SLC25A12 expression is associated with neurite outgrowth and is upregulated in the prefrontal cortex of autistic subjects

Article metrics


Autism is a neurodevelopmental disorder with a strong genetic component, probably involving several genes. Genome screens have provided evidence of linkage to chromosome 2q31–q33, which includes the SLC25A12 gene. Association between autism and single-nucleotide polymorphisms in SLC25A12 has been reported in various studies. SLC25A12 encodes the mitochondrial aspartate/glutamate carrier functionally important in neurons with high-metabolic activity. Neuropathological findings and functional abnormalities in autism have been reported for Brodmann's area (BA) 46 and the cerebellum. We found that SLC25A12 was expressed more strongly in the post-mortem brain tissues of autistic subjects than in those of controls, in the BA46 prefrontal cortex but not in cerebellar granule cells. SLC25A12 expression was not modified in brain subregions of bipolar and schizophrenic patients. SLC25A12 was expressed in developing human neuronal tissues, including neocortical regions containing excitatory neurons and neocortical progenitors and the ganglionic eminences that generate neocortical inhibitory interneurons. At mid-gestation, when gyri and sulci start to develop, SLC25A12 molecular gradients were identified in the lateral prefrontal and ventral temporal cortex. These fetal structures generate regions with abnormal activity in autism, including the dorsolateral prefrontal cortex (BA46), the pars opercularis of the inferior frontal cortex and the fusiform gyrus. SLC25A12 overexpression or silencing in mouse embryonic cortical neurons also modified dendrite length and the mobility of dendritic mitochondria. Our findings suggest that SLC25A12 overexpression may be involved in the pathophysiology of autism, modifying neuronal networks in specific subregions, such as the dorsolateral prefrontal cortex and fusiform gyrus, at both pre- and postnatal stages.


Infantile autism was first characterized by Kanner in 1943, but the pathophysiological mechanisms underlying this disorder remain largely unknown.1 Autism is a developmental disorder characterized by social deficits, communication abnormalities, repetitive and stereotyped behavior and impaired planning and attention.2 It is associated with various degrees of mental retardation in about three-quarters of cases and with epilepsy in a third.2, 3 Various biological factors, primarily genetic factors, are thought to be involved in autism, and multiple genetic loci have already been implicated.4

Several studies have identified a potential autism susceptibility locus on chromosome 2q24–q33.4 Two single-nucleotide polymorphisms (SNPs) in the SLC25A12 gene, located on 2q31, have recently been associated with autism, conferring a genotype relative risk of up to 4.8.5 SNPs in SLC25A12 have been screened in four different family cohorts and association was confirmed in two of these studies, providing further evidence for the involvement of this gene in autism.6, 7, 8, 9 SLC25A12 encodes the AGC1/Aralar1 protein, a calcium-binding solute carrier located in the inner mitochondrial membrane, critical for the production of adenosine triphosphate and expressed principally in adult human muscle, heart and brain.10

We selected two brain regions—prefrontal cortex Brodmann's area (BA) 46 and cerebellum hemisphere lobule VI—for analysis by quantitative reverese transcription (RT)-PCR based on two criteria: energy requirements and functional dysregulation in autism. We hypothesized that SLC25A12 dysregulation would be detected in the brain subregions in which neurons have the highest energy demand. The two highest costs associated with neuronal activity are those relating to dendritic tree functioning and the propagation of axon potential along the axon, which is predicted to consume much of the energy of the neuron.11 Comparative studies of pyramidal neurons in granular prefrontal cortex have shown that human neurons display much larger dendritic trees and, on average, up to 20 times as many dendritic spines as those in the primary cortexes.12, 13 The maintenance of such dendritic complexity probably requires high levels of energy consumption.14 We thought that SLC25A12 expression might be modified in prefrontal cortex pyramidal neurons in adult brain. Several post-mortem studies have highlighted areas of anatomical abnormality in the autistic brain, with consistent findings observed in the limbic system and cerebellum.15, 16 Abnormal brain functioning has been reported in autistic subjects, particularly in the cortical networks involved in social interaction, such as the medial and inferior frontal cortex, the medial and ventral temporal lobes, superior temporal sulcus, amygdala and cerebellum.17, 18 We investigated the possible deregulation of SCL25A12 expression in autism, using mRNA extracted from post-mortem tissues from autistic individuals and controls, focusing on two regions: frontal cortex BA46 and cerebellum hemisphere lobule VI granule cells.

We also analyzed SLC25A12 expression during human brain development, searching for possible SLC25A12 molecular gradients in the brain gyri forming these neocortical regions at postnatal stages, corresponding to possible subregions with high-energy demand. We focused on two main developmental stages: 8 weeks and mid-gestation (19–23 weeks). At 8 weeks, neuronal progenitors proliferate rapidly in the ventricular zone and the first postmitotic neurons have migrated to generate the cortical plate. In humans, neocortical interneurons are generated from two different compartments, with 65% arising from the ventricular zone, a lineage specific to primates, and 35% from the ganglionic eminence, a ventral telencephalic region that also generates subcortical neurons, such as striatal neurons.19 It is therefore possible to determine whether a gene is expressed in one or both of these two key compartments: the ventricular zone and the ganglionic eminence. At mid-gestation (19–23 weeks), the human neocortex starts to form regions, with the genesis of sulci and gyri. The cortical plate also becomes more organized, with the formation of cell layers IV–VI.20 Neocortex parcellation into sulci and gyri is likely to require large amounts of energy, which may lead to differences in SLC25A12 expression in neocortical subregions. We used quantitative in situ hybridization for SLC25A12, to investigate molecular differences during this parcellation process.

Energy demands are likely to increase in gyri building local neuronal circuits during neocortex development. The functional activity of these neuronal circuits involves the elongation of interneuronal processes, such as dendrites.21 Mitochondria play a key role in these differentiation events, and a high level of dynamism is observed in dendritic mitochondrial distribution, regulated by synaptic activity and correlated with synapse morphogenesis.22 This raises the question of the possible involvement of SLC25A12 in dendritic outgrowth, which can be directly addressed by manipulating SLC25A12 expression levels by overexpression or silencing in mouse neocortical primary cultures and by measuring dendrite length and number.

We show here that SLC25A12 is more strongly expressed in the prefrontal cortical BA46 region in the brains of autistic patients than in controls. SLC25A12 expression was detected in the neocortical cortical plate containing the excitatory neurons. It was also detected in the ventricular zone and ganglionic eminences, two regions of the telencephalon known to generate the inhibitory interneurons of the neocortex. We identified a gradient of SL25A12 expression in the fetal medial and inferior frontal cortex. The inferior frontal cortex generates a region containing the mirror neurons involved in social cognition, which is known to be altered in autism. A second gradient was quantified in the fetal ventral temporal cortex, which generates the fusiform area, a region known to be involved in face recognition and to be functionally abnormal in autism. Finally, we showed that this overexpression of SCL25A12 resulted in an increase in dendrite length early in neocortical neuron development.

Materials and methods

Human brain post-mortem samples

Frozen post-mortem samples of human frontal cortex (BA46) and of cerebellum (lobule VI) from nine autistic patients and eight controls were provided by the Autism Tissue Program. The age, sex and post-mortem interval of individuals are presented in Supplementary Table 1.

