Autism Speaks, Science, Latest Research

The special symposium was organized by Autism Speaks and supported by the Wellcome Trust. It was attended by 50 people including researchers at the cutting edge of this frontier in autism research and related fields and by representatives from Autism Speaks in the US and UK; the Wellcome Trust and other funding organizations (including the UK's Medical Research Council and the Simons Foundation); and pharmaceutical companies with an active R&D interest in autism and related disorders. The participants travelled from a number of countries in Europe including the UK, and from Canada and the United States.

The scientific program was focused and intensive and there was a high level of knowledge exchanged across different research disciplines, concentrating on data published within the last 12 months or in press. The symposium opened with a description of the clinical characteristics of autism including both impairment and behavioral problems and the special skills associated with the disorder. The current status of our knowledge regarding the emergence of early behavioral signs in very young children was reviewed and it was noted that this knowledge was acquired by studying the younger baby brothers and sisters of children already diagnosed with autism. By following a group of siblings it has been possible to show that atypical behaviors such as reduced eye contact or reduced social smiling begin to emerge around 6 months of age. These studies are relevant towards building the knowledge base for early detection and diagnosis, leading to opportunities for even earlier intervention.

The symposium also reviewed the current status of what is known about changes in autism brain - both in living individuals and findings using post-mortem brain tissue. It is clear that changes can be seen at the structural, functional, organizational, neurochemical and cellular levels. In general the brain alterations in autism are complex, and the observed change can depend on age and be confined to specific brain areas. For example, it was discussed that in early life, head circumference can differ from normal but the deviation depends on age. For some individuals increased head size was coincident with enlarged body size and more impaired adaptive behaviors, but less so with impaired IQ scores and verbal language development. So increased head size may reflect a sub-group or sub-phenotype of autism.

The data presented on autism genetics were very powerful and generally reflected the major investment made in this sector. Like other psychiatric and neurological disorders a number of genetic autism susceptibility factors have been detected in a proportion of individuals affected by the disorder (perhaps 5-10%). Some genetic mutations have been detected and although these are rare and not generalizable to all those affected by autism, nevertheless the findings are an important breakthrough since the genes may converge in a biochemical pathway. At this point, most of the identified genes code for proteins that function at the neuronal synapse. This exciting discovery is beginning to unravel the complex neurobiology of autism and these findings, together with other presentations on how genes function to ‘normally' regulate the neuronal synapse, prompted a discussion on whether autism is a ‘synaptic disorder' and how this scientific question might be addressed. ‘Clock' genes that regulate circadian rhythm and sleep patterns were also discussed on the basis that some people with autism show altered melatonin levels and sleep problems.

The talks on genetics also focused on another type of genetic work using cutting edge technologies to study Copy Number Variations (CNVs, i.e submicroscopic additions and deletions in DNA). The data showed that CNVs are important in autism: in families with only one individual affected by autism (simplex families), CNVs are common and these CNVs tend not to be inherited from the parents but arise de novo or spontaneously. Whereas in families with more than one affected individual (multiplex families), CNVs are less common but tend to be inherited from the parents. Very recently a CNV on chromosome 16 (position 16p11.2) has been shown to be associated with language impairment in autism and other language disorders. This is highly important because it opens up the possibility that a young child at risk for autism could be given a CNV test and interventions could be provided to specifically address language. More work is needed to bring this discovery to the point of being a reliable diagnostic test that could be included as part of a clinical management program for young children.

The genetic findings are beginning to unravel the complexity of the altered neurobiology in autism, and animal models will be pivotal for use in the laboratory so that researchers can experimentally test the role of genes and their impact on brain function, for example pinning down how changes in genes actually cause the behaviors seen in autism. Good progress has been made in developing valuable mouse ‘models' of autism and these are based on observing modifications to natural mouse behaviors as analogues of the human symptoms. These mouse models are being optimized to provide a laboratory system to test the effects of implicated gene targets and to evaluate the efficacy of novel treatments.

Importantly mouse models of disorders related to autism have already provided important and converging lines of evidence, regarding the plasticity of neurons and the brain retaining the ability to remodel itself even in maturity. The study of mouse models of disorders such as Rett syndrome, Fragile X and Tuberous Sclerosis produced some remarkable findings: in mice which received a single genetic manipulation to induce neurological and behavioral symptoms that mimic these disorders, it was possible to rescue the phenotypes and reverse some if not all of the neurological symptoms either by replacement of the defective gene or by administration of a pharmacological agent and it was possible to do this in juvenile and mature animals. This prompts the highly controversial suggestion that in these mouse models, neurons that were once defective, are not damaged irrevocably and can retain the ability to repair both in juvenile and mature animals. While mice are not men, these findings should be further considered since they have important implications for adolescents and adult affected by neurodevelopmental disorders i.e. that potential therapy could address the heart of the disorder rather than just offering symptomatic relief.

Environmental contributions to autism were also discussed, most specifically with respect to immune stimulation. The origin of potential immune dysfunction in autism was examined in terms of a pre-natal origin that could persist into adulthood and how this might impact on the developing and mature brain through disrupted signaling pathways and formation of oxidative stress.

The symposium was closed by a family perspective on living with autism and it was clear that although the neurobiology research will provide answers regarding cause and consequence, there are issues that families deal with on a daily basis (such as problematic associated behaviors rather than core autism deficits) that research also needs to address.

Overall the symposium achieved the following:

reviewed the outstanding science in this arena of neurobiology and autism research;
highlighted some of the research challenges and roadblocks that need to be addressed (such as the post-mortem neuropathology being generally studied in brains donated by individuals with more severe autism), gaps in the knowledge base (for example lack of MRI data on brain structure in young children) and neglected areas (such as the special skills found in some individuals with autism;
identified the key scientific questions, in particular those amenable to experimental investigation with today's available resources and scientific methodologies;
promoted cross-disciplinary interactions and networking opportunities
attracted new high quality and capable scientists to the field;
provided insight into the future directions of autism neurobiology research and an opportunity to work jointly with the Wellcome Trust as a prelude to other potential interactions.
showcased neurobiology research in autism to other major funders

DETAILED SUMMARY

The symposium on the neurobiology of autism took place at the new Wellcome Trust conference facilities in London on 9^th and 10^th January 2008 and was attended by 50 people including researchers at the cutting edge of this frontier in autism research and related fields and by representatives from Autism Speaks in the US and UK; the Wellcome Trust; other funding organizations (including the UK's Medical Research Council and the Simons Foundation); and pharmaceutical companies with an active R&D interest in autism and related disorders. The participants travelled from France, Germany, Italy, the Netherlands, Canada, and a number of research centers in the United States and United Kingdom.

This was the first formal occasion on which Autism Speaks collaborated with the Wellcome Trust. The objectives set for the symposium were to:

to review the outstanding science in the neurobiological arena of autism research;
gain insight into the directions of autism neurobiology research over the next few years and identify roadblocks;
identify the key scientific questions, in particular those amenable to experimental investigation with today's available resources and scientific methodologies;
promote cross-discipline interactions and networking opportunities;
attract new high quality and capable scientists to the field;
to work jointly with the Wellcome Trust as a prelude to other potential interactions;
to showcase neurobiology research in autism to other major funders.

The symposium was chaired by Professor Martin Raff (University College, London) and Dr. Christopher Walsh (Harvard Medical School). The scientific program was focused and intensive. Some areas of autism research although important (e.g. the link between neurobiology, cognition and psychology (such as Executive Function and Theory of Mind)) were not included due to time constraints.

On Day 1, the first session was on the Clinical Characteristics of Autism

and the symposium was opened by a review of ‘Autism: Clinical features and research challenges' by the distinguished Professor Sir Michael Rutter (Institute of Psychiatry, London and Patron of Autism Speaks in the UK). Professor Rutter reviewed the definition and heterogeneity of autism symptoms, and noted the first signs manifest at 12-24 months of age - although for regression (which may account for up to one third of cases) the signs emerge later at 18-30 months. The key features of autism persist throughout adult life and there is huge variation in the extent of impairment. He talked about the genetic features of the disorder and familial risk but also discussed causal factors of non-genetic origin. In summarizing the neurobiological alterations in autism, he noted there are few consistent neuropathological changes and no localized structural brain abnormalities but rather impaired integration across brain systems. In reviewing the research challenges, he noted the focus on the deficits of autism and the neglect of the special skills some with autism show; and that research needs to address the issue of the studying sub-groups and the generalizability of the research findings. For example, post mortem brain research most often uses brain material donated by individuals more severely affected by autism and with co-morbid epilepsy.

This was followed by a talk from Dr. Lonnie Zwaigenbaum (University of Alberta) on ‘Early developmental and behavioral trajectories in autism: insights from infant siblings' studies.' He reviewed the need for prospective studies (such as of the ‘baby sibs' of children already diagnosed with autism) to monitor the onset and developmental course of the early behavioral features of autism and ASD (autism spectrum disorders). He reviewed the emerging hypotheses regarding underlying developmental processes and how this might fit with other features such as increased head circumference and brain size. He showed there is progressive onset of behavioral symptoms which increase and evolve over the first few years of life. For instance some behaviors (atypical visual tracking, eye contact, orienting to name, social interest and affect, imitation, reactivity, repetitive sensory-oriented behaviors) emerge at 6 months of age and become exaggerated at 12 months of age. This was most readily apparent in the children who later received an autism diagnosis and to a lesser extent in siblings of children with other ASDs or developmental delay. Dr. Zwaigenbaum was mindful of the need for reliable interpretation of such findings and specificity for autism (versus developmental delay). He showed the onset of symptoms follows different trajectories. For example differences in the AOSI scores (Autism Observation Scale for Infants, a standardized, interactive, play-based measure of early signs of autism), emerge at 6 months and follow a different trajectory for children with ASD and intellectual impairment. Cognitive and language development is atypical - for example of the nine baby siblings who received an ASD diagnosis and were followed prospectively, six had reduced IQ and language scores from 12-36 months, concurrent with progressive worsening of ASD symptoms and the three with average or near-average language skills, followed a steady trajectory despite symptoms. Dr. Zwaigenbaum talked about regression in autism and discussed whether this was slow down or reduced progress in the acquisition of new skills and the possibility that regressive autism represents the extreme of a continuum of developmental trajectories.

Neuropathology, neuroimaging
Dr. Eric Courchesne (University of California, San Diego) reviewed the ‘Three phases of brain growth pathology in autism: Postmortem and in vivo imaging evidence.' He immediately identified a gap in the knowledge base of MRI imaging data on brain structure in younger children (both autistic and neurotypical). He went on to talk about evidence for altered brain structure in adolescents and adults affected by autism, such as reduced neuron size and volume, different patterns of neuronal organization within specific brain structures and the signs of neuroinflammation. He questioned whether the findings from brain studies in ‘older' individuals with autism could be extrapolated to a younger age group. Compared to neurotypical, there are alterations in head circumference in young children who later receive an autism diagnosis- with head circumference being reported as normal or even smaller at birth in autism, followed a phase of accelerated growth and larger head circumference at 12 months of age (with brain enlargements seen particularly in the frontal and temporal lobes). After years 1 or 2, this overgrowth slows and at 6 years of age, the size of autism brain and head circumference is similar to control. At a functional level, imaging has shown enlargements are present in brain areas associated with social and language skills. It is important to note that brain overgrowth is phased and follows a variety of trajectories and is not present in all autism cases.

