Molecular Psychiatry (2011) 16, 478–490 & 2011 Macmillan Publishers Limited All rights reserved 1359-4184/11 www.nature.com/mp FEATURE REVIEW The role of neurotrophic factors in T Nickl-Jockschat and TM Michel Department of Psychiatry and Psychotherapy, RWTH Aachen University, Aachen, Germany

Autism spectrum disorders (ASDs) are pervasive developmental disorders that frequently involve a triad of deficits in social skills, communication and language. For the underlying neurobiology of these symptoms, disturbances in neuronal development and synaptic plasticity have been discussed. The physiological development, regulation and survival of specific neuronal populations shaping neuronal plasticity require the so-called ‘neurotrophic factors’ (NTFs). These regulate cellular proliferation, migration, differentiation and integrity, which are also affected in ASD. Therefore, NTFs have gained increasing attention in ASD research. This review provides an overview and explores the key role of NTFs in the aetiology of ASD. We have also included evidence derived from neurochemical investigations, association studies and animal models. By focussing on the role of NTFs in ASD, we intend to further elucidate the puzzling aetiology of these conditions. Molecular Psychiatry (2011) 16, 478–490; doi:10.1038/mp.2010.103; published online 12 October 2010 Keywords: autism; autism spectrum disorders; BDNF; ; neurotrophic factors; NT-3

Introduction Although the aetiology of ASD is still not fully understood, twin and adoption studies suggest a Approximately one child per 145 babies born in the strong genetic role in the manifestation of the United States will be diagnosed with some form of disorder.8–11 Monozygotic twins show concordance autism spectrum disorder (ASD) throughout their rates of approximately 70–90%, whereas they are only 1 lifespan according to new estimates. The differences 0–10% in dizygotic twins.12,13 However, in addition noted in individuals with autism relate to develop- to genetic risk factors, there is evidence for an associa- ment deficiencies of language, social interaction tion with a variety of other conditions, including skills, repetitive and stereotyped movements and prenatal exposure to noxae such as alcohol, behaviours, hyperactivity, sensory disturbances, thalidomide or sodium-valproate as well as obstetric 2,3 restricted interests and more rarely self-injury. complications, pre- and post-natal brain damage, Additional challenging aspects of ASD are comorbid chromosomal aberrations such as fragile X-syndrome, 4 disorders, such as epilepsy, gastrointestinal pro- tuberous sclerosis and 22q11.2 deletion syndrome 5 6 blems and sleep disorder. The clinical presentation among others.14–16 of autism is complex and variable, and therefore the One of the most consistent findings in autism is an term ASD is often used in this context. According to increased brain size during development.17–20 Chil- the ICD-10 (International Statistical Classification of dren with autism seem to undergo an abnormally Diseases and Related Health Problems, Tenth Revi- accelerated brain growth during development.18,19,21–23 sion), ASD is described as a pervasive developmental It has been suggested that the formation of neuronal disorder. Differentiation from other developmental connections or the elimination of inappropriate disorders such as , which show over- connections does not proceed in the typical manner. lapping symptoms and a similar neurobiology to Striking neuropathological findings, such as fewer ASD, is not always easy, especially in the very early Purkinje cells, smaller neuronal size and decreased stages. Rett syndrome is a comparatively rare disorder dendritic branches in subjects diagnosed with ASD, with clinical proximity to ASD, but in contrast to have been reported for various brain regions, such as ASD, a mono-genetic background can be defined in a the cerebellum,24–28 the hippocampus and the amyg- 7 majority of cases. Rett syndrome, which affects dale.25,26 Magnetic resonance imaging studies have girls only, is characterized by a profound learning also shown grey and white matter changes such as disability following early normal development, with a reduced neuronal and axodendendritic pruning in consistent cluster of clinical features. autism.29–34 Lately, inflammatory processes have also been discussed in the aetiopathology of autism.28 A Correspondence: Dr TM Michel, Department of Psychiatry and vast body of research has been undertaken in the last Psychotherapy, RWTH Aachen University, Pauwelsstrasse 30, decade to elucidate the underlying pathophysiology Aachen, D-52074, Germany. E-mail: [email protected] of autism. Received 13 January 2010; revised 31 August 2010; accepted 6 The development and maintenance of the central September 2010; published online 12 October 2010 nervous system is influenced by several different Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 479 mechanisms, one of the most prominent being low affinity,59–62 but can be regulated by receptor neurotrophic factors (NTFs).35–41 dimerization, structural modifications or association NTFs do not only have a key role in processes such with the p75 receptor.63,64 as brain development and maintenance of neurons Substrate binding leads to Trk dimerization and, and their dentrites throughout life, but they also therefore, to trans-autophosphorylation at two tyrosine critically influence the formation and elimination of residues (Tyr490 and Tyr785) in the cytoplasmatic neuronal connections. They have been discussed in domain of the receptor, activating the intracellular many studies as promising candidates responsible for signalling cascade.62 A key molecule interacting with part of the alterations seen in autism.35–43 these phosphotyrosine motifs is the Src homologous NTFs comprise a range of different super- and collagen-like (Shc) adaptor protein. Shc links families. At least six factors belong to the neurotro- Trk signalling to two major signalling pathways: the phin family: nerve growth factor (NGF), brain-derived Ras/Raf/ERK and the phosphatidylinositol-3 kinase growth factor (BDNF), neurotrophin (NT)-3, NT-4, (PI3K) pathway. NT-5 and NT-6.44–47 The neurotrophins are a family of After activation, Shc interacts with Grb2. Regarding closely related that were first identified as the Ras/Raf/ERK pathway, Grb2 then binds to SOS survival factors for sympathetic and sensory neu- (son of sevenless). The Shc/Grb2/SOS complex rons.48 The core functions of neurotrophins during attaches itself to the membrane through the inter- neurodevelopment include regulation of cell prolif- action of Shc with the phosphorylated receptor, eration, migration and survival. They also modulate and mediates signalling to the Ras/mitogen-activated axonal and dendritic outgrowth, synapse formation protein kinase pathway.65,66 The neurotrophins’ and other neuroplastic processes.49 ability to activate Raf depends on Rap1, a small Further neurotrophic superfamilies include neuro- endosomal G-protein.67 The active Ras/Raf/ERK path- kines such as ciliary NTF (CNTF) and leukaemia way influences, for example, of the inhibitory factor (LIF), insulin-like growth factors cyclic AMP-response element (CREB) transcription (for example, IGF-1 and IGF-2), as well as the vast factor. Overall, this pathway has effects on the cell transforming growth factor-b (TGF-b) superfamily (for cycle, neurite outgrowth and synaptic plasticity.68 example, TGF-b1, -b2 and -b3), and their distant Alternatively, Shc-bound Grb2 can activate PI3K relative glial cell line-derived neurotrophic factor through the Gab1 (Grb2-associated binder-1). PI3K (GDNF). Additionally, there are a variety of other can propagate cellular survival through Akt (protein proteins that can at least partly exert an influence on kinase B) activities,65 and also mitogenic signalling, neurotrophic functions, although they are not con- cell survival, cytoskeletal remodelling and vesicular sidered as NTFs themselves. trafficking.69 Furthermore, individual NTFs and their receptors Binding of phospholipase-Cg to activate Trk con- are quite distinct and are subject to considerable stitutes an additional, Shc-independent pathway, changes in the course of neural development.50,51 resulting in the release of inositol phosphates and The proteins of the neurotrophin family have a activation of protein kinase C.65,70 molecular weight of 13 kDa (NGF) to 27 kDa (BDNF, Through a different set of adaptor proteins, p75 NT-3) and high isoelectric points (9–10.5). BDNF is signalling results in increases in Jun N-terminal one of the most abundant NTFs in mammalian brains. kinase, nuclear factor-kB and ceramide.71 P75 is The neurotrophin genes, like other peptide growth capable of mediating apoptosis in a pathophysio- factors, encode a precursor peptide.52,53 Thus, the logical context, for example, after seizure or inflam- protein forms of neurotrophins exist in the human mation72,73 and—in oligodendrocytes—after spinal brain, both in the mature and in its precursor form injury,74 and also physiologically during neurodeve- (for example, proBDNF). The precursor form is lopment.75 secreted in both a basal and an activity-dependent Surprisingly, pro-neurotrophins display higher fashion and is processed extracellularly to its mature affinities for the p7556 and are more effective form by proteolytic cleavage.54–56 inductors of p75-dependent apoptosis.56,74 Thus, Only one frequent, nonconservative polymorphism proteolytic cleavage seems to be an important med- in the human BDNF gene (dbSNP number rs6265) has iator of neurotrophin biological activity, partly decid- been identified. This single-nucleotide polymorph- ing on cellular fate. ism (SNP) at nucleotide 196 (G/A) produces an amino Neurotrophins have been shown to produce in- acid substitution (valine to methionine) at codon 66 creases in excitatory postsynaptic currents in hippo- (val66met). campal neurons.76,77 Both BDNF and NT-3 are able to Mature neurotrophins form stable, non-covalent induce long-term potentiation (LTP) in hippocampal dimers and bind with lower affinity to the p75 neurons, a pivotal factor in the consolidation of receptor and with higher affinity to receptors of the memory and learning. Vice versa, BDNF or NT-4 tyrosine kinase family (TrkA, B, C): BDNF and NT-4/5 knockout mice show impairments of LTP in the bind TrkB and NT-3 binds TrkC and to a lesser extent hippocampus.78,79 TrkA.46,47,57,58 These interactions have generally been The molecular mechanisms of LTP induction by considered to be of high affinity. However, the neurotrophins are still only poorly understood. binding of NGF to TrkA, and of BDNF to TrkB, is of Protein kinase inhibitors lead to impairments in

