BLOG

International Child Development Resource Center

January 11, 2008

• IMFAR CONFERENCE ABSTRACT – SLIDE SESSION 3

S3.1.1
PARENTS OF CHILDREN WITH AUTISTIC DISORDER: SIGNIFICANT OBSESSIVE COMPULSIVE SYMPTOMS RK Abramson,* K.M. Wieduwilt, S.A. Ravan, K.A. Decena, H.H. Wright, and M.L. Cuccaro. Dept. Neuropsych & Behav. Sci., Univ. South Carolina Sch. Of Med., Columbia, SC 29203

Objective: Repetitive behaviors and stereotyped patterns are part of the Autism Diagnostic Inventory (ADI-R) diagnostic algorithm for Autistic Disorder (AD). This study examines OC symptoms in parents of AD children as a first step in relating OC symptoms in parents and restrictive/repetitive behaviors of the ADI-R diagnostic algorithm in AD children. Methods: The Yale-Brown Obsessive Compulsive Behavior Scale (Y-BOCS), (Goodman, 1986), a widely used self-report scale which identifies the nature and extent of OC symptoms was sent to parents of AD children participating in the Duke/USC genetic study of AD. Family history data on psychiatric illness was available for each family. Results: Family history data for 131 families at the USC site indicated that 34.4% of families had one or more first and second degree relative with probable or confirmed diagnosis of OCD and 11.5% with confirmed diagnosis. Males (n=28) represented 45.2% and females (n=34) 54.8% of Y-BOCs respondents. Clinically significant Y-BOCS scores were present in 34% of parents. The clinical Y-BOCS group differed significantly from the subclinical Y-BOCS group in the endorsement of obsession and compulsion items on the Y-BOCS. Checking compulsions and rituals were most significant (p=0.001) along with obsessions relating to sound/noises (p=0.005). These compulsions and rituals compare to Cuccaroís (2000) ADI-R derived Factor 2: problems with minor changes, compulsions and rituals in AD children. Conclusions: The family history data is similar to that reported by Bolton (1998). The high number of parents that endorsed Y-BOCS OC symptoms could be explained by which parents returned the Y-BOCS. Parents with clinical Y-BOCS scores differed significantly than parents with subclinical scores in the obsessions and compulsions that they endorsed.

S3.1.2
THE PSYCHIATRIC STATUS OF PARENTS OF INDIVIDUALS WITH AUTISM; A META ANALYTIC STUDY. N. Yirmiya*, M. Shaked, O. Izhaki, O. Duphlat and O. Erel. The Hebrew University of Jerusalem, Department of Psychology, Mount Scopus, Jerusalem, ISRAEL 91905.

The comparison between parents of individuals with autism/PDD and parents of typically developing individuals and parents of individuals with diagnoses other than autism included 75 effect sizes representing 18 independent published studies. Of the 75 effect sizes, 17 effect sizes suggest that parents of individuals with autism compared to other parents show more psychiatric difficulties, 54 effect sizes suggest no difference, and only 3 effect sizes suggest that parents of individuals with autism reveal less psychiatric difficulties. The effect sizes ranged from 1.62 to 1.75. The composite weighted mean effect size of .21 with its 95% confidence interval of .16-.26 implies that parents of individuals with autism show more psychiatric difficulties compared to parents of other individuals. According to Cohen (1977) criteria of an effect size (in the metric of d) this composite weighted mean effect size is of small magnitude. Several potential moderator variables including type of comparison group employed, gender (of parent), type of psychiatric outcomes and measures employed, and level of functioning (proband) were examined. Results indicated that “type of comparison group” (parents of individuals with Down syndrome, MR, Unknown etiology, learning disabilities, and psychiatric diagnoses) and “gender of parent” are important moderator variables. The results are discussed in terms of the genetic liability for autism.

S3.1.3
THE NEUROPSYCHOLOGICAL ENDOPHENOTYPE IN RELATIVES OF INDIVIDUALS WITH AUTISTIC DISORDER Fritz Poustka & Sven Bˆlte Department of Child and Adolescent Psychiatry, J.W.G.-University Frankfurt/M., Germany

Based on twin and family studies, there is evidence today that idiopathic autism is closely associated with hereditary factors. In addition, there is a mounting body of literature showing that autistic disorders are often accompanied by executive dysfunction, weak central coherence and problems in theory of mind. Aside from core autism and other diagnosis in the pervasive developmental disorder spectrum, there is the notion of a broader or extended phenotype of autism. Several family investigations indicate the existence of subclinical, mild expressions of the autistic condition in relatives. However, it is still considerably controversial which social, communicative and stereotypic features definitely constitute the broader phenotype. The same is true for a possible genetic liability for the neuropsychological deficits found in autism. In this ongoing study, we compare the performance on various crucial neuropsychological tests for executive function (WCST, TMT, ToH), weak central coherence (EFT, Block Design) and theory of mind (recognition of facial affect) in parents and siblings of subjects with autism (singleton & multiplex families), schizophrenia or mental retardation without autism. Preliminary results are presented and critically discussed in the light of contemporary research on the genetics of autism and the definition of the broader phenotype of the disorder.

S3.1.4
QUANTITATIVE ASSESSMENTS OF AUTISM SYMPTOMS IN PROBANDS AND FAMILY MEMBERS: BROADER PHENOTYPE AUTISM SCALE. G. Dawson, G. Schellenberg, E. Wijsman, J. Munson, A. Estes. Univ. of Washington, Seattle, WA 98195

Studies assessing the broader phenotype of autism typically have used general measures of intellectual and language ability, and qualitative measures of autism-related symptoms based on family history interviews and clinical observation. Advances in statistical genetics (i.e., quantitative trait locus analysis, QTL) have allowed researchers to use quantitative measurements of phenotype. As symptoms in autism vary along continua of severity, QTL analysis may prove to be an important tool. The present study describes the development of a rating scale of autism symptoms that can be used to quantify the broader autism phenotype in parents and siblings of children with autism (Broader Phenotype Autism Scale ≠ BPAS). The BPAS combines interview and direct observation; each proband and family member is rated on autism symptoms by a clinically trained examiner. Ratings range from severely impaired to above average functioning which adds statistical power to QTL analyses by identifying individuals highly unlikely to have the autism trait. Based on a sample of parents (N=228) and probands (N=209) in multiple-incidence families, inter-rater reliability estimates (intraclass corr.) for each of the traits ranged from .60-.91 (mean .79). Symptom domains based on theoretically derived categories demonstrated good internal consistency (mean Cronbach alpha = .81). Importantly, both probands and family members displayed a range of scores, with overlap between distributions for probands and family members. Comparisons with the Autism Family History Interview are being conducted to examine how dimensional versus categorical definitions of broader phenotype in autism correlate.

S3.1.5
Maternal age effect observed in mothers of children with autistic disorder and a de novo, maternally derived, isodicentric chromosome 15 (Idic 15) Wolpert CM, Donnelly SL, Laskowski L, Elston L, Ravan S, Decena K, Abramson RK, Wright H, Cuccaro ML, Gilbert J, Pericak-Vance MA

In 1979 Wisniewski reported a maternal age effect for inverted duplication of chromosome 15 (now known as idic 15) in 5 cases. We hypothesized that we would observe the same effect in cases of individuals with both AutD and idic 15 (N=7). Here we report the parental age findings and compare them to a larger data set of cases ascertained for an AutD genetic linkage study. All individuals enrolled in the study were ascertained through a clinical diagnosis of AutD. Qualifying individuals had their diagnosis confirmed using the ADI-R. Seven individuals from this data set had both AutD and a maternally derived, de novo, idic 15. The mean maternal age in this subset of cases is 36 (std = 3.33). In the remaining data set of individuals with AutD (n=360) the mean maternal age is 30 (std = 5.28). This difference in maternal age between the two groups is significantly different (p-value = 0.003). The mean paternal age in the idic 15 subset is 37 (std = 3.92). In the remaining data set (N=350) the mean paternal age is 32 (std = 5.75). This difference in paternal age between the two groups is significantly different (p-value = 0.0347). These data both support Wisniewskiís findings of a parental age effect as well as extend the diagnosis to include children with both AutD and isodicentric chromosome 15.

S3.1.6
CHROMOSOMES 2 & 7 IN AUTISM: PHENOTYPIC SUBTYPES S. Donnelly1, C. Wolpert1, L. Elston1, S. Ravan2, K. Decena2, R. Abramson2, H. Wright2, M. Cuccaro1, M.A. Pericak-Vance1. 1) Cent for Human Genetics, Duke Univ Med Cent, Durham, NC 2) WS Hall Psych Inst, Univ of South Carolina, Columbia, SC.

Autism (AutD) is a neurodevelopmental disorder characterized by social-communicative impairments and repetitive behaviors. Multiple genes may have etiologic relevance to AutD. Chr7 and Chr2 have demonstrated the strongest linkage when phrase speech delay (PSD) has been used to subset families. We used ordered subset analysis (OSA), a statistical method in which a covariate is used to determine a homogeneous group of families that contribute to linkage, to identify Chr2 and Chr7 family subsets. Using PSD (phrase speech onset >36 months) as a covariate, OSA identified 29 families for Chr7 and Chr2. All Chr7 families were in the Chr2 subset (n=17); a subset of Chr2 families were only linked to Chr2 (n=12). We compared the Chr2 and Chr7/2 subsets on the Autism Diagnostic Interview-Revised (ADI-R), Aberrant Behavior Checklist (ABC), and Vineland Adaptive Behavior Scales (VABS). There were no group differences in ADI-R social or repetitive behaviors or on the ABC Lethargy factor. However, the Chr2 group had significantly higher scores in VABS social (p=.0004), communication, (p=.019), daily living skills (p=.0006) domains and overall adaptive functioning (p=.0004). The Chr2 group had lower scores on ADI-R measures: gestures (p=.003) and social imitative behaviors (p=.001). These findings suggest that 1) Chr2 and Chr7/2, despite being stratified on PSD, are clinically differentóeven in communicative function. 2) The Chr7/2 subset is more impaired than the individuals in the Chr2 subset alone. 3) Chr2 individuals may show a different trajectory of language and communication development. S3.2.1STABILIZATION OF PHOSPHOLIPID MEMBRANE ARCHITECTURE IN AUTISTIC SPECTRUM DISORDER P.C. KANE, PH.D., J.S. FOSTER, M.D.*, A. CARTAXO, M.D. WellSpring Clinic, 250 W. Lancaster Ave., Wayne, PA 19087 and Thomas Jefferson University, Philadelphia, PA.

