Pre-Treatment Gene Expression and Risperidone Associated Weight Gain in Children with Autism Spectrum Disorders (ASD)

Background: Atypical antipsychotics have been shown to be beneficial in decreasing behavioral disturbances, such as aggression and anxiety, in children with ASD. However, previous studies have shown that these drugs are not effective for all children. In addition, they also have the potential for serious short-term and long-term side effects, including excessive appetite, weight gain, and diabetes. In our recently published study, we demonstrated that peripheral blood gene expression before treatment with the atypical antipsychotic risperidone may predict improvements in severe behavioral symptoms (Lit et al, 2011). We identified 5 genes that were differentially expressed between the subjects who imporved the most and the least to risperidone. In this study, our aim was to apply the same method to identify genes that predict severe weight gain.

Objectives: To determine if a peripheral blood gene expression profile prior to 8 weeks of treatment with risperidone will predict weight gain in children.

Methods: We used data from 41 subjects children who had an initial Aberrant Behavior Checklist Irritability (ABC-I) subscale rating of ≥18. The 8-week dosing schedule mirrored that used in two recent positive trials of risperidone. Affymetrix GeneChip Human Exon 1.0 ST Arrays were used to obtain gene expression values. Raw Affymetrix CEL files were imported into Partek Genomics Suite 6.6 Beta. Probe-level expression was summarized and normalized using the robust multichip average technique. Weight gain was represented through change in z-scores of anthropomorphically-adjusted BMI distributions. Subjects with a decrease in weight (n=4) were excluded from analyses. ANOVA was performed on pre-treatment exon expression levels across high, medium, and low weight gain groups based on change in BMI z-score pre-post treatment controlling for age and gender. Benjamini-Hochberg FDR method (α <0.05) was used to correct for multiple comparisons. Expression of exons within genes that had an expression fold change of > |1.5| between groups were chosen.

Results: We found one differentially expressed gene, PLP2 (A4 protein), that had a 1.88 underfold expression in high weight gain subjects (change in z-score > 1) compared to low weight gain subjects (change in z-score < 0.25) (p=1.73x10^-7).  We found no differentially expressed genes between the high and medium or the medium and low weight gain groups.

Conclusions: A recent study by Lee et al (2004) has associated the PLP2/A4 gene to the CCR1 signaling pathway, which is involved in immune and inflammatory processes. A separate study by Zhang et al (2007) associated this gene with X-linked mental retardation. The genes identified in our current study did not replicate those found in a study of risperidone treatment of autism done by Correia et al (2009). The difficulty of finding similar genes expressed across studies highlights the challenge of predicting personalized responses to medication without knowing the specific “causative genes” and the need to recruit large enough samples to control for multiple gene effects.