Neurochemistry International 61 (2012) 1325–1332

Contents lists available at SciVerse ScienceDirect

Neurochemistry International

journal homepage: www.elsevier .com/locate /nci

Aralar mRNA and protein levels in neurons and astrocytes freshly isolatedfrom young and adult mouse brain and in maturing cultured astrocytes

Baoman Li, Leif Hertz, Liang Peng ⇑Department of Clinical Pharmacology, China Medical University, Shenyang, PR China

a r t i c l e i n f o

Article history:Received 31 July 2012Received in revised form 4 September 2012Accepted 13 September 2012Available online 24 September 2012

Keywords:AralarAstrocyteGlutamateNeuronProtein expressionAstrocyte culture

0197-0186/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.neuint.2012.09.009

⇑ Corresponding author. Address: Department of CMedical University, No. 92 Beier Road, Heping Distri+86 24 23256666x5130; fax: +86 24 23251769.

E-mail address: hkkid08@yahoo.com (L. Peng).

a b s t r a c t

Intense glucose-based energy metabolism and glutamate synthesis by astrocytes require malate–aspartate-shuttle (MAS) activity to regenerate NAD+ from NADH formed during glycolysis, since brainlacks significant glycerophosphate shuttle activity. Aralar is a necessary aspartate/glutamate exchangerfor MAS function in brain. Based on cytochemical immunoassays the absence of aralar in adult astrocyteswas repeatedly reported. This would mean that adult astrocytes must regenerate NAD+ by producing lac-tate from pyruvate, eliminating its use by oxidative and biosynthetic pathways. We alternatively usedastrocytes and neurons from adult brain, freshly isolated by fluorescence-activated cell sorting, to deter-mine aralar protein by a specific antibody and its mRNA by real-time PCR. Both protein and mRNAexpressions were identical in adult neurons and astrocytes and similar to whole brain levels. The samelevel of aralar expression was reached in well-differentiated astrocyte cultures, but not until late devel-opment, coinciding with the late-maturing brain capability for glutamate formation and degradation.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

During the last 10 years in vivo magnetic resonance spectro-scopic (13C-NMR) assays of metabolism of 13C-labeled glucose oracetate have demonstrated that astrocytes in adult brain have arate of oxidative metabolism of glucose in gray matter correspond-ing to 20–30 percent of the total, i.e., at least similar to neurons cal-culated per volume (reviewed by Hertz, 2011b). Even experimentsusing incorporation of radioactive or stable isotopes of the astro-cyte-specific substrate acetate into neuronal glutamate show thishigh percentage (Cruz and Cerdán, 1999; Blüml et al., 2002; Lebonet al., 2002; Deelchand et al., 2009; Boumezbeur et al., 2010; Patelet al., 2010; Lanz et al. 2012), with any differences in absolute ratesdue to species differences and/or different use of anesthetics.Glucose-derived pyruvate is needed by astrocytes for two majorpurposes, (i) to supply ATP from oxidative pathways for energy-consuming processes, such as uptake of potassium ions (K+)(Somjen et al., 2008; MacAulay and Zeuthen, 2012; Wang et al.,2012) and glutamate (Danbolt, 2001) from extracellular fluid,and (ii) to produce glutamate from glucose via the anapleroticpathway using pyruvate carboxylase, which is absent in neurons(reviewed by Hertz et al., 2007; Hertz, 2011b). This pathway isneeded for de novo synthesis of glutamine, which in the brain

ll rights reserved.

linical Pharmacology, Chinact, Shenyang, PR China. Tel.:

serves as an essential precursor for the neurotransmitters gluta-mate and GABA.

Pyruvate is generated by the glycolytic pathway in the cytosol,and its production involves one oxidative process, formation ofdiphosphoglycerate from glyceraldehyde 3-phosphate. In this reac-tion, NAD+ is reduced to H+ + NADH, which is unable to cross themitochondrial membrane for its re-oxidation. NAD+ must be regen-erated for glycolysis to continue, and this can be accomplished by aredox shuttle system that transfers ‘reducing equivalents’ to themitochondria or by the cytoplasmic lactate dehydrogenase (LDH)reaction. Because conversion of pyruvate to lactate by LDH elimi-nates pyruvate as an oxidative-biosynthetic substrate, astrocyticredox shuttling is required to generate pyruvate for ATP and gluta-mate production, and probably also during glutamate degradation(Hertz, 2011a; Bauer et al., 2012).

There are two major intracellular redox shuttle systems, theglycerol-phosphate shuttle and malate–aspartate shuttle (MAS).Both cytosolic and mitochondrial glycerol-3-phosphate dehydro-genases are present in brain, but the importance of this shuttlein brain is probably negligible because these two enzymes are ex-pressed in different cell types (Nguyen et al., 2003; LaNoue et al.,2007). The main pathway for re-synthesis of cytosolic NAD+ inbrain is the MAS. As illustrated in Fig. 1, the MAS transfers reducingequivalents from cytoplasm to mitochondria by means of coupledreactions that carry out oxidation–reduction and transaminationreactions, utilizing oxaloacetate (OAA), aspartate, malate, a-ketoglutarate (a-KG), and glutamate as participants in the shuttle(for details, see Fig. 1 and its legend). Cycling of these compounds

http://dx.doi.org/10.1016/j.neuint.2012.09.009mailto:hkkid08@yahoo.comhttp://dx.doi.org/10.1016/j.neuint.2012.09.009http://www.sciencedirect.com/science/journal/01970186http://www.elsevier.com/locate/nci

The Malate-Aspartate Shuttle Transfers Reducing Equivalents from Cytoplasm to Mitochondria

Aspartate

OAA

NAD+ NADH

Malate

OAA

NAD+ NADH

Malate

CytosolMitochondrion

Aspartate

α-KG

α-KG

3 ATP via electrontransport chain

Glutamate

Glutamate

AATm

AATc

MDHm

MDHc

Fig. 1. In the malate–aspartate shuttle (MAS) cytosolic malate dehydrogenase(MDHc) oxidizes NADH and converts oxaloacetate (OAA) to malate (top right offigure), which enters the mitochondria in exchange with a-ketoglutarate (a-KG).The mitochondrial malate dehydrogenase (MDHm) re-oxidizes malate to OAA,which is transaminated to aspartate by the mitochondrial aspartate aminotrans-ferase (AATm). Aspartate leaves the mitochondria in exchange with glutamate,requiring ACG (aralar or citrin). In the mitochondria glutamate conversion to a-KGis essential for AATm activity forming aspartate from OAA and delivering a-KG formitochondrial export. The glutamate imported into the mitochondria had beenformed by cytosolic aspartate aminotransferase (AATc) from a-KG after its entryinto the cytosol. Without MAS activity NADH formed in the cytosol duringglycolysis would have been unable to enter the mitochondria for oxidation.Reprinted from Hertz and Dienel (2002), with permission.

