The Genetics of Childhood Disintegrative Disorder
Objectives: To identify rare variants that are associated with CDD, to determine if the candidate genes affected by these variants are coexpressed and/or belong to a common molecular pathway, and to compare these genes to those identified in an independent cohort of subjects affected by autistic regression.
Methods: We performed whole exome sequencing and copy number variant analyses of 18 families affected by CDD, which included 18 probands, 18 unaffected sibling controls, and their parents, to identify three types of rare (novel or found at most once across online and in-house databases) protein-changing variants: (1) de novo, (2) homozygous, and (3) hemizygous (mother-to-son transmission on chrX). We used the human BrainSpan exon-array transcriptome dataset to plot the brain expression profile of CDD candidate genes and conduct co-expression analysis.
Results: We identified one or more rare variants for all but one proband. CDD candidate genes did not overlap with high-confidence ASD genes. One gene, SUPT20HL2, which may play a role in transcription, was affected in two unrelated probands. Candidate genes that were found to be most conserved at variant position and most intolerant of variation (TRRAP, ZNF236, KIAA2018) also play a role or may be involved in transcription. A significant number of CDD candidate genes were co-expressed (P=0.0059). Overall, the candidate genes are more highly expressed in non-neocortical regions (amygdala, cerebellum, hippocampus, thalamus, striatum) versus neocortical regions. Moreover, there are increasing levels of expression in amygdala, striatum, and hippocampus during 1-6 years of age, the range that encompasses symptom onset. We compared the difference in median expression levels between non-neocortical and neocortical regions for genes affected by nonsynonymous and synonymous variants in CDD probands and their unaffected siblings, as well as ASD probands from the Simons Simplex Collection (SSC) both with (SSC+R) and without (SSC-R) regression. The expression profile of CDD candidate genes is qualitatively distinct from all comparison gene sets except for that of nonsynonymous variants in SSC+R.
Conclusions: There are important areas of overlap and difference between the genetics of CDD and ASD, perhaps reflecting the similarities and differences between the clinical features of these two disorders. Although there is no overlap between the list of candidate genes, both contain genes which play roles in transