2004 Progress Report: Environmental Factors in the Etiology of Autism; Molecular Biomakers CoreEPA Grant Number: R829388C003
Subproject: this is subproject number 003 , established and managed by the Center Director under grant R829388
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
Center: CECEHDPR - University of California at Davis Center for the Study of Environmental Factors in the Etiology of Autism
Center Director: Pessah, Isaac N.
Title: Environmental Factors in the Etiology of Autism; Molecular Biomakers Core
Investigators: Hagerman, Paul , Denison, Michael , Gregg, Jeffrey , Rocke, David
Current Investigators: Hagerman, Paul , Gregg, Jeffrey , Sharp, Frank
Institution: University of California - Davis
EPA Project Officer: Hahn, Intaek
Project Period: September 30, 2001 through September 29, 2002
Project Period Covered by this Report: September 30, 2003 through September 29, 2004
RFA: Centers for Children's Environmental Health and Disease Prevention Research (2001) RFA Text | Recipients Lists
Research Category: Health , Children's Health , Health Effects
The principal objective of the Molecular Biomarkers Core is to identify patterns of altered gene expression that form a significant association with autism in human populations, or which are coupled to specific environmental factors in animal models. During this period the core has established a microarray facility that utilizes Affymetrix GeneChip technology. An essential component of the Core’s current efforts are to develop methods of analysis for microarray data in an effort to both better define, and reduce, the errors associated with the interpretation of microarray results (Geller et al 2002, in preparation).
The principal effort of the CORE over the past year has continued to focus on microarray analysis related to human samples, and on the development of better tools for analysis of microarray data. Blood samples are being provided mainly from two sources at present: samples from Project I (CHARGE samples), both directly to Core III and through Project III/Core II, and from the M.I.N.D. Institute clinic for fragile X samples. Total RNA extracted from blood samples are banked for batch analysis by microarray. During the course of the past year, RNA extraction methods have changed, from TRIZOL to PAX-based technologies, in an effort to improve the stability of the RNA samples, a particularly significant issue due to variations in the time of transport of bloods. We have also initiated a series of single-gene polymorphism analyses for a number of genes with possible involvement in autism. Finally, methods of analysis of the microarray data continue to be improved.
Utilization of lymphoblastoid cell lines (CAN/AGRE) from families with autism to evaluation of the utility of blood-based microarray discovery in autism - To provide a basic foundation for the “blood genomics” approach in Aim 2, we performed statistical analysis of gene expression profiles using lymphoblastoid cell lines derived from children with autism and their families. The goal was to assess the feasibility of using lymphoblastoid cell lines in identifying genes associated with autism. Replicate experiments using Affymetrix GeneChip arrays demonstrated that gene expression data from the cell lines was consistent and highly reproducible. We then performed further gene expression studies and identified differentially expressed genes between cell lines derived from children with autism and cell lines derived from their normally developing siblings. Using these identified genes (n=48), we were able to cluster the two groups into their respective phenotype classes. In addition, the identified genes were utilized in order to identify potential pathways that may be involved in autism. Our approach provides insights into the potential use of lymphoblastoid cell lines and by implication “blood genomics” as a viable tool in identifying genes associated with autism. This work is being submitted for publication.
Association of autism with fragile X syndrome provides an approach that utilizes the fragile X gene (FMR1) to identify additional, epistatic genes that give rise to the autism phenotype - Fragile X syndrome (FXS) is a single-gene disorder with a strong association with autism (AUT). Approximately 4 to 6% of children with autism have fragile X syndrome, and approximately 15 to 33% of children with FXS have autism. FXS may thus hold clues to the identification of a small number of “second” genes that act in a synergistic fashion with the FMR1 gene to give rise to the autism phenotype. Our working hypothesis is therefore that one or more genes, from a pool of autism susceptibility genes, are responsible for the autism phenotype with FXS. To test this hypothesis, we are performing microarray analysis on peripheral blood leucocytes from children with FXS, divided into two groups, depending on their autism status (FXS ± AUT).