1326 B. Li et al. / Neurochemistry International 61 (2012) 1325–1332

between the cytoplasm and mitochondria also requires two carrierproteins, a malate/a-ketoglutarate exchanger (OGC – Slc25a10)and a glutamate/aspartate exchanger (AGC – Slc25a12 [aralar] orSlc25a13 [citrin]). There are two AGC forms in adult brain, predom-inantly AGC1 or aralar, with small clusters of citrin only in a fewneurons (Contreras et al., 2010). In contrast, hepatocytes only ex-press citrin (Del Arco et al., 2002). Adult brain astrocytes thereforeneed coordinated activities of OGC, aralar, mitochondrial and cyto-solic aspartate aminotransferases, and malate dehydrogenases forMAS function (Fig. 1), and thus for both energy metabolism andtransmitter synthesis.

Expression of OGC and both the mitochondrial and cytosolicaspartate aminotransferases in brain is well established (Fonnum,1968; Horio et al., 1988; McKenna et al., 2000), and aspartate ami-notransferase activity is high in cultured astrocytes (Schousboeet al., 1977; Erecińska et al., 1993). However, the operation ofMAS in astrocytes in adult brain in situ has been questioned be-cause aralar (or citrin) was barely detectable in immunocytochem-ical assays (Ramos et al., 2003; Berkich et al., 2007). The highestlevel of cytochemically determined aralar expression was obtainedwhen an antigen retrieval technique was used in a study by Pardoet al. (2011). However, electron microscopic analysis of aralarlocalization determined by immunogold-particle labeling of neuro-nal and astrocytic mitochondria indicated that astrocytic mito-chondria contained only about 7% of the total number of labeledparticles. Astrocytic processes were identified by their irregularshape and filamentous membranes that often surround both axonsand dendrites and/or by the formation of gap junctions, but thisdetermination may not include the abundant astrocytic mitochon-dria in fine peripheral processes of astrocytes (Lovatt et al., 2007).Complete lack of astrocytic aralar would make oxidative metabo-lism of glucose and glutamate biosynthesis in astrocytes impossi-ble. However, based on the MAS turnover rates determined byBerkich et al. (2007) and the contribution of astrocytes to cerebro-cortical volume, Hertz (2011a) concluded that the total amount ofaralar in the experiments by Pardo et al. (2011) was enough to sus-tain known rates of astrocytic oxidative metabolism. Nevertheless,

the low mitochondrial expression is worrisome, and higher MASactivity may be needed for glutamate and GABA turnover. At leastthe immunocytochemical assays that detected even lower levels ofaralar, if any, in adult astrocytes (Ramos et al., 2003; Berkich et al.,2007) accordingly appear to be discordant with the readily-detectable, high rates of glucose oxidation and anaplerotic activityconclusively determined by in vivo MRS studies (see Section 1).Together with the positive cytochemical study by Pardo et al.(2011) this raises the possibility that the other immunocytochem-ical assays may have failed to detect aralar antigen in astrocytes.Failure to detect astrocytic aralar in the studies that did not usean antigen-retrieval procedure (Ramos et al., 2003; Berkich et al.,2007) could have arisen from incomplete antigen exposure to theantibody. In fact, Nishino and Nowak (2004) showed that antigenretrieval substantially enhanced signals from heat shock protein(HSP) 72 and glial fibrillary acidic protein (GFAP), attenuatedsignals from HSP27, and did not alter the strong MAP2 signal.Alternatively, a false-negative result could arise from lack of anti-genicity upon tissue fixation or tissue processing (Fritschy, 2008).

At least as high mRNA expression of aralar in adult astrocytes asin neurons has been shown in cells separated by fluorescence-activated cell sorting (FACS) from brains of transgenic mice co-expressing one fluorescent signal with an astrocytic marker and adifferent fluorescent signal with a neuronal marker (Lovatt et al.,2007). Although these findings open the possibility that astrocytesmight also translate aralar mRNA to protein, they do not necessar-ily prove it. Protein expression was not studied, because rathersmall amounts of cells are obtained by the cell-sorting techniques,and many genes were investigated. Similar to a subsequent studyby Cahoy et al. (2008) mRNA was therefore determined by micro-array analysis.

In the present study, the cell separation technique employed byLovatt et al. (2007) was scaled up by using several transgenic ani-mals, so that enough material was obtained to determine aralarprotein levels by Western blotting and mRNA by real-time poly-merase chain reaction (PCR). Cellular extracts were prepared fromneurons and astrocytes isolated from transgenic mice at 14 and35 days of age, and the results were compared to whole brain ex-tracts from 70-day-old adult CD-1 mice. Similar comparisons weremade in developing cultured cerebral cortical astrocytes but not inneurons, which cannot be maintained for a sufficiently long time inculture (Peng et al., 1991).

2. Material and methods

2.1. Animals and cell preparation

Male and female CD-1 or FVB/NTg(GFAP GFP)14Mes/J or B6.Cg-Tg(Thy1-YFPH)2Jrs/J mice (from The Jackson Laboratory, Bar Har-bor, ME) were housed as previously detailed (Fu et al., 2012). Thetransgenic mice combine expression of Thy1, a marker of large pro-jection neurons (see Lovatt et al., 2007), with a specific fluorescentsignal, and the astrocyte marker GFAP, with a fluorescent signal ata different wavelength. All experiments were carried out in accor-dance with the USA National Institute of Health Guide for the Careand Use of Laboratory Animals (NIH Publications No. 80–23) re-vised 1978, and all experimental protocols were approved by theInstitutional Animal Care and Use Committee of China MedicalUniversity.