Professor Declan Murphy (Institute of Psychiatry, London) talked about ‘Neuroimaging and autism; multidisciplinary approaches relating biology to symptoms, causes, and treatments.' He posed the question as to what new research avenues and directions the known neurobiology of autism opens up. He noted that brain imaging studies contribute to the debate on brain overgrowth in young children and provide evidence to support alterations in brain structure and functional connectivity of brain systems and integration in autism. He reviewed some specific examples generated by the research team at the Institute of Psychiatry and concluded that brain alterations are age and ‘phenotype' dependent, that the observed alterations are complex and can be regionally specific. He reviewed the evidence for altered brain neurochemistry, particularly in the serotonergic and glutaminergic systems which offer in-roads into the development of tractable novel therapeutic approaches for treatment of the autism core deficits and particularly problematic behaviors such as repetitive behaviors.

The next session focused on Genetics .

Professor Anthony Monaco (University of Oxford and principal investigator in the Autism Genome Project) gave an overview of ‘The genetics of autism.' He noted that autism is a complex genetic disorder and so far a number of mutations have been identified, although these mutations are present only in a small proportion of affected individuals and families (and indeed non-affected individuals) and therefore not generalizable to the whole autism population. Recently mutations in neuroligin 4, neurexin and Shank 3 genes have been reported and although these are rare mutations, together the findings are pivotal and represent converging evidence for a biochemical pathway altered in autism - these ‘scaffold' proteins are functionally linked and clustered at the neuronal synapse. Professor Monaco talked about the consortia of genetic researchers and their activities and reviewed the Autism Genome Project (AGP). He showed that the AGP has generated a large dataset with genome wide SNP analysis that confirmed some previously reported candidate gene areas and also identified new candidate areas. Phase 1 of AGP focused on Copy Number Variation (CNV, i.e. submicroscopic chromosomal duplications or deletions), linkage analysis, association analysis and trait subset analysis. De novo CNV (i.e. not inherited from the parents or very rare found in the general population) were more commonly found in simplex families (i.e. with only one affected offspring) and he highlighted the recent finding of a microdeletion in chromosome 16p11.2 in approximately 1% of affected families. Phase II of the AGP will continue with whole genome association and CNV analysis, as well as re-sequencing candidate genes. The AGP is coordinating its efforts with other research groups with similar interests, including a formal collaboration on a whole genome association study with the Broad Institute, a member of the Boston-based Autism Consortium.

Dr. Christopher Walsh (Harvard Medical School) talked on ‘Diverse genetic mechanisms of autism in diverse populations.' He showed that 10% of simplex families have CNVs and this finding is striking and could potentially be a genetic based test to complement diagnosis or evaluate autism susceptibility. Indeed his Boston-based Autism Consortium recommended that some children should be offered chromosomal microarray testing in addition to the standard karityping and Fragile X screening. He talked about CNVs and how these were derived showing that de novo CNVs are more common in simplex families whereas inherited CNVs and DNA mutations are more common for multiplex (i.e. more than one affected offspring) and consanguineous families. He noted that of the recently identified CNV in chromosome 16p11.2, some were de novo while others inherited.

Dr. Daniel Geschwind (University of California, Los Angeles) spoke about ‘Autism genetics and the challenge of heterogeneity.' He highlighted the pivotal role the AGRE (Autism Genetic Resource Exchange) has played in enabling genetics studies. Despite this, the AGRE sample size is not large enough. Yet, he noted that increasing the sample size does not necessarily deliver a proportional increase in the statistical power. An alternative approach to increase statistical power is to use a sub-population of DNA donors with a narrower, well defined phenotype. He went on to discuss the inherent challenge of ‘knowing the phenotype' since autistic traits extend and merge into ‘normality' and how this relates to common genetic variants. He highlighted recent findings that specific autistic traits can be linked to specific gene candidates, for example CNTNAP2 (Contactin Associated Protein-Like 2) is associated with language impairments in males with autism and also other language based disorders.

Overall the genetics talks were very powerful and reflected the state of the field and the major investment made in this sector. However, it remains the case that, as with most other psychiatric and neurological disorders, the known genetic autism susceptibility factors occur in a small number of families. Nevertheless, such knowledge about genetic risk and where in the brain these genes are active and exert their effects, are important steps in unraveling the altered neurobiological pathways and brain structures in autism. Importantly identification of genetic targets provides opportunity for the generation of mouse models of autism and pinning down how genes impact on behavior and a laboratory setting to test potential new interventions.

