Doi:10.1016/j.neulet.2007.03.004

Aberrant amino acid transport in fibroblasts from children with autism Elisabeth Fernell , Aristea Karagiannakis , Gunnar Edman , Lars Bjerkenstedt , Frits-Axel Wiesel , Nikolaos Venizelos a Department of Neuropaediatrics, Astrid Lindgren Children’s Hospital, Karolinska University Hospital, SE 171 76 Stockholm, Sweden b Department of Clinical Medicine, Biomedicine, ¨ c Department of Psychiatry, R& D Section, Danderyd’s Hospital, SE-182 87 Danderyd, Sweden d Department of Clinical Neuroscience, Karolinska University Hospital, SE-171 76 Stockholm, Sweden e Department of Neuroscience, Psychiatry, Uller˚aker, Uppsala University Hospital, SE-750 17 Uppsala, Sweden Received 17 October 2006; received in revised form 20 February 2007; accepted 1 March 2007 Abstract
Autism is a developmental, cognitive disorder clinically characterized by impaired social interaction, communication and restricted behaviours.
The present study was designed to explore whether an abnormality in transport of tyrosine and/or alanine is present in children with autism. Skinbiopsies were obtained from 11 children with autism (9 boys and 2 girls) fulfilling the DSM-IV diagnostic criteria for autistic disorder and 11healthy male control children. Transport of amino acids tyrosine and alanine across the cell membrane of cultured fibroblasts was studied bythe cluster tray method. The maximal transport capacity, Vmax and the affinity constant of the amino acid binding sites, Km, were determined.
Significantly increased Vmax for alanine (p = 0.014) and increased Km for tyrosine (p = 0.007) were found in children with autism. The increasedtransport capacity of alanine across the cell membrane and decreased affinity for transport sites of tyrosine indicates the involvement of two majoramino acid transport systems (L- and A-system) in children with autism. This may influence the transport of several other amino acids across theblood–brain-barrier. The significance of the findings has to be further explored.
2007 Elsevier Ireland Ltd. All rights reserved.
Keywords: Autism; Amino acid transport; Fibroblasts; Tyrosine; Alanine Autism is a developmental disorder characterized by severely tive ability and behaviour becomes evident. This specific form impaired social interaction, communication and restricted of autism is referred to as the early regressive type behaviours. Specific cognitive dysfunctions are linked to this In line with other developmental disorders several underly- triad of symptoms The vast majority of the patients have ing etiologies have been suggested. Prenatal causes of various an overall low intellectual function with IQs below the normal kinds dominate However, in most cases, available neuro- variation. Attention deficits are usually present v- logical assessments cannot reveal the underlying cause. Autism ity and stereotypes are common. Other behavioural disturbances, is a disorder with a strong genetic component Positive such as self-injuries, tantrums and outbursts may occur, entail- associations with autism were reported for more than 15 genes ing a severe clinical picture. Treatment with post-synaptic dopamine-blockers, i.e. classical neuroleptic drugs, has been Dopamine (DA) acts as a powerful regulator of different shown to reduce the symptom level in some patients The aspects of cognitive functions and is a key neurotransmitter in the onset of clinical symptoms varies and a considerable number brain. Several studies have shown its regulatory role for motor of children develop relatively normally until the age of 18–24 and cognitive/executive functions Already in 1987 Gill- months. At that time deterioration with respect to communica- berg and Svennerholm the major DA metabolitehomovanillic acid (HVA) in the cerebrospinal fluid (CSF) inunmedicated children with autism, in children with other devel-opmental disorders and in a healthy comparison group. Children ∗ Corresponding author. Tel.: +46 19 30 10 28; fax: +46 19 30 37 78.
with autism were found to have increased levels of HVA and E-mail address: (N. Venizelos).
URL: http://www.oru.se/templates/oruExtNormal 35718.aspx the authors suggested that a hyperfunction of the dopaminergic system in the brain stem and the mesolimbic system could be 0304-3940/$ – see front matter 2007 Elsevier Ireland Ltd. All rights reserved.
doi: E. Fernell et al. / Neuroscience Letters 418 (2007) 82–86 part of the pathogenesis in autism. However, in another study nia, however, her cognitive problems later became evident, first of HVA levels in the CSF in children with autism, no signifi- as a mental retardation, and eventually at the age of 6 years cant differences were found between children with autism and autism was diagnosed. (Etiologically a specific syndrome was controls link between autism, disintegrative disorder, i.e.
autism with late onset (Heller’s syndrome; schizophre- Eight of the 11 children had undergone a magnetic resonance nia was proposed by Honjo also discussed by Schieveld tomography or a computerized tomography with normal results.