After decapitation cerebral hemispheres minus olfactory bulbs,hippocampi, and basal ganglia were immediately removed andused either for cell culture and whole-brain studies (CD-1 mice)or cell sorting (transgenic mice). For the latter, cerebral hemi-spheres were placed in cold Hanks’ buffer containing the glutamatereceptor antagonists DNQX (3 lM) and APV (100 lM). A cell

B. Li et al. / Neurochemistry International 61 (2012) 1325–1332 1327

suspension was prepared as previously described (Lovatt et al.,2007). The cerebral hemispheres were cut into small pieces, and di-gested with 8 U/ml papain in Ca2+/Mg2+-free PIPES/cysteine buffer,pH 7.4, for 1 h at 37 �C/5%CO2. After washing, the tissue was fur-ther digested with 40 U/ml DNase I in Mg2+-containing minimumessential medium (MEM) with 1% bovine serum albumin (BSA)for 15 min at 37 �C/5% CO2, carefully triturated in cold MEM with1% BSA, centrifuged over a 90% Percoll gradient to collect all cellsat and above the lipid layer. This solution was further diluted fivetimes and centrifuged to collect the pellet. The cells were re-suspended in cold MEM with 1% BSA and 4 mg ml�1 propidiumiodide (PI).

Immediately thereafter the cells were sorted into cold MEMwith 1% BSA, using the BD FACSAria Cell Sorting System (35 psisheath pressure, FACSDiva software S/W 2.2.1; BD Biosciences,San José, CA) as described by Lovatt et al. (2007). GFP and YFP wereexcited by a 488 nm laser, and emissions were collected by 530 nmdiscrimination filters. mRNA expression of cell markers of astro-cytes (Connexin 30, GFAP, Glt-1 and Fgfr3), neurons (Gabra-1,KCC2, Snap25 and synaptotagmin) and oligodendrocytes (Connex-in 47, Mag, Mog and Mbp) were determined (Supplementary Fig. 1[from Fu et al., 2012], with further details in legend). No contami-nation with neurons and oligodendrocytes was found in the astro-cyte samples or of astrocytes and oligodendrocytes in the neuronalsamples (Fu et al., 2012).

2.2. Cultures

Primary cultures of astrocytes were prepared from newbornmale or female mice as previously described (Hertz et al., 1978,1998; Hertz 2012) with minor modifications. The neopallia of thecerebral hemispheres were aseptically isolated as described above,freed of meninges, dissociated by vortexing, filtered through nylonmeshes, diluted in culture medium, and planted in Falcon Primariaculture dishes. The culture medium was a Dulbecco’s Medium with7.5 mM glucose, 20% horse serum, and the cultures were incubatedat 37oC in a humidified atmosphere of CO2/air (5:95%). The med-ium was exchanged with fresh medium of similar compositionon day 3, and subsequently every 3–4 days. At day 3, the serumconcentration was reduced to 10%, and after the age of 2 weeks,0.25 mM dibutyryl cyclic AMP (dBcAMP) was included in the med-ium. This compound increases intracellular cyclic AMP and pro-motes differentiation in astrocyte cultures derived from newbornbrain (Hertz, 1990, 1993; Meier et al., 1991; Schubert et al.,2000). The age of 2 weeks for its addition has been determinedexperimentally, and is consistent with the finding by Moonenand Sensenbrenner (1976) that astrocytes need a certain stage ofdevelopment in order to respond to dBcAMP, and that by Lodinet al. (1979) that astrocytes de-differntiate in vitro, unless treatedwith this compound. The statement by Fedoroff et al. (1984) thatthe dBcAMP-treated cells correspond to reactive astrocytes hasproven incorrect (Wandosell et al., 1993), However, unfortunatelyit may have influenced most researchers using cultured astrocytesfor almost 30 years, and in the process damaged the reputation ofcultured astrocytes (Kimelberg, 2010). The similarities betweennot only levels but also development of aralar protein and mRNAexpression in the cultured astrocytes and in freshly isolated astro-cytes, which will be shown in ‘Results’, support the validity of cul-tured astrocytes obtained using these procedures as valid modelsof their in vivo counterparts.

2.3. Real-time PCR

Although results for reverse transription PCR were alreadyavailable, mRNA expression was re-determined by real time poly-merase chain reaction (RT-PCR or qPCR). A cell suspension was

prepared by collecting cells in Trizol. The RNA pellet was precipi-tated with isopropyl alcohol, washed with 70% ethyl alcohol, anddissolved in distilled water. Primers for aralar (fwd: 50-CCTCACCTCAGTTTGGTGTCACTC-30; rev: 50-GTGGCCGTGGCAAGTCTGTA-30) and TATA-binding protein (TBP), used as a referencegene (fwd: 50-GCCTTCCTTCTTGGGTATG-30; rev: 50-GAGGTCTTTA-CGGATGTCAAC-30) were generated by TaKaRa Biotechnology(Dalian, China) and optimized to an equal annealing temperatureof 60 �C. The 179 bp product has no similarity with any citrin se-quence, as shown by checking the Fwd, Rev primer and the whole175 bp sequence on line (blast.ncbi.nlm.nih.gov). It is also differentfrom a primer previously used for reverse transcription PCR andthat used by Lovatt et al., 2007, both of which were also testedand provided comparable results. Only those observed by real-timePCR will be presented in Section 3. However, the classical PCRamplification, which was performed in a Robocycler thermocycler:3 min at 95 �C, followed by 40 cycles at 95 �C for 5 s and 60 �C for30 s, then 95 �C for 3 min, 55 �C for 1 min, followed by PCR productseparation by 1% agarose gel electrophoresis resulted in a singleband with desired length for both aralar and TBP. SYBR Green-based real-time PCR with an Mx 3000P instrument and the GreenQuantitative RT-PCR Kit from Agilent Technologies (Cary, NC, USA)was performed using the optimized protocol. The final PCR mixturecontained 2 ll each of forward and reverse primers (1 lM), 2 ll ofFast Start DNA Master SYBR Green I (2�), 0.3 ll of Ret Dye (1 lM),2 ll of cDNA template, and it was made up to 20 ll with nucleasefree water (Pérez et al., 2012). Reactions were performed in dupli-cate. Real-time PCR efficiency (E) for each pair of primers andtarget gene was determined using 5-fold serial dilutions of RTproduct (1 lg, 200, 40, 8 and 1.6 ng). The number of cycles (Ct) nec-essary to obtain a threshold fluorescent signal of target genes andreference gene, TBP was determined with 1 lg cDNA. E and Ct werecalculated from MxPro QPCR Software (Agilent Technologies, Cary,NC, USA). The relative expression ratio (ratio) of a target gene wascalculated based on E and Ct as follows (Pfaffl, 2001):

Ratio ¼ ð1þ EtargetÞ½DCttargetðcontrol� sampleÞ�=ð1þ Eref Þ� ½DCtref ðcontrol� sampleÞ�

2.4. Statistical analysis

The differences between multiple groups were analyzed byone-way analysis of variance (ANOVA) followed by Fisher’s LSDmultiple comparison test for unequal replications. The level ofsignificance was set at P < 0.05.