Brain samples of BA46 from bipolar patients (n=33), schizophrenia (SZ) patients (n=34) and controls (n=34) were provided by the Stanley Foundation and have been described elsewhere.23, 24 Additional samples from bipolar (n=7), SZ (n=7) and controls (n=7) were obtained for BA11, BA22 and BA46, and have been described elsewhere.25

RNA extraction

We cut 20 sections of each individual brain, from BA46 (50 μm thick) and the cerebellum (8 μm thick), with a LEICA M3050S cryostat. Granule cells from lobule VI of the cerebellum samples were microdissected with the assistance of a laser, using an Arcturus PixCell instrument (Labstrade Inc., Miami, FL, USA). LMD 4.1 and IM1000 machines (Leica) were used for laser sectioning and image acquisition, respectively. Samples were collected and homogenized in Trizol reagent (GIBCO-BRL Life Technologies, Grand Island, NY, USA), treated with DNase I (Ambion, Austin, TX, USA) and processed according to the manufacturer's instructions.

Quantitative real-time RT-PCR analysis

Reverse transcription was carried out by incubating 20 μl of RT mixture containing 2.5 μl total RNA from human post-mortem tissue or 1 μg of total RNA from N18 cells, 200 U M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, USA), 0.5 mM of each dNTP and 6.25 μM random hexamers (Invitrogen) at 37 °C for 50 min. The cDNAs obtained were amplified by real-time PCR, using either Applied Biosystems probes (Hs00196245_m1 for NFL and Hs00186535_m1 for SLC25A12) or custom-made probes (PROLIGO) labeled at their 5′ ends with a fluorogenic reporter dye (FAM) and at their 3′ ends with a quencher dye (TAMRA). The sequences of the PCR primers and probes are given in Supplementary Table 2. PCR assays had a final reaction volume of 20 μl, and contained 2 U of Taq polymerase (Applied Biosystems, Foster City, CA, USA), 10 μM primers and fluorogenic probe. PCR was carried out over 40 cycles of 95 °C for 15 s, 60 °C for 1 min and 50 °C for 1 min. We used the Opticon2 sequence detection system (MJ Research/Biorad), with the Opticon Monitor software for data analysis. The NFL RNA sequence was amplified and used for the normalization of specific mRNA levels. Concentration ranges were constituted using serial dilutions of cDNAs from the whole human embryonic brain (8 weeks) from 0.1953 to 50 ng for both specific mRNAs and NFL mRNAs. Standard curves were constructed by PCR on these dilutions, to calculate the relative amounts of NFL mRNA and specific mRNA sequences in the experimental samples included in a given amplification reaction.

Prenatal human samples

Tissue samples from fetuses were provided by elective nontherapeutic abortion, with the informed consent of the parents, according to the recommendations of the local ethics committee. Tissues were fixed in 4% paraformaldehyde, embedded in paraffin and sectioned (5 μm). Four developmental stages were studied, corresponding to 8, 19, 21 and 23 weeks of gestation.

Probe synthesis and in situ hybridization

We used human and mouse cDNA clones from the RZPD Library corresponding to SLC25A12 (IRAKp961J0927Q2 and IRAKp961I1617Q2), AUTS2 (IRALp962I2117Q2) and PRKCB1 (IMAGp998D0311490Q1). We synthesized 35S-labeled riboprobes for SLC25A12 (human probe of 1990 bp and mouse probe of 1829 bp, linearized with BglII), AUTS2 (probe of 1171 bp linearized with BamHI) and PRKCB1 (probe of 608 bp linearized with SacII), using the P1460 riboprobe in vitro transcription system (Promega, Madison, WI, USA). Hybridization was carried out with both antisense and sense riboprobes, in the conditions described in a previous study.26 No signal was obtained with sense riboprobes. Expression was quantified with a Biospace Micro Imager, using Betavision analysis software (Biospace Instruments).27 Adjacent sections were stained with toluidine blue for histological examination.

SLC25A12-GFP constructs and shRNA vectors

The coding region of the SLC25A12 cDNA (2034 bp) was amplified by PCR from RZPD clone IRAKp961I1617Q2 using the following primers: forward 5′-IndexTermGAGACTCGAGATGCGCGTCAAGGTGCATACAACC-3′ (XhoI site underlined) and reverse 5′-IndexTermGAGAGGATCCCGTTGGGCTGCTGCTGCCGCCTTGGG-3′ (BamHI site underlined). The XhoI/BamHI digestion product of the amplicon was inserted into the multiple cloning site of the pEFGP-N1 (Clontech, Mountain View, CA, USA) expression vector, under control of the CMV promoter. Clones were verified by sequencing.

We silenced mouse Slc25a12 with shRNA vectors described elsewhere.28 shRNA Slc25a12 and control shRNA (scrambled) vectors were obtained from the SIGMA clone library. The shRNA Slc25a12 used in this study was the TRCN0000069910 clone.

N18 cell line cultures and transfection

The mouse neuroblastoma N18 cell line29 was cultured in Dulbecco's modification of Eagle's medium (DMEM) supplemented with 10% fetal calf serum, in 35 mm dishes. N18 cells were transfected with 800 ng of Slc25a12-enhanced green fluorescence protein (EGFP) construct or with 800 ng of Slc25a12-EGFP construct and 2400 ng of shRNA Slc25a12 vector, as described above. Immunocytochemistry analysis and RNA extraction were performed 24 h after transfection.

Primary cell culture, transfection and neurite analysis

E13.5 mouse telencephalic neurons were dissociated enzymatically (0.25% trypsin, DNase), mechanically triturated with a flamed Pasteur pipette and plated on 35 mm dishes (8 × 105 cells per dish) coated with poly-DL-ornithine (Sigma, St Louis, MO, USA), in DMEM (Invitrogen) supplemented with 10% fetal calf serum. Four hours after plating, DMEM was replaced with Neurobasal medium (Invitrogen) supplemented with 2 mM glutamine and 2% B27 (Invitrogen). In one experiment, primary neuronal cultures were transfected after 1 day in culture and analyzed on days 2, 3 or 4. In a second experiment, neurons were transfected after 2 or 3 days in culture and analyzed 24 h after transfection. Cells were transfected with the constructs described, above using Lipofectamine (Invitrogen), as described by the manufacturer. Neurite analysis was carried out with ImageJ software (Wayne Rasband, NIH).

Image acquisition and quantification of mitochondrial movements

Mitochondrial movements were analyzed with mouse cortical neurons cultured as described above, cotransfected with Slc25a12-EGFP and DsRedMito or with pEGFP control vector and DsRedMito after 5 days in culture, with analysis on day 6. For time-lapse videomicroscopy, the medium was replaced with fresh phenol red-free DMEM medium. Cell images were taken with a cooled charge-coupled device (CCD), CoolSnap HQ (Roper Scientific), mounted on the Leica DM IRE2 microscope and a × 100 oil immersion objective lens (N.A. 1.4, Zeiss). The cells were illuminated with a shutter DG4 illuminator. The intensity of DsRedMito-derived fluorescence along the processes of each neuron was measured, using a user-defined threshold, with MetaMorph software (Universal Imaging, West Chester, PA, USA). Cells were then maintained at 37 °C on the stage of Leica inverted microscope in an incubator containing 5% CO2 (Saur). Fluorescent images were captured every 5 s, for 15 min, with the CCD camera. We analyzed the movement of individual mitochondria manually, using the SpotTracker plugin of ImageJ (Wayne Rasband, NIH). We used Mander's coefficient (R) for colocalization analysis using Mander's coefficient plugin of ImageJ (Wayne Rasband, NIH).


Cells were fixed by incubation at room temperature for 20 min, in 4% paraformaldehyde in phosphate-buffered saline (PBS). They were permeabilized by incubation for 15 min in 0.5% Triton in PBS and incubated for 40 min with 3% bovine serum albumin 0.5% Tween-20 in PBS. The cells were then incubated for 2 h with primary antibody (mouse monoclonal anti-MAP2, 1 of 250, Sigma) in 3% bovine serum albumin and 0.5% Tween-20 in PBS. Cells were washed three times with 0.1% Tween-20 in PBS and incubated for 1.5 h with the secondary antibody (Cy3-anti-mouse immunoglobulin G, 1 of 400, Sigma) in 0.5% Tween-20 in PBS. Finally, the cells were washed three times. Immunofluorescence was observed with a TCS_4D confocal imaging system (Leica Instruments), using standard FITC filter sets.

Organotypic hippocampal slice culture, transfection, image acquisition and analysis

Rat organotypic hippocampal slice cultures were prepared as previously described.30 Cultures were maintained for 11–12 days in vitro and were transfected with a gene gun (Bio-Rad, Hercules, CA, USA), according to the manufacturer's instructions. Plasmids, at a ratio of 20 μg pcDNA3.1-EGFP to 40 μg pSport-SLC25A12, or shRNA-SLC25A12 construct or empty vector were precipitated onto 10 mg of 1.6 μm gold microcarriers. We reproducibly obtained the transfection of 10–30 cells per slice culture and expression was stable until 7–8 days. Control experiments showed that all cells cotransfected with EGFP and pDsRed constructs expressed green and red fluorescence (n=7). Two days after transfection, hippocampal organotypic cultures were submerged in a chamber perfused with an extracellular medium containing 124 mM NaCl, 1.6 mM KCl, 2.5 mM CaCl2, 1.5 mM MgCl2, 24 mM NaHCO3, 1.2 mM KH2PO4, 10 mM glucose and 2 mM ascorbic acid, saturated with 95% O2 and 5% CO2, pH 7.4, at 33 °C. Dendritic segments were imaged with a × 60 water immersion objective, using a Visitron (Puchheim, Germany) spinning disk confocal system. Control experiments showed that repetitive imaging under these conditions did not affect the morphology or viability of transfected cells. Z-stacks of 50–200-μm-long dendritic segments were taken on the secondary and tertiary dendrites of pyramidal neurons and the characteristics of dendritic protrusions were assessed with Metamorph software (Universal Imaging Corporation, Downingtown, PA, USA). The parameters analyzed included spine density (included all protrusions) and length (measured from the limit of the dendrite to the tip of the protrusion), size of spine head (taken as the largest diameter).

Statistical analysis

Results for quantitative real-time PCR were analyzed using Student's t-test, analysis of variance and linear regression.


Upregulation of SLC25A12 expression in the post-mortem prefrontal cortex

Quantitative real-time RT-PCR was performed on human post-mortem samples from a frontal cortical region (BA46) and on the granule cells of cerebellum hemisphere lobule VI to determine the levels of expression of SLC25A12, L1CAM and NFL. We used the mRNA level for NFL, a gene with bona fide neuron-specific expression, to normalize the mRNA levels of the other two genes (Figures 1a–d). Relative transcript levels for nine autistic brain samples were compared with those for eight controls. SLC25A12 was more strongly expressed (P=0.028) in the post-mortem frontal cortex of autistic subjects (241.2±30.9) than in controls (157.5±10.6) (Figure 1a; Supplementary Table 2). No significant differences were observed in the relative levels of L1CAM mRNA (autistic subjects: 145.4±35.6; controls: 118.4±35.1). No difference in the level of expression of SLC25A12 in the granule cells of cerebellum lobule VI was found between autistic subjects (96.9±25.7) and controls (96.1±21.5) (Figure 1c; Supplementary Table 3). As for BA46, no significant differences were observed in the relative levels of L1CAM mRNA, (autistic subjects: 139.5±29.8 and controls: 111.1±38.8) in cerebellum lobule VI. The two groups differed significantly in age distribution but regression analysis showed that the levels of expression of the three genes were independent of age, sex and post-mortem interval (Supplementary Table 2). All the patients were homozygous for the risk SNPs rs2056202 (G/G) and rs2292813 (G/G), whereas two of the eight controls were heterozygous for these two polymorphisms. This finding is consistent with a causal role for autism-associated SLC25A12 allelic variants. These results suggest that SLC25A12 is upregulated in a specific area of the brain in autistic subjects.

Figure 1

SLC25A12 and L1CAM mRNA levels in post-mortem human prefrontal cortex (BA46) and cerebellum (lobule VI) granule cell specimens from autistic patients and controls. We quantified mRNA levels by quantitative real-time RT-PCR (QRT-PCR). Granule cells were microdissected from lobule VI as described in the ‘Materials and methods’. SLC25A12 (a) and L1CAM (b) mean mRNA levels in the BA46 prefrontal cortex of subjects with autism (n=9) and controls (n=8). SLC25A12 (c) and L1CAM (d) mean mRNA level in the lobule VI granule cells of subjects with autism (n=9) and controls (n=8). *P<0.05.

Genes encoding mitochondrial proteins have been shown to be deregulated in both bipolar disorder (BP) and SZ.25, 31, 32 Together with SLC25A12 upregulation in autism, these findings suggested that SLC25A12 might be deregulated in BP and SZ. We used previously studied samples23, 24, 25 to investigate SLC25A12 expression in three neocortical regions: orbitofrontal and mid-frontal gyri (BA11), mid- and inferior-frontal gyri (BA46) and the Wernicke's area of the temporal lobe (BA22). No statistically significant differences were detected by quantitative real time RT-PCR between the BP and SZ samples and controls (Supplementary Tables 4 and 5).

We also investigated whether this upregulation was specific to SL25A12 or whether other genes encoding mitochondria proteins were also deregulated in autism. We studied AK2, CASQ1 and NDUFV2, which encode mitochondrial proteins; NDUFV2 has been shown to be deregulated in BP and SZ.25, 32 No significant differences in transcript levels for these three genes in BA46 and cerebellum were found between autistic and control samples (Supplementary Tables 2 and 3).

SLC25A12 expression in the human embryonic telencephalon

Quantitative in situ hybridization was used to investigate the pattern of SLC25A12 expression during human prenatal development. This approach can be used to detect differences in expression between regions, and gradients of expression in a given structure. At 8 weeks, SLC25A12 expression was observed almost exclusively in the brain (Figure 2). The SLC25A12 transcript was detected in the central nervous system, including the telencephalon, hippocampal anlagen, mesencephalon and cerebellar anlagen (Figures 2a and b). Interestingly, quantification of the radioactive signal in the telencephalon indicated a specific gradient of expression, decreasing from posterior (28.69±1.90 c.p.m. mm−2) to anterior (15.15±4.05 c.p.m. mm−2) (Figure 2a; Supplementarys Figure S1a1–a3). SLC25A12 was equally strongly expressed in the cortical plate containing the postmitotic neurons and in the ventricular zone, in which precursor neurons divide (Figures 2c and d). SLC25A12 was widely expressed in ganglionic eminences at this developmental stage (Figures 2a and b; Supplementary Figure S1a1). The SLC25A12 transcript was also detected in the spinal cord, adrenal medulla, kidney and liver (Figures 2a and b). The pattern of expression of SLC25A12 in the central nervous system is consistent with its function of providing ATP—the chief energy source of the neurons—and an indirect role in both proliferating progenitors and postmitotic neurons.

Figure 2

SLC25A12 mRNA levels in the human central nervous system during early development. Sections of 8-week-old human embryos hybridized with SLC25A12 antisense radioactive riboprobe (a, c) or stained with toluidine blue (b, d). (a) SLC25A12 transcripts were detected in the telencephalon (tel), hippocampus anlagen (hipp), cerebellum anlagen (Cb), ganglionic eminence (ge), liver (li), the adrenal medulla (ad), kidney (k) and spinal cord (sc), but not in the heart (h) and lung (lu) (sagittal sections). (c) SLC25A12 transcripts are present in equal amounts in the cortical plate (cp) and ventricular zone (vz) of the telencephalon (transverse sections). Scale bars indicate 1 mm.

We compared the expression pattern of SLC25A12 with that of other autism candidate genes, AUTS2 and PRKCB1. AUTS2, located on 7q11.2, has been identified as a candidate autism susceptibility gene, as it was found to be disrupted in a pair of autistic twins.33 Rare variants of this gene have been identified in cases of autism.34 Haplotypes of the PRKCB1 gene, located on 16p11.2 and encoding a protein kinase c-β, have recently been reported to be associated with autism.35 PRKCB1 is expressed in cerebellar Purkinje cells and is involved in the induction of long-term depression in these neurons.36 Like SLC25A12, AUTS2 and PRKCB1 were found to be expressed in the telencephalon, ganglionic eminence, cerebellum anlagen and, more weakly, in the medulla oblongata (Supplementary Figure S1). We compared the expression of each gene between the different subregions (Table 1). In contrast to the gradient of SLC25A12 expression along the antero–posterior axis (anterior-low, posterior-high), AUTS2 and PRKCB1 were uniformly expressed throughout the telencephalon (Supplementary Figure S1a1–c4). SLC25A12 expression in the posterior telencephalon was five times stronger than PRKCB1 expression and 30 times stronger than AUTS expression (Supplementary Figure S2). Like SLC25A12, AUTS2 was strongly expressed, to equivalent levels, in both the cortical plate and ventricular zone. In contrast, PRKCB1 expression was restricted to the ventricular zone. SLC25A12, AUTS2 and PRKCB1 were all expressed in the cerebellum anlagen.

Table 1 Quantified mRNA expression of SLC25A12, PRKCB1 and AUTS2 at 8 weeks

SLC25A12 gradients in the lateral frontal cortex and in the ventral temporal lobe at mid-gestation

We investigated SLC25A12 expression in subregions of the neocortex, by examining human fetal brain sections taken at mid-gestation (19–23 weeks) (Figure 3; Supplementary Figure S3). At 19 weeks, SLC25A12 expression was restricted to specific telencephalic compartments, including the frontal and temporal neocortex, caudate nucleus and putamen (Figure 3). It was also maintained in the ganglionic eminences (Figure 3) and found in the cerebrocerebellum (Supplementary Figure S4). SLC25A12 expression was strongest in the central part of the prefrontal cortex, between the superior frontal sulcus anlagen and the anterior rim of the Sylvian sulcus (Figure 3a). Quantification of the radioactive signal indicated a specific gradient of expression: high in the lateral prefrontal cortex region (2.19±0.28 c.p.m mm−2; P<0.05), low in the cingular cortex (0.49±0.02 c.p.m. mm−2), in the superior prefrontal cortex (0.93±0.19 c.p.m. mm−2), in the insular cortex (0.93±0.27 c.p.m. mm−2) and in the lateral temporal cortex (0.87±0.08c.p.m. mm−2 (Figure 3a).

Figure 3

SLC25A12 molecular gradients in human neocortex at mid-gestation. Schematic diagram and brain positions of 19-week-old human fetal brain coronal sections (Aa, Ba). Sulci present at 19 week are indicated. Coronal sections of brains hybridized with SLC25A12 antisense radioactive riboprobe (Ab, Bb), stained with toluidine blue (Ac, Bc), and quantification (Ad, Bd). A(ad) Identification of a molecular gradient in the lateral prefrontal cortex. SLC25A12 transcripts were also detected in the ganglionic eminence (ge), caudal nucleus (nc), putamen (p) and internal capsule (ic). B(a-d) Identification of a molecular gradient in the future fusiform gyrus of the ventral temporal cortex. SLC25A12 transcripts were also detected in hippocampus anlagen (hipp) and future parahippocampal cortex (para hipp). Note peaks of gradients in lateral prefrontal neocortex (Ad) and in the future fusiform gyrus (Bd), respectively. *P<0.01; **P<0.001.

A second gradient of SLC25A12 expression was observed in the ventral temporal cortex, running from the collateral sulcus to the lateral temporal cortex (Figure 3b). Quantification of the radioactive signal indicated a specific gradient of expression: high in the future fusiform gyrus (5.12±0.70 c.p.m. mm−2; P<0.001), low in the future parahippocampal gyrus and hippocampal anlagen (2.65±0.25 c.p.m. mm−2), in the lateral temporal cortex (1.50±0.25 c.p.m. mm−2) and in the frontal/parietal cortex (near the future central sulcus) (1.08±0.10 c.p.m. mm−2) (Figure 3b). Surprisingly, the SLC25A12 transcript was detected in the hippocampus anlagen at 8 and 19 weeks, but not in the hippocampus at 23 weeks (Supplementary Figure S3b2; c2).

We again compared the pattern of expression of SLC25A12 with those of AUTS2 and PRKCB1 (Supplementary Figure S3b3–b4; c3–c4 and Table 2). Unlike SLC25A12, which was not expressed in the hippocampus, both PRKCB1 and AUTS2 were expressed in the dentate gyrus, CA1 and CA3 pyramidal cell subregions (Supplementary Figure S3c3, c4). PRKCB1 expression also followed a gradient, from the lateral temporal lobes to the ventral temporal regions generating the fusiform gyrus and parahippocampal gyrus, reminiscent of the gradient described for SLC25A12 at 19 weeks (compare Supplementary Figure S3b3 with Figure 3b). In contrast, AUTS2 expression was weak in these regions. Like SLC25A12, PRKCB1 and AUTS2 were expressed in the ganglionic eminences, caudate nucleus and putamen nuclei, which were identifiable at this developmental stage (Figures 3a and b; Supplementary Figure S3b2–b4). We therefore observed two molecular gradients of SLC25A12 expression: a twofold difference between the superior frontal sulcus anlagen and the anterior rim of Sylvian sulcus (insular cortex) and a threefold difference between the future fusiform gyrus of the temporal cortex and the lateral temporal region.

Table 2 Quantified mRNA expression gradient of SLC25A12, PRKCB1 and AUTS2 in ventral temporal lobe at mid-gestation

Pattern of Slc25a12 expression during mouse brain development

We compared the patterns of SLC25A12 expression in human and mouse brains. Embryonic day 14 (E14) in mice has been likened to the developmental stage corresponding to 8 weeks of gestation in humans.37 At E14, Slc25a12 was strongly expressed in both the cortical plate and ganglionic eminence of the mouse telencephalon (Supplementary Figure S5a). These data are fully consistent with the pattern of expression at 8 weeks in humans (Figure 2 and Supplementary Figure). The hippocampus matures postnatally in rodents, whereas it develops prenatally in primates.38, 39, 40 We observed Slc25a12 expression in postnatal mouse hippocampus (P21), whereas the Slc25a12 transcript was not detected in the 23 weeks human hippocampus (Supplementary Figure S5b), consistent with the relative developmental periods. In addition, no molecular gradient of Slc25a12 expression was observed in the mouse brain.