Thus far, the gene expression profiles of fourteen patients (5 patients with FXS+AUT and 9 patients with FXS) have been analyzed. Cognitive assessments and ADOS-G were performed for each case to confirm AUT. The data were transformed and normalized as described by Geller et al (2003) (see: Aim 5). A two-sample t-test was performed between the groups for each probe and probeset. A method proposed by Storey was employed that simultaneously controlled the false discovery rate (FDR) and maintained a high power.
Using a probeset-level analysis, 22,283 probe sets were tested (U133A) and 36 differentially expressed transcripts were identified after FDR corrections. However, using a probe-level analysis, 247,965 probes were tested resulting in 648 differentially expressed probes. There were 10 transcripts that had differentially expressed probes and were differentially expressed at the probeset level. These include: DC12 protein, NADH dehydrogenase (ubiquinone) Fe-S protein 3, PFTAIRE protein kinase 1, T cell receptor delta locus (identified twice), GNAS complex locus, KIAA1194, p53 regulated PA26 nuclear protein, and hypothetical protein MGC29761. Each of the differentially expressed transcripts will be separately validated/rejected using quantitative RT-PCR.
We are also utilizing molecular pathway analysis tools to associate differentially expressed transcripts and to explore relationships for the development of hypotheses that can be tested using cell biological techniques. As an example, we have used the Ingenuity Pathway tool to identify affected molecular pathways. Of the 648 differentially expressed probes, 603 probes mapped to known genes and 279 were found to be focus genes for network generation. Using an analysis of 1.3 million manually curated molecular interactions; the analysis tool identified 61 networks that contain these focus genes. These methods will be pursued as we collect larger patient datasets. We found a significant number of differentially expressed transcripts that are involved in neuronal migration. As an example, it was reported that in cerebral cortex, a protein-protein complex consisting of mouse Integrin alpha 3 [Itga3] and mouse Integrin beta1 [Itgb1] is necessary for the migration of neurons that involves REELIN [RELN] protein (Dulabon et al., 2000).
Finally, the analysis tool allows the interrogation of specific cellular signaling pathways to find associated differentially expressed genes within each of them. This can be very valuable when designing experiments to validate the role of each differentially expressed gene for functional outcomes of a particular difference. An understanding of the alterations in gene expression in blood samples from children with autism and fragile X syndrome, in terms of cellular signaling, will provide the best correlation to neuronal cell biology. These correlations can then be tested using cellular models of neuronal development. A particularly interesting signaling pathway thought to be important for neuronal development is the growth factor PI3kinase/AKT pathway. We have identified multiple transcripts involved in this pathway. Interestingly, the major differences appear to most substantially affect the WNT pathway.
We are currently replicating the preliminary results described above using an additional group of 30 subjects. Recruitment is continuing for these very interesting patients. Our current goal is to study a total of 50 FXS patients and 50 with both FXS and autism.
CHARGE samples under Project I – We are receiving samples from children with autism, general population, and MR/DD (non-autistic) samples, both from Northern and Southern California; the utilization of samples from Southern California has been made possible by converting from a TRIZOL-based method for RNA extraction to the use of PAX tubes, which preserve blood RNA at the time/site of draw. We are in the process of analyzing the first 22 samples (AUT and GP only).
Examination of the relative distributions of polymorphism frequencies for single genes within autism and non-autism subjects - While there is strong evidence that epistatic interactions take place between several susceptibility genes, genetic analysis has not identified any specific gene known to consistently cause autism. Mapping susceptibility genes that may be involved in complex disorders such as autism is especially difficult due to the small effect of a single gene. Association studies with polymorphisms in candidate genes could thus be an attractive strategy to study gene-gene interactions. To explore the impact of allelic variation and expression on autism, we have been examining the association between autism and single-gene polymorphisms at six loci, each of which having been implicated as an autism-associated gene; they are: Adenosine Deaminase, Serotonin transporter (5-HT), Glutathione S-transferase, Reelin, dopamine beta-hydroxylase, and monoamine oxidase A. To date, we have performed PCR-based studies on 103 patients with autism, 102 normal controls, and 19 patients with fragile X and autism. Chi-squared analyses of the PCR data did not reveal any significant associations at the p=0.05 level for any of the genes examined thus far: adenosine deaminase 2 (ADA2) allele (p=0.33); null allele in the Glutathione transferase T (GSTT1) gene (p=0.69); null allele in the Glutathione transferase M (GSTM1) gene (p=0.07); serotonin transporter alleles (p=0.18); and dopamine B-hydroxylase DBH alleles (p=0.38). We are currently setting up and running the tests for Reelin and monoamine oxidase.