In both conditions cognitive dysfunction is an important The four children with epilepsy were treated with antiepilep- characteristic. Interestingly, in patients with schizophrenia, we tics; and one of these children had also received a low dose of have previously found a connection between tyrosine transport risperidone. Two other children were treated with a low dose of and cognitive functioning Tyrosine is the precursor of risperidone and serotonin reuptake inhibitor, respectively. Rou- dopamine synthesis and a limitation of its access to the brain tine plasma amino grams had been performed on all the children will influence dopamine content and functioning Amino acids play an important role in brain development and cog- The Ethics Committee at Karolinska Hospital approved the nitive functioning, drastically demonstrated in children with study. The parents were introduced to the study procedure and phenylketonuria (PKU). In this context it should be of interest gave informed consent to let their child participate in the study.
to investigate tyrosine transport in children with autism.
Skin biopsies were taken in connection with surgery for Tyrosine transport is mediated from plasma to brain through hypospadia in 11 boys between the ages of 1 and 13 years old membranes mainly via the L-system, named from its func- (mean 4 years). All were healthy and there was no suspicion of tion to transport large, neutral, branched and aromatic amino developmental disorders in any of these children.
acids (LNAA), i.e. phenylalanine, tyrosine, tryptophan, methio- All growth media, antibiotics, fetal calf serum and Amnio- nine, leucine, isoleucine, valine and histidine. The L-system maxTM were purchased from Gibco Invitrogen cell culture, is sodium-independent and consists of four isoforms; LAT1, Sweden. Phosphate buffered saline (PBS) from Statens LAT2, LAT3 and LAT4 L-system is widely expressed in Veterin¨armedicinska Anstalt (SVA), Uppsala, Sweden. [U- the body and occurs in the blood–brain-barrier, kidney and small 14C]-labelled L-tyrosine (specific activity 434 mCi/mmol) and intestine n fibroblasts only the total tyrosine transport has [U-14C]-labelled L-alanine (specific activity 128 mCi/mmol) been studied, without discriminating between the isoforms of were purchased from Moravek Biochemical Inc., California, the L-system To a smaller extent, tyrosine is also trans- U.S.A. Unlabelled L-tyrosine and L-alanine were purchased ported from plasma to brain by the sodium-dependent A-system (“A” is an abbreviation for alanine) Three variants of sys- Fibroblasts were cultured from skin biopsies using Eagle’s tem A, designed ATA1, ATA2 and ATA3, have been identified Minimum Essential Medium (EMEM) supplemented with fetal ransport of short-chain neutral amino acids such as alanine calf serum (10%, vol/vol), penicillin V (125 U/ml), streptomycin is mainly mediated by the transport A-system ut also by (125 ␮g/ml), l-glutamine (2 mmol/l) and Amnio-MaxTM. Stock L-system TA2 is widely expressed in mammalian tissues cultures of individual fibroblast stains were cultured in 75 cm2 occurs like the L-system in the blood–brain-barrier Costar plastic tissue culture flasks in a humidified atmosphere of 5% CO2 at 37 ◦C. Fibroblast lines were used experimentally The aim of this study was to test our working hypothesis that children with autism have changes in amino acid transport Amino acids transport assays were performed according mechanisms, based on the profound importance of amino acids to the fibroblast model rapid measurement of in brain development and for cognitive functioning.
amino acid flux in adherent intact fibroblast cells. Approxi- A 2 mm skin punch biopsy was taken after anaesthetizing the mately 5 × 104 cells were seeded in a 24 multiwell tray (2 cm mid-forearm under aseptic conditions. The tissue was placed diameter; Costar Europe Ltd., Costar, NY) and grown to con- immediately in tubes containing complete culture medium.