Effect of Slc25a12 overexpression and silencing on dendrites at pre- and postnatal stages

In studies of post-mortem BA46 sections from autistic patients and controls, we found that SLC25A12 was significantly more strongly expressed in autistic patients than in controls, and we observed SLC25A12 expression in the cortical plates of humans and mice, at equivalent stages (8 weeks in human and E14 in mouse). We assessed the possible role of SLC25A12 in dendrite outgrowth by overexpressing Slc25a12 in neurons derived from embryonic day 13 (E13) embryonic mice. Primary cultures of cortical neurons can be classified as pyramidal or nonpyramidal on the basis of morphological criteria.41 Pyramidal cells have a longer apical dendrite and basal dendrites. We focused our analysis on pyramidal neurons (Figures 4a–c). In the first experiment, cortical neurons were plated on E13, transfected on day 1 in culture 1 (DIC1) and analyzed on DIC2, DIC3 and DIC4. We analyzed the effects of overexpression (Figure 4d) or silencing (Figure 4e). On DIC2, neurons overexpressing Slc25a12 had significantly (P<0.05) longer (68.4±7.5 μm, n=24) apical dendrites than neurons transfected with the control EGFP vector (39.8±5.5 μm, n=11). On DIC3, neurons overexpressing Slc25a12 had significantly (P<0.05) shorter (239.7±49.0 μm, n=7) apical dendrites than neurons transformed with the control green fluorescent protein (GFP) vector (467.2±73.0 μm, n=12) (Figures 4b and d). On DIC4, neurons overexpressing Slc25a12 did not survive whereas neurons transfected with the control GFP vector had apical dendrites three times longer than on DIC3 (1682.3±139.9 μm, n=6). For silencing experiments, we selected an efficient shRNA plasmid on the basis of Slc25a12 expression change in a mouse neuroblastoma cell line (N18). Slc25a12-GFP was overexpressed in these cells and we quantified changes in Slc25a12 transcript levels and EGFP fluorescence intensity as an indicator of Slc25a12-EGFP fusion protein levels (Supplementary Figure S6). On DIC2 and DIC3, silenced neurons had significantly (P<0.05) shorter (107.3±17.9 μm, n=10 and 371.5±73.5 μm, n=11, respectively) apical dendrites than neurons transfected with the control scrambled shRNA (205.5±28.1 μm, n=12 and 660.8±107.6, n=11, respectively) (Figures 4c and e). In contrast, 3 days after transfection with shRNA (DIC4), silenced and control neurons did not differ significantly in apical dendrite length (shRNA value: 765.5±95.3 μm, n=12; scrambled value: 657.9±105.7, n=6) (Figure 4e). In a second experiment, cells were transfected on DIC2 or DIC3. Both the overexpression and silencing of Slc25a12 resulted in a decrease in apical dendrite length. For Slc25a12 overexpression, apical dendrite length was significantly shorter on DIC3 (195.3±13.6 μm, n=12 vs 382.7±52.5 μm, n=14 with P<0.01) (Figure 4f), whereas for Slc25a12-silenced neurons, it was significantly shorter on DIC4 (567.9±62.6 μm, n=14 vs 984.5±194.7 μm, n=7 with P<0.05) (Figure 4g).

Figure 4

Slc25a12 overexpression and silencing in dendrites at pre- and postnatal stages. A(ad)–C(a, b) Slc25a12 overexpression and silencing in primary cultures of cortical neurons maintained for 1 day in vitro, transfected on day 1 in culture 1 (DIC1) and analyzed after 24 h (DIC2), 48 h (DIC3) and 72 h (DIC4). Left upper panel represents the construct of the Slc25a12-EGFP vector and right upper panel represents the first experimental paradigm. Confocal slices of neurons transfected with EGFP control vector (Aa, Ba) or Slc25a12-EGFP vector, visualized by EGFP fluorescence (green) (Ac, Bb) MAP2 immunofluorescence (red) (Ab) and the merged image (Ad). Scale bar indicates 10 μm. Confocal slices of neurons transfected with scrambled shRNA vector (Ca) or shRNA Slc25a12 vector (Cb). Scale bar indicates 10 μm. (D, E) Quantitative analysis of EGFP-labeled apical dendrites in neurons overexpressing Slc25a12-EGFP (D) or with shRNA-Slc25a12 silencing (E), expressed as a percentage, according to the ratio of pEGFP and scrambled shRNA controls, respectively. The differences between Slc25a12-EGFP and EGFP samples, and between shRNA-Slc25a12 and scrambled shRNA were significant on DIC2 and DIC3 (*P<0.05). On DIC4, Slc25a12 overexpression was toxic and Slc25a12 repression lost its effect. (F, G) Representation of the second experiment. Note that neurons were transfected on DIC2 or DIC3 and analyzed 24 h after transfection (DIC3 and DIC4, respectively). Quantitative analysis of EGFP-labeled apical dendrites in neurons overexpressing Slc25a12-EGFP (F) or with shRNA-Slc25a12 silencing (G), expressed as a percentage, according to the ratio of pEGFP and scrambled shRNA controls, respectively. The difference between Slc25a12-EGFP and EGFP samples is significant on DIC3 (**P<0.01) and the difference between shRNA-Slc25a12 and scrambled shRNA is significant on DIC4 (*P<0.05). (H) Slc25a12 silencing but not overexpression modify dendritic spine morphology in hippocampal pyramidal neurons. Illustration of the spine morphology of pyramidal neurons transfected EGFP with either empty vector or the Slc25a12 construct or shSlc25a12 (Ha). Scale bar indicates 5 μm. Quantitative analysis of the number of protrusions, spine length and head width (Hb). Significant elongation of spine demonstrated by the right shift of the cumulative distribution of spine length. No statistical differences were identified in protrusion number and spine width. Bars represent the s.e.m. for 451–1310 protrusions in 5–17 experiments.

Finally, we also investigated the effects of Slc25a12 overexpression and Slc25a12 knockdown on the ex vivo postnatal cortical network constituted by hippocampal organotypic slices from postnatal day 6 rats.30 Two days after the transfection of slices, neuronal morphology was studied by confocal imaging. In the majority of pyramidal neurons overexpressing Slc25a12, no detectable change in the organization of dendritic arbors was found. However a small proportion of Slc25a12-transfected cells showed signs of toxicity, such as enlargement of soma, dendritic swelling and excessive protrusion growth (25%, 3 of 12 neurons). Neurons in slices transfected with wild-type Slc25a12 (n=9) had similar numbers of dendritic spines and these were similar in size to those observed in EGFP-transfected control slices (n=7) (protrusion density: 0.98±0.08 per micrometer for Slc25a12 vs 0.99±0.05 for EGFP only; spine length: 1.13±0.03 μm vs 1.04±0.05 μm; spine head width: 0.66±0.02 μm vs 0.6±0.02 μm). In contrast, shRNA-Slc25a12 expression modified the dendritic spine morphology without obvious dendritic arbor changes. The length of spine was significantly increased by Slc25a12 knockdown (n=5; 1.38±0.17 μm) compared to EGFP cotransfected with empty vector (n=10; 1.12±0.02 μm, P<0.05). The spine density (0.91±0.11 per micrometer vs control: 0.82±0.04 per micrometer) and the spine head diameter (0.68±0.04 μm vs control: 0.66±0.01 μm) were however not changed significantly by sh-Slc25a12 transfection.

Functional analysis of Slc25a12 overexpression on mitochondrial trafficking in cortical dendrites

We investigated the functional deregulation induced by SLC25A12 overexpression in autism, by studying the effects of Slc25a12-EGFP on mitochondrial trafficking in the dendrites of neocortical neurons in primary culture. Slc25a12-EGFP fully colocalized with DsRedMito, a marker targeted to the mitochondrial matrix (Mander's coefficient R=0.972±0.011 for DsRedMito vs control EGFP, with n=14; R=0.995±0.020 for DsRedMito vs Slc25a12-EGFP, with n=20; P<0.05) (Figure 5a). As previously described in primary cultures of cortical neurons, dendritic mitochondria displays processive and oscillatory movements in both directions in dendrites, at a speed of 6.5±0.5 nm s−1.42

Figure 5

Slc25a12 overexpression modifies mitochondria trafficking in cortical dendrites. DIC (Aa), DsRedMito (Ab), Slc25a12-EGFP (Ac) and colocalization (Ad) pictures of a cotransfected neuron. Scale bar: 10 μm. (B) Time-lapse movies of the rectangular areas for Ab, illustrating two examples of mitochondrial movement for 140 s with 20 s intervals between frames. Scale bars: 1 μm. (C) Merged picture of pEGFP control vector (green) and DsRedMito (red) for cotransfected neurons (Ca). Scale bar: 10 μm. We show a time-lapse movie of the rectangular area in the right panel (Cb) illustrating the movement of the mitochondria for 140 s with 20 s intervals between frames. Scale bar: 1 μm. (D) Changes in the mitochondrial speeds of neurons cotransfected with Slc25a12-EGFP and DsRedMito. Note that mitochondrial speed is significantly higher in Slc25a12-overexpressing neurons than in control neurons (*P<0.05).