Red cell lipid analysis of 500 pediatric patients with ASD to assess fatty acid depletion or imbalance in relation to severity of presentation with application of oral supplementation. Red Cell Lipid analysis was performed through Kennedy Krieger Institute Peroxisomal Diseases Laboratory in Baltimore, MD. Laboratory studies revealed moderate to severe lipid abnormalities in all subjects involving imbalance/deficiency of essential fatty acids, elevation of renegade fatty acids as saturated very long chain and odd chain lipids, peroxisomal disturbance, suppression of DMAs as myelination markers, compromised membrane integrity, and low total red cell lipid content. Repletion of essential lipids, supplementation of catalysts and control of phospholipase A2 through a modified ketogenic diet, has yielded positive outcomes of our patient population in two satellite clinics.

S3.2.2
A VOXEL-BASED MORPHOMETRY STUDY OF GREY MATTER IN AUTISM. A.H. Hardan*, O. Yorbik, N. J. Minshew, V.A. Diwadkar, and M.S. Keshavan Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA 15213.

Voxel-based morphometry (VBM) permits the exploration of differences in gray matter density at a microstructural level using statistical parametric mapping analyses. In this study, VBM was applied to MRI scans obtained from 28 non-mentally retarded autistic male children and adults and 28 individually matched, healthy community volunteers, to identify gray matter differences between the two groups. Compared to controls, autistic subjects demonstrated increased gray matter density in the following regions: left inferior parietal lobule, right frontal lobe sub-gyral, right anterior cingulate, left cingulate gyrus, and left superior frontal gyrus. In contrast, gray matter density was decreased in the autistic subjects in the right insula, left precentral gyrus, left middle frontal gyrus, left superior temporal gyrus, left precuneus, right middle frontal gyrus, amygdala, and right cuneus. No significant increase in gray matter density was observed with age in the two groups. However, decrease in gray matter density was noted with age in the left insula, right postcentral gyrus, and left superior frontal gyrus in the autistic group, and in the left frontal gyrus, left precuneus, right lingual gyrus (occipital lobe) and the left middle temporal gyrus in the control group. Results from this study are consistent with previous neuroimaging studies, but more importantly, they indicate the existence of gray matter differences in brain regions where previous anatomic studies have failed to show any abnormalities.

S2.3.3
WHITE MATTER INCREASES IN AUTISM ARE LARGELY IN SUPERFICIAL RADIATE REGIONS. M.R. Herbert*, D.A. Ziegler, N. Makris, H.A. Sanders, J.J. Normandin, C. Deutsch, D.N. Kennedy, V.S. Caviness. Center Morphometric Anal, MGH, Charlestown, MA 02130 & Shriver Center, Waltham, MA 02254.

PURPOSE: To assess the regional distribution of white matter (WM) volume increases in autistic brains METHOD: 1. MRI-based whole brain gray-white segmentation and cortical parcellation; 2. Algorithmic parcellation of cerebral WM into radiate, sagittal and bridging p compartments; 3. Analysis of volumes and of volumes adjusted (covaried) for total cerebral volume. SUBJECTS: 32 school-aged boys (17 non-retarded autistic; 15 control). RESULTS: 1) Volumes: Frontal, occipital and temporal lobe radiate WM and fornix were significantly larger in the autistic group; basal forebrain was significntly smaller. 2) Adjusted volumes: Radiate frontal and occipital lobe WM was significantly larger while internal capsule and corpus callosum were significantly smaller in the autistic group. CONCLUSIONS: WM volume increases in autism are concentrated in radiate or superficial white matter, which accounts for about æ of WM volume. In bridging regions fornix was somewhat larger, while compartments containing interhemispheric connections, corticothalamic tracts and multiple other corticofugal connections were relatively smaller. IMPLICATIONS: Volume increases are largely in frontal, temporal and occipital sub-cortical U-fibers and corona radiata, which include short-cortico-cortical connections as well as links to deeper tracts. Areas containing longer cortico-cortical connections are relatively smaller in the autistic brain. This may be consistent with hyper-focus and context-insensitivity seen in autism.

S2.3.4
REGION-SPECIFIC CEREBELLAR ABNORMALITIES IN AUTISM: PROTEIN AND GENE EXPRESSION. E. M. Sajdel-Sulkowska, Dept. Psychiatry, Harvard Medical School; Brigham and Womenís Hospital, Boston, MA 02115.

Histological evidence suggests region-specific abnormalities in the cerebella of autistic individuals affeccting neocerebellar lobules V;X. To address the causative mechanisms of these region-specific differences, present studies compared the expression of selected gene products in cerebellar samples derived from the vermal (section 1) and neocerebellar region (section 9) of three autistic (age: 9, 16.5, 34 years, postmortem interval (pmi): 12, 21, 19 hours) and three age- and pmi-matched controls. RNA extracted from frozen cerebellar samples was analyzed by RPA using 32 P-labeled riboprobes. Proteins were analyzed by quantitative western blots of lysates prepared from the same cerebellar samples. Decreased synaptophysin gene expression of more than 30% was most consistently observed in autistic cerebella. The decrease was more pronounced in the neocerebellum (over 40%)than in vermis. Such a decrease may be consistent with reduced synaptic density. These results suggest that altered gene expression in the neocerebellum may contribute to the histological abnormalities observed in this region. A more extensive gene analysis, now in progress, addresses the molecular mechanisms of cerebellar abnormalities in autism. We thank ATP and specifically the U. Maryland Brain and Tissue Bank for providing the cerebellar samples.