3. Results

3.1. Selectivity and linear range of anti-aralar Western blots

The anti-aralar antibody recognized a single band with thesame molecular weight (70 kDa) as aralar in extracts of adultmouse brain and 3-week-old cultured astrocytes, but the proteinwas not detected in liver, which is known to only express citrin(entire blot with some debris shown in Fig. 2A). Scanning of 3 indi-vidual experiments showed that the expression of aralar relative tob-actin was similar in whole brain and in the cultured astrocytes(Fig. 2B). When the amounts of protein loaded on the gel were var-ied over the range 10–100 lg (Fig. 2C), the signal intensities forboth aralar and b-actin were nearly linear up to about 80 mg(Fig. 2D). The routinely-loaded protein amount (50 lg) is thus wellwithin the rectilinear range (Fig. 2B), indicating that increases ordecreases in the amount of aralar will be reliably quantified.

aralar

β-actin

10 25 50 75 100

Brain Tissue (μg)

0

0.1

0.2

0.3

0.4

0.5

Brain Astrocytes Liver Neurons

Rat

io o

f ar

lar/

β-ac

tin

80 kDa

60 kDa

50 kDa

40 kDa

30 kDa

100 kDa

150 kDa

aralar

β-actin

Brain Astrocytes Liver Neurons

MW

Int

ensi

ty

Brain Tissue (μg)

β-actin

aralar

A

B

C

D

Fig. 2. Protein expression of aralar in brain, liver and cultured astroccytesdetermined with the antibody sc-271056, specific to aralar, and showing theentire gel. (A) A representative immunoblot showing protein expression for aralarand b-actin, used as a house-keeping protein. The staining of HRP-labeled 2ndantibody was photographed by fluorescent imaging system, but the molecularweight (MW) was photographed by neutral light, and is thus not visible in theFig. at the same time. The size of aralar is 70 kDa, and that of b-actin 46 kDa.Similar results were obtained from three independent experiments. (B) Mean-s ± SEM (n = 3) of scanned ratios between aralar and b-actin. (C) A representativeimmunoblot showing protein expression for aralar, determined with the antibodysc-271056, and b-actin, used as a house-keeping protein, in intact CD-1 mousebrain with applied of protein amounts between 10 and 100 lg. (D) Intensity of theexpressions of aralar and b-actin at different amounts of applied protein, measuredby scanning.

1328 B. Li et al. / Neurochemistry International 61 (2012) 1325–1332

3.2. Aralar protein and mRNA levels in adult astrocytes and neuronsobtained by FACS

Fig. 3A shows an individual Western blot demonstrating selec-tive labeling of aralar and distinct increases in protein expressionof aralar between the ages of 14 and 35 days in dissociated astro-cytes, neurons, and whole brain when assayed in the same gelsand blots using the same amounts of protein per lane (50 lg).The amount of aralar relative to b-actin increased with age from14 to 35 days. At each age, aralar levels were similar in astrocytesand neurons freshly-isolated from whole brain and equivalent tothat in whole brain (Fig. 3B). Expression of aralar mRNA infreshly-dissociated astrocytes and neurons was similar at 14 days,but by 35 days its level was much higher in astrocytes comparedwith neurons and similar to that in whole brain (Fig. 3C and D).Thus, in the brain in vivo ontogenetic development of aralar is sim-ilar in both cell types (protein) or becomes eventually expressed toa greater extent in astrocytes (mRNA).

3.3. Aralar protein and mRNA levels in cultured astrocytes

Early developmental increases in aralar protein with time incultured astrocytes were modest. Two-week-old astrocyte culturesas well as 3-week-old cultures which had not received the routinedifferentiating treatment with dBcAMP from the age of 2 weekshad similar levels (Fig. 4A). However, dBcAMP treatment morethan doubled astrocytic aralar expression, raising it to the levelof that in intact brain. Thus, differentiation is a critical aspect ofaralar expression in cultured astrocytes similar to what has beenshown for many other astrocyte characteristics (Hertz, 1990; Meieret al., 1991).

Aralar mRNA expression in the cultured astrocytes was lowestat 1 week, intermediate at 2 weeks, and highest at 3 weeks afterdBcAMP-treatment (Fig. 4B). In this context it should be re-emphasized that astrocytes need to reach a certain developmentalstage, before they can respond to dBcAMP (Mooonen andSensenbrenner, 1976). At this point in time aralar mRNA expres-sion reached a similarly high level as in brain, i.e., a higher levelthan in freshly isolated neurons. This slow development is remark-able, as will be discussed below.

4. Discussion

The question whether mature astrocytes express an astrocyte-glutamate carrier, an essential component of the malate–aspartateredox shuttle system, was first raised by Ramos et al. (2003) and fol-lowed up in subsequent studies by Berkich et al. (2007) and Pardoet al. (2011). This is a tremendously important issue because itbrings attention to the question how astrocytes obtain the pyruvatethey require for oxidative metabolism and glutamate synthesis. Invivo NMR studies in many laboratories have established very signif-icant rates of oxidation of [13C]glucose in astrocytes, which appearsto be at odds with at least some immunocytochemical data.Although the amount of protein reflects capacity, not biologicalactivity, extremely low aralar protein levels in astrocytes or evenabsence of aralar (Ramos et al., 2003; Berkich et al. 2007) are notcompatible with a functioning MAS. Moreover, two studies in cul-tured astrocytes correlate MAS activity with function in astrocytes:Fitzpatrick et al. (1988) showing inhibited glucose metabolism, butuninhibited pyruvate metabolism after MAS inhibition, and Amaralet al. (2011) demonstrating direct relationship between pyruvateoxidation rate and MAS flux. The low and declining activity in cul-tured astrocytes reported by Ramos et al. (2003) can probably beexplained by differences in culturing technique.