We calculated that the mobility of mitochondria stained with dsRedMito following dsRedMito/Slc25a12-EGFP cotransfection was 14.9±3 μm s−1 (n=16), whereas that after dsRedMito/pEGFP control vector transfection was 8±1.3 μm s−1 (n=22). Slc25a12 overexpression therefore significantly increased the velocity of the mitochondria (P=0.03) (Figure 5b–d).


Autism-specific SLC25A12 overexpression in the BA46 prefrontal cortex region

SLC25A12 was overexpressed in the BA46 of the prefrontal cortex of autistic subjects. The levels of expression of the genes studied were independent of age, sex and post-mortem interval, in patients, controls and both groups considered together. The expression of only one of these genes, SLC25A12, was significantly higher in post-mortem prefrontal BA46 cortex specimens from nine autistic patients than in eight controls. Samples from the same individuals showed no significant difference in expression of this gene in the cerebellar granule cells isolated by laser-assisted microdissection. These results are limited by the number of cases that we were able to study and by the restricted number of brain regions studied. Despite these limitations, our results support the hypothesis that SLC25A12 is more strongly expressed in an adult brain subregions involved in social interaction in autistic subjects than in controls. Our quantitative real time RT-PCR results for human post-mortem brain tissues suggest that SLC25A12 is overexpressed in only part of this system.

We analyzed possible changes in the expression of AK2, CASQ1 and NDUFV2, encoding mitochondrial proteins, to discriminate between a primary change in SLC25A12 expression linked to a particular allelic form of the SLC25A12 gene in autistic patients and a compensatory increase in SLC25A12 transcription due to an abnormal demand for cortical energy in autistic brains. No changes were detected, suggesting that SLC25A12 overexpression may be linked to a particular allelic form of this gene. No SLC25A12 upregulation was detected in brain samples (BA11, BA22 and BA46 subregions) from bipolar and SZ patients, consistent with SLC25A12 overexpression being specific to autistic disorders.

This overexpression in BA46 of the dorsolateral prefrontal cortex (DLPFC) may be associated with the changes in prefrontal cortex structure and function reported in patients with autism. Regionally, the frontal lobes are the most enlarged in autistic patients.43 Cell columns are more numerous in the brains of autistic patients, and are smaller and less compact in their cellular configuration in BA9 of the DLPFC.44 Functional magnetic resonance imaging has shown that these prefrontal structures are associated with impaired executive functioning, relating to working memory and attention in particular.45, 46 Diffusion tensor imaging has demonstrated differences in long-distance frontoparietal connections between autistic subjects and controls.47 These results are consistent with the local deregulation of networks involved in working memory and attention, with a possible increase in local connectivity and a selective loss of long-range connectivity.48

Different expression patterns for human SLC25A12 and its mouse ortholog

We found that SLC25A12 was expressed mostly in the brain during human embryonic and fetal development. However, this restricted pattern of expression changes after birth, as high levels of SLC25A12 expression have been reported in skeletal muscle, heart, pancreas and kidney in adults, with low levels of expression observed in the brain.10 Furthermore, the neuronal regions expressing SLC25A12 (or Slc25a12, the mouse ortholog) seem to differ in mice and humans. In particular, we observed no expression of this gene in the human hippocampus at 23 weeks, whereas strong expression was detected in the mouse hippocampus both before and after birth.49 Our results and published data show that the spatio-temporal regulation of human SLC25A12 expression evolved from that for the mouse ortholog, as already described for other encoding neuron-specific proteins.50

A prenatal pattern of expression in regions involved in social cognition

Quantitative ISH revealed the existence of three gradients of SLC25A12 expression in the developing neocortex, which we were able to measure. At 8 weeks, a decreasing gradient of expression from the posterior to the anterior of the telencephalon was observed, similar to that reported for several transcription factor genes, such as TBR1, EMX1 and LHX2, and for genes encoding the EphA family of receptor tyrosine kinases in fetal macaque brain.51

At mid-gestational stages, two molecular gradients of SLC25A12 expression were observed, in the lateral frontal cortex and in the ventral temporal cortex. These two gradients were not found in expression analysis for two other candidate autism susceptibility genes, AUTS2 and PRKCB1, demonstrating the specificity of these gradients. The formation of higher-order cortical areas and circuits is a robust feature specific to the primate brain and not observed in mouse brain.52 The lateral frontal cortex gradient was located in a region responsible for generating the DLPFC, including the BA46 and pars opercularis of the inferior frontal gyrus. The DLPFC is functionally abnormal in children with autism spectrum disorders (ASDs).45, 46, 47 The pars opercularis of the inferior frontal contains mirror neurons.53 Mirror neurons, activated during goal-directed actions and the observation of such actions performed by others, are impaired in individuals with ASDs.54 The ventral temporal cortex gradient of SLC25A12 expression is located in the area that generates the fusiform gyrus. This region is known to mediate social perception in the visual domain and to be differentially activated in autistic patients as a function of gaze fixation.55, 56 High levels of SLC25A12 expression in some gyri may help to optimize cortical circuits, in agreement with the strong expression of Slc25a12 in tonic active neurons in mice.49

Dendritic phenotypes in relation to SLC25A12 contribution to autism

Gradients of expression for genes encoding proteins involved in energy demand may be important for the development of cortical networks. We observed that overexpression or silencing of SLC25A12 had a direct impact on dendritic outgrowth, respectively, increasing or decreasing the mean length of the apical dendrite, on mitochondria traffic in dendrites and dendritic spines morphology at postnatal stages. Cytoskeletal proteins play a critical role in neuronal development, including roles in growth cone motility and dendrite outgrowth.57, 58 Microtubules, neurofilaments and associated proteins together regulate dynamic polymerization and transport to generate axon, dendrites and branch outgrowth. Microtubules and actin filaments require ATP for their polymerization, especially at the motile end of the growth cone. The mitochondria provide the ATP that leads to dendrite outgrowth. This ATP production is impacted by SLC25A12 expression in cell cultures and in mice where tonic active neurons strongly expressed Slc25a12 and increased ATP production.49, 59, 60 The complex array of cytoskeletal proteins maintain dendritic morphology and allow the movement of organelles, including mitochondria, within the cytoplasm from the soma to the periphery of cell. It is important to note that we observed a significant increase velocity of mitochondria in the context of neurons overexpressing SLC25A12.

Silencing experiments indicate that shRNA can reduce at least by half the Slc25a12 transcript levels, mimicking Slc25a12 haploinsufficiency. In contrast, Slc25a12 overexpression cannot be easily measured in single neurons. We cannot exclude that part of the observed effects are indirect and mediated by cell pathology mechanisms induced by disruption of the protein organization of the inner membrane of mitochondria where SLC25A12 products are inserted. Further experiments will be needed to address these questions. Whatever the involved mechanisms, our results point to the importance of a normal SLC25A12 gene dosage in cortical neurons.