fluence in 1.5 ml of MEM for 5 days. The cells were washed The study group consisted of 11 children with autism, twice with 1ml phosphate buffered saline (PBS) and directly according to the DSM-IV criteria 9 boys and 2 girls, pre-incubated with 1 ml PBS containing 1% glucose for 1 h at aged 5–11 years (mean 8 years). All were all patients 37 ◦C, 5% CO2 to deplete endogenous amino acids pools. After at the Neuropaediatric Outpatient Clinic at the Department removal of the pre-incubation medium, the initial rate of a par- of Neuropaediatrics, Karolinska University Hospital, Stock- ticular amino acid uptake was measured during incubation for holm and had been assessed and diagnosed by experienced 60 s at 37 ◦C in 0.2 ml PBS, containing a constant amount of neuropaediatric-neuropsychiatric teams. Ten had mental retar- 14C-labelled amino acid (0.2 ␮Curie) and 12 different concen- dation (U.K. learning disabilities) and one had an IQ in the lower trations (varying between 0.01 and 1.5 mmol/l for tyrosine and normal area. Four of the 11 children also had epilepsy, and in at between 0.02 and 6 mmol/l for alanine) of unlabelled amino least 2 children without clinical epilepsy, the EEG had revealed acids. Amino acid transport assay was terminated by rapidly epileptiform discharges. All boys and one of the girls had an washing the cells twice with 2 ml of ice-cold PBS. After wash- uneventful neonatal period and first year of life and clinically ing, the wells were drained and 230 ␮l of trichloroacetic acid they were classified as being of the early regressive type. No (10%, w/v) was added to each well for 20 min at room tempera- specific disease had earlier been demonstrated. The second girl ture. Two hundred microlitres of the radioactive soluble amino had a different clinical picture, starting with muscular hypoto- acid extract was removed for liquid scintillation counting. The E. Fernell et al. / Neuroscience Letters 418 (2007) 82–86 trays were drained and the precipitated proteins were dissolved the other hand the Km [t (13.9) = 1.05, p = 0.311] for alanine did in 1 mol/l NaOH and assayed by the modified Lowry method not differ significantly between the two groups.
using bovine serum albumin as standard.
Vmax and Km for tyrosine and alanine for the four chil- The amino acid kinetic parameters were calculated dren with autism and epilepsy were compared with those of from the Lineweaver–Burke plot equation [1/V0 = (Km/Vmax the autistic children without epilepsy. There were no significant [S]) + (1/Vmax)], by using computerized software as described differences between the groups (data not shown).
previously by Flyckt et al. in 2001. V0 is the initial transport Fibroblast cell cultures offer a model for experimental stud- velocity and [S] is the transport substrate concentration, Vmax ies on the transport of amino acids across cell membranes. The is the maximal uptake rate for the carrier-mediated process kinetic parameters Vmax and Km were studied, reflecting the (nmol/min × mg protein) and Km is the affinity constant (the maximal transport capacity of tyrosine/alanine and the affin- concentration at half-saturation; ␮mol/l).
ity of tyrosine/alanine for the binding sites of the transporters, Each experiment was performed at 12 different concentra- tions in duplicate in the same incubation for both tyrosine and The main finding of this study was that fibroblasts from chil- dren with autism had an elevated Vmax for alanine and a higher The tyrosine transport was studied on two occasions at differ- Km for tyrosine. The higher Km corresponds with a decreased ent time points, to diminish cell artefacts that can be caused by affinity between the transport protein and the substrate tyro- the condition of the cells, and could influence the low Vmax and sine. The lower affinity for tyrosine indicates that a higher Km tyrosine values. Alanine was analyzed only once since the concentration of extracellular tyrosine is required to reach the experiment precision is higher than tyrosine (Vmax twice and Km maximal transport capacity, resulting in a decreased competition five times larger than tyrosine). Differences in Vmax and Km for between tyrosine and other amino acids. Alanine on the other tyrosine between autistic and normal children were analyzed by hand showed a higher transport capacity in this group of chil- parametric analysis of variance (two-way ANOVA) for repeated dren with autism, but the Km did not differ significantly between measurements (group × measurement).
the two groups. This indicates that fibroblasts from the children The group differences in alanine transport were analyzed with with autism in this study group had an elevated expression of alanine transporting protein. Hence, our main result shows that The coefficients of variation for the tyrosine determinations there is a change in the kinetics of tyrosine and alanine transport measured on two occasions in 11 patients and 11 controls cell in this study group of children with autism.
lines were 15%, respectively, 16% for Vmax and 20% for Km for Recently, we have investigated the two major transporters of tyrosine and alanine (L- and A-systems) in fibroblast from All variables were summarized by standard descriptive statis- patients with schizophrenia and controls study pro- tics (mean and standard deviations). The distributions were vided evidence that 90% of the total uptake of tyrosine and 65% scrutinized for deviations from normal. Vmax and Km did not of alanine were transported through the sodium-independent deviated from normal distribution. In all analyses, a significance L-system, whereas 10% of tyrosine and 75% of alanine was transported by the A-system. Consequently, alanine inhibited Maximal transport capacity (Vmax) and mean affinity of bind- tyrosine uptake by 65% in fibroblasts in both patients with ing site (Km) of tyrosine and alanine in children with autism and schizophrenia and controls. Therefore, there seems to be a competition between alanine and tyrosine for transport across The Vmax [F (1, 20) = 2.60, p = 0.123] for tyrosine did not differ significantly between the two groups, but the children with Brain tyrosine is dependent on the influx of this amino acid autism had a significantly higher Km [F (1, 20) = 8.90, p = 0.007] across the blood–brain-barrier. Tyrosine is critical for main- taining adequate levels of dopamine and therefore of vital The Vmax for alanine was significantly higher [t (20) = 2.69, importance for adequate cognitive functions. A major reason for p = 0.014] in the children with autism compared to controls, on studying tyrosine transport is the association between PKU and Table 1Kinetic parameters (Vmax, Km) of tyrosine and alanine transport in cultured fibroblasts from children with autism and controls The results are presented as mean (M) and standard deviation (S.D.).
Vmax, indicates maximal transport capacity (nmol/min × mg protein).
Km, affinity of binding sites for a specific amino acid (␮mol/l).
a M ± S.D. for tyrosine are based on the average of the two repeated uptake measurements.
E. Fernell et al. / Neuroscience Letters 418 (2007) 82–86 autism. PKU is a well-known developmental disorder, includ- The number of children with epilepsy in our study is in agree- ing mental retardation, autism and epilepsy. In PKU there is ment with the reported frequency of epilepsy in autism, being a restricted influx of tyrosine into the brain. The biochemical about one-third According to Tuchman and Rapin defect is a failure to convert the amino acid phenylalanine to epilepsy and autism are both heterogeneous clinical disorders tyrosine due to inadequate functioning of the enzyme, phenylala- associated with an array of etiologies and pathologies and the nine hydroxylase. This disorder has been practically eliminated present evidence suggests that there are common pathophysi- due to perinatal screening and affected children are given a ological mechanisms that account for both the autism and the phenylalanine-restricted diet. Despite treatment lower cognitive epilepsy. We have no idea whether the proportion of children functioning, especially with respect to executive functions, has with both autism and epilepsy in our study group influences our been demonstrated in patients with PKU These executive results since the relation of clinical and subclinical epilepsy to dysfunctions are especially linked to the prefrontal cortex, which autistic behaviour and early regression is unsettled.
is rich in dopaminergic projection and thus dependent on tyro- In conclusion, our findings may indicate an elevated access of sine supply to the brain. The role of tyrosine supplementation alanine and less availability of tyrosine in the brain, resulting in in PKU is questionable and Kalsner et al., suggested that the a lower dopaminergic activity. On the other hand, in vitro data is appropriate strategy might be to normalize all brain amino acid hard to translate to in vivo situations with reciprocal interactions involving many complex systems. Our findings demonstrate that In this study, an aberrant transport of tyrosine and ala- there may be disturbances in transport mechanisms for amino nine across the fibroblast cell membrane was demonstrated.
acids, at least at the membrane level. A further exploration of this We speculate that there might be a corresponding disturbance research field concerning disturbances of amino acid transport in of amino acid transport across the blood–brain-barrier. The children with autism includes molecular investigations, looking increased transport velocity of alanine across the cell mem- for polymorphisms in gene loci, and further transport studies, brane may indicate involvement of the LAT2-transport system which will hopefully provide more valid answers.
in which there is a competition with tyrosine. LAT2 is oneof the four isoforms that constitute the L-system. LAT2 has Acknowledgements
a broader substrate specificity, including tyrosine and alanineand is widely distributed in the body Such competition, The authors are grateful to Associate Professor Agneta Nor- due to increased transport of alanine across the blood–brain- denskj¨old, Department of Pediatric Surgery and Astrid Lindgren barrier, and accordingly a low tyrosine influx into the brain, Children’s Hospital for valuable help with material from the may have different functional consequences. Neurons, and control children. The study was supported by grants from particularly dopaminergic neurons, may be more susceptible S¨allskapet Barnav˚ard, Astrid Lindgren Children’s Hospital, to disturbances in tyrosine transport resulting in a decreased Stiftelsen Frimurare Barnhuset in Stockholm (EF), Research dopamine synthesis in these cells. This could consequently lead Council (Nr 8318), Magnus Bergvall Foundation, Ingrid Thuring to compensatory changes in dopaminergic transmission and and S¨oderstr¨om-K¨oningska Foundation.
Both in schizophrenia and autism the synaptic and connec- References
tivity hypotheses have been discussed In a recent report,Bjerkenstedt et al. that the dysfunctional tyro- [1] A. Albers, A. Broer, C.A. Wagner, I. Setiawan, P.A. Lang, E.U. Kranz, F.
sine transport from plasma to brain, revealed in patients with Lang, S. Broer, Na+ transport by the neural glutamine transporter ATA1, schizophrenia, would affect the dopaminergic pathways and [2] G. Allen, E. Courchesne, Attention function and dysfunction in autism, cause abnormal connectivity, secondary to the aberrant tyrosine transport. Accordingly, the significance of our present findings, [3] American Psychiatric Association, Diagnostic and Statistical Manual of indicating a decreased affinity for transport sites of tyrosine in Mental Disorders (DSM-IV), American Psychiatric Press, Washington, children with autism, might be that such ensuing dopamine dys- regulation will result in abnormal neural connectivity also in this [4] M.K. Belmonte, G. Allen, A. Beckel-Mitchener, L.M. Boulanger, R.A.
Carper, S.J. Webb, Autism and abnormal development of brain connectivity, There are certain limitations with this study: the patient group [5] I.N. Bespalova, J.D. Buxbaum, Disease susceptibility genes for autism, is small, only 11 children with autism, resulting in a low power of the statistic analyses. On the other hand, the precision of in vitro [6] L. Bjerkenstedt, L. Farde, L. Terenius, G. Edman, N. Venizelos, F.A.
studies is high and fewer patients are needed. Six children with Wiesel, Support for limited brain availability of tyrosine in patients withschizophrenia, Int. J. Neuropsychopharmacol. 9 (2006) 247–255.
autism were medicated with drugs acting on the CNS. However, [7] S. Bodoy, L. Martin, A. Zorzano, M. Palacin, R. Estevez, J. Bertran, Iden- the present experiments were performed on cells that had grown tification of LAT4, a novel acid transporter with system L activity, J. Biol.
for several generations in vitro. It seems unlikely that the results could be influenced by factors such as the condition of the patient [8] S. Danielsson, I.C. Gillberg, E. Billstedt, C. Gillberg, I. Olsson, Epilepsy in at the time of biopsy or prior medication by drugs acting on the young adults with autism: a prospective population-based follow-up studyof 120 individuals diagnosed in childhood, Epilepsia 46 (2005) 918–923.
CNS. Rather, the findings observed in the cells from patients [9] J.C. Dobson, E. Kushida, M. Williamson, E.G. Friedman, Intellectual per- with autism may reflect a hereditary trait that can be transmitted formance of 36 phenylketonuria patients and their nonaffected siblings, E. Fernell et al. / Neuroscience Letters 418 (2007) 82–86 [10] L. Flyckt, N. Venizelos, G. Edman, L. Bjerkenstedt, L. Hagenfeldt, F.A.
[22] Y. Kanai, H. Endou, Functional properties of multispecific amino acid trans- Wiesel, Aberrant tyrosine transport across the cell membrane in patients porters and their implications to transporter-mediated toxicity, J. Toxicol.
with schizophrenia, Arch. Gen. Psychiatry 58 (2001) 953–958.
[11] W.G. Frankle, J. Lerma, M. Laruelle, The synaptic hypothesis of [23] B. Mackenzie, J.D. Erickson, Sodium-coupled neutral amino acid (System schizophrenia, Neuron 39 (2003) 205–216.