S2.3.5
SIMILAR BRAIN ASYMMETRIES IN BOYS WITH AUTISM AND DEVELOPMETNAL LANGUAGE DISORDER. D.A.Ziegler*, M.R. Herbert, J.M. Goldstein, N. Makris, D.N. Kennedy, V.S. Caviness. Cntr Morphometric Anal, MGH, Charlestown, MA 02130 & Mass Ment Health Cntr, Boston, MA 02115.

Subgroups of children with developmental language disorder (DLD) and autism (AUT) often show similarities in language impairment, while AUT is further defined by abnormal socialization and repetitive behaviors. Here, we compare these groups in previously implicated brain regions. SUBJECTS: 48 non-retarded school-aged boys (17 AUT; 16 DLD; 15 control (CTR)). METHODS: 1) Anatomically-based MRI gray-white matter segmentation of major cortical/subcortical brain structures 2) neocortical gyral-based parcellation 3) calculation of volumes and symmetry indices [SI = 2(L – R)/(L + R)] for 17 brain regions (cortical: frontal, temporal, parietal cortex; subcortical: basal ganglia, hippocampus and amygdala), that are associated with language, affective arousal, or repetitive behaviors. RESULTS: AUT and DLD differed from CTR, but were similar to each other, in asymmetries in brain regions associated with language and repetitive behaviors. Among other differences, they showed a loss of normal leftward asymmetry in Broca’s area, but exaggerated leftward asymmetry in planum temporale. No significant differences were found between the three groups in asymmetries of regions associated with affective arousal. CONCLUSIONS: AUT and DLD boys appear to vary systematically and similarly from CTR in terms of brain asymmetry. These data, along with our recent report of increased white matter in both AUT and DLD, suggest that networks associated with language and repetitive behavior may have similarly altered circuitry in both disorders.

S2.3.6
SYNAPSES AND PURKINJE CELLS IN CEREBELLUM IN AUTISM. G. Y. Wen and J. Chen. Laboratory of Ultrastructure and Immunohistology, Institute for Basic Research in Developmental Disabilities, Staten Island, NY 10314.

Autism is a neurodevelopmental disorder that results in severe impairments in language, cognition, and communication. Although this syndrome has been recognized for over half a century, its etiology is still unknown. Limited neuropathological studies have been carried out on autism brain. This study determined the level of reduction in Purkinje cells (PCs) and correlated this level with the alteration in synaptic density in cerebellum in autism. Approximately 1-mm-thick slices of formalin-fixed cerebellum were obtained from four autism and four control cases, rinsed in phosphate-buffered saline (PBS); and further fixed with 4% glutaraldehyde in PBS. Proceeding with standard protocols of electron microscopy (EM), we made tissue blocks in Spurr medium. Sections at a thickness of 0.3 mm were cut from these tissue blocks and stained with 1% toluidine blue for quantitative analysis by light microscopy of linear PC density/mm in the cerebellum. Ultrathin sections at a thickness of 80 nm were stained with uranyl acetate and lead citrate for EM examination to quantify the synaptic density/mm2. On average, there was a 22% reduction (n = 8) of linear PC density/mm in the cerebellum in autism. However, there was a 39% gain (n = 2) in synaptic density/mm2 in the upper region of the cerebellar molecular layer, on the basis of one 9-year-old-male with autism and one 7-year-old normal control male. Results from this preliminary study indicated a possible inverse relationship between reduction in PC density and synaptic density in the cerebellum of individuals with autism. S3.3.1Autism and vaccines; no association with thiomersal; no evidence of immune suppression following MMR. Brent Taylor(1), Elizabeth Miller(2), Nick Andrews(2). (1)University College London and (2)Centre for Disease Surveillance Control, Public Health Laboratory Service, London UK

Vaccine safety is of great concern to parents, to health professionals and to politicians. A rise in cases of autism has been recorded in the USA and the UK, although our recent study (IMFAR 2001) suggested that rates have levelled off since 1992. Never-the-less the apparent rise, at least in a susceptible subgroup, has been blamed on MMR vaccine, which is hypothesized to cause ‘new-variant’ autism with developmental regression. MMR vaccine is believed to overwhelm the immature immune system, rendering the child susceptible to persistent measles virus infection, leading to chronic disease such as bowel problems and autism. The mercury-containing preservative, thiomersal, used in some vaccines (diphtheria, pertussis, tetanus Hepatitis B and influenza) has also been incriminated as a cause or primer of autism and other neurodevelopmental problems. We have developed a system which enables rapid assessment of such possible causal associations, by linking individual hospital or primary care diagnoses to population-based immunisation records. MMR vaccine did not increase the risk of hospitalisation with invasive bacterial infection in the 3 months after vaccination; rather there was a protective effect. These results provide no support for the concept of “immunological overload” induced by multiple-antigen vaccinations, nor calls for single antigen vaccines. Analysis of data on over 100,000 children in the UK general practice research data-base showed no relationship between culmulative dosage of thiomersal and the development of autism, communication problems, behaviour problems or other developmental disorders. A consistent trend was towards more such problems in unvaccinated children and least in children fully vaccinated on schedule. These findings have been independently confirmed by analysis of data from ALSPAC, a detailed longitudinal birth cohort study around Bristol UK. These results should provide further reassurance to families concerned that they might have caused their child’s autism by complying with recommended vaccination programmes. They also confirm the usefulness of active surveillance systems for investigating possible adverse events following vaccination.