Jacqueline Crawley, an expert on animal behavior with the United States National Institute of Mental Health, opened a session on Mouse Models with ‘Modeling the Behavioral Symptoms of Autism in Mice: Assays to Test Hypotheses and Evaluate Therapeutics.' The important point was made that whilst we should not ‘anthropomorphize', some of the natural behaviors of mice are analogous to and can be used to mimic aspects of autistic phenotype such as sociability, communication and repetition. Firstly for modeling sociability, measurement of behaviors such as social approach, grooming, nose-to-nose sniffing, pushing past other mice or exploration is useful. Dr. Crawley noted the International Mouse Phenome Project had revealed a useful mouse strain (BTBR) which exhibits social deficits which are present in juveniles and adult animals. Interestingly the BBTR strain shows abnormal brain structure and specifically the absence of the corpus callosum (the nerve bridge between the two hemispheres) which has been implicated as part of the impaired connectivity of brain systems in autism). Repetition can be modeled using, for example, repeated self grooming, or resistance to change of a learned pattern of behavior or of routine in searching corridors for hidden food. The greatest challenge remains in modeling animal behaviors relevant to communication, and the development of assays based on olfaction (mouse social odours) or auditory signals such as ultrasonic vocalization on maternal separation is underway. The development of all these mouse behavioral assays has been mindful of reliability, replicability, quantification and automation. The optimization of the phenotyping assays is on-going and is an iterative process in discovering the causes of autism since mouse models are pivotal in testing hypotheses about the role of specific genes and as tools for translating and evaluating the efficacy of novel treatments.

Randall Carpenter (Seaside Therapeutics) talked about ‘Fragile X and the pathophysiology of autism.' In contrast to autism the genetics of Fragile X are simple: a single mutation in the gene coding for FMR1 (Fragile X Mental Retardation) protein underlies the Fragile X syndrome - the most common form of heritable mental retardation and one of the leading identified causes of autism. Studying the biology of a monogenic disorder may provide a means of unraveling the complex neurobiology of autism. Fragile X is becoming regarded as a disorder of synaptic function. FMRP is a transcription regulator (i.e. normally dampens the transcription of RNA to protein), and the mutations seen in Fragile X prevent FMRP from functioning normally - the impact of the mutation is to remove the brake on transcription. Although FMRP is promiscuous and dysfunction can increase synthesis of many different proteins, it seems that over expression of the mGlu5 receptor protein is pivotal which implies that blockade of the mGlu5 receptor could be an effective therapeutic strategy to overcome the molecular defects. Dr. Carpenter highlighted a recent study from Mark Bear's (Massachusetts IThg) group which showed that in transgenic mice carrying the mutated FMR1 protein, when a second genetic manipulation is introduced to reduce the function of the mGlu5 receptor gene, the Fragile X-like symptoms were abolished. mGlu5 receptors are a tractable pharmacological target and indeed small molecule antagonists exist as development candidates (as spin-offs from R&D activity in other therapeutic areas such as pain and anxiety). Dr. Carpenter went on to review studies showing that pharmacological antagonism of the mGluR5 receptor can overcome the molecular deficits of mutated FMRPhg in several model systems, both vertebrate and non-vertebrate. He detailed the properties of STX107 a novel mGluR5 receptor antagonist currently under development by Seaside Therapeutics as a novel therapeutic approach in Fragile X.

More insight on potential recovery of function was given by Dr. Adrian Bird (University of Edinburgh) who described animal studies relating to Rett syndrome (one of the five Pervasive Developmental Disorders) which affects 1 in 10,000 girls. In his talk ‘Reversibility of Rett syndrome-like symptoms in a mouse model' he discussed that Rett syndrome results from a mosaic expression of mutated copies of the MECP2 gene. The MECP2 protein binds to methylated sites in genomic DNA and facilitates gene silencing and is highly abundant in neurons of the mature nervous system. In mice, knock down of the MECP2 gene from the embryonic stage, caused Rett-like symptoms and a severe neuronal phenotype with abnormal gait, irregular breathing and hind limb clasping. The symptoms first appeared in mice aged 4-12 months and the onset was associated with reduced long term potentiation in hippocampal brain slices (an electrophysiological measure of learning). The animals had reduced brain size, decreased dendritic arborization of neurons but signs of cell death or degeneration were absent. He showed that in mice, reactivation of the MECP2 gene progressively reversed the neurological symptoms (i.e. phenotype reversal) and normalized neuronal function (as measured by LTP). These findings in immature and mature animals imply the brain is ‘plastic', that MECP2 is involved in neuronal maintenance rather than development and defective neurons that are not irrevocably damaged retain the ability to recover; how these remarkable findings of rescue in experimental animals relate to the human situation is not known.

Daniela Toniolo (Instituto Scientifico San Raffaele, Milano) presented on ‘Cognitive impairments in Gdi1 knockout mice and mental retardation – from human gene to animal model' (mental retardation is often co-morbid with autism). She reviewed the diverse genetic findings in mental retardation with over 30 genes being implicated. She then focused on the GDi1 gene which regulates Rab GTPase activity and is involved in buffering synaptic vesicles (which store neurochemicals in readiness to be released on synaptic activation). GDi1 is involved in the recycling or regeneration of active Rab GTPase. Transgenic mice in which the GDi1 gene has been deleted show altered behaviors - less aggression and, in the males only, impaired working memory and fear conditioning, altered social behaviors and enhanced sensitivity to pro-convulsants. In these transgenic mice, electron microscope studies of the ultrastructure of the CA1 region of the hippocampus showed that synapses and synaptic vesicles appear normal. Electrophysiological studies in hippocampal slices, showed normal function of the readily releasable pool of synaptic vesicles but the reserve (or topping up) vesicular pool was reduced and became depleted following application of long trains of electrical stimuli. Together these findings indicate a loss of synaptic strength. The hippocampus is important for learning and memory which is impaired in mental retardation and these mice may offer a suitable laboratory based test system.

Luis Parada (University of Texas Southwestern Medical Center) closed the session with talk on ‘Mouse models of P13-kinase signaling in the CNS and social interaction deficits.'

Dr. Parada showed that PTEN (phosphatase and tensin homologue on chromosome ten) is selectively expressed in specific regions of the brain and reviewed innovative studies to determine the specific effect of regional knock down of PTEN localized to either cortical or hippocampal neurons (note only about 30% of neurons in each region were affected). In the hippocampal PTEN-/- knock down experiments, it was found that from 4 weeks old some of the mice showed macroencephaly (with specific hypertrophy in the dentate gyrus region of the hippocampus), seizures, social interaction deficits (hypoactive when in group of animals yet hyperactive when the animal is alone,) and elevated anxiety. By 10 weeks symptoms were present in all mice. Chronic administration of rapamycin suppressed the enlargement of the dentate gyrus and dendritic hypertrophy. Rapamycin was effective when administered before or after the onset of symptoms of hippocampal PTEN knock down. In the hippocampal knockdown animals, rapamycin was ineffective versus the seizures (which are a secondary effect). When PTEN-/- knock down was localized to the cortex, the cortex neurons showed anatomical features similar to those seen in mice in which the Tuberous Sclerosis genes had been deleted (TSC-/-). In the cortical PTEN-/- knock downs, animals treated with rapamycin showed normal nesting activity, improvements in 16 other behaviors and reduced duration and amplitude of seizures.

On Day 2, the first session examined some basic science around the synapse and posed the question ”is autism a synaptic disorder?”

Thomas Bourgeron (Pasteur Institute, Paris) explored ‘The interplay between synaptic and clock genes in autism spectrum disorders.' This presentation focused on the finding that some individuals with ASD show reduced plasma levels of melatonin. Melatonin is involved in regulating circadian rhythm and importantly sleep patterns. It is synthesized in the pineal gland and derived from tryptophan with the last step in the synthetic pathway being catalyzed by the enzyme acetyl serotonin methyltransferase (ASMT). The ASMT gene is altered in a small number of individuals with autism (with a splice variant mutation causing truncation in the promoter region) and this results in abolition of the normal diurnal variation in melatonin levels. Physiologically melatonin has effects (via GABA and glutamate) in the suprachiasmatic nucleus which is part of the hypothalamus. The hypothesis that supplementation of melatonin might have therapeutic benefit is currently being tested in a large multicentre randomized trial in the UK, known as the MENDS trial (The use of Melatonin in children with Neurodevelopmental Disorders and impaired Sleep). Dr. Bourgeron also noted that in terms of genetics, a gene dosage effect had important consequences on expression of the phenotype. For example, deletions in the Shank3 gene are associated with autism whereas duplication of Shank 3 is associated with ADHD.

Nils Brose (Max Planck Institute of Experimental Medicine Göttingen, Germany) continued with ‘Genetic dissections of neuroligin function - how aberrant synapse function in the brain might cause autism.' He explained that neuroligins are transmembrane proteins, found in most synapses in the brain, yet surprisingly are not required for synapse formation since synaptic density is unaltered in the absence of neuroligins. Neuroligins are involved in synaptic maintenance and contribute to use-dependent formation of neural circuits. Selective deletion of the neuroligin (NLG) genes has differential effects on different types of synapse - with NLG1 being associated with excitatory (glutamate) synapses and regulating the ratio of NMDA/AMPA receptors whereas NLG2 is associated exclusively with inhibitory (GABA and glycine) synapses. He reviewed evidence from mouse studies of how synapses function in the absence of the NLG4 gene and showed that whilst NLG4 knock out mice were apparently normal they show some altered social behaviors and a subtle reduction in brain size. Preliminary data suggest NLG4 knock out mice have AMPA receptor defects.

Michael Greenberg (Harvard University) the theme on basic synapse biology by examining ‘Signal transduction networks that regulate synapse development and cognitive function.' Out of the 600 genes that are known to regulate synapses, he focused on two genes (MEF2 and NPAS4), which are transcriptional regulators, and used hippocampal neurons as a model system. The action of MEF2 is to reduce synapse number by turning on other genes. MEF2 is particularly active at excitatory synapses and its activity increases upon functional activation of these synapses (i.e. MEF2 activation restricts excitatory synapses). On the other hand, in mice knockdown of the NPAS4 gene causes a reduction in inhibitory synaptic density especially in some inhibitory interneurons. Mice with NPAS4 deletion show a behavioral phenotype associated with anxiety, hyperactivity and seizures. In hippocampal slices silencing of NPAS4 function by a siRNA, caused a decrease in the frequency of inhibitory post-synaptic currents (IPSCs). Speculatively, BDNF, which has been implicated in autism, is a downstream transcriptional target of NPAS4.

The second session of Day 2 focused on Genes and Environment . Paul Patterson (California Institute of Technology) opened the session with a talk on the ‘Immune involvement in autism and schizophrenia: etiology and animal models.' He began by reviewing the evidence of immune system involvement in schizophrenia. He noted that an important issue when considering genes and the environment is that about one third of monozygotic twins do not share the same placenta, which means that they would have different in utero environments. He reviewed the scientific literature regarding the involvement of the immune system in autism and noted that post mortem brain studies showed chronic microglial activation (in brain material obtained from donors aged 5-44 years of age). He examined the possible origins of permanent CNS immune dysregulation and examined the evidence obtained from animal studies regarding a possible pre-natal insult or trigger for post-natal neuroinflammation. He showed that following lipopolysaccaride (LPS) administration to pregnant mice, the offspring had an altered behavioral profile with some ‘autistic like' characteristics such as increased anxiety to mild stress and reduced working memory. Neuroanatomical examination showed alterations in the cerebellum (reduced Purkinje cell number from day P11 and in adulthood) and altered migration of cortical neurons. He has carried out an elegant set of experiments implicating the cytokine IL-6 as one of the immune mediators of this maternal immune activation model. Further investigation is needed.