0

0.5

1

1.514 days

35 days

Rat

io o

f ar

alar

/β-a

ctin

Astrocytes (GFAP) Neuron (YFP) Brain

** *

14 days 35 days 14 days 35 days 14 days 35 days

Astrocytes (GFAP) Neuron (YFP) Brain (Control)

aralar

β-actin

0.0

0.5

1.0

1.5

Astrocytes (GFAP) Neurons (YFP) Brain

14 days

35 days

Rel

ativ

e E

xpre

ssio

n R

atio

*

**

Threshold

A

B

C

D

Fig. 3. Protein and mRNA expression of aralar in astrocytes and neurons isolated by FACS from cerebral hemispheres of 14- and 35-day old astrocyte-labeled (FVB/NTg(GFAPGFP)14Mes/J or neuron-labeled B6.Cg-Tg(Thy1-YFPH)2Jrs/J) mice and in intact brain of adult CD-1 mice. (A) A representative immunoblot showing protein expression foraralar and b-actin, used as a house-keeping protein. The size of aralar is 70 kDa, and of b-actin 46 kDa. Similar results were obtained from three independent experiments.(B) Means ± SEM of scanned ratios between aralar and b-actin. ⁄Statistically significant (P < 0.05) difference from the same preparation from 14-day old animals. (C)A representative amplification plot of aralar mRNA expression, determined by real-time PCR in astrocytes and neurons isolated by FACS from cerebral hemispheres of 14- and35-day old astrocyte-labeled (FVB/NTg(GFAP GFP)14Mes/J or neuron-labeled B6.Cg-Tg(Thy1-YFPH)2Jrs/J) mice and in intact brain of adult CD-1 mice. Similar results wereobtained from three independent experiments. (D) Means ± SEM (n = 3) of the relative expression ratio (ratio) of aralar. ⁄Statistically significant (P < 0.05) difference from thesame preparation from 14-day old animals.

B. Li et al. / Neurochemistry International 61 (2012) 1325–1332 1329

The present study circumvented the possibility that the re-ported absence or reduced expression of aralar protein in adultastrocytes compared to neurons found with immunocytochemicalassays might underestimate aralar expression in astrocytic

mitochondria. It used a different approach that avoided the com-plexities of immunoassays and of recognition of all astrocytic mito-chondria in intact tissue. Extracts of freshly-isolated astrocytes andneurons obtained from brain and separated by FACS were assayed

Threshold

0

0.5

1

1.5

2

2.5

Rat

io o

f ar

alar

/β-a

ctin

*

** **

Aralar

TBP

Brain

1 week 2 week 3 week

Astrocytes

dBcAMP+

0.0

0.5

1.0

1.5

Rel

ativ

e E

xpre

ssio

n R

atio

1 week

Astrocytes

dBcAMP

+

2 week 3 week Brain

****

*

1 week

Astrocytes

dBcAMP+

2 week 3 week Brain

A

B

C

D

Fig. 4. Protein and mRNA expression of aralar in primary cultures of astrocytes and in intact brain of adult CD-1 mice. Astrocytes were cultured for 1 or 2 weeks, and for3 weeks with (dBcAMP) or without (3 week) addition to the medium of 0.25 mM dibutyryl cAMP from the age of 2 weeks. (A). A representative immunoblot showing proteinexpression for aralar and b-actin, used as a house-keeping protein. Similar results were obtained from three independent experiments. (B) Means ± SEM (n = 3) of scannedratios between aralar and b-actin. ⁄Statistically significant (P < 0.05) difference from 2 and 3 weeks groups in astrocytes. ⁄⁄Statistically significant (P < 0.05) difference from allother groups, but not from each other. (C) A representative amplification plot of aralar. mRNA expression of aralar determined by real-time PCR in primary cultures ofastrocytes and in intact brain of adult CD-1 mice. Similar results were obtained from three independent experiments. (D) Means ± SEM (n = 3) of the relative expression ratio(ratio) of aralar. ⁄Statistically significant (P < 0.05) difference from 2 and 3 weeks groups in astrocytes. ⁄⁄Statistically significant (P < 0.05) difference from all other groups, butnot from each other.

1330 B. Li et al. / Neurochemistry International 61 (2012) 1325–1332

by Western blotting and qRT-PCR. Aralar protein expression levelswere similar in these two cell types in both suckling and youngadult mice, whereas mRNA levels were higher in astrocytes, con-

firming previous findings by Lovatt et al. (2007) of at least as highmRNA levels in astrocytes as in neurons. Moreover, the possibilityof deficient translation in astrocytes was discounted by showing

B. Li et al. / Neurochemistry International 61 (2012) 1325–1332 1331

that aralar protein levels were as high in astrocytes as in neurons.The remarkably high mRNA levels in astrocytes may indicate areadiness to respond to potential metabolic stimuli.

Similar assays in cultured astrocytes showed that aralar proteinand mRNA levels in astrocytes doubled between 1 and 2 weeks toeventually reach or bypass the level in neurons. The doubling ofaralar protein after astrocytic differentation by dBcAMP, broughtthe levels up to those observed in astrocytes obtained by FACS inadult brain. These Western blot immunoassays were obtainedwithin the linear ranges for signal intensity as a function of aralarand b-actin protein amount. They detected aralar, not citrin, sinceno antibody response was found in liver cells. Moreover, the aralar-specific primer for mRNA would not recognize citrin. Equivalentexpression of aralar in astrocytes and neurons isolated from brainshould remove a significant hurdle for acceptance of high rates ofoxidative metabolism in brain astrocytes and in at least some typesof cultured cells (reviewed by Hertz et al., 2007; Hertz 2011b).