A possible pathophysiological autism model involving SLC25A12 misregulation

SLC25A12 overexpression in specific brain gyri may affect development and cognitive functions in the brain. Energy provision via SLC25A12 is involved in three key mechanisms of normal brain development—sulcus formation, dendritic outgrowth and synaptic plasticity—which might be particularly important in autism frequently considered as a neurodevelopmental disease.61 However, cognitive brain disorders such as autism, are typically pleiotropic, and multiple genes are thought to be involved.4 The identification of subtle interactions between several genetic effectors will require additional approaches, such as the generation of genetically modified mice.62, 63 The generation of mouse models with different levels of expression of genes of interest, including SLC25A12, should make it possible to identify the phenotypic consequences of candidate gene deregulation for simple phenotypic traits potentially related to autism.


  1. 1

    Kanner L . Autistic disturbances of affective contact. Nerv Child 1943; 2: 217–250.

  2. 2

    Volkmar FR, Pauls D . Autism. Lancet 2003; 362: 1133–1141.

  3. 3

    Tuchman R, Rapin I . Epilepsy in autism. Lancet Neurol 2002; 1: 352–358.

  4. 4

    Veenstra-VanderWeele J, Cook Jr EH . Molecular genetics of autism spectrum disorder. Mol Psychiatry 2004; 9: 819–832.

  5. 5

    Ramoz N, Reichert JG, Smith CJ, Silverman JM, Bespalova IN, Davis KL et al. Related linkage and association of the mitochondrial aspartate/glutamate carrier SLC25A12 gene with autism. Am J Psychiatry 2004; 161: 662–669.

  6. 6

    Blasi F, Bacchelli E, Carone S, Toma C, Monaco AP, Bailey AJ, et al., International Molecular Genetic Study of Autism Consortium (IMGSAC). SLC25A12 and CMYA3 gene variants are not associated with autism in the IMGSAC multiplex family sample. Eur J Hum Genet 2006; 14: 123–126.

  7. 7

    Segurado R, Conroy J, Meally E, Fitzgerald M, Gill M, Gallagher L . Confirmation of association between autism and the mitochondrial aspartate/glutamate carrier SLC25A12 gene on chromosome 2q31. Am J Psychiatry 2005; 162: 2182–2184.

  8. 8

    Rabionet R, McCauley JL, Jaworski JM, Ashley-Koch AE, Martin ER, Sutcliffe JS et al. Lack of association between autism and SLC25A12. Am J Psychiatry 2006; 163: 929–931.

  9. 9

    Turunen J, Tero Ylisaukko-oja T, Kilpine H, Rehnström K, Kempas E, Vanhala R et al. Association Analysis of SLC25A12 and EN2 in the Finnish Families with Autism-Spectrum Disorders. WCPG: Cagliari, Italy, 2006, abstracts.

  10. 10

    del Arco A, Satrustegui J . Molecular cloning of Aralar, a new member of the mitochondrial carrier superfamily that binds calcium and is present in human muscle and brain. J Biol Chem 1998; 273: 23327–23334.

  11. 11

    Attwell D, Laughlin SB . An energy budget for signaling in the grey matter of the brain. J Cereb Blood Flow Metab 2001; 21: 1133–1145.

  12. 12

    Elston GN . Pyramidal cells of the frontal lobe: all the more spinous to think with. J Neurosci 2000; 20: RC95.

  13. 13

    Elston GN . Cortex, cognition and the cell: new insights into the pyramidal neuron and prefrontal function. Cereb Cortex 2003; 13: 1124–1138.

  14. 14

    Sherwood CC, Stimpson CD, Raghanti MA, Wildman DE, Uddin M, Grossman LI et al. Evolution of increased glia-neuron ratios in the human frontal cortex. Proc Natl Acad Sci USA 2006; 103: 13606–13611.

  15. 15

    Bauman ML, Kemper TL . Neuroanatomic observations of the brain in autism: a review and future directions. Int J Dev Neurosci 2005; 23: 183–187.

  16. 16

    Palmen SJ, van Engeland H, Hof PR, Schmitz C . Neuropathological findings in autism. Brain 2004; 127: 2572–2583.

  17. 17

    Baron-Cohen S, Belmonte MK . Autism: a window onto the development of the social and the analytic brain. Annu Rev Neurosci 2005; 28: 109–126.

  18. 18

    DiCicco-Bloom E, Lord C, Zwaigenbaum L, Courchesne E, Dager SR, Schmitz C et al. The developmental neurobiology of autism spectrum disorder. J Neurosci 2006; 26: 6897–6906.

  19. 19

    Letinic K, Zoncu R, Rakic P . Origin of GABAergic neurons in the human neocortex. Nature 2002; 417: 645–649.

  20. 20

    Rakic P . Specification of cerebral cortical areas. Science 1988; 241: 170–176.

  21. 21

    Whitford KL, Dijkhuizen P, Polleux F, Ghosh A . Molecular control of cortical dendrite development. Annu Rev Neurosci 2002; 25: 127–149.

  22. 22

    Li Z, Okamoto K, Hayashi Y, Sheng M . The importance of dendritic mitochondria in the morphogenesis and plasticity of spines and synapses. Cell 2004; 119: 873–887.

  23. 23

    Ohnishi T, Yamada K, Ohba H, Iwayama Y, Toyota T, Hattori E et al. A promoter haplotype of the inositol monophosphatase 2 gene (IMPA2) at 18p11.2 confers a possible risk for bipolar disorder by enhancing transcription. Neuropsychopharmacology 2007; 32: 1727–1737.

  24. 24

    Aoki-Suzuki M, Yamada K, Meerabux J, Iwayama-Shigeno Y, Ohba H, Iwamoto K et al. A family-based association study and gene expression analyses of netrin-G1 and -G2 genes in schizophrenia. Biol Psychiatry 2005; 57: 382–393.

  25. 25

    Nakatani N, Hattori E, Ohnishi T, Dean B, Iwayama Y, Matsumoto I et al. Genome-wide expression analysis detects eight genes with robust alterations specific to bipolar I disorder: relevance to neuronal network perturbation. Hum Mol Genet 2006; 15: 1949–1962.

  26. 26

    Al Halabiah H, Delezoide AL, Cardona A, Moalic JM, Simonneau M . Expression pattern of NOGO and NgR genes during human development. Gene Expr Patterns 2005; 5: 561–568.

  27. 27

    Charpak G, Dominik W, Zaganidis N . Optical imaging of the spatial distribution of beta-particles emerging from surfaces. Proc Natl Acad Sci USA 1989; 86: 1741–1745.

  28. 28

    Moffat J, Grueneberg DA, Yang X, Kim SY, Kloepfer AM, Hinkle G et al. A lentiviral RNAi library for human and mouse genes applied to an arrayed viral high-content screen. Cell 2006; 124: 1283–1298.

  29. 29

    Amano T, Richelson E, Nirenberg M . Neurotransmitter synthesis by neuroblastoma clones (neuroblast differentiation-cell culture-choline acetyltransferase-acetylcholinesterase-tyrosine hydroxylase-axons-dendrites). Proc Natl Acad Sci USA 1972; 69: 258–263.

  30. 30

    Stoppini L, Buchs PA, Muller D . A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 199; 37: 173–182.

  31. 31

    Konradi C, Eaton M, MacDonald ML, Walsh J, Benes FM, Heckers S . Molecular evidence for mitochondrial dysfunction in bipolar disorder. Arch Gen Psychiatry 2004; 61: 300–308.

  32. 32

    Iwamoto K, Bundo M, Kato T . Altered expression of mitochondria-related genes in postmortem brains of patients with bipolar disorder or schizophrenia, as revealed by large-scale DNA microarray analysis. Hum Mol Genet 2005; 14: 241–253.

  33. 33

    Sultana R, Yu CE, Yu J, Munson J, Chen D, Hua W et al. Identification of a novel gene on chromosome 7q11.2 interrupted by a translocation breakpoint in a pair of autistic twins. Genomics 2002; 80: 129–134.

  34. 34

    Richler E, Reichert JG, Buxbaum JD, McInnes LA . Autism and ultraconserved non-coding sequence on chromosome 7q. Psychiatr Genet 2006; 16: 19–23.