N/A) transporters of the SLC38 gene family, Pflugers Arch. 447 (2004) [12] G.C. Gazzola, V. Dall’Asta, R. Franchi-Gazzola, M.F. White, The cluster- tray method for rapid measurement of solute fluxes in adherent cultured [24] M. Narayan, S. Srinath, G.M. Anderson, D.B. Meundi, Cerebrospinal fluid cells, Anal. Biochem. 115 (1981) 368–374.
levels of homovanillic acid and 5-hydroxyindoleacetic acid in autism, Biol.
[13] R.F. Gazzola, R. Sala, O. Bussolati, R. Visigalli, V. Dall’Asta, V. Ganapathy, G.C. Gazzola, The adaptive regulation of amino acid transport system A is [25] A. Nieoullon, Dopamine and the regulation of cognition and attention, associated to changes in ATA2 expression, FEBS Lett. 490 (2001) 11–14.
[14] C. Gillberg, M. Coleman, The Biology of the Autistic Syndromes, third [26] E. Olsson, F.A. Wiesel, L. Bjerkenstedt, N. Venizelos, Tyrosine transport ed., MacKeith Press, Cambridge, UK, 2000.
in fibroblasts from healthy volunteers and patients with schizophrenia, [15] C. Gillberg, L. Svennerholm, CSF monoamines in autistic syndromes and Neurosci. Lett. 393 (2006) 211–215.
other pervasive developmental disorders of early childhood, Br. J. Psychi- [27] I. Rapin, Autistic regression and disintegrative disorder: how important the role of epilepsy? Semin. Pediatr. Neurol. 2 (1995) 278–285.
[16] L. Hagenfeldt, N. Venizelos, L. Bjerkenstedt, F.A. Wiesel, Decreased tyro- [28] J.N. Schieveld, Case reports with a child psychiatric exploration of catato- sine transport in schizophrenic patients, Life Sci. 41 (1987) 2749–2757.
nia, autism, and delirium, Int. Rev. Neurobiol. 72 (2006) 195–206.
[17] T. Hatanaka, W. Huang, H. Wang, M. Sugawara, P.D. Prasad, F.H. Leibach, [29] S. Shea, A. Turgay, A. Carroll, M. Schulz, H. Orlik, I. Smith, F. Dunbar, V. Ganapathy, Primary structure, functional characteristics and tissue Risperidone in the treatment of disruptive behavioral symptoms in children expression pattern of human ATA2, a subtype of amino acid transport with autistic and other pervasive developmental disorders, Pediatrics 114 system A, Biochim. Biophys. Acta 1467 (2000) 1–6.
[18] E.L. Hill, U. Frith, Understanding autism: insights from mind and brain, [30] M. Sugawara, T. Nakanishi, Y.J. Fei, W. Huang, M.E. Ganapathy, F.H.
Philos. Trans. R Soc. Lond. B Biol. Sci. 358 (2003) 281–289.
Leibach, V. Ganapathy, Cloning of an amino acid transporter with func- [19] S. Honjo, A case in which diagnosis between autism. Heller’s syndrome and tional characteristics and tissue expression pattern identical to that of childhood schizophrenia is difficult, Jpn. J. Psychiatry Neurol. 48 (1994) system A, J. Biol. Chem. 275 (2000) 16473–16477.
[31] R. Tuchman, I. Rapin, Epilepsy in autism, Lancet Neurol. 1 (2002) [20] R. Hyde, P.M. Taylor, H.S. Hundal, Amino acid transporters: roles in amino acid sensing and signalling in animal cells, Biochem. J. 373 (2003) [32] T.H. Wassink, L.M. Brzustowicz, C.W. Bartlett, P. Szatmari, The search for autism disease genes, Ment. Retard. Dev. Disabil. Res. Rev. 10 (2004) [21] L.R. Kalsner, F.J. Rohr, K.A. Strauss, M.S. Korson, H.L. Levy, Tyrosine supplementation in phenylketonuria: diurnal blood tyrosine levels and pre- [33] F.A. Wiesel, G. Edman, L. Flyckt, ˚ sumptive brain influx of tyrosine and other large neutral amino acids, J.
L. Bjerkenstedt, Kinetics of tyrosine transport and cognitive functioning in schizophrenia, Schizophr. Res. 74 (2005) 81–89.

Source: http://www.autism.se/RFA/uploads/stockholm/aminosyretransportautism.pdf