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 480 LTP, indicating that intracellular protein phosphor- sensitive bead-based assay (Luminex, Austin, TX, ylation as well as phosphatidylinositol lipids and USA). The concentration of BDNF in maternal mid- inositol-1,4,5 receptors are important for LTP.80 pregnancy and neonatal specimens was similar across all three study groups, questioning a connection NTFs in autism identified by serum level and between ASD pathogenesis and BDNF levels. genetic studies Recently, further analysis using Luminex techno- logy indicated developmental trajectories for the BDNF serum levels in autism concentrations of NTFs in ASD.40 Concentrations of It has been suggested that early BDNF hyperactivity a set of neurotrophins, and cytokines may be involved in autism early in life. Nelson et al.36 were analysed from archived neonatal blood of were the first to report higher BDNF levels in neonatal children with later-diagnosed autism, Down syn- blood of children with ASD. The study compared drome, very preterm birth and term control infants a set of neurotrophin and serum levels as healthy controls. Results showed inhibitory in a group of children with ASD, another group substances in eluates from archived bloodspots, of children with mental retardation (with no comor- especially marked for measurement of BDNF. Concen- bidity of ASD) and a healthy control group, using trations in control subjects differed by age: BDNF rose recycling immunoaffinity chromatography. Along markedly with age, whereas NT-3 and NT-4/5 con- with neonatal concentrations of vasoactive intestinal centrations were lower in adults than in newborn peptide, calcitonin gene-related peptide and NT-4/5, infants. Interleukin-8 concentrations were higher in BDNF concentrations were higher in children with newborn infants, preterm and term, than in adults. ASD and in those with mental retardation without In samples from ASD children, NT-3 levels were ASD than in control children. In 99% of children with significantly lower than controls and an increase in autism and 97% with mental retardation, levels of at vasoactive intestinal peptide approached statistical least one of these molecules exceeded those of all significance. Concentrations of NT-4/5 and calcito- control children.36 nin-gene-related peptide were correlated in infants Concordant with these results, Connolly et al.38 with ASD but not in Down syndrome or controls.40 showed mean BDNF levels to be elevated in children The results of another recent study suggested with autism (mean age 5.9±3.9) compared with disturbances of BDNF level trajectories in ASD controls. Notably, both IgG and IgM autoantibodies patients.82 In healthy controls, serum BDNF concen- to BDNF were also elevated in children with ASD tration was found to be increased over the first several compared with controls. The latter findings were not years and decreased slightly after reaching adulthood. idiosyncratic for ASD children: children with child- No gender differences were reported. In the autism hood disintegrative disorder and those with epilepsy cases, mean levels were significantly lower in also showed elevated IgG and IGM BDNF autoanti- children 0–9 years old compared with teenagers or bodies. adults, or with age-matched healthy controls, indicat- An overall increase of BDNF levels in autistic ing a delayed BDNF increase during development. individuals was shown in a study by Miyazaki et al.37 The study also investigated circadian and seasonal However, when interpreting the results of this study, fluctuations of BDNF serum concentrations. Circa- one has to take into account that the age range was dian changes, but not seasonal changes, were found much larger in the autistic individuals (between 3 in serum levels of BDNF. In a separate study with and 27 years) than in the control group (22–24 years). adult rats, a circadian change in serum BDNF was A later study by Hashimoto et al.39 showed that in found to be similar to that in the cortex, indicating a young autistic male adults (age 21.1±2.1 years), the possible association with cortical functions. BDNF concentration was significantly lower than in In contrast to the variety of studies on serum levels neurotypical controls (22.2 years). The serum levels of of NTFs, only few studies have examined cerebro- BDNF in patients with autism were significantly spinal fluid (CSF) levels. Riikonen and Vanhala35 lower than those of normal controls. Nevertheless, focussed on NGF levels in the CSF of patients with no correlations between BDNF levels and clinical ASD and Rett syndrome, and used enzyme-linked variables in autistics were detected. immunosorbent assay for the determination of NGF. As BDNF was found to be frequently altered in ASD In ASD patients, mainly normal CSF NGF levels were patients, one study focussed on its usability as an found, whereas there were low to negligible values in early biological marker for autism.81 Besides a group Rett syndrome. with ASD patients, two comparison groups were The cause for altered neurotrophin serum and CSF enroled: children with mental retardation or devel- levels in ASD patients remains yet to be clarified. opmental delay and unaffected controls randomly Given the putatively strong role of genetic factors in sampled from the California birth certificate files. the aetiopathogenesis of ASD, it is necessary to take They were frequency matched to autism cases on sex, this into account when discussing the potential role birth year and birth month. BDNF concentrations of neurotrophins in this context. A study by Lang were measured in archived mid-pregnancy and et al.83 indeed showed that the well-characterized neonatal blood specimens drawn during routine Val66Met SNP was shown to be associated with incre- prenatal and newborn screening using a highly ased BDNF serum concentrations. This polymorphism,

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 481 leading to a Met substitution for Val at codon 66 changes of GDNF levels in Rett syndrome patients in the prodomain for BDNF, is common in human were observed.87 populations, with an allele frequency of 20–30% in Besides members of the neurotrophin protein Caucasian populations.84 Therefore, further detailed family as typical exponents of NTFs, serum levels of investigations are needed to shed light on the role TGF-b1 have also been examined in ASD patients. of a potential interplay between the Met allele and TGF-b1 is a multifunctional ubiquitious polypeptide elevated serum BDFN levels in ASD. cytokine that was initially named for its ability to However, studies on neurotrophin gene variants induce transformation to a mesenchymal phenotype and autism are scarce. in epithelial cultured cells.88 In the human brain, Nishimura et al.41 reported in their study that three, TGF-b signalling is known to be involved in a variety four- and five-SNP haplotypes were significantly of biological processes, including specification of associated with ASD. Val66Met is associated with developmental fate in embryogenesis for both neu- alterations in brain anatomy and memory, but its rons and glial cells and suppression of immune relevance to clinical disorders is unclear. Never- reactions.89,90 theless, it did not show any significant correlation To the best of our knowledge, only two studies have with the disorder in their study. Moreover, the examined TGF-b1 serum levels in autism. Okada expression of BDNF mRNA in peripheral blood et al.91 were the first to investigate TGF-b1 serum lymphocytes might differ from that in neurons. BDNF levels in ASD patients. Serum levels of TGF-b1in19 mRNA levels in the drug-naive ASD patients were male adult patients with autism and 21 age-matched increased compared with the healthy control group. male healthy subjects were measured using enzyme- Some clues for an involvement of the BDNF gene in linked immunosorbent assay. The serum levels of the aetiopathogenesis of ASD come from a study on TGF-b1 in the patients with autism were significantly the so-called ‘WAGR’ syndrome. This is a rare lower than those of normal controls. However, there genomic disorder caused by deletion of the 11p14– were no marked or significant correlations between p12 region. It comprises the following serum TGF-b1 levels and other clinical variables, symptom clusters: Wilms tumour, aniridia, genito- including ADI-R (Autism Diagnostic Interview- urinary malformations and mental retardation. A Revised) scores, Y-BOCS (Yale–Brown Obsessive- majority of WAGR patients have mental retardation Compulsive Scale), aggression, Theory of Mind, and and psychiatric comorbidities such as attention- Intellectual Quotient in patients. deficit/hyperactivity disorder and obsessive compul- A second study was able to find similar results in sive disorder, and > 20% of the patients also have children with ASD.92 Plasma levels of active TGF-b1 features of autism.85 Although the Wilms tumour/ were evaluated in 75 children with ASD compared genitourinary anomalies and aniridia are caused by with 68 controls. Children with ASD had significantly deletion of Wilms tumour 1 (WT1) and paired box 6 lower plasma TGF-b1 levels when compared with (PAX6) genes, respectively, the genomic cause of typically developing controls and also with children mental retardation and autism in WAGR syndrome with developmental disabilities other than ASD, remains unknown. Using oligonucleotide arrays, a after adjusting for age and gender. In addition, there study aimed to characterize the 11p14–p12 deletions were significant correlations between psychological and to identify all the genes involved in each measures and TGF-b1 levels, such that lower deletion.86 Besides the already described WT1 and TGF-b1 levels were associated with lower adaptive PAX6 deletions, a majority of patients enroled behaviours. showed a deletion of the BDNF gene. It is tempting On a molecular genetic level, so far only one study to speculate that the fact that the BDNF gene showed a has investigated correlations of TGF-b1 SNPs and deletion in many, but not all, patients might at least ASD.93 No significant SNP or haplotypic associations partly explain the fact that autism is a frequent, but were observed for TGF-b1. not mandatory, part of the WAGR syndrome. It is important to mention that both studies Insulin-like growth factors included only comparatively small sample sizes IGF-1 and IGF-2 are members of the insulin gene (n = 104 in the patient group of Nishimura et al.41 family, which stimulate cellular proliferation and and just 34 patients in Xu et al.86). So far, association differentiation during embryonic and postnatal devel- studies with larger sample sizes or even with opment.94–96 Both IGF-1 and IGF-2 are widely ex- independent replications showing a correlation pressed in the central nervous system and exert their between BDNF gene variants and autism do not exist. mitogenic influence through the membrane-bound In this context, the study results cited above should tyrosine kinase IGF-1 receptor.97 be read with care and can serve at the most as first IGF-1 and IGF-2 are bound to six different binding hints of an involvement of BDNF gene variants in the proteins. These proteins are thought to have a role in pathogenesis of autism. the modulation of IGF action and targeting of IGFs to certain tissues. Especially, IGF-1 seems to be impor- Cytokines and TGF-b tant for proper neurodevelopment: mice lacking IGF-1 For GDNF, so far, only data for CSF levels in Rett or IGF-1 receptor have decreased brain volumes,98 syndrome patients are available. No significant whereas IGF-1 overexpression increases brain growth

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 482 with increased width of myelin sheath.99 As overall male subjects were measured. The serum levels of brain volume in ASD patients is increased,100 an EGF in subjects with high-functioning autism were involvement of IGF-1 in the pathogenesis of autism is significantly lower than those of normal control plausible. subjects. There were no correlations between serum Riikonen et al.99 determined IGF-1 and IGF-2 levels EGF levels and clinical variables in the subjects with from CSF using radioimmunoassay in 25 children autism.104 with autism and with a median age of 5 years 5 Platelet-derived growth factor (PDGF) is mitogenic months, and in 16 age-matched comparison children for many cell types.105–107 The PDGF family is without disability. IGF-1 and IGF-2 concentrations composed of four different polypeptide chains were further correlated with age of patients and head (A–D), encoded by four different genes.108–110 PDGF size. CSF IGF-1 concentration was significantly lower chains assemble to a disulphide bond, homo- or in patients with autism than in the control group. In hetero-dimer, thus forming the biologically active children with autism < 5 years of age, the CSF PDGF protein.111 In neuronal stem cells, PDGF has concentrations were significantly lower than their been shown to direct neuronal differentiation, and it age-matched comparisons. The head circumferences also supports neuronal cell survival and prevents cell correlated with CSF IGF-1 in children with autism, death.112 PDGF mediates its effects by binding to the but no such correlation was found in the comparison high-affinity tyrosine kinase transmembrane recep- group. There was no difference between the two tors (PDGFa receptor and PDGFb receptor), leading in groups in CSF IGF-2 concentrations. It is important to consequence to the activation of a variety of signal mention that no patients with autism had macro- generators such as PI3K and ras.113 cephaly. Up to now, only one study has investigated PDGF Another study determined serum levels of IGF-1 subtype serum levels, comparing autistic male chil- and IGF-2, IGFBP-3 (IGF binding protein-3), GHBP dren aged 6–19 with an age-matched control group.114 (growth hormone binding protein) and of dehydro- Serum levels of the PDGF-BB subtype were found to epiandrosterone and its sulphate in 71 boys between be significantly higher in the patient group. Remark- 4 and 8 years with ASD or autism and 59 age-matched ably, in the patient group, PDGF serum levels were control boys.101 In contrast to the study of Riikonen significantly and positively correlated with the et al.,99 subjects in the autism/ASD group had domain C scores of the ADI-R, which represent significantly greater head circumferences than con- stereotyped patterns of behaviour. The same study trols, and also significantly higher weights than found no association of altered vascular endothelial controls, whereas height did not differ significantly growth factor serum levels and autism. between groups. Levels of IGF-1, IGF-2, IGFBP-3 and GHBP in the group with autism/ASD were all significantly higher than in controls. Non-NTF susceptibility genes for autism and their Urinary IGF-1 levels correlate with serum levels.102 influence on NTF levels and function A study used this noninvasive method, investigating For many years, research has focussed on associating the daily urinary excretion of IGF-1, epidermal personality traits, behaviour and brain function with growth factor (EGF) and IGFBP-3 in autistic children specific genetic variants.115–120 However, compelling (n = 34, age 2–5 years) and age-matched control molecular genetic evidence for an involvement of children (n = 29). The mean urinary IGF-1 level was neurotrophins in ASD comes from functional ana- lower in the autism group than the control group. lyses of established ASD susceptibility gene variants. Although these genetic variants do not directly affect Other NTFs genes of the NTF family, they often lead to alterations EGF is a common mitogenic factor with neurotrophic of NTF levels or to impairments of NTF functions.121 effects on cultured cortical neurons, stimulating neurite outgrowth in dopaminergic cells.103 Caps2 One study was able to show an association between Ca2 þ -dependent activator protein 2 (CAPS2, alter- an EGF haplotype and autism.93 The authors focussed natively CADPS2) is a promising ASD candidate gene on the genes of three different cytokines: TGF-b1, EGF in the chromosomal 7q31 region, a replicated linkage and hepatocyte growth factor. In a trio association region for ASD.122 The CAPS2 protein is enriched on study using DNA samples from families recruited into vesicular structures in the parallel fibre terminals the Autism Genetic Resource Exchange, 252 trios of granule cells. CAPS2-asscociated vesicles contain with a male offspring scored for autism were selected BDNF and NT-3 in large quantity. Remarkably, for the study. Transmission disequilibrium tests induced overexpression of CAPS2 increases depolar- revealed significant haplotypic association of EGF ization-induced BDNF and NT-3 release.123 Thus, with autism. No association was found for TGF-b1 impaired CAPS2 function might lead to altered BDNF and hepatocyte growth factor. and NT-3 levels in autism. Another study found that serum levels of EGF were Most data suggest that disruption of the CAPS2 altered in adult subjects with high-functioning aut- gene in ASD patients is due to copy number ism. Serum levels of EGF in 17 male subjects with variations.122 In contrast to the more extensively high-functioning autism and 18 age-matched healthy studied SNPs, which affect only one base in the

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 483 sequence of a given gene, copy number variations are cascade, leading in consequence to a disturbance of larger mutations sized from one kilobase to several neuronal activity-dependent maturation. Hence, this megabases. hypothesis yet remains to be proven and does not A study by Sadakata et al.124 reported an aberrant, offer further insight into the exact cause of this alternatively spliced CAPS2 mRNA that lacks exon disturbance. 3 (Caps2deltaexon3) in some patients (n = 16) with ASD. In a subsequent animal study using a mouse JARID1C model, Sadakata et al.125 tried to further characterize A single case report describes a novel missense CAPS2 splice variants and identified six alternative mutation in the Jumonji AT-rich interactive domain splicing variants, CAPS2a–f. CAPS2c showed slightly 1C (JARID1C) gene.138 Recent novel molecular decreased BDNF-releasing activity, which might be evidence suggests that JARID1C is a histone demethy- likely due to the C-terminal truncation of the PH lase specific for di- and trimethylated H3K4 (histone 3 domain in CAPS2. Thus, the previously mentioned lysine 4) and functions as a regulator of neuronal human CAPS2 splicing variant found no homologue gene transcription.139,140 The JARID1C-regulated in mice. genes include BDNF. It has to be noted that up to As CAPS2 is considered a candidate gene for ASD now, only 14 cases of JARID1C mutations have been and has been characterized as a neuronal activity- reported in ASD cases. dependent modulator of neurotrophin release, CAPS2 All previous findings taken together, there is gene mutations might provide a contribution to mounting evidence for altered neurotrophin serum altered neurotrophin activity in ASD patients. How- levels in ASD. Specific variants of ASD susceptibility ever, other ASD susceptibility gene variants also seem genes could provide at least some explanations for the to impact neurotrophin activity. alterations found in the neurotrophin serum levels. Most animal models investigated suggest an interplay MECP2 between BDNF and the studied behaviour. As the Rett syndrome (see above) has the advantage of Nonetheless, several important questions still need a well-defined genetic cause in a majority of cases and to be addressed. Neurotrophin levels are subject to its clinical presentation is close to autism, the study developmental trajectories over the course of life- of this disorder might also provide valuable insight time.40,51,141 Apparently, more research on the course into ASD pathophysiology. In most patients, Rett of neurotrophin trajectories in ASD should be carried syndrome is caused by mutations in the X-linked out. Moreover, there seems to be a circadian rhythmi- methyl-CpG-binding protein 2 (MECP2).7,126 MeCP2 city of at least BDNF secretion.82 It still needs to be is a chromatin-associated protein127 that can both clarified whether ASD patients show the same activate and repress gene transcription,128,129 and it is circadian secretion patterns as healthy controls. capable of both suppressing and inducing BDNF Therefore, the subjects have to be matched not only transcription.120–132 Neuronal activity triggers MeCP2 for age, but also for the time of day at which the blood phosphorylation, therefore leading to BDNF gene sample was obtained. processing, whereas prevention of MeCP2 phosphor- Besides these central nervous action mechanisms of ylation interferes with the ability of MeCP2 to BDNF, altered peripheral modes of action also deserve regulate dendritic patterning, spine morphogenesis to be mentioned,142 although they are not a central and the activity-dependent induction of BDNF tran- point of view of this review. scription.131 These results suggest a central role for MeCP2 in neural activity-dependent brain maturation Neuropsychology and imaging studies that at least partly exerts its influence on synaptic plasticity through its effects on BDNF. Altered levels of NTFs and the functional and Molecular MeCP2/BDNF-related pathologies in neuroanatomical consequences: lessons from autism are more complex than in Rett syndrome. neuropsychology and brain imaging Whereas MeCP2 mutations seem to be rare in ASD,133 Changes in both anatomical and/or functional cere- MeCP2 mRNA levels were found to be reduced in the bral connectivity have been reported frequently in absolute majority of ASD cortex samples.134 Although ASD patients.121,143,144 As neurotrophins are known to a comprehensive explanation for reduced MeCP2 modulate axonal and dendritic outgrowth, synapse transcription rates still remains to be found, disrupted formation and other neuroplastic processes,49 neuro- co-regulation of MeCP2 and EGR2 (early growth trophin dysfunction might well exert an impact on response gene 2) leads to an impaired positive both anatomical and functional aspects of brain feedback loop on transcription rates of both genes.135 connectivity. Thus, these changes in brain structure EGR2 is one of the so-called immediate activity genes, and function might reflect important pathophysio- a class of genes rapidly and transiently induced by logical consequences at least partially caused by neuronal activation and other cellular or extra- altered neurotrophin levels and functionality. Find- cellular stimuli without the necessity for de novo ings in these fields will be discussed here as far as protein synthesis.136 In turn, BDNF signalling was they relate to the topic of this review. shown to enhance EGR2 transcription.137 These data Until now, only very few studies on white matter suggest a disordered MeCP2/EGR2/BDNF signalling abnormalities have been published, with a focus on

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 484 regions possibly involved in ASD pathophysiology.145 Not all human behaviours can be translated to animal In one of the first studies using diffusion tensor models. However, by analytically observing and imaging to determine anatomical connectivity in ASD dissecting behaviour, it is possible to derive mouse patients, Barnea-Goraly et al.146 found reduced anato- models that reflect aberrant social interaction. mical connectivity in brain regions of social cognition Consequently, study groups aimed at establishing such as the fusiform gyrus and the superior temporal behavioural tests thought to assess behavioural sulcus as well as in regions associated with theory of abnormalities in mouse models paralleling core mind tasks (ventromedial prefrontal cortex, anterior autistic features such as aberrant social responses, cingulate, temporo-parietal junction and amygdala). repetitive behaviour or restricted interests.163 Testing Subsequent studies found reduced anatomical con- screens for social responses usually include assays for nectivity of the callosum,147 the frontal lobe148 and the social approach and preference for social novelty, in superior temporal gyrus.149 Notably, a study enroling which mice have a choice between social and non- a large samples of ASD patients between the ages of social stimuli of different types.164 Behavioural assays 10 and 35 years found an impairment of anatomical exist for both ‘lower-order’ motoric stereotypy and connectivity for the whole age range in white matter ‘higher-order’ responses reflecting, for example, structures involved in long-range intra- and inter- restricted interests and strong resistance to environ- hemispheric communication (anterior corona radiate, mental change. Whereas the ‘lower-order’ compo- right retrolenticular part of internal capsule and nents can be tested, for example, by home cage corpus callosum).150 observation,165 ‘higher-order’ motor functions can be Although studies on anatomical connectivity focus evaluated, for example, in T-maze or water maze on changes in white matter structures, functional tasks.164 As symptoms in autism spectrum disorders magnetic resonance imaging studies can be used to emerge early in childhood, valid mouse models of investigate co-activation patterns during a given ASD should display behavioural phenotypes recapi- paradigm. Executive function has been shown to be tulating the time course of disease onset.166–168 One of decreased in ASD,151 which appears to be present in the most common assays to assess behavioural adulthood as well as throughout development.152–155 abnormalities in early stages of development is the Several studies examined functional connectivity measurement of ultrasonic vocalization in mouse in ASD patients during higher-order information pups separated from their mothers as a test for altered processing, describing a pattern of reduced fronto- emotional behaviour.169–171 parietal connectivity during response inhibition,156 Social behaviour in the mouse model seems to be sentence comprehension156,157 and working memory influenced by neurotrophins. Mice raised in a socially tasks.158,159 Although initially hypothesized,157 these enriched environment display higher propensity to changes do not seem to reflect a general functional interact socially. Interestingly, it has been shown that underconnectivity in the brains of ASD patients, but these animals show higher NGF levels. These are an inability to effectively integrate cerebral networks further influenced by social status and higher BDNF to support complex information processing in a levels in the brain.163 Chen et al.172 investigated the cognitive task.143 influence of the Val66Met polymorphism on anxiety- Strikingly, variations in the BDNF genes have been related behaviour in a mouse model. Consistent with shown to influence age-related decline in executive the findings mentioned above, mice with Met/Met function, although results that show that decline allele (that consecutively leads to decreased neuronal happens more rapidly in Met allele carriers are secretion of BDNF) display more anxious behaviour admittedly not unanimous.160,161 Animal experiments in stressful settings. However, it has to be acknowl- hint at an impaired synaptic plasticity in the edged that other behavioural paradigms—for exam- prefrontal cortex as a potential neurobiological ple, sociability or repetitive behaviour—were not mechanism mediating BDNF effects on executive tested in this study. A more detailed characterization functions.162 of the behavioural phenotypes relating to ASD- Altogether, results derived from brain imaging and specific traits seems to be a promising goal for future neuropsychology hint at an involvement of BDNF research in this field. dysfunction in the impairment of executive functions in the context of a reduced anatomical and functional Summary and future work connectivity. Owing to their pivotal role not only during neuro- Preclinical models development, but also in the regulation of mature synapto- and neuroplasticity, NTFs have been the Besides the research strategies mentioned above, the focus of various studies on neuropsychiatric disor- use of animal models is of great importance for ders and ASD.51,141,173,174 Most investigations in the determining the genotype–phenotype relationship of context of NTFs in ASD have concentrated on the neurotrophins in the context of ASD. Advanced neurotrophin family. Various studies were able to transgenic and recombinant technologies as well as show alterations in neurotrophin serum levels. Non- the sequencing of the mouse genome have made the neurotrophic ASD susceptibility genes were shown mouse to be the model of choice for many geneticists. not only to influence BDNF expression, but also to be

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 485 regulated by BDNF signalling in a positive feedback However, only two studies prompt speculations loop. It could be speculated that genetic vulnerability concerning altered TGF-b1 levels in ASD patients.91,92 for ASD might exert its influence at least in part A third was not able to show an association of TGF-b1 through altered neurotrophin expression patterns. gene variants with ASD.93 Additional studies with Reduced neurotrophin serum levels might in turn be larger sample sizes are needed before drawing final responsible for typical features of cellular morpho- conclusions. logy and impaired neuronal functioning in ASD. Similarly, more in-depth studies on IGFs have to be Some methodological limitations still have to be carried out before interpreting the few studies on this taken into consideration before interpreting of the molecule in the context of ASD. However, investiga- results of the studies. Few studies have actually tions on IGF-1, IGF-2 and the IGF-binding proteins reported to control for circadian variation or neuro- have reported first promising results. Two studies developmental trajectories of neurotrophin secre- successfully showed an association between elevated tion.82 In fact, to our knowledge, studies dealing IGF-1 levels and head circumference,99,101 whereas a with the question of whether ASD patients show any third even detected even lower IGF-1 levels in the disturbances in the circadian rhythmicity of neuro- autism group.102 However, the discrepancies might be trophin secretion do not exist. Hence, with the due to methodological limitations: Riikonen et al.99 currently available data, an exact characterization of measured CSF levels and Mills et al.101 serum levels, neurotrophin serum level alterations cannot be given. whereas Fall et al.102 analysed urine samples. Head As shown above, not only mature neurotrophin circumference in the ASD groups might also influ- proteins but also their unconverted pro-forms are able ence the results. As IGF-1 levels have been shown to to interact with Trk and p75 receptors. Due to the be associated with total brain volume in animal higher affinity of pro-neurotrophins to the p75 experiments,98,99 elevated IGF-1 levels might not receptor, which is capable of mediating apoptosis,56,74 necessarily be associated with autism but with head the ratio between mature and immature neuro- circumference/macrocephaly. trophins might influence cellular fate. Consequently, A majority of NTFs have not been investigated for pathologies of the neurotrophin system might not their role in autism. Hence, the specificity of the only involve the overall level of the converted reported alterations of NTF levels is hard to judge. proteins, but also the level of the unconverted pro- Also, comparisons of reported NTF level changes in forms, whereas dysfunctions of their converting ASD with NTF levels in other psychiatric disorders enzymes might be key mediators of this process. are still scarce. Consequently, altered neurotrophin serum levels are Finally, NTFs have a different impact on different hard to interpret without knowledge of their pro-form neuronal sub-populations. Studies on NTF serum levels. Even normal neurotrophin serum levels might or CSF levels represent valuable exploratory ap- not rule out pathologies of the neurotrophin system. proaches. Nonetheless, they can only provide sparse Accordingly, studies on the regional cerebral levels of insight into the complex pathomechanisms on a mature and immature neurotrophins might help to cellular and histological level. elucidate the pathomechanisms here. The combination of both neuroimaging and mole- Moreover, solidification of evidence for significant cular biology has already shown some promising interactions between ASD susceptibility genes and, results in other areas of psychiatric research.118,120,121 for example, BDNF is still required. Histopathological Due to its noninvasive character, a comparatively studies might yield additional evidence. large sample size can be included. Moreover, long- Comparatively few studies have examined the role itudinal studies are possible as well. A correlation of non-neurotrophic NTFs and ASD. Because of its between genetics, neurochemistry and in vivo studies double role as potent immune suppressor and its on brain structure and function may yield additional influence on specification of developmental fate in insight on the pathogenesis of autism. embryogenesis,89,90 the TGF-b1 molecule is another The emerging understanding of the role of NTFs in promising candidate in this context. TGF-b1 might autism will most likely affect diagnostic and ther- exert its effect by both functions. Immune suppres- apeutic processes. Molecular screening might become sion by the TGF-b is of high importance for the a new valid diagnostic tool for early identification of protection of the developing tissues; knock-out mice ASD patients.179 lacking one of the three TGF-b isoforms usually only Both conventional and experimental therapies survive several days or weeks.175–177 Reduced TGF-b1 might benefit from an increased knowledge about levels in autism might lead to defective immune NTF-related pathophysiology. A recent study has suppression and in consequence to autoimmune- shown the BDNF Val66Met polymorphism to be induced neuronal loss. The influence of TGF-b associated with prolactine elevation caused by risper- signalling pathways on a cellular phenotype also idone treatment in autistic patients.180 Thus, knowl- seem to fit into another model of ASD pathophysiol- edge about the BDNF genotype might help to limit ogy: the Ras/Raf/ERK pathway—besides the growth- prolactine elevation-associated side effects. arrest inducing Smad pathway the second-best de- The development of NTF-specific therapeutic scribed cascade—has been described as a potent agents might provide causal therapeutic approaches. inductor of change to cellular morphology.89,178 Based on the hypothesis of a hyperactivation of TrkB

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 486 signalling in autism, the use of a TrkB partial agonist patients with autistic disorder: an MRI study. Neurology 2001; 57: has been proposed as a novel therapy.181 245–254. Although up to now, to our best knowledge, clinical 18 Courchesne E, Carper R, Akshoomoff N. Evidence of brain overgrowth in the first year of life in autism. JAMA 2003; 290: studies on the potential of therapeutic agents inter- 337–344. acting with NTF signalling have not been conducted, 19 Hazlett HC, Poe M, Gerig G, Smith RG, Provenzale J, Ross A et al. NTF-related signalling pathways appear as a promis- Magnetic resonance imaging and head circumference study of ing candidate for therapeutic intervention in autism. brain size in autism: birth through age 2 years. Arch Gen In summary, there is good evidence for the hypo- Psychiatry 2005; 62: 1366–1376. 20 Sparks BF, Friedman SD, Shaw DW, Aylward EH, Echelard D, thesis that NTFs have an important role in the Artru AA et al. Brain structural abnormalities in young children pathophysiology of autism. Further investigations in with autism spectrum disorder. Neurology 2002; 59: 184–192. not only the molecular mechanisms, but also correla- 21 Dementieva YA, Vance DD, Donnelly SL, Elston LA, Wolpert tions of molecular parameters with histopathology CM, Ravan SA et al. Accelerated head growth in early develop- ment of individuals with autism. Pediatr Neurol 2005; 32: and neuroimaging might be worthwhile pursuing for 102–108. a more comprehensive understanding of the patho- 22 Dissanayake C, Bui QM, Huggins R, Loesch DZ. Growth in stature mechanisms underlying the complex ASDs. and head circumference in high-functioning autism and Asperger disorder during the first 3 years of life. Dev Psychopathol 2006; 18: 381–393. 23 Dawson G, Munson J, Webb SJ, Nalty T, Abbott R, Toth K. Rate of Conflict of interest head growth decelerates and symptoms worsen in the second year of life in autism. Biol Psychiatry 2007; 61: 458–464. The authors declare no conflict of interest. 24 Bailey A, Luthert P, Dean A, Harding B, Janota I, Montgomery M et al. A clinicopathological study of autism. Brain 1998; 121: 889–905. References 25 Kemper TL, Bauman M. Neuropathology of infantile autism. J Neuropathol Exp Neurol 1998; 57: 645–652. 1 Di-Cicco-Bloom E, Lord C, Zwaigenbaum L, Courchesne E, Dager 26 Kemper TL, Bauman M. Neuropathology of infantile autism. Mol SR, Schmitz C et al. The developmental neurobiology of autism Psychiatry 2002; 7: S12–S13. spectrum disorder. J Neurosci 2006; 26: 6897–6906. 27 Lee M, Martin-Ruiz C, Graham A, Court J, Jaros E, Perry R et al. 2 Volkmar FR, Pauls D. Autism. Lancet 2003; 362: 1133–1141. Nicotinic receptor abnormalities in the cerebellar cortex in 3 Minishew NJ, Williams DL. The new neurobiology of autism: autism. Brain 2002; 125: 1483–1495. cortex, connectivity, and neuronal organization. Arch Neurol 28 Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo 2007; 64: 945–950. CA. Neuroglial activation and neuroinflammation in the brain of 4 Volkmar FR, Nelson DS. Seizure disorders in autism. J Am Acad patients with autism. Ann Neurol 2005; 57: 67–81. Child Adolesc Psychiatry 1990; 29: 127–129. 29 Ben Bashat D, Kronfeld-Duenias V, Zachor DA, Ekstein PM, 5 Quigley EM, Hurley D. Autism and the gastrointestinal tract. Am J Hendler T, Tarrasch R et al. Accelerated maturation of white Gastroenterol 2000; 95: 2154–2156. matter in young children with autism: a high b value DWI study. 6 Richdale AL, Prior MR. The sleep/wake rhythm in children with Neuroimage 2007; 37: 40–47. autism. Eur Child Adolesc Psychiatry 1995; 4: 175–186. 30 Bloss CS, Courchesne E. MRI neuroanatomy in young girls: a 7 Shahbazian MD, Zoghbi HY. Rett syndrome and MeCP2: linking preliminary study. J Am Acad Child Adolesc Psychiatry 2007; 46: epigenetics and neuronal function. Am J Hum Genet 2002; 71: 515–523. 1259–1272. 31 Carper RA, Moses P, Tigue ZD, Courchesne E. Cerebral lobes in 8 Coon H. Current perspectives on the genetic analysis of autism. autism: early hyperplasia and abnormal age effects. Neuroimage Am J Med Genet Semin Med Genet 2006; 142C: 24–32. 2002; 16: 1038–1051. 9 Landa RJ, Holman KC, Garrett-Mayer E. Social and communica- 32 Hazlett HC, Poe M, Gerig G, Smith RG, Piven J. Cortical white and tion development in toddlers with early and later diagnosis grey brain tissue volume in adolescents and adults with autism. of autism spectrum disorders. Arch Gen Psychiatry 2007; 64: Biol Psychiatry 2006; 59: 1–6. 853–864. 33 Kates WR, Burnette CP, Eliez S, Strunge LA, Kaplan D, Landa R et 10 Landa RJ. Diagnosis of autism spectrum disorders in the first 3 al. Neuroanatomic variation in monozygotic twin pairs discor- years of life. Nat Clin Pract Neurol 2008; 4: 138–147. dant for the narrow phenotype in autism. Am J Psychiatry 2004; 11 Lauritsen MB, Pedersen CB, Mortensen PB. Effects of familial risk 161: 539–546. factors and place of birth on the risk of autism: a nationwide 34 Palmen SJ, Hulshoff Pol HE, Kemner C, Schnack HG, Durston S, register-based study. J Child Psychol Psychiatr 2005; 46: Lahuis BE et al. Increased grey-matter volume in medication- 963–971. naive high-functioning children with autism spectrum disorder. 12 Bailey A, LeCouteur A, Gottesman I, Bolton P, Simonoff E, Yuzda Psychol Med 2005; 35: 561–570. E et al. Autism as a strongly genetic disorder: evidence from a 35 Riikonen R, Vanhala R. Levels of cerebrospinal fluid nerve- British twin study. Psychol Med 1995; 25: 63–77. growth factor differ in infantile autism Rett syndrome. Dev Med 13 Steffenburg S, Gillberg C, Hellgren L, Andersson L, Gillberg IC, Child Neurol 1999; 41: 148–152. Jakobsson G et al. A twin study of autism in Denmark, Finland, 36 Nelson KB, Grether JK, Croen LA, Dambrosia JM, Dickens BF, Iceland, Norway and Sweden. J Child Psychol Psychiatr 1989; 30: Jelliffe LL et al. Neuropeptides and neurotrophins in neonatal 405–416. blood of children with autism or mental retardation. Ann Neurol 14 Tierney E, Nwokoro NA, Porter FD, Fruend LS, Ghuman JK, 2001; 49: 597–606. Kelley RI. Behavior phenotype in the RSH/Smith-Lemli-Opitz 37 Miyazaki K, Narita N, Sakuta R, Miyahara T, Naruse H, Okado N syndrome. Am J Med Genet 2001; 98: 191–200. et al. Serum neurotrophin concentrations in autism and mental 15 Bandim JM, Ventura LO, Miller MT, Almeida HC, Costa AE. retardation: a pilot study. Brain Dev 2004; 26: 292–295. Autism and Mobius sequence: an exploratory study of children in 38 Connolly AM, Chez M, Streif EM, Keeling RM, Golumbek PT, northeastern Brazil. Arq Neuropsiquiatr 2003; 61: 181–185. Kwon JM et al. Brain-derived neurotrophic factor and autoanti- 16 Kolevzon A, Gross R, Reichenberg A. Prenatal and perinatal risk bodies to neural antigens in sera of children with autistic factors for autism: a review and integration of findings. Arch spectrum disorders, Landau-Kleffner syndrome, and epilepsy. Pediatr Adolesc Med 2007; 161: 326–333. Biol Psychiatry 2006; 59: 354–363. 17 Courchesne E, Karns CM, Davis HR, Ziccardi R, Carper RA, 39 Hashimoto K, Iwata Y, Nakamura K, Tsujii M, Tsuchiya KJ, Tigue ZD et al. Unusual brain growth patterns in early life in Sekine Y et al. Reduced serum levels of brain-derived

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 487 neurotrophic factor in adult male patients with autism. Prog sympathetic and sensory ganglia. Mol Cell Neurosci 1995; 6: Neuropsychopharmacol Biol Psychiatry 2006; 30: 1529–1531. 544–566. 40 Nelson PG, Kuddo T, Song EY, Dambrosia JM, Kohler S, 62 Chao MV. The p75 neurotrophin receptor. J Neurobiol 1994; 25: Satyanarayana G et al. Selected neurotrophins, neuropeptides, 1373–1385. and cytokines: developmental trajectory and concentrations in 63 Arevalo JC, Conde B, Hempstead BL, Chao MV, Martin-Zanca D, neonatal blood of children with autism or Down syndrome. Int J Perez P. TrkA immunoglobulin-like ligand binding domains Dev Neurosci 2006; 24: 73–80. inhibit spontaneous activation of the receptor. Mol Cell Biol 41 Nishimura K, Nakamura K, Anitha A, Yamada K, Tsujii M, 2000; 20: 5908–5916. Iwayama Y et al. Genetic analyses of the brain-derived neuro- 64 Esposito D, Patel P, Stephens RM, Perez P, Chao MV, Kaplan DR et trophic factor (BDNF) gene in autism. Biochem Biophys Res al. The cytoplasmic and transmembrane domains of the p75 and Commun 2007; 356: 200–206. Trk A receptors regulate high affinity binding to nerve growth 42 Maisonpierre PC, Belluscio L, Squinto S, Ip NY, Furth ME, factor. J Biol Chem 2001; 276: 32687–32695. Lindsay RM et al. Neurotrophin-3: a neurotrophic factor related 65 Kaplan DR, Miller FD. Neurotrophin signal transduction in the to NGF and BDNF. Science 1990; 247: 1446–1451. nervous system. Curr Opin Neurobiol 2000; 10: 381–391. 43 Airaksinen MS, Titievsky A, Saarma M. GDNF family neuro- 66 Ravichandran KS. Signaling via Shc family adapter proteins. trophic factor signaling: four masters, one servant? Mol Cell Oncogene 2001; 20: 6322–6330. Neurosci 1999; 13: 313–325. 67 York RD, Yao H, Dillon T, Ellig CL, Eckert SP, McCleskey EW 44 Barde YA. The nerve growth factor family. Prog Growth Factor et al. Rap1 mediates sustained MAP kinase activation induced by Res 1990; 2: 237–248. nerve growth factor. Nature 1998; 392: 622–626. 45 Go¨tz R, Ko¨ster R, Winkler C, Raulf F, Lottspeich F, Schartl M 68 Lonze BE, Ginty DD. Function and regulation of CREB family et al. Neurotrophin-6 is a new member of the nerve growth factor transcription factors in the nervous system. Neuron 2002; 35: family. Nature 1994; 372: 266–269. 605–623. 46 Lindsay RM, Wiegand SJ, Altar CA, DiStefano PS. Neurotrophic 69 Wymann MP, Pirola L. Structure and function of phosphoinosi- factors: from molecule to man. Trends Neurosci 1994; 17: tide 3-kinases. Biochim Biophys Acta 1998; 1436: 127–150. 182–190. 70 Chao MV. Neurotrophins and their receptors: a convergence 47 Barbacid M. Neurotrophic factors and their receptors. Curr Opin point for many signalling pathways. Nat Rev Neurosci 2003; 4: Cell Biol 1995; 7: 148–155. 299–309. 48 Skaper SD. The biology of neurotrophins, signalling pathways, 71 Roux PP, Barker PA. Neurotrophin signaling through the p75 and functional peptide mimetics of neurotrophins and their neurotrophin receptor. Prog Neurobiol 2002; 67: 203–233. receptors. CNS Neurol Disord Drug Targets 2008; 7: 46–62. 72 Dowling P, Ming X, Raval S, Husar W, Casaccia-Bonnefil P, Chao 49 Avital A, Goshen I, Kamsler A, Segal M, Iverfeldt K, Richter-Levin M et al. Up-regulated p75NTR neurotrophin receptor on glial G et al. Impaired interleukin-1 signaling is associated with cells in MS plaques. Neurology 1999; 53: 1676–1682. deficits in hippocampal memory processes and neural plasticity. 73 Roux PP, Colicos MA, Barker PA, Kennedy TE. p75 neurotrophin Hippocampus 2003; 13: 826–834. receptor expression is induced in apoptotic neurons after seizure. 50 Altar CA, Criden MR, Lindsay RM, DiStefano PS. Characteriza- J Neurosci 1999; 19: 6887–6896. tion and topography of high-affinity 125I-neurotrophin-3 binding 74 Beattie MS, Harrington AW, Lee R, Kim JY, Boyce SL, Longo FM to mammalian brain. J Neurosci 1993; 13: 733–743. et al. ProNGF induces p75-mediated death of oligodendrocytes 51 Durany N, Michel T, Kurt J, Cruz-Sa´nchez FF, Cervo´s-Navarro J, following spinal cord injury. Neuron 2002; 36: 375–386. Riederer P. Brain-derived neurotrophic factor and neurotrophin-3 75 Majdan M, Miller FD. Neuronal life and death decisions levels in Alzheimer’s disease brains. Int J Dev Neurosci 2000; 18: functional antagonism between the Trk and p75 neurotrophin 807–813. receptors. Int J Dev Neurosci 1999; 17: 153–161. 52 Seidah NG, Benjannet S, Pareek S, Chre´tien M, Murphy RA. 76 Kang H, Schuman EM. A requirement for local protein synthesis Cellular processing of the neurotrophin precursors of NT3 and in neurotrophin-induced hippocampal synaptic plasticity. BDNF by the mammalian proprotein convertases. FEBS Lett 1996; Science 1996; 273: 1402–1406. 379: 247–250. 77 Levine ES, Dreyfus CF, Black IB, Plummer MR. Brain-derived 53 Seidah NG, Benjannet S, Pareek S, Savaria D, Hamelin J, Goulet B neurotrophic factor rapidly enhances synaptic transmission in et al. Cellular processing of the nerve growth factor precursor by hippocampal neurons via postsynaptic tyrosine kinase receptors. the mammalian pro-protein convertases. Biochem J 1996; 314: Proc Natl Acad Sci USA 1995; 92: 8074–8077. 951–960. 78 Korte M, Carroll P, Wolf E, Brem G, Thoenen H, Bonhoeffer T. 54 Michalski B, Fahnestock M. Pro-brain-derived neurotrophic Hippocampal long-term potentiation is impaired in mice lacking factor is decreased in parietal cortex in Alzheimer’s disease. brain-derived neurotrophic factor. Proc Natl Acad Sci USA 1995; Brain Res Mol Brain Res 2003; 111: 148–154. 92: 8856–8860. 55 Peng S, Wuu J, Mufson EJ, Fahnestock M. Precursor form of brain- 79 Xie CW, Sayah D, Chen QS, Wei WZ, Smith D, Liu X. Deficient derived neurotrophic factor and mature brain-derived neuro- long-term memory and long-lasting long-term potentiation in trophic factor are decreased in the pre-clinical stages of mice with a targeted deletion of neurotrophin-4 gene. Proc Natl Alzheimer’s disease. J Neurochem 2005; 93: 1412–1421. Acad Sci USA 2000; 97: 8116–8121. 56 Lee R, Kermani P, Teng KK, Hempstead BL. Regulation of 80 Yang F, He X, Feng L, Mizuno K, Liu XW, Russell J et al. PI-3 cell survival by secreted proneurotrophins. Science 2001; 294: kinase and IP3 are both necessary and sufficient to mediate NT3- 1945–1948. induced synaptic potentiation. Nat Neurosci 2001; 4: 19–28. 57 Rodrı´guez-Te´bar A, Dechant G, Go¨tz R, Barde YA. Binding of 81 Croen LA, Goines P, Braunschweig D, Yolken R, Yoshida CK, neurotrophin-3 to its neuronal receptors and interactions with Grether JK et al. Brain-derived neurotrophic factor and autism: nerve growth factor and brain-derived neurotrophic factor. maternal and infant peripheral blood levels in the Early Markers EMBOJ 1992; 11: 917–922. for Autism (EMA) Study. Autism Res 2008; 1: 130–137. 58 Dechant G, Rodrı´guez-Te´bar A, Barde YA. Neurotrophin recep- 82 Katoh-Semba R, Wakako R, Komori T, Shigemi H, Miyazaki N, Ito tors. Prog Neurobiol 1994; 42: 347–352. H et al. Age-related changes in BDNF protein levels in human 59 Dechant G, Biffo S, Okazawa H, Kolbeck R, Pottgiesser J, Barde serum: differences between autism cases and normal controls. Int YA. Expression and binding characteristics of the BDNF receptor J Dev Neurosci 2007; 25: 367–372. chick trkB. Development 1993; 119: 545–558. 83 Lang UE, Hellweg R, Sander T, Gallinat J. The Met allele of the 60 Mahadeo D, Kaplan L, Chao MV, Hempstead BL. High affinity BDNF Val66Met polymorphism is associated with increased nerve growth factor binding displays a faster rate of association BDNF serum concentrations. Mol Psychiatry 2009; 14: 120–122. than p140trk binding. Implications for multi-subunit polypeptide 84 Shimizu E, Hashimoto K, Iyo M. Ethnic difference of the BDNF receptors. J Biol Chem 1994; 269: 6884–6891. 196G/A (val66met) polymorphism frequencies: the possibility to 61 Schro¨pel A, von Schack D, Dechant G, Barde YA. Early explain ethnic mental traits. Am J Med Genet B Neuropsychiatr expression of the nerve growth factor receptor ctrkA in chick Genet 2004; 126B: 122–123.

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 488 85 Fischbach BV, Trout KL, Lewis J, Luis CA, Sika M. WAGR 106 Westermark B, Wasteson A. The response of cultured human syndrome: a clinical review of 54 cases. Pediatrics 2005; 116: normal glial cells to growth factors. Adv Metab Disord 1975; 8: 984–988. 85–100. 86 Xu S, Han JC, Morales A, Menzie CM, Williams K, 107 Reddy UR, Pleasure D. Expression of platelet-derived growth Fan YS. Characterization of 11p14-p12 deletion in WAGR factor (PDGF) and PDGF receptor genes in the developing rat syndrome by array CGH for identifying genes contributing to brain. J Neurosci Res 1992; 31: 670–677. mental retardation and autism. Cytogenet Genome Res 2008; 122: 108 Li WM, Huang WQ, Huang YH, Jiang DZ, Wang QR. Positive and 181–187. negative hematopoietic cytokines produced by bone marrow 87 Riikonen R. Neurotrophic factors in the pathogenesis of Rett endothelial cells. Cytokine 2000; 12: 1017–1023. syndrome. J Child Neurol 2003; 18: 693–697. 109 Bergsten E, Uutela M, Li X, Pietras K, Ostman A, Heldin CH et al. 88 Moses HL, Branum EL, Proper JA, Robinson RA. Transforming PDGF-D is a specific, protease-activated ligand for the PDGF beta- growth factor production by chemically transformed cells. receptor. Nat Cell Biol 2001; 3: 512–516. Cancer Res 1981; 41: 2842–2848. 110 LaRochelle WJ, Jeffers M, McDonald WF, Chillakuru RA, Giese 89 Shi Y, Massague´ J. Mechanisms of TGF-beta signaling from cell NA, Lokker NA et al. PDGF-D, a new protease-activated growth membrane to the nucleus. Cell 2003; 113: 685–700. factor. Nat Cell Biol 2001; 3: 517–521. 90 Nickl-Jockschat T, Arslan F, Doerfelt A, Bogdahn U, Bosserhoff A, 111 Fredriksson L, Li H, Eriksson U. The PDGF family: four gene Hau P. An imbalance between Smad and MAPK pathways is products form five dimeric isoforms. Cytokine Growth Factor Rev responsible for TGF-beta tumor promoting effects in high-grade 2004; 15: 197–204. gliomas. Int J Oncol 2007; 30: 499–507. 112 Kwon YK. Effect of neurotrophic factors on neuronal stem cell 91 Okada K, Hashimoto K, Iwata Y, Nakamura K, Tsujii M, Tsuchiya death. J Biochem Mol Biol 2002; 35: 87–93. KJ et al. Decreased serum levels of transforming growth factor- 113 Sung JY, Lee SY, Min DS, Eom TY, Ahn YS, Choi MU et al. beta1 in patients with autism. Prog Neuropsychopharmacol Biol Differential activation of phospholipases by mitogenic EGF and Psychiatry 2007; 31: 187–190. neurogenic PDGF in immortalized hippocampal stem cell lines. 92 Ashwood P, Enstrom A, Krakowiak P, Hertz-Picciotto I, Hansen J Neurochem 2001; 78: 1044–1053. RL, Croen LA et al. Decreased transforming growth factor beta1 in 114 Kajizuka M, Miyachi T, Matsuzaki H, Iwata K, Shinmura C, autism: a potential link between immune dysregulation and Suzuki K et al. Serum levels of platelet-derived growth factor BB impairment in clinical behavioral outcomes. J Neuroimmunol homodimers are increased in male children with autism. Prog 2008; 204: 149–153. Neuropsychopharmacol Biol Psychiatry 2010; 34: 154–158. 93 Toyoda T, Nakamura K, Yamada K, Thanseem I, Anitha A, Suda S 115 Weijers HG, Wiesbeck GA, Wodarz N, Keller H, Michel T, Bo¨ning et al. SNP analyses of growth factor genes EGF, TGFbeta-1, and J. Gender and personality in alcoholism. Arch Womens Ment HGF reveal haplotypic association of EGF with autism. Biochem Health 2003; 6: 245–255. Biophys Res Commun 2007; 360: 715–720. 116 Wiesbeck GA, Weijers HG, Wodarz N, Hermann M, Johann M, 94 Mozell RL, McMorris FA. Insulin-like growth factor I stimulates Keller H et al. Dopamin D 2 (DAD 2) and dopamin D 3 (DAD 3) oligodendrocyte development and myelination in rat brain receptor gene polymorphismus and treatment outcome in alcohol aggregate cultures. J Neurosci Res 1991; 30: 382–390. dependence. J Neural Transm 2003; 110: 813–820. 95 Barres BA, Hart IK, Coles HS, Burne JF, Voyvodic JT, Richardson 117 Wiesbeck GA, Weijers HG, Wodarz N, Keller HK, Michel TM, WD et al. Cell death and control of cell survival in the Herrmann MJ et al. Serotonin transporter gene polymorphism oligodendrocyte lineage. Cell 1992; 70: 31–46. and personality traits in primary alcohol dependence. World J 96 Werther GA, Russo V, Baker N, Butler G. The role of the insulin- Biol Psychiatry 2004; 5: 42–45. like growth factor system in the developing brain. Horm Res 118 Ettinger U, Kumari V, Collier DA, Powell J, Luzi S, Michel TM et 1998; 49: 37–40. al. Catechol-O-methyltransferase (COMT) Val 158 Met genotype 97 D0Mello SR, Borodezt K, Soltoff SP. Insulin-like growth factor and is associated with BOLD response as a function of task potassium depolarization maintain neuronal survival by distinct characteristic. Neuropsychopharmacology 2008; 33: 3046–3057. pathways: possible involvement of PI 3-kinase in IGF-1 signaling. 119 Sheldrick AJ, Krug A, Markov V, Leube D, Michel TM, Zerres K J Neurosci 1997; 17: 1548–1560. et al. Effect of COMT val158met genotype on cognition and 98 Beck KD, Powell-Braxton L, Widmer HR, Valverde J, personality. Eur Psychiatry 2008; 23: 385–389. Hefti F. Igf1 gene disruption results in reduced brain size, CNS 120 Nickl-Jockschat T, Rietschel M, Kircher T. Korrelation zwischen hypomyelination, and loss of hippocampal granule and Risikogenvarianten fu¨ r Schizophrenie und Hirnstrukturanoma- striatal parvalbumin-containing neurons. Neuron 1995; 14: lien. Nervenarzt 2009; 80: 40–53. 717–730. 121 Nickl-Jockschat T, Michel TM. Genetische und hirnstrukturelle 99 Riikonen R, Makkonen I, Vanhala R, Turpeinen U, Kuikka J, Anomalien bei Autismus Spektrum-Sto¨rungen: eine Bru¨ cke zum Kokki H. Cerebrospinal fluid insulin-like growth factors IGF-1 Versta¨ndnis der A¨ tiopathogenese. Nervenarzt 2010; e-pub first: and IGF-2 in infantile autism. Dev Med Child Neurol 2006; 48: doi:10.1007/so0115-010-2989-5. 751–755. 122 Autism Genome Project Consortium. Mapping autism risk loci 100 Courchesne E, Pierce K, Schumann CM, Redcay E, Buckwalter using genetic linkage and chromosomal rearrangements. JA, Kennedy DP et al. Mapping early brain development in Nat Genet 2007; 39: 319–328. autism. Neuron 2007; 56: 399–413. 123 Sadakata T, Mizoguchi A, Sato Y, Katoh-Semba R, Fukuda M, 101 Mills JL, Hediger ML, Molloy CA, Chrousos GP, Manning- Mikoshiba K et al. The secretory granule-associated protein Courtney P, Yu KF et al. Elevated levels of growth-related CAPS2 regulates neurotrophin release and cell survival. hormones in autism and autism spectrum disorder. J Neurosci 2004; 24: 43–52. Clin Endocrinol (Oxf) 2007; 67: 230–237. 124 Sadakata T, Washida M, Iwayama Y, Shoji S, Sato Y, Ohkura T 102 Fall CH, Clark PM, Hindmarsh PC, Clayton PE, Shiell AW, Law et al. Autistic-like phenotypes in Cadps2-knockout mice and CM. Urinary GH and IGF-I excretion in nine year-old children: aberrant CADPS2 splicing in autistic patients. J Clin Invest 2007; relation to sex, current size and size at birth. Clin Endocrinol 117: 931–943. (Oxf) 2000; 53: 69–76. 125 Sadakata T, Kakegawa W, Mizoguchi A, Washida M, Katoh- 103 Yamada M, Ikeuchi T, Hatanaka H. The neurotrophic action and Semba R, Shutoh F et al. Impaired cerebellar development and signalling of epidermal growth factor. Prog Neurobiol 1997; 51: function in mice lacking CAPS2, a protein involved in neuro- 19–37. trophin release. J Neurosci 2007; 27: 2472–2482. 104 Suzuki K, Hashimoto K, Iwata Y, Nakamura K, Tsujii M, Tsuchiya 126 Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U, Zoghbi KJ et al. Decreased serum levels of epidermal growth factor in HY. Rett syndrome is caused by mutations in X-linked MECP2, adult subjects with high-functioning autism. Biol Psychiatry encoding methyl-CpG-binding protein 2. Nat Genet 1999; 23: 2007; 62: 267–269. 185–188. 105 Kohler N, Lipton A. Platelets as a source of fibroblast growth- 127 Kumar A, Kamboj S, Malone BM, Kudo S, Twiss JL, Czymmek KJ promoting activity. Exp Cell Res 1974; 87: 297–301. et al. Analysis of protein domains and Rett syndrome mutations