  35. 35

    Philippi A, Roschmann E, Tores F, Lindenbaum P, Benajou A, Germain-Leclerc L et al. Haplotypes in the gene encoding protein kinase c-beta (PRKCB1) on chromosome 16 are associated with autism. Mol Psychiatry 2005; 10: 950–960.

  36. 36

    Hirono M, Sugiyama T, Kishimoto Y, Sakai I, Miyazawa T, Kishio M et al. Phospholipase Cbeta4 and protein kinase Calpha and/or protein kinase CbetaI are involved in the induction of long term depression in cerebellar Purkinje cells. J Biol Chem 2001; 276: 45236–45242.

  37. 37

    Kaufman M . Mouse and human embryonic development: a comparative overview. In: Strachan T, Lindsay S, Wilson DI (eds). Molecular Genetics of Early Human Development. T. BIOS Scientific Publishers Ltd.: Oxford, UK, 1997, pp 77–110.

  38. 38

    Rakic P, Nowakowski RS . The time of origin of neurons in the hippocampal region of the rhesus monkey. J Comp Neurol 1981; 196: 99–128.

  39. 39

    Nowakowski RS, Rakic P . The site of origin and route and rate of migration of neurons to the hippocampal region of the rhesus monkey. J Comp Neurol 1981; 196: 129–154.

  40. 40

    Khazipov R, Esclapez M, Caillard O, Bernard C, Khalilov I, Tyzio R et al. Early development of neuronal activity in the primate hippocampus in utero. J Neurosci 2001; 21: 9770–9781.

  41. 41

    Whitford KL, Marillat V, Stein E, Goodman CS, Tessier-Lavigne M, Chedotal A et al. Regulation of cortical dendrite development by Slit–Robo interactions. Neuron 2002; 33: 47–61.

  42. 42

    Li Z, Okamoto K, Hayashi Y, Sheng M . The importance of dendritic mitochondria in the morphogenesis and plasticity of spines and synapses. Cell 2004; 119: 873–887.

  43. 43

    Carper RA, Courchesne E . Localized enlargement of the frontal cortex in early autism. Biol Psychiatry 2005; 57: 126–133.

  44. 44

    Casanova MF, Buxhoeveden DP, Switala AE, Roy E . Minicolumnar pathology in autism. Neurology 2002; 58: 428–432.

  45. 45

    Luna B, Minshew NJ, Garver KE, Lazar NA, Thulborn KR, Eddy WF et al. Neocortical system abnormalities in autism: an fMRI study of spatial working memory. Neurology 2002; 59: 834–840.

  46. 46

    Silk TJ, Rinehart N, Bradshaw JL, Tonge B, Egan G, O'Boyle MW et al. Related articles, links visuospatial processing and the function of prefrontal-parietal networks in autism spectrum disorders: a functional MRI study. Am J Psychiatry 2006; 163: 1440–1443.

  47. 47

    McAlonan GM, Cheung V, Cheung C, Suckling J, Lam GY, Tai KS et al. Mapping the brain in autism. A voxel-based MRI study of volumetric differences and intercorrelations in autism. Brain 2005; 128: 268–276.

  48. 48

    Belmonte MK, Allen G, Beckel-Mitchener A, Boulanger LM, Carper RA, Webb SJ . Autism and abnormal development of brain connectivity. J Neurosci 2004; 24: 9228–9231.

  49. 49

    Ramos M, del Arco A, Pardo B, Martinez-Serrano A, Martinez-Morales JR, Kobayashi K et al. 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 2003; 143: 33–46.

  50. 50

    Fougerousse F, Bullen P, Herasse M, Lindsay S, Richard I, Wilson D et al. Human-mouse differences in the embryonic expression patterns of developmental control genes and disease genes. Hum Mol Genet 2000; 9: 165–173.

  51. 51

    Donoghue MJ, Rakic P . Molecular gradients and compartments in the embryonic primate cerebral cortex. Cereb Cortex 1999; 9: 586–600.

  52. 52

    Sur M, Rubenstein JL . Patterning and plasticity of the cerebral cortex. Science 2005; 310: 805–810.

  53. 53

    Rizzolatti G, Craighero L . The mirror-neuron system. Annu Rev Neurosci 2004; 27: 169–192.

  54. 54

    Dapretto M, Davies MS, Pfeifer JH, Scott AA, Sigman M, Bookheimer SY et al. Understanding emotions in others: mirror neuron dysfunction in children with autism spectrum disorders. Nat Neurosci 2006; 9: 28–30.

  55. 55

    Pierce K, Haist F, Sedaghat F, Courchesne E . The brain response to personally familiar faces in autism: findings of fusiform activity and beyond. Brain 2004; 127: 2703–2716.

  56. 56

    Dalton KM, Nacewicz BM, Johnstone T, Schaefer HS, Gernsbacher MA, Goldsmith HH et al. Gaze fixation and the neural circuitry of face processing in autism. Nat Neurosci 2005; 8: 519–526.

  57. 57

    Dent EW, Gertler FB . Cytoskeletal dynamics and transport in growth cone motility and axon guidance. Neuron 2003; 40: 209–227.

  58. 58

    Hirokawa N, Takemura R . Molecular motors and mechanisms of directional transport in neurons. Nat Rev Neurosci 2005; 6: 201–214.

  59. 59

    Lasorsa FM, Pinton P, Palmieri L, Fiermonte G, Rizzuto R, Palmieri F . Recombinant expression of the Ca(2+)-sensitive aspartate/glutamate carrier increases mitochondrial ATP production in agonist-stimulated Chinese hamster ovary cells. J Biol Chem 2003; 278: 38686–38692.

  60. 60

    Pardo B, Contreras L, Serrano A, Ramos M, Kobayashi K, Iijima M et al. Essential role of aralar in the transduction of small Ca2+ signals to neuronal mitochondria. J Biol Chem 2006; 281: 1039–1047.

  61. 61

    Geschwind DH, Levitt P . Autism spectrum disorders: developmental disconnection syndromes. Curr Opin Neurobiol 2007; 17: 103–111.

  62. 62

    Hemann MT, Fridman JS, Zilfou JT, Hernando E, Paddison PJ, Cordon-Cardo C et al. An epi-allelic series of p53 hypomorphs created by stable RNAi produces distinct tumor phenotypes in vivo. Nat Genet 2003; 33: 396–400.

  63. 63

    Mager J, Bartolomei MS . Strategies for dissecting epigenetic mechanisms in the mouse. Nat Genet 2005; 37: 1194–1200.

Download references


This work was supported by INSERM, Fondation France-Télécom and Fédération pour la Recherche sur le Cerveau. AMLB received a fellowship from Fondation des Treilles. We thank Dr Jane Pickett at the Autism Tissue Program for facilitating tissue collection. Tissue samples were provided by the National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders at the University of Miami, from Director Dr Carol Petito; the Harvard Brain Tissue Resource Center, from Director Dr Francine Benes. Bipolar and schizophrenia tissue samples were donated by the Stanley Foundation (Bethedsa, MD, USA). We thank Dr Manuel Rojo, INSERM U582, for providing the DsRedMito plasmid and advice, and Christophe Chamot and Tristan Piolot, Imagerie, Institut Jacques Monod for assistance with videomicroscopy analysis.

Author information

Correspondence to M Simonneau.

Additional information

Supplementary Information accompanies the paper on the Molecular Psychiatry website (

Supplementary information

Rights and permissions

Reprints and Permissions

About this article


  • SLC25A12
  • human embryo
  • fetal brain
  • in situ hybridization
  • expression pattern
  • cortical parcellation
  • gene overexpression
  • dendrite length

Further reading