Gene expression profiling differentiates autism case-controls and phenotypic variants of autism spectrum disorders: evidence for circadian rhythm dysfunction in severe autism

Abstract

Autism spectrum disorders (ASD) are neurodevelopmental disorders characterized by delayed/abnormal language development, deficits in social interaction, repetitive behaviors and restricted interests. The heterogeneity in clinical presentation of ASD, likely due to different etiologies, complicates genetic/biological analyses of these disorders. DNA microarray analyses were conducted on 116 lymphoblastoid cell lines (LCL) from individuals with idiopathic autism who are divided into three phenotypic subgroups according to severity scores from the commonly used Autism Diagnostic Interview-Revised questionnaire and age-matched, nonautistic controls. Statistical analyses of gene expression data from control LCL against that of LCL from ASD probands identify genes for which expression levels are either quantitatively or qualitatively associated with phenotypic severity. Comparison of the significant differentially expressed genes from each subgroup relative to the control group reveals differentially expressed genes unique to each subgroup as well as genes in common across subgroups. Among the findings unique to the most severely affected ASD group are 15 genes that regulate circadian rhythm, which has been shown to have multiple effects on neurological as well as metabolic functions commonly dysregulated in autism. Among the genes common to all three subgroups of ASD are 20 novel genes mostly in putative noncoding regions, which appear to associate with androgen sensitivity and which may underlie the strong 4:1 bias toward affected males.

Figures

Figure 1

Gene expression analyses of LCL from controls and autistic individuals in which the samples from the autistic individuals are shown in no specific order (A) and grouped according to ASD phenotype derived from cluster analyses of ADIR scores (B). Top 530 significant genes identified by a 2-class SAM analysis with a false discovery rate (FDR) <5%, which compared gene expression between controls and all autistic individuals. The expression level of each gene (row) in this “heat map” is shown relative to that of RNA from a Universal human RNA standard (Stratagene). cDNA from the reference standard was labeled with Cy-5, while the cDNA of the samples was labeled with Cy-3. The heat map represents the log2(ratio) of Cy5/Cy3 fluorescence, with green showing decreased gene expression in Cy5 relative to Cy3 and red indicating the reverse. Thus, a change in color from green to red in this figure represents a decrease in relative expression.

Figure 2

(A) Gene expression matrix of the top 123 significant genes from 4-class SAM analysis of gene expression data from the three ASD groups and nonautistic control group, with an expression cutoff of log2(ratio)≥±0.3. A more stringent FDR (<0.1%) was employed in this analysis to maximize gene discrimination among the groups although the separation of groups is still obvious with 673 genes (Supplemental Fig. 1). Color coding: Turquoise = nonautistic individuals; red = ASD with severe language impairment; blue = mild ASD; yellow = ASD with savant skills. Again, the heat map represents the log2(ratio) of Cy5/Cy3 fluorescence, with green showing decreased gene expression in Cy5 relative to Cy3 and red indicating the reverse. (B) Principal components analysis of samples on the basis of the 123 significant differentially expressed genes identified by the 4-class SAM analyses described above. The one savant individual (yellow) that clearly associates with the language subgroup (red) also has a severe language impairment.

Figure 3

Separation of phenotypic variants of ASD from controls based on principal components analysis (PCA) of the differentially expressed genes identified by 2-class SAM analyses for each of the ASD subgroups and controls. (A) Separation of individuals with severe language impairment (L) and controls (C) based on 602 differentially expressed genes (FDR<0.1%). Turquoise-colored points represent control samples in all figures, with the second color representing the ASD samples. (B) Separation of individuals with mild ASD (M) and controls based on 502 differentially expressed genes (FDR<0.5%). (C) Separation of ASD “savants” (S) and controls based on 127 differentially expressed genes (FDR<0.5%). Differentially expressed genes for the respective ASD groups are listed in Supplemental Tables V–VII.

Figure 4

Venn diagram showing overlap of differentially expressed genes relative to controls among three subgroups of ASD. Significant genes were determined for each group by 2-class SAM analysis of the data from each ASD group and that of the nonautistic controls, with FDR≤5%. Because of the quantitative differences in gene expression that appeared to correlate with severity of the phenotype, no expression cutoff was applied in order to observe the maximum overlap of differentially expressed genes among the ASD groups. The complete list of differentially expressed genes between the L group and controls is given in Supplemental Table VIII. [Color figure can be viewed online at www.interscience.wiley.com]

Figure 5

(A) Top multigene network associated with the data set of differentially expressed genes (FDR≥5%) in both the severely language-impaired (L) and mild (M) ASD groups, as determined by Ingenuity Pathway Analysis. An expression cutoff of log2(ratio)≥±0.3 was applied prior to the analysis. (B) qRT-PCR analyses on five representative samples each from the L and M subgroups as well as the control group, focusing on hub genes in Figure 5A as well as genes of relevance to inflammation (NFKB1) and epigenetic regulation (MBD2), which are both implicated in autism. Data represent fold-changes±SE calculated using the delta-delta-Ct method and with respect to average Ct values obtained from the control group (n = 5). *P<0.05

Figure 6

Response of seven novel transcripts to 10 nM dihydroxytestosterone (DHT). The data shows the relative expression of the transcripts in DHT-treated cells to that in vehicle (ethanol-treated) cells after a 2-day exposure to the hormone. P-values for triplicate analyses were all <0.05.

Figure 7

Differentially expressed circadian rhythm genes in the severe ASD phenotype affect multiple processes commonly associated with ASD pathology. The color coding of the entities within this relational gene/molecular network are as follows: Red—genes that show increased expression in autistic individuals on average relative to controls; Green—genes that show decreased expression in autistic individuals on average relative to controls; Pink—additional circadian regulators inserted by Pathway Studio 5. Blue—small molecules including melatonin, serotonin, testosterone, and nitric oxide (NO); Yellow—cell processes; Lavender—disorders associated with these genes.

Figure 8

Association of differentially expressed genes in the combined autistic sample (A) with quantitative trait loci for ASD identified by genetic analyses. For in silico mapping of our differentially expressed genes in autism-related QTLs described in the literature, a QTL interval was defined as two LOD units around the linkage peak, provided that mapping information on multiple genetic markers in the vicinity of the peak was given. Otherwise, the interval was defined by 10−15 megabases on either side of a lone genetic marker defining the QTL. Red: differentially expressed genes within identified QTL; Blue: differentially expressed genes not associated with QTL. Chi square analyses were used to identify significant enrichment of differentially expressed genes on a particular chromosome. ***P<0.0005; **P<0.015; *P<0.025. P-values for enrichment of differentially expressed genes in QTLs on specific chromosomes for the ASD subgroups are provided in the text.