The developmental increases of brain aralar protein and mRNAlevels between 2 and 4 weeks of age are consistent with a compa-rable increase in fluxes in glutamatergic and GABAergic neuronalTCA cycles. Tricarboxylic acid cycle function matures early, asshown by maximum ability of stimuli of energy production to en-hance brain slice oxidation around postnatal day 15 (Holtzmanet al., 1982). Increases in the activity of the glutamate (GABA)-glutamine cycle between postnatal days 10 and 30 (Chowdhuryet al., 2007) occur in parallel with the rise in aralar protein andmRNA levels between postnatal days 14 and 35 (Fig. 3). Other glu-cose-metabolizing enzymes (e.g., hexokinase, aldolase, LDH, andpyruvate dehydrogenase) also exhibit large increases in their activ-ity between the ages of �15 and �30 days (Leong and Clark, 1984;Land et al., 1977). Synaptic mitochondria mature earlier (Almeidaet al., 1995) than non-synaptic mitochondria (Bates et al., 1994),the fraction that would include astrocytic mitochondria. The cyto-solic malate dehydrogenase (MDHc) which operates in the MAS,but not in the TCA cycle, has a slow developmental increase innon-synaptic mitochondria, whereas the mitochondrial malatedehydrogenase (MDHm), which functions both in the MAS (Fig. 1)and in the TCA cycle, matures much faster (Malik et al., 1993).Additional strong evidence that the late development is relatedto the maturation of glutamatergic and GABAergic signaling isthe demonstration by Patel and Balázs (1970) that incorporationof 14C from glucose into amino acids in the rat brain in vivo in-creases sharply between postnatal days 10 and 20, reaching itsmaximum around day 25, and that the maximum increase in glu-tamine/glutamate specific activities (a sign of metabolic compart-mentation) may even occur a few days later. The activity ofglutamine synthetase, an astrocytic enzyme required to providethe precursor for neurotransmitter glutamate and GABA for neu-rons, also increases steeply during all of the first 3 weeks of devel-opment in cultured astrocytes and in brain in vivo (Hertz et al.,1978; Patel et al., 1982). Together these observations suggest thata considerable part of the late increase in aralar expression in bothneurons and astrocytes reflects the increase of glutamate produc-tion, astrocyte-to-neuron transfer of glutamate via glutamine,GABA synthesis, and astrocytic glutamate and GABA degradationthat are essential for glutamatergic and GABAergic transmitteractivity.

In conclusion, using the fluorescence-based technique for braincell separation developed by Lovatt et al. (2007), we confirmedtheir observation that freshly-isolated astrocytes and neurons haveat least similar levels of aralar mRNA expression. These findingswere extended to demonstrate large increases in expression ofnot only aralar mRNA but also its protein in both cell types be-tween postnatal days 14 and 35 and similar protein levels in themature cells. The findings suggest a developmental correlationnot only with TCA cycle activity, but also with formation and,

perhaps, degradation of glutamate and GABA (Hertz, 2011a, b).This is consistent with the need for astrocytic MAS activity not onlyfor energy metabolism but also for these processes. The newestand best performed immunocytochemical study (Pardo et al.,2011) is qualitatively, although not quantitatively in agreementwith the observations made.

Acknowledgements

This study was supported by Grants 31171036 to L.P. and No.31000479 to B.L. from the National Natural Science Foundationof China.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.neuint.2012.09.009.

References

Almeida, A., Brooks, K.J., Sammut, I., Keelan, J., Davey, G.P., Clark, J.B., Bates, T.E.,1995. Postnatal development of the complexes of the electron transport chainin synaptic mitochondria from rat brain. Dev. Neurosci. 17, 212–218.

Amaral, A.I., Teixeira, A.P., Håkonsen, B.I., Sonnewald, U., Alves, P.M., 2011. Acomprehensive metabolic profile of cultured astrocytes using isotopic transientmetabolic flux analysis and C-labeled glucose. Front Neuroenerg. 3, 5.

Bates, T.E., Heales, S.J., Davies, S.E., Boakye, P., Clark, J.B., 1994. Effects of 1-methyl-4-phenylpyridinium on isolated rat brain mitochondria: evidence for a primaryinvolvement of energy depletion. J. Neurochem. 63, 640–648.

Bauer, D.E., Jackson, J.G., Genda, E.N., Montoya, M.M., Yudkoff, M., Robinson, M.B.,2012. The glutamate transporter, GLAST, participates in a macromolecularcomplex that supports glutamate metabolism. Neurochem. Int. 61, 566–574.

Berkich, D.A., Ola, M.S., Cole, J., Sweatt, A.J., Hutson, S.M., LaNoue, K.F., 2007.Mitochondrial transport proteins of the brain. J. Neurosci. Res. 85, 3367–3377.

Blüml, S., Moreno-Torres, A., Shic, F., Nguy, C.H., Ross, B.D., 2002. Tricarboxylic acidcycle of glia in the in vivo human brain. NMR Biomed. 15, 1–5.

Boumezbeur, F., Petersen, K.F., Cline, G.W., Mason, G.F., Behar, K.L., Shulman, G.I.,Rothman, D.L., 2010. The contribution of blood lactate to brain energymetabolism in humans measured by dynamic 13C nuclear magnetic resonancespectroscopy. J. Neurosci. 30, 13983–13991.

Cahoy, J.D., Emery, B., Kaushal, A., Foo, L.C., Zamanian, J.L., Christopherson, K.S., Xing,Y., Lubischer, J.L., Krieg, P.A., Krupenko, S.A., Thompson, W.J., Barres, B.A., 2008.A transcriptome database for astrocytes, neurons, and oligodendrocytes: a newresource for understanding brain development and function. J. Neurosci. 28,264–278.

Chowdhury, G.M., Patel, A.B., Mason, G.F., Rothman, D.L., Behar, K.L., 2007.Glutamatergic and GABAergic neurotransmitter cycling and energymetabolism in rat cerebral cortex during postnatal development. J. Cereb.Blood Flow Metab. 27, 1895–1907.

Contreras, L., Urbieta, A., Kobayashi, K., Saheki, T., Satrústegui, J., 2010. Low levels ofcitrin (SLC25A13) expression in adult mouse brain restricted to neuronalclusters. J. Neurosci. Res. 88, 1009–1016.

Cruz, F., Cerdán, S., 1999. Quantitative 13C NMR studies of metaboliccompartmentation in the adult mammalian brain. NMR Biomed. 12, 451–462.

Danbolt, N.C., 2001. Glutamate uptake. Prog. Neurobiol. 65, 1–105.Deelchand, D.K., Nelson, C., Shestov, A.A., Uğurbil, K., Henry, P.G., 2009.

