), traveled distance (d), resting time between displacements (e), and the number of zone entries (f) in the center and periphery of the arena are valutated in WT and Aralar-hemizygote mice. Results are expressed as the mean ± SEM (= 7). Statistical analysis was performed by one-way anova followed by Newman–Keuls test. ** 0.01.

Hemizygous-Aralar mice presented also an enhanced sensitivity to the locomotor stimulating effects of amphetamine (Fig. 6c–f). Amphetamine provoked a dose-dependent increase in average speed in both genotypes, with a more pronounced effect in HT mice (Fig. 6c), reaching a plateau at 5.0 mg/kg. At 3.0 mg/kg dose of intraperitoneal amphetamine, a higher increment in traveled distance (Fig 6d), with the consequent decrease in resting time (Fig. 6e), in relation to that observed at sham administration was found in HT as compared with WT. A similar course is observed in the number of entries in center and periphery (Fig. 6f). VMAT2 was only slightly decreased to 73% in striatum of Aralar-HT adult mice as compared with control littermates (Fig. 6b). However, a higher decrease in VMAT2 content was found in striatum from Aralar-KO mice (Fig. 4e). As amphetamine is thought to release newly formed, mostly cytoplasmic rather than vesicular, catecholamines (Itier et al. 2003), these findings suggest that the reduction of Aralar-MAS activity alter the handling of intracytoplasmic neurotransmitters as DA.

Discussion

Aralar-KO mice present a short lifespan, dying at PND22–23, with generalized tremor, severe hypomyelination (Jalil et al. 2005) and modifications in cortical projections but with no apparent neuronal cell death (Sakurai et al. 2010; Ramos et al. 2011). Herein, we show that Aralar-KO mice exhibit a higher exploratory activity, hyperactivity, hyperreactivity, and anxious-like behavior with aversive conditions as well as an increase in rearing; parameters that are known to be sensitive to interferences with the DAergic system (Bernardi et al. 1981). In addition, Aralar-KO mice have a loss in motor coordination and alterations in the gait pattern and equilibrium, deficits that have been extensively associated to striatal DAergic damage (Menéndez et al. 2006; Taylor et al. 2011). Failure to perform homing test, in Aralar-KO mice, constitutes a reliable indicator of reduced motivation associated to dopamine deficiency (Palmiter 2008). According to the behavior observations, KO mice (PND20) showed depletion of DA in DA projection-rich areas, striatum, and limbic system, where levels of DA and its metabolites are highest [60-fold higher DA in striatum than brainstem (Table 2)]; and DA mishandling was reflected by significant increased DOPAC/DA ratio in Aralar-KO striatum. DA mishandling was also found in striatum of adult (18-months old) and healthy Aralar-hemizygote mice (Fig. 6). The KO mice also showed a marked decrease in 5-HT and its metabolite 5-HIAA in brainstem and diencephalon, the regions with higher 5-HT levels in control animals. An increase in locomotion and DA metabolism could be associated to acceleration of cellular DA uptake (Husain et al. 1994), as supported by the higher DAT/VMAT2 ratio found in Aralar-KO striatum as compared with controls. However, hyperactivity in open field might be also related to anxiety-dependent behavior, because of DA (Zweifel et al. 2011) and 5HT depletions (for review, Fernandez and Gaspar 2011), as it was only detected in novelty-related experimental situations.

Our results indicate that the nigrostriatal DAergic system and striatum are preferentially vulnerable to Aralar-MAS deficiency associated to DA mishandling; and previously, we showed that the lack of Aralar-MAS induces a decrease in the maximal respiratory capacity in neurons (Gómez-Galán et al. 2011). These observations are well in agreement with the notion that striatum is highly susceptible to mitochondrial metabolic dysfunctions (Zeevalk et al. 1997; Watabe and Nakaki 2008; Pickrell et al. 2011; Sterky et al. 2012). Together with these findings, fall in striatal GABA and a pronounced decline in glutamate and glutamine were found. Our results now reveal two new defects related to Aralar deficiency in striatum: (i) Inability of GABAergic striatal neurons to achieve a mature phenotype. These neurons remain immature as reflected by the spared DARPP32 (a protein that mediates DAergic neurotransmission in almost all the medium spiny neurons), with little variations in GABA or GAD65 expression (Lauder et al. 1986); and (ii) loss of DA and DA mishandling reflected by increased DOPAC/DA ratio. The significant reduction in striatal VMAT2 of Aralar-KO mice supports that the presynaptic DAergic nerve endings are damaged but still present as DAergic markers as TH and DAT were unaffected. In contrast to striatum, in regions enriched in DAergic somata as brainstem and diencephalon, DA content was found to be increased, perhaps because of an attempt to compensate for any dysfunction in the surviving DAergic neurons, as occurs in the very early stages of Parkinson′s disease (PD) (Hefti et al. 1980; Altar et al. 1987; Hornykiewicz and Kish 1987).