Examination of the expression of Ca channels in dendritic cells – We are examining the expression of Ca channels in immature and mature dendritic cells from C57/Bl6 mice. Results to date have demonstrated that the spatial distribution of the three calcium channels are rather remarkable and point to an important functional relationship. We have verified these 'hits', identified with the gene expression microarrays at the transcriptome and at the proteome, using immunocytochemistry.
Analysis of microarray data – We are continuing our process of developing an integrated approach to the analysis of microarray data. The first paper on this subject is now published. We have developed methods and software for determination of differential expression using Affymetrix arrays for a broad variety of experimental designs: treatment vs. control, multiple treatments, and factorial designs. We have developed methods for estimation of the correct transformation scale on which to perform such analyses; the software will be ready for testing by the end of June 2003.
We have proposed a method to relate gene expression profiles with censored phenotypes such as patient survival time or time to cancer recurrence. We have proposed a dimension reduction strategy, which combines principal components analysis and sliced inverse regression, to identify linear combinations of genes, that both account for the variability in the gene expression levels and preserve the phenotypic information. A predictive survival model is then built upon the extracted gene components. The proposed method is shown to provide a good predictive performance for patient survival, as demonstrated by both the significant survival difference between the predicted risk groups, and the receiver operator characteristics analysis.
Finally, we have proposed a strategy that combines pharmacogenomics modeling with dimension reduction that can naturally handle both continuous and discrete phenotypes, and can deal with the high dimensionality of the microarray gene expression data. The method is shown to provide informative data visualization that reveals various association patterns between the pharmacological response and genomic features. It also helps establish models that exhibit good prediction accuracy.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
|Other subproject views:||All 4 publications||2 publications in selected types||All 2 journal articles|
|Other center views:||All 146 publications||134 publications in selected types||All 133 journal articles|
||Geller SC, Gregg JP, Hagerman P, Rocke DM. Transformation and normalization of oligonucleotide microarray data. Bioinformatics 2003;19(14):1817-1823.||
Supplemental Keywords:RFA, Health, Scientific Discipline, PHYSICAL ASPECTS, ENVIRONMENTAL MANAGEMENT, Health Risk Assessment, Chemistry, Risk Assessments, Susceptibility/Sensitive Population/Genetic Susceptibility, Disease & Cumulative Effects, Physical Processes, Children's Health, genetic susceptability, Biology, Risk Assessment, chemical exposure, neurotoxic, xenobiotics, biomarkers, gene-environment interaction, neurodevelopment, pesticides, exposure, halogenated aromatics, children, neurobehavioral, neurodevelopmental, neurotoxicity, etiology, susceptibility, human exposure, neurobehavioral effects, autism, biological markers, mechanisms, exposure assessment, neurological development, biomarker, synergistic interactions
Progress and Final Reports:Original Abstract
Main Center Abstract and Reports:R829388 CECEHDPR - University of California at Davis Center for the Study of Environmental Factors in the Etiology of Autism
Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
R829388C001 Environmental Factors in the Etiology of Autism; Analytic Biomakers (xenobiotic) Core
R829388C002 Environmental Factors in the Etiology of Autism; Cell Activation/Signaling Core
R829388C003 Environmental Factors in the Etiology of Autism; Molecular Biomakers Core
R829388C004 Environmental Factors in the Etiology of Autism; Childhood Autism Risks from Genetics and the Environment (The CHARGE Study)
R829388C005 Environmental Factors in the Etiology of Autism; Animal Models of Autism
R829388C006 Environmental Factors in the Etiology of Autism; Molecular and Cellular Mechanisms of Autism