Simultaneous measurement of neuronal and glial metabolism in rat brainin vivo using co-infusion of [1,6–13C2]glucose and [1,2–13C2]acetate. J. Magn.Reson. 196, 157–163.

del Arco, A., Morcillo, J., Martínez-Morales, J.R., Galián, C., Martos, V., Bovolenta, P.,Satrústegui, J., 2002. Expression of the aspartate/glutamate mitochondrialcarriers aralar1 and citrin during development and in adult rat tissues. Eur. J.Biochem. 269, 3313–3320.

Erecińska, M., Pleasure, D., Nelson, D., Nissim, I., Yudkoff, M., 1993. Cerebralaspartate utilization: near-equilibrium relationships in aspartateaminotransferase reaction. J Neurochem. 60, 1696–1706.

Fedoroff, S., McAuley, W.A., Houle, J.D., Devon, R.M., 1984. Astrocyte cell lineage. V.Similarity of astrocytes that form in the presence of dBcAMP in cultures toreactive astrocytes in vivo. J. Neurosci. Res. 12, 14–27.

Fitzpatrick, S.M., Cooper, A.J., Hertz, L., 1988. Effects of ammonia and beta-methylene-DL-aspartate on the oxidation of glucose and pyruvate by neuronsand astrocytes in primary culture. J Neurochem. 51, 1197–1203.

Fonnum, F., 1968. The distribution of glutamate decarboxylase and aspartatetransaminase in subcellular fractions of rat and guinea-pig brain. Biochem. J.106, 401–412.

Fritschy, J.M., 2008. Is my antibody-staining specific? How to deal with pitfalls ofimmunohistochemistry. Eur. J. Neurosci. 28, 2365–2370.

http://dx.doi.org/10.1016/j.neuint.2012.09.009http://dx.doi.org/10.1016/j.neuint.2012.09.009

1332 B. Li et al. / Neurochemistry International 61 (2012) 1325–1332

Fu, H., Li, B., Hertz, L., Peng, L., 2012. Contributions in astrocytes of SMIT1/2 andHMIT to myo-inositol uptake at different concentrations and pH. Neurochem.Int. 61, 187–194.

Hertz, L., 1990. Dibutyryl cyclic AMP treatment of astrocytes in primary cultures asa substitute for normal morphogenic and ‘functiogenic’ transmitter signals. In:Privat, A., Giacobini, E., Timiras, P., Vernadakis, A. (Eds.), Molecular Aspects ofDevelopment and Aging in the Nervous System. Plenum, N.Y., pp. 227–243.

Hertz, L., 1993. Metabolic interactions between neurons and astrocytes. In:Fedoroff, S., Doucette, R., Juurlink, B.H.J. (Eds.), Biology and Pathology ofAstrocyte–Neuron Interactions. Plenum. New. Y., pp. 1–13.

Hertz, L., 2011a. Brain glutamine synthesis requires neuronal aspartate: acommentary. J. Cereb. Blood Flow Metab. 31, 384–387.

Hertz, L., 2011b. Astrocytic energy metabolism and glutamate formation–relevancefor 13C-NMR spectroscopy and importance of cytosolic/mitochondrialtrafficking. Magn. Reson. Imaging 29, 1319–1329.

Hertz, L., 2012. Isotope-based quantitation of uptake, release, and metabolism ofglutamate and glucose in cultured astrocytes. Methods Mol. Biol. 814, 305–323.

Hertz, L., Bock, E., Schousboe, A., 1978. GFA content, glutamate uptake and activityof glutamate metabolizing enzymes in differentiating mouse astrocytes inprimary cultures. Dev. Neurosci. 1, 226–238.

Hertz, L., Dienel, G.A., 2002. Energy metabolism in the brain. Int. Rev. Neurobiol. 51,1–102.

Hertz, L., Peng, L., Dienel, G.A., 2007. Energy metabolism in astrocytes: high rate ofoxidative metabolism and spatiotemporal dependence on glycolysis/glycogenolysis. J. Cereb. Blood Flow Metab. 27, 219–249.

Hertz, L., Peng, L., Lai, J.C., 1998. Functional studies in cultured astrocytes. Methods16, 293–310.

Holtzman, D., Olson, J., Zamvil, S., Nguyen, H., 1982. J. Neurochem. 39, 274–276.Horio, Y., Tanaka, T., Taketoshi, M., Uno, T., Wada, H., 1988. Rat cytosolic aspartate

aminotransferase: regulation of its mRNA and contribution to gluconeogenesis.J. Biochem. 103, 805–808.

Kimelberg, H.K., 2010. Functions of mature mammalian astrocytes: a current view.Neuroscientist 16, 79–106.

Land, J.M., Booth, R.F., Berger, R., Clark, J.B., 1977. Development of mitochondrialenergy metabolism in rat brain. Biochem. J. 164, 339–348.

LaNoue, K.F., Carson, V., Berkich, D.A., Hutson, S.M., 2007. Mitochondrial/CytosolicInteractions via Metabolite Shuttles and Transporters, in Handbook ofNeurochemistry and Molecular Neurobiology. In: Lajtha, A., Gibson, G.E.,Dienel, G.A. (Eds.). Springer-Verlag, Berlin, pp. 5616–5689, Vol. 2.

Lanz, B., Uffmann, K., Wyss, T., Weber, M., Buck, B., Buck, A., Gruetter, R., 2012. Atwo-compartment mathematical model of neuroglial metabolism using [1-11C]acetate. J. Cereb. Blood Flow Metab. 32, 548–559.

Lebon, V., Petersen, K.F., Cline, G.W., Shen, J., Mason, G.F., Dufour, S., Behar, K.L.,Shulman, G.I., Rothman, D.L., 2002. Astroglial contribution to brain energymetabolism in humans revealed by 13C nuclear magnetic resonancespectroscopy: elucidation of the dominant pathway for neurotransmitterglutamate repletion and measurement of astrocytic oxidative metabolism. J.Neurosci. 22, 1523–1531.

Leong, S.F., Clark, J.B., 1984. Regional enzyme development in rat brain. Enzymesassociated with glucose utilization. Biochem. J. 218, 131–138.