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 489 indicate that multiple regions influence chromatin-binding temporal gyrus and temporal stem in autism. Neurosci Lett 2007; dynamics of the chromatin-associated protein MECP2 in vivo. 424: 127–132. J Cell Sci 2008; 121: 1128–1137. 150 Keller TA, Kana RK, Just MA. A developmental study of the 128 Yasui DH, Peddada S, Bieda MC, Vallero RO, Hogart A, Nagarajan structural integrity of white matter in autism. Neurosci Lett 2007; RP et al. Integrated epigenomic analyses of neuronal MeCP2 424: 127–132. reveal a role for long-range interaction with active genes. Proc 151 Dawson G, Webb S, Schellenberg GD, Dager S, Friedman S, Natl Acad Sci USA 2007; 104: 19416–19421. Aylward E et al. Defining the broader phenotype of autism: 129 Chahrour M, Jung SY, Shaw C, Zhou X, Wong ST, Qin J et al. genetic, brain, and behavioral perspectives. Dev Psychopathol MeCP2, a key contributor to neurological disease, activates and 2002; 14: 581–611. represses transcription. Science 2008; 320: 1224–1229. 152 Bennetto L, Pennington BF, Rogers SJ. Intact and impaired 130 Chen WG, Chang Q, Lin Y, Meissner A, West AE, Griffith EC et al. memory functions in autism. Child Dev 1996; 67: 1816–1835. Derepression of BDNF transcription involves calcium-dependent 153 Luna B, Doll SK, Hegedus SJ, Minshew NJ, Sweeney JA. phosphorylation of MeCP2. Science 2003; 302: 885–889. Maturation of executive function in autism. Biol Psychiatry 131 Zhou Z, Hong EJ, Cohen S, Zhao WN, Ho HY, Schmidt L et al. 2007; 61: 474–481. Brain-specific phosphorylation of MeCP2 regulates activity- 154 Ozonoff S, Cook I, Coon H, Dawson G, Joseph RM, Klin A et al. dependent Bdnf transcription, dendritic growth, and spine Performance on Cambridge Neuropsychological Test Automated maturation. Neuron 2006; 52: 255–269. Battery subtests sensitive to frontal lobe function in people with 132 Martinowich K, Hattori D, Wu H, Fouse S, He F, Hu Y et al. DNA autistic disorder: evidence from the Collaborative Programs of methylation-related in activity-dependent Excellence in Autism network. J Autism Dev Disord 2004; 34: BDNF gene regulation. Science 2003; 302: 890–893. 139–150. 133 Abrahams BS, Geschwind DH. Advances in autism genetics: 155 Volkmar F, Chawarska K, Klin A. Autism in infancy and early on the threshold of a new neurobiology. Nat Rev Genet 2008; 9: childhood. Annu Rev Psychol 2005; 56: 315–336. 341–355. 156 Kana RK, Keller TA, Minshew NJ, Just MA. Inhibitory control in 134 Nagarajan RP, Hogart AR, Gwye Y, Martin MR, LaSalle JM. high-functioning autism: decreased activation and underconnec- Reduced MeCP2 expression is frequent in autism frontal cortex tivity in inhibition networks. Biol Psychiatry 2007; 62: and correlates with aberrant MECP2 methylation. 198–206. Epigenetics 2006; 1: e1–11. 157 Just MA, Cherkassky VL, Keller TA, Minshew NJ. Cortical 135 Swanberg SE, Nagarajan RP, Peddada S, Yasui DH, activation and synchronization during sentence comprehension LaSalle JM. Reciprocal co-regulation of EGR2 and MECP2 is in high-functioning autism: evidence of underconnectivity. Brain disrupted in Rett syndrome and autism. Hum Mol Genet 2009; 2004; 127: 1811–1821. 18: 525–534. 158 Koshino H, Carpenter PA, Minshew NJ, Cherkassky VL, Keller 136 Sng JC, Taniura H, Yoneda Y. A tale of early response genes. TA, Just MA. Functional connectivity in an fMRI working Biol Pharm Bull 2004; 27: 606–612. memory task in high-functioning autism. Neuroimage 2005; 24: 137 Glorioso C, Sabatini M, Unger T, Hashimoto T, Monteggia LM, 810–821. Lewis DA et al. Specificity and timing of neocortical transcrip- 159 Koshino H, Kana RK, Keller TA, Cherkassky VL, Minshew NJ, tome changes in response to BDNF gene ablation during Just MA. fMRI investigation of working memory for faces in embryogenesis or adulthood. Mol Psychiatry 2006; 11: 633–648. autism: visual coding and underconnectivity with frontal areas. 138 Adegbola A, Gao H, Sommer S, Browning M. A novel mutation in Cereb Cortex 2008; 18: 289–300. JARID1C/SMCX in a patient with autism spectrum disorder 160 Erickson KI, Kim JS, Suever BL, Voss MW, Francis BM, Kramer (ASD). Am J Med Genet A 2008; 146A: 505–511. AF. Genetic contributions to age-related decline in executive 139 Christensen J, Agger K, Cloos PA, Pasini D, Rose S, Sennels L function: a 10-year longitudinal study of COMT and BDNF et al. RBP2 belongs to a family of demethylases, specific for tri- and polymorphisms. Front Hum Neurosci 2008; 2:11. dimethylated lysine 4 on histone 3. Cell 2007; 128: 1063–1076. 161 Raz N, Rodrigue KM, Kennedy KM, Land S. Genetic and 140 Iwase S, Lan F, Bayliss P, de la Torre-Ubieta L, Huarte M, Qi HH vascular modifiers of age-sensitive cognitive skills: effects of et al. The X-linked mental retardation gene SMCX/JARID1C COMT, BDNF, ApoE, and hypertension. Neuropsychology 2009; defines a family of histone H3 lysine 4 demethylases. Cell 2007; 23: 105–116. 128: 1077–1088. 162 Sakata K, Woo NH, Martinowich K, Greene JS, Schloesser RJ, 141 Durany N, Michel T, Zochling R, Boissl KW, Cruz-Sanchez FF, Shen L et al. Critical role of promoter IV-driven BDNF transcrip- Riederer P et al. Brain-derived neurotrophic factor and neuro- tion in GABAergic transmission and synaptic plasticity in the trophin 3 in schizophrenic psychoses. Schizophr Res 2001; 52: prefrontal cortex. Proc Natl Acad Sci USA 2009; 106: 5942–5947. 79–86. 163 Moy SS, Nadler JJ, Perez A, Barbaro RP, Johns JM, Magnuson TR 142 Enstrom A, Onore C, Tarver A, Herzt-Picciotto I, Hansen R, Croen L et al. Sociability and preference for social novelty in five inbred et al. Peripheral blood leukocyte production of BDNF following strains: an approach to assess autistic–like behavior in mice. mitogen stimulation in early onset and regressive autism. Am J Genes Brain Behav 2004; 3: 287–302. Biochem Biotechnol 2008; 4: 121–129. 164 Moy SS, Nadler JJ, Young NB, Perez A, Holloway LP, Barbaro RP 143 O0Hearn K, Asato M, Ordaz S, Luna B. Neurodevelopment et al. Mouse behavioral tasks relevant to autism: phenotypes of 10 and executive function in autism. Dev Psychopathol 2008; 20: inbred strains. Behav Brain Res 2007; 176: 4–20. 1103–1132. 165 Lewis MH, Tanimura Y, Lee LW, Bodfish JW. Animal models of 144 Mu¨ ller RA. From loci to networks and back again: anomalies in restricted repetitive behavior in autism. Behav Brain Res 2007; the study of autism. Ann NY Acad Sci 2008; 1145: 300–315. 176: 66–74. 145 Verhoeven JS, De Cock P, Lagae L, Sunaert S. Neuroimaging of 166 Branchi I, Ricceri L. Transgenic and knock-out mouse pups: the autism. Neuroradiology 2010; 52: 3–14. growing need for behavioral analysis. Genes Brain Behav 2002; 1: 146 Barnea-Goraly N, Kwon H, Menon V, Eliez S, Lotspeich L, 135–141. Reiss AL. White matter structure in autism: preliminary 167 Branchi I, D0Andrea I, Sietzema J, Fiore M, Di Fausto V, Aloe L evidence from diffusion tensor imaging. Biol Psychiatry 2004; et al. Early social enrichment augments adult hippocampal 55: 323–326. BDNF levels and survival of BrdU-positive cells while increasing 147 Alexander AL, Lee JE, Lazar M, Boudos R, DuBray MB, Oakes TR anxiety- and ‘depression’-like behavior. J Neurosci Res 2006; 83: et al. Diffusion tensor imaging of the corpus callosum in 965–973. AUTISM. Neuroimage 2007; 34: 61–73. 168 Ricceri L, Moles A, Crawley J. Behavioral phenotyping of mouse 148 Sundaram SK, Kumar A, Makki MI, Behen ME, Chugani HT, models of neurodevelopmental disorders: relevant social Chugani DC. Diffusion tensor imaging of frontal lobe in autism behavior patterns across the life span. Behav Brain Res 2007; spectrum disorder. Cereb Cortex 2008; 18: 2659–2665. 176: 40–52. 149 Lee JE, Bigler ED, Alexander AL, Lazar M, DuBray MB, Chung 169 Moy SS, Nadler JJ. Advances in behavioral genetics: mouse MK et al. Diffusion tensor imaging of white matter in the superior models of autism. Mol Psychiatry 2008; 13: 4–26.

Molecular Psychiatry Neurotrophic factors in autism T Nickl-Jockschat and TM Michel 490 170 Picker JD, Yang R, Ricceri L, Berger-Sweeney J. An altered palate in mice lacking TGF-beta3 indicates defects of epithelial- neonatal behavioral phenotype in Mecp2 mutant mice. mesenchymal interaction. Nat Genet 1995; 11: 415–421. NeuroReport 2006; 17: 541–544. 176 Kulkarni AB, Karlsson S. Transforming growth factor-beta 1 171 Moles A, Kieffer BL, D0Amato FR. Deficit in attachment behavior knockout mice. Am J Pathol 1993; 143:3–9. in mice lacking the mu-opioid receptor gene. Science 2004; 304: 177 Sanford LP, Ormsby I, Gittenberger-de Groot AC, Sariola H, 1983–1986. Friedman R, Biovin GP et al. TGF-b2knockoutmicehavemultiple 172 Chen ZY, Jing D, Bath KG, Ieraci A, Khan T, Siao CJ et al. Genetic developmental defects that are non-overlapping with other TGF-b variant BDNF (Val66Met) polymorphism alters anxiety-related knockout phenotypes. Development 1997; 124: 2656–2670. behavior. Science 2006; 314: 140–143. 178 Arslan F, Bosserhoff AK, Nickl-Jockschat T, Doerfelt A, Bogdahn 173 Thome J, Nara K, Foley P, Michel T, Gsell W, Retz W et al. CNTF U, Hau P. The role of versican isoforms V0/V1 in glioma genotypes: influence on choline acetyltransferase (ChAT) and migration mediated by transforming growth factor-beta2. acetylcholine esterase (AChE) activities and neurotrophin 3 Br J Cancer 2007; 96: 1560–1568. (NT3) concentration in human post mortem brain tissue. J Brain 179 Brkanac Z, Raskind WH, King BH. Pharmacology and genetics of Res 1997; 38: 443–445. autism: implications for diagnosis and treatment. Per Med 2008; 174 Michel TM, Frangou S, Camara S, Thiemeyer D, Jecel J, Tatschner 5: 599–607. T et al. Altered glial cell line-derived neurotrophic factor (GDNF) 180 Correia CT, Almeida JP, Santos PE, Sequeira AF, Marques CE, concentrations in the brain of patients with depressive disorder: Miguel TS et al. Pharmacogenetics of risperidone therapy in a comparative post-mortem study. Eur Psychiatry 2008; 23: autism: association analysis of eight candidate genes with drug 413–420. efficacy and adverse drug reactions. Pharmacogenomics J 2009. 175 Kaartinen V, Voncken JW, Shuler C, Warburton D, Bu D, 181 Tsai SJ. TrkB partial agonists: potential treatment strategy for Heisterkamp N et al. Abnormal lung development and cleft major depression. Med Hypotheses 2007; 68: 674–676.

Molecular Psychiatry