The impairment of the dopaminergic system, particularly in the dopaminergic striatal terminals of the Aralar-KO mouse, is probably related to an increase in oxidative stress caused by the lack of Aralar-MAS activity which adds to the known vulnerability of these terminals to oxidative stress. This increased oxidative stress was reflected, in the very significant decrease in the GSH/GSSG ratio specifically in Aralar-KO striatum. The decrease in GSH/GSSG in Aralar-KO striatum most likely reflects an increased oxidative stress in the cytosol which has the largest GSH pool and/or the mitochondria that contains a much smaller GSH pool (Spina and Cohen 1989; Murphy 2012). Although we believe that the initial decrease in GSH is mitochondrial (see below), H2O2 escaped from mitochondria to the cytosol probably contributes to the decrease in cytosolic GSH/GSSG.

Mitochondria produce O2 and H2O2 which are detoxified due to GSH and thioredoxin together with a number of enzymes that ultimately use one of these two thiol molecules as redox agents (Murphy 2012). The regeneration of the reduced forms of glutathione and thioredoxin requires NADPH. There are three systems which produce NADPH in brain mitochondria, NADP-isocitrate dehydrogenase, malic enzyme, and energy dependent nicotinamide nucleotide transhydrogenase (NNT) (Andres et al. 1980; Albracht et al. 2011). The third of these systems, NNT, utilizes mitochondrial NADH and the proton electrochemical gradient to produce NADPH. There is evidence that NNT is important in supplying NADPH for mitochondrial detoxification, as the lack of NNT increased O2/H2O2 production in mitochondria of beta cells (Freeman et al. 2006), and impairs cellular redox homeostasis and energy metabolism in human adrenocortical (Meimaridou et al. 2012) and pheochromocytoma cells (Yin et al. 2012).

NADPH production through NNT may be limited by mitochondrial NADH production. In brain, which utilizes glucose as main energy source, mitochondria produce NADH from pyruvate in the tricarboxylic acid cycle. The lack of Aralar results in a pronounced decrease in MAS, the major NADH shuttle in brain (Jalil et al. 2005), resulting in an increased lactate-to-pyruvate ratio (Pardo et al. 2011), and in a limitation in pyruvate supply to mitochondria as reflected in a reduced maximal respiration rate in intact neurons (Gómez-Galán et al. 2011). This scenario is one in which mitochondrial NADH production is clearly limited, and this will result in a lack of inactivation of O2 and H2O2 which will cause oxidative damage to Aralar-KO mitochondria, and the escape of H2O2 to the cytosol (Han et al. 2003), causing oxidative stress in this cellular compartment.

ROS formation is further potentiated in Aralar-KO striatum because VMAT2 deficiency resulting in an increase in non-vesicular DA that might favor both MAO-mediated oxidation and autooxidation of non-protected cytosolic DA. These two processes lead to the formation of ROS, such as hydrogen peroxide, and reactive quinone and semi-quinone species produced by DA autooxidation (Graham 1978; Maker et al. 1981). Furthermore, DACHR (o-quinone dopaminochrome, a product of DA oxidation) has been reported to increase, in a dose-dependent way, the production of H2O2 constitutively observed at Complex I of the mitochondrial respiratory chain (Zoccarato et al. 2005). Our data suggest that the rate of production of ROS evoked by Aralar deficiency in striatum override cellular mechanisms for reducing GSSG and might have important consequences in DA neuronal physiology, which is particularly sensitive to oxidative stress (Zeevalk et al. 1997; Drechsel and Patel 2008).