S3.3.2
Acting as Medical Expert in litigation A. WAKEFIELD

We have described a temporal association between MMR exposure and onset of regressive autism plus a novel entero-colitis. This study examined the possibility of a dose-response effect for an MCV. Patients & Methods: ìExposedî -; children with normal early development & regressive autism who had received more than one MCV (N=28) – were compared with age and sex matched ìunexposedî – children with normal early development & with regressive autism who had received only one MCV (N=28) but otherwise similar baseline characteristics to the exposed group. Comparisons included: secondary (2o) developmental/behavioural regression; 2o physical deterioration, prospective, observer-blinded scores of endoscopic & histological disease severity, and growth trajectories. Results: Exposed scored significantly higher than unexposed for: 2o regression on the basis of (i) parental history (relative risk [RR] 20.00 (2.89-138.26) p S3.3.3
THIMEROSAL AND MERCURY POTENTLY INTERFERE WITH REGULATION OF METHIONINE SYNTHASE BY INSULIN-LIKE GROWTH FACTOR-1. R.C. Deth* and M. Waly. Dept. Pharmaceutical Sciences, Northeastern Univ., Boston, MA 02115.

Thimerosal is implicated as a potential cause of autism, but a specific target that is relevant to autism has yet to be identified. In human neuroblastoma cells, we have shown that folate-dependent, methionine synthase-mediated homocysteine (HCY) methylation is stimulated by insulin-like growth factor-1 (IGF-1). Via this action IGF-1 lowers the concentrations of HCY and S-adenosylmethionine, a powerful inhibitor of DNA methylation, formed from HCY and adenosine. Abnormalities involving folate and adenosine metabolism have been previously linked to autism and Rett Syndrome is caused by mutations affeecting DNA methylation events Moreover, dopamine-stimulated folate-dependent phospholipid methylation (PLM) has been proposed to be essential in attention. IGF-1-stimulation required Cu2+, evidenced by the loss of activity after pretreatment with penicillamine. Cu+, Zn2+, Cr3+ and Hg2+ selectively inhibited IGF-1-stimulated activity. Studies with Hg2+ revealed potent inhibition, with a threshold at 1 nM and complete loss of IGF-1 activity at 10 mM. Presence of Cu2+ (10 mM) abolished the effect of Hg2+. Thimerosal was more than two orders of magnitude more potent than Hg2+, with a threshold of 10 pM and an IC50 of 1 nM. Our data indicate extreme sensitivity of this IGF-1 response to mercury and thimerosal. This action of IGF-1 is critical for both DNA methylation-dependent development and for attention, which are impaired in autism. We propose that mercury and thimerosal contribute to autism by interfering with this IGF-1 mechanism. Deficits in related pathways may predispose particular individuals to the risk associated with thimerosal exposure.

S3.3.4
MATERNAL INFLUENZA VIRUS INFECTION CAUSES A CEREBELLAR DEFECT RESEMBLING THAT IN AUTISM. P. H. Patterson* and L. Shi, Biology Division, California Institute of Technology, Pasadena, CA 91125.

Maternal viral infection is implicated as a risk factor in autism. In developing an animal model based on this risk factor, we found that respiratory infection with influenza virus of pregnant mice at mid-gestation leads to behavioral abnormalities in the adult offspring. These include deficits in open field and novel object exploration, prepulse inhibition and social interaction. We now report that these adult mice also display a type of neuropathology repeatedly found in autism ñ a deficit in Purkinje cells, specifically in lobule VII of the cerebellar vermis. We also occasionally see misplaced Purkinje cells in the white matter of this lobule. In addition, the molecular layer of lobule VII appears to be thinner in these mice. This pathology may be related to the behavior of these mice, as Pierce and Courchesne (2001) have shown a strong correlation between the amount of deficit in vermis lobules VI-VII and the lack of exploration of novel objects in autistic subjects. Conversely, lobules VI-VII are enlarged in Williams syndrome, where subjects display hypersocial behavior and excessive linguistic affect (Schmitt et al., 2001).

S3.3.5
ABNORMAL THROAT AND GUT FLORA IN CHILDREN WITH REGRESSIVE AUTISM. (1)Rosseneu SLM, (2)van Saene HKF, (1)Heusckel R,(1)Murch SH. (1)Centre of Paediatric Gastroenterology, Royal Free University Hospital, London, United Kingdom, (2)Department of Medical Microbiology, University of Liverpool, United Kingdom.