Lodin, Z., Faltin, J., Korínková, P., 1979. The effect of dibutyryl cyclic AMP oncultivated glial cells from corpus callosum of 30-day-old rats. Physiol.Bohemoslov. 28, 105–111.

Lovatt, D., Sonnewald, U., Waagepetersen, H.S., Schousboe, A., He, W., Lin, J.H., Han,X., Takano, T., Wang, S., Sim, F.J., Goldman, S.A., Nedergaard, M., 2007. Thetranscriptome and metabolic gene signature of protoplasmic astrocytes in theadult murine cortex. J. Neurosci. 27, 12255–12266.

MacAulay, N., Zeuthen, T., 2012. Glial K+ clearance and cell swelling: key roles forcotransporters and pumps. Neurochem. Res. Feb 26. [Epub ahead of print].

Malik, P., McKenna, M.C., Tildon, J.T., 1993. Regulation of malate dehydrogenasesfrom neonatal, adolescent, and mature rat brain. Neurochem. Res. 18, 247–257.

McKenna, M.C., Stevenson, J.H., Huang, X., Hopkins, I.B., 2000. Differentialdistribution of the enzymes glutamate dehydrogenase and aspartateaminotransferase in cortical synaptic mitochondria contributes to metaboliccompartmentation in cortical synaptic terminals. Neurochem. Int. 37, 229–241.

Meier, E., Hertz, L., Schousboe, A., 1991. Neurotransmitters as developmentalsignals. Neurochem. Int. 19, 1–15.

Moonen, G., Sensenbrenner, M., 1976. Effects of dibutyryl cyclic AMP on culturedbrain cells from chick embryos of different ages. Experientia 32, 40–42.

Nguyen, N.H., Bråthe, A., Hassel, B., 2003. Neuronal uptake and metabolism ofglycerol and the neuronal expression of mitochondrial glycerol-3-dehydrogenase. J. Neurochem. 85, 831–842.

Nishino, K., Nowak Jr., T.S., 2004. Time course and cellular distribution of hsp27 andhsp72 stress protein expression in a quantitative gerbil model of ischemicinjury and tolerance. thresholds for hsp72 induction and hilar lesioning in thecontext of ischemic preconditioning. J. Cereb. Blood Flow Metab. 24, 167–178.

Pardo, B., Rodrigues, T.B., Contreras, L., Garzón, M., Llorente-Folch, I., Kobayashi, K.,Saheki, T., Cerdan, S., Satrústegui, J., 2011. Brain glutamine synthesis requiresneuronal-born aspartate as amino donor for glial glutamate formation. J. Cereb.Blood Flow Metab. 31, 90–101.

Patel, A.B., de Graaf, R.A., Rothman, D.L., Behar, K.L., Mason, G.F., 2010. Evaluation ofcerebral acetate transport and metabolic rates in the rat brain in vivo using 1H-[13C]-NMR. J. Cereb. Blood Flow Metab. 30, 1200–1213.

Patel, A.J., Balázs, R., 1970. Manifestation of metabolic compartmentation during thematuration of the rat brain. J. Neurochem. 17, 955–971.

Patel, A.J., Hunt, A., Gordon, R.D., Balázs, R., 1982. The activities in different neuralcell types of certain enzymes associated with the metabolic compartmentationglutamate. Brain Res. 256, 3–11.

Peng, L.A., Juurlink, B.H., Hertz, L., 1991. Differences in transmitter release,morphology, and ischemia-induced cell injury between cerebellar granule cellcultures developing in the presence and in the absence of a depolarizingpotassium concentration. Dev. Brain Res. 63, 1–12.

Pérez, L.J., Díaz de Arce, H., Cilloni, F., Salviato, A., Marciano, S., Perera, C.L., Salomoni,A., Beato, M.S., Romero, A., Capua, I., Cattoli, G., 2012. An SYBR Green-based real-time RT-PCR assay for the detection of H5 hemagglutinin subtype avianinfluenza virus. Mol. Cell Probes. 26, 137–145.

Pfaffl, M.W., 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 29, e45.

Ramos, M., del Arco, A., Pardo, B., Martínez-Serrano, A., Martínez-Morales, J.R.,Kobayashi, K., Yasuda, T., Bogónez, E., Bovolenta, P., Saheki, T., Satrústegui, J.,2003. Developmental changes in the Ca2+-regulated mitochondrial aspartate-glutamate carrier aralar1 in brain and prominent expression in the spinal cord.Brain Res. Dev. Brain Res. 143, 33–46.

Schousboe, A., Svenneby, G., Hertz, L., 1977. Uptake and metabolism of glutamate inastrocytes cultured from dissociated mouse brain hemispheres. J. Neurochem.29, 999–1005.

Schubert, P., Morino, T., Miyazaki, H., Ogata, T., Nakamura, Y., Marchini, C., Ferroni,S., 2000. Cascading glia reactions: a common pathomechanism and itsdifferentiated control by cyclic nucleotide signaling. Ann. N. Y. Acad. Sci. 903,24–33.

Somjen, G.G., Kager, H., Wadman, W.J., 2008. Computer simulations of neuron-gliainteractions mediated by ion flux. J. Comput. Neurosci. 25, 349–365.

Wandosell, F., Bovolenta, P., Nieto-Sampedro, M., 1993. Differences betweenreactive astrocytes and cultured astrocytes treated with di-butyryl-cyclicAMP. J. Neuropathol. Exp. Neurol. 52, 205–215.

Wang, F., Smith, N.A., Xu, Q., Fujita, T., Baba, A., Matsuda, T., Takano, T., Bekar, L.,Nedergaard, M., 2012. Astrocytes modulate neural network activity by Ca2+-dependent uptake of extracellular K+. Sci. Signal 5, ra26.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

Aralar mRNA and protein levels in neurons and astrocytes freshly isolated from young and adult mouse brain and in maturing cultured astrocytes1 Introduction2 Material and methods2.1 Animals and cell preparation2.2 Cultures2.3 Real-time PCR2.4 Statistical analysis

3 Results3.1 Selectivity and linear range of anti-aralar Western blots3.2 Aralar protein and mRNA levels in adult astrocytes and neurons obtained by FACS3.3 Aralar protein and mRNA levels in cultured astrocytes

4 DiscussionAcknowledgementsAppendix A Supplementary dataReferences