Antonio Persico (University "Campus Bio-Medico" Rome) talked about ‘Genes, environmental factors and oxidative stress.' He described how his research has uncovered a previously undetected autism subgroup. Studies have consistently found that large head size, or macrocephaly, is more frequent in people with autism. Dr. Persico discovered that individuals with large heads share other similar features, including macrosomy, or abnormally enlarged body size. Using subjects and data from the Autism Genetic Resource Exchange (AGRE), the researchers analyzed numerous morphological, clinical, and biochemical features of 241 people with autism to see if any features co-occurred with macrocephaly. In addition to finding that head circumference was highly correlated with height and weight (macrosomy), they found that macrocephaly was associated with more impaired adaptive behaviors, but with less impaired IQ scores and verbal language development. Most surprisingly, macrocephaly was also associated with self-reported immune issues (particularly atypical food ‘allergies') in either the patient or their first degree relatives.

Although the associations between head circumference, height, and weight in autism are only correlations and do not indicate that one causes the other, their convergence in this study potentially identifies a "macrosomic" endophenotype. People with this endophenotype may have a similar cause for their autism, and future research could test whether they share similar genetic or exposure anomalies.

Jane Westley (Autism Speaks) closed the formal presentations with ‘Living with autism – a family's perspective' describing life with her autistic son George. Her openness reminded everyone of the purpose of autism research, how it should be directed and exploited to meet the needs of families living with the disorder and how basic science discoveries should be translated into something tangible and meaningful that lessens the impact of the disorder and improves quality of life for all those affected - including interventions to deal with co-morbidities associated with autism which often cause troublesome behavioral problems. She highlighted how individual needs change as a toddler grows into childhood and adolescence.

Scientific Outcomes

The symposium closed with a lively and informed discussion on the following:

whether the existing neurobiological evidence supports autism being a ‘synaptic' disorder and whether mutations in ‘scaffold' proteins could cause autism; can and how do mutant synaptic proteins relate to autism, mental retardation and epilepsy, and where is the specificity underlying these phenotypes?;
that refinement or validation of the synaptic hypothesis is experimentally tractable and steps could include: sequencing all the synaptic-protein genes; investigating the ‘autistic' synapse within a functional neuronal circuit or brain slice; or using stem cells particularly those derived from post-mortem brains to model synaptogenesis in autism;
that research findings need to be translated across clinical observations and mouse behaviors;
that the understanding of the neuropathology of the human brain needs to be expanded and the problems of performing this type of research need to be addressed, principally the scarcity of biomaterials;
if we are to move the mouse field forward, there needs to be discussion on the use of a standardized mouse model (e.g. back crossed) versus the value of using a mixed genetic background.
the relationship between the synapse, cognition and behavior; the need to investigate the link between the cognitive theories of autism and the neurobiology will require cross discipline collaboration;
the research benefits of being able to sub-divide autism into phenotypes (possibly based on cognitive endophenotypes) and whether this could be achieved via development of new selective psychometric tests. Phenotypes and psychometric tests could be adapted for use in mouse models;
animal studies in disorders related to autism (such as Rett or Fragile X syndrome) may be informative, although it is not clear if these disorders are truly the same as idiopathic autism.
the fact that the phenotype in mouse models of autism-associated genetic syndromes could be reversed or rescued even in adulthood is an incredibly unexpected and exciting finding; however, the question remains how this relates to humans and to autism
are autism and these other medical syndromes truly neurodevelopmental disorders that ‘evolve' and changes with age perhaps through accumulation of deficits and downstream impairments, or are they only “neurodevelopmental” in that they occur during early development but can otherwise be reversed without accrual of damage?

It was concluded at the end of the meeting that the objectives set for it had been achieved.