Background: Recent data including non-specific entero-colitis and complex immuno-mediated gastrointestinal pathology suggest that the gut may play a role in regressive autism. Chronic underlying disease influences gut flora and promotes carriage of abnormal aerobic Gram-negative bacilli (AGNB) including Enterobacter and Pseudomonas species. Study objectives: to asess whether children with regressive autism carry AGNB in throat and gut. Patients: between April 1st and mid-June 2002, all children with gastrointestinal symptoms and diagnosed with regressive autism were consecutively enrolled in this study. None of the children were treated with antibiotics in the 6 weeks prior to enrolment. Measurements and results: throat and rectal swabs were obtained in this single survey culture study. All swabs were processed using the 4-quadrant method combined with enrichment broth to detect low level carriage. Of 53 children, 27 (51%) were positive for abnormal AGNB, 19 and 14 at oropharyngeal and rectal level, respectively. The 2 main microbes isolated were Enterobacter and Citrobacter. 30 (58%) and 19 (36%) patients were carriers of sensitive Staphylococcus aureus and Candida albicans, respectively. Both AGNB and S.aureus were present in overgrowth. Conclusion: this single sample survey shows that half of the children with gastrointestinal symptoms and diagnosed with regressive autism carried abnormal gut flora. A multiple culture survey is underway to confirm persistent carriage.

S3.3.6
SUSCEPTIBILITY OF MICE TO DISTURBANCES OF BEHAVIOR AND BRAIN ARCHITECTURE FOLLOWING POSTNATAL THIMEROSAL EXPOSURE PARALLELS STRAIN SENSITIVITY TO AUTOIMMUNE DISEASE. M. Hornig*, D. Chian, and W.I. Lipkin. Center for Immunopathogenesis and Infectious Diseases, Mailman School of Public Health, Columbia University, New York, NY 10032.

Increases in reported cases of autism spectrum disorders are ascribed to broader definition of the disorder, altered case ascertainment, or pre- or postnatal exposure to environmental factors such as microbes, stress, or toxins. Specific concern is focused on the possibility that ethylmercury in the vaccine preservative, thimerosal, results in aberrant brain development and behavioral features reminiscent of autism in immature, susceptible hosts. Early postnatal administration of thimerosal using doses and timing that mimic the childhood immunization schedule induces mouse strain-specific effects on weight gain, locomotor and exploratory activity, stereotypic behaviors, and size of CA regions of hippocampus. SJL/J mice, a strain with heightened sensitivity to autoimmune disease, show the most prominent behavioral and neuropathologic effects. In this strain, male gender is associated with a more severe outcome. C57BL/6J mice demonstrate an intermediate phenotype, and BALB/cJ mice have minimal deficits. These findings suggest that brain architecture and function may be altered in genetically susceptible hosts following postnatal thimerosal exposure, and support the utility and relevance of this model as a tool for identifying genetic and maturational factors underlying vulnerability to toxin-induced CNS injury and understanding the pathogenesis of human neurodevelopmental disorders. S3.4.1A SOCIAL ADJUSTMENT ENHANCEMENT INTERVENTION FOR AUTISM SPECTRUM DISORDERS. M. Solomon* and B. L. Goodlin-Jones. M.I.N.D. Institute, U.C. Davis Health System, Sacramento, CA 95817.

Treatment plans for children with autism spectrum disorders (ASDs) frequently recommend participation in a social skills training group. This presentation reports the findings of a twenty week intervention using cognitive behavioral and psychodynamic techniques to enhance the social adjustment of high functioning children with ASDs. This social skills curriculum was designed to address three areas hypothesized to be deficient in persons with ASDs including emotion recognition and understanding, theory of mind, and real-life-type problem solving. The program also included a weekly parent psychoeducational meeting. Twenty boys aged 8 to 12 with diagnoses of high functioning autism, Asperger Syndrome, and PDD-NOS were recruited to participate in the study. Children were matched based on age and level of cognitive function. One member of each matched pair was then randomly selected to receive intervention (a social skills group) first or to serve as waiting list control group subject. After 20 weeks, there was a crossover of the two groups, and children serving as waiting list control subjects first were invited to participate in a social skills group. Modest improvements in facial expression recognition, some measures of theory of mind, and real-life-type problem solving were reported for intervention group subjects. Parent and child depression scores also decreased marginally while children were receiving intervention, and there were trend level improvements in parent’s assessments of their own competency when handling problem behaviors in their children. Results are discussed in the context of the effects of providing comprehensive and collaborative services to families of children with ASDs, and to the potential role of anxiety in influencing child outcomes.

S3.4.2
PROPOSED USE OF TWO-PART INTERACTIVE MODELING AS A MEANS TO ENGENDER EMPATHY IN CHILDREN WITH AUTISTIC SPECTRUM DISORDER. I.M. Pepperberg and D. Sherman. Psychology Dept, Brandeis University, Waltham, MA 02454; MIT Media Lab, Cambridge, MA 02139; New-Found Therapies, Monterey, CA 93940.

Many behavior modification/intervention programs are based on operant procedures developed for animal subjects, but few use modeling procedures in which one student observes interactions between two proficient trainers. We show how such procedures, which successfully trained Grey parrots (Psittacus erithacus) to produce and comprehend elements of human language, can be adapted for use with children with autistic spectrum disorders to engenger empathetic interactions.

S3.4.3
AGE AND NUMBER OF DOSES ARE DETERMINANTS OF THE RESPONSE TO SECRETIN IN YOUNG CHILDREN WITH AUTISM. W. Herlihy, J. Rusche, P. Rioux, and the Secretin Phase 2 Clinical Trial Group. Repligen Corp., Waltham, MA 02453.