Decreased VMAT2 expression was found exclusively in striatum, but not in limbic system, of Aralar-KO mice. A similar finding was reported as a key pathogenic event preceding nigrostriatal dopamine neurodegeneration and clinical manifestations in a primate model of PD and attributed to an association of VMAT2 with α-synuclein aggregates induced by oxidative stress as a result of (1-methyl-4-phenyl-1,2,3,6-tetrahydropiridine) MPTP treatment (Chen et al. 2008a). A direct interaction between VMAT2 and α-synuclein with disruption of synaptic vesicle dynamics has been proposed by other groups (Lotharius and Brundin 2002; Mosharov et al. 2006; Guo et al. 2008; for references, Taylor et al. 2011). The loss of VMAT2 in the Aralar-KO mice might be also because of sequestration with α-synuclein aggregates but this remains an open question. The mishandling of DA via reduced VMAT2, associated to an increased striatal DOPAC/DA (Fig. 4a), and GSSG/GSH ratios (Fig. 4b), might be sufficient to cause DA-mediated toxicity and neurodegeneration in the nigrostriatal DA system (Mooslehner et al. 2001; Caudle et al. 2007; Chen et al. 2008b; for review, see Taylor et al. 2011). The specific vulnerability of nigrostriatal DA neurons might be explained because they have a higher ROS formation than those DA neurons in limbic system (Surmeier et al. 2011) and also these A9 DA neurons are specially prone to ROS attack, that is, the scarce proportion of glial cells surrounding DA neurons in the substantia nigra (for review, Mena et al. 2002), the presence of neuromelanin pigment in subpopulations of DA-containing mesencephalic neurons (Hirsch et al. 1988) and the low content of mitochondria in DA neurons of the substantia nigra pars compacta (Liang et al. 2007) might be mentioned between other characteristics.

The results demonstrate that AGC1-MAS deficiency in mice targets monoaminergic brain systems in striatum. DA neurotransmission is also altered in mice with mutations of α-synuclein, parkin or DJ-1, considered suitable models for PD studies. These mice, as reported herein for Aralar-KO and previously for VMAT2-KO mice (Colebrooke et al. 2006), do not display loss of midbrain DA neurons, the hallmark of Parkinson's disease (Dawson et al. 2010; for references). Perhaps, DA neurons do not degenerate in the mouse models because there are compensatory mechanisms that prevent the loss of DA neurons during their short lifespan. Also, it is worth to take into account that mice possess low neuromelanin and high GSH content what might render them specially resistant to DA degeneration compared with primates and humans (Hirsch et al. 1988; Itier et al. 2003, for references). These factors would contribute to underestimate the putative detrimental effects of Aralar-MAS deficiency on the DA system in humans. Deficiencies or failure in the operation of the Aralar-MAS pathway, resulting in a limited mitochondrial NADH formation, NNT function and ROS detoxifying capacity, might constitute an important factor at the origin of DA degeneration and its implication in human pathologies as Parkinson′s and Huntington′s diseases might be thoughtfully explored.

Acknowledgements

This study was supported by grants from the Ministerio de Educación y Ciencia BFU2008-04084/BMC (to JS), and Ciencia e Innovación (SAF2010-16427 to MD), Comunidad de Madrid S-GEN-0269-2006 MITOLAB-CM (to JS), European Union Grant LSHM-CT-2006-518153 (to J.S.), and CureFXS E-Rare. EU/FIS PS09102673, Spanish Ministry of Health (PI 082038 to MD), Marató TV3, Jerome Lejeune (JMLM/AC/08-044) to MD, Fundación Médica Mutua Madrileña (to BP), and by an institutional grant from the Fundación Ramón Areces to the CBMSO. CIBERER is an initiative of the ISCIII.

The authors thank to Isabel Manso, Barbara Sesé and Dolores Cano for technical support. IL-F is a recipient of a predoctoral contract from the Comunidad de Madrid, Spain.

Conflict of interest

None declared.

Ancillary