There has been limited success with pharmacologic interventions in autism and no agent has been shown to improve the core deficits in reciprocal social interaction. We have completed a placebo controlled, double-blind study of three doses of human synthetic secretin in 126 autistic children aged 3 years to 6 years 11 months. Children were treated at three week intervals with either 2 CU/kg of secretin or a placebo and assessed at baseline and two weeks after the third treatment. A primary determinant of response was patient age with 3 and 4 year old patients responding more robustly to secretin vs. placebo than older patients. The 3 and 4 year olds showed a significant improvement in ADOS reciprocal social interaction scores (p S3.4.4
VIDEO MODELING TO IMPROVE THE PERCEPTION OF EMOTION IN CHILDREN WITH AUTISM. B.A. Corbett and E.V. Larsson. M.I.N.D. Institute, University of California, Davis, Sacramento, CA 95817 and Families for Effective Autism Treatment of Minnesota, Minneapolis, MN 55455.

Video modeling utilizes the medium of video to teach a variety of skills to participants who watch and then imitate behaviors performed in short, videotaped scenes. Video modeling is based in behavioral principles and preliminary results suggest that the procedures can assist the child to become a better observational learner. We have used these procedures to facilitate language acquisition, adaptive skills, play behavior, and social skills. The current investigation will extend our work by showing case studies using video modeling procedures to improve the perception of emotion in children with autism spectrum disorders. The treatment includes the participant being exposed to video taped scenes of typically developing peers engaged in a variety of age-appropriate social and play scenes in which they display basic emotions (happy, sad, angry, fearful). Thus, the primary aim of the video modeling treatment is to test the hypothesis that children with autism spectrum disorders, who demonstrate impairment in the processing of emotional and social information, can learn to identify, interpret and use emotional information in a socially appropriate, flexible and adaptive way. Preliminary results will be presented along with videotaped examples of the intervention.

S3.4.5
SYSTEMATIC ANALYSIS OF SOCIAL INITIATIVE POTENTIAL (SIP) AS A PIVOTAL VARIABLE IN PROGNOSIS FOR CHILDREN WITH AUTISM. L.A. Vismara, R.L. Koegel* and L.K. Koegel. Autism Research and Training Center, University of California at Santa Barbara, Santa Barbara, CA 93106-9490.

Recent advances in research, concerning the identification of prognostic indicators, suggest that more positive and long-term gains are associated with children who display specific behavioral characteristics. In order to determine whether certain variables are associated with good prognoses for children with autism, parents and children with autism participated in parent education programs at University of California, Santa Barbara. All children received pivotal response treatment based on applied behavior analysis (ABA) procedures in a Pivotal Response Teaching (PRT) format. Such approaches focus on the pivotal area of motivation to increase engagement and responsiveness in social communication interaction and joint attention and provide intervention in key pivotal areas that appear to lead to widespread positive effects on many other behaviors. While some of the children achieved particularly excellent outcomes, others showed much smaller improvements. According to the preliminary data, the differences in outcome might be related to the presence of specific prognostic indicators present at pre-intervention that are associated with exceptionally favorable outcomes. Specifically, if the child engaged in any behaviors focused primarily on obtaining social reinforcement, such as exhibiting a gesture, behavior, or communicative attempt towards the caregiver for the purposes of social reinforcement, as opposed to requesting an object primarily to obtain the object, this variable, coined social initiative potential (SIP), might be associated with children attaining promising outcomes. Therefore, in this study, social initiative potential will be systematically investigated through a programmatic line of research to determine whether it might be influential in predicting why some children with autism attain extremely favorable outcomes and others exhibit poor outcomes despite intervention efforts. Equally important, this study will attempt to examine whether children with autism might be taught to exhibit social initiative potential in order to improve their intervention outcomes.

S3.4.6
DYSFUNCTION OF DOPAMINERGIC SYSTEM IN INFANTILE AUTISM: DOUBLE-BLIND PLACEBO-CONTROLLED PET STUDY WITH TETRAHYDROBIOPTERIN TREATMENT Y. Watanabe, K. Takahashi, R. Torstenson, T. Danfors, O. Eeg-Olofsson, A-L. von-Knorring, R. Moulder, H. Engler, P. Hartvig, and B. Langstrom Department of Physiology, Osaka City University Graduate School of Medicine, Osaka 545-8585, Japan, Uppsala University PET Centre, Department of Child Neurology & Psychiatry, Uppsala University, UAS, S-751 85 Uppsala, Sweden.

Eighteen children with autistic disorder (AD) were recruited for treatment with 6R-L-tetrahydrobiopterin (BH4), a cofactor for tyrosine and tryptophan hydroxylases, and also a cofactor for nitric oxide synthase. A considerable portion of children with AD showed a low level of BH4 in their CSF. Inclusion criteria (diagnosed with DSM-IV) were a relatively low level of BH4 in their CSF, no asthma, no epilepsy, and between 3 and 7 years of age. For clinical evaluation, the Childhood Autism Rating Scale (CARS) was mainly used every third month. The cross-over double-blind manner was applied: three groups were medicated with BH4 (3 mg/kg/day, per os) each for six months during the different time period. PET studies, using three radiotracers L-[11C]DOPA, N [11C]methylspiperone (NMS), [15O]H2O, for measuring dopamine turnover, dopamine D2 receptor binding and regional cerebral blood flow, respetively, were performed at three time-points, before and sixth and twelve months after treatment. Dopamine turnover in the anterior cingulate gyrus and thalamus significantly increased and NMS binding decreased in the thalamus in the treatment group compared with placebo group. Dopamine turnover in the amygdala correlated negatively with CARS. These findings suggest that some dysfunction of the dopaminergic system can be related to the pathophysiology of the infantile autism.

SLIDE SESSION 3.2: NEUROANATOMY

S3.2.1
STABILIZATION OF PHOSPHOLIPID MEMBRANE ARCHITECTURE IN AUTISTIC SPECTRUM DISORDER P.C. KANE, PH.D., J.S. FOSTER, M.D.*, A. CARTAXO, M.D. WellSpring Clinic, 250 W. Lancaster Ave., Wayne, PA 19087 and Thomas Jefferson University, Philadelphia, PA.

PURPOSE: To assess the regional distribution of white matter (WM) volume increases in autistic brains METHOD: 1. MRI-based whole brain gray-white segmentation and cortical parcellation; 2. Algorithmic parcellation of cerebral WM into radiate, sagittal and bridging p compartments; 3. Analysis of volumes and of volumes adjusted (covaried) for total cerebral volume. SUBJECTS: 32 school-aged boys (17 non-retarded autistic; 15 control). RESULTS: 1) Volumes: Frontal, occipital and temporal lobe radiate WM and fornix were significantly larger in the autistic group; basal forebrain was significntly smaller. 2) Adjusted volumes: Radiate frontal and occipital lobe WM was significantly larger while internal capsule and corpus callosum were significantly smaller in the autistic group. CONCLUSIONS: WM volume increases in autism are concentrated in radiate or superficial white matter, which accounts for about 3/4 of WM volume. In bridging regions fornix was somewhat larger, while compartments containing interhemispheric connections, corticothalamic tracts and multiple other corticofugal connections were relatively smaller. IMPLICATIONS: Volume increases are largely in frontal, temporal and occipital sub-cortical U-fibers and corona radiata, which include short-cortico-cortical connections as well as links to deeper tracts. Areas containing longer cortico-cortical connections are relatively smaller in the autistic brain. This may be consistent with hyper-focus and context-insensitivity seen in autism.

SLIDE SESSION 3.3: TERATOGENS & ETIOLOGY

S3.3.1
AUTISM AND VACCINES; NO ASSOCIATION WITH THIOMERSAL; NO EVIDENCE OF IMMUNE SUPPRESSION FOLLOWING MMR. Brent Taylor(1), Elizabeth Miller(2), Nick Andrews(2). (1)University College London and (2)Centre for Disease Surveillance Control, Public Health Laboratory Service, London UK

S3.3.2
ACTING AS MEDICAL EXPERT IN LITIGATION A. WAKEFIELD

We have described a temporal association between MMR exposure and onset of regressive autism plus a novel entero-colitis. This study examined the possibility of a dose-response effect for an MCV. Patients & Methods: “Exposed” -; children with normal early development & regressive autism who had received more than one MCV (N=28) – were compared with age and sex matched “unexposed” – children with normal early development & with regressive autism who had received only one MCV (N=28) but otherwise similar baseline characteristics to the exposed group. Comparisons included: secondary (2o) developmental/behavioural regression; 2o physical deterioration, prospective, observer-blinded scores of endoscopic & histological disease severity, and growth trajectories. Results: Exposed scored significantly higher than unexposed for: 2o regression on the basis of (i) parental history (relative risk [RR] 20.00 (2.89-138.26) p S3.3.3
THIMEROSAL AND MERCURY POTENTLY INTERFERE WITH REGULATION OF METHIONINE SYNTHASE BY INSULIN-LIKE GROWTH FACTOR-1. R.C. Deth* and M. Waly. Dept. Pharmaceutical Sciences, Northeastern Univ., Boston, MA 02115.

S3.3.4
MATERNAL INFLUENZA VIRUS INFECTION CAUSES A CEREBELLAR DEFECT RESEMBLING THAT IN AUTISM. P. H. Patterson* and L. Shi, Biology Division, California Institute of Technology, Pasadena, CA 91125.

Maternal viral infection is implicated as a risk factor in autism. In developing an animal model based on this risk factor, we found that respiratory infection with influenza virus of pregnant mice at mid-gestation leads to behavioral abnormalities in the adult offspring. These include deficits in open field and novel object exploration, prepulse inhibition and social interaction. We now report that these adult mice also display a type of neuropathology repeatedly found in autism – a deficit in Purkinje cells, specifically in lobule VII of the cerebellar vermis. We also occasionally see misplaced Purkinje cells in the white matter of this lobule. In addition, the molecular layer of lobule VII appears to be thinner in these mice. This pathology may be related to the behavior of these mice, as Pierce and Courchesne (2001) have shown a strong correlation between the amount of deficit in vermis lobules VI-VII and the lack of exploration of novel objects in autistic subjects. Conversely, lobules VI-VII are enlarged in Williams syndrome, where subjects display hypersocial behavior and excessive linguistic affect (Schmitt et al., 2001).

SLIDE SESSION 3.4: TREATMENT & OUTCOMES 2

S3.4.1
A SOCIAL ADJUSTMENT ENHANCEMENT INTERVENTION FOR AUTISM SPECTRUM DISORDERS. M. Solomon* and B. L. Goodlin-Jones. M.I.N.D. Institute, U.C. Davis Health System, Sacramento, CA 95817.

-TOP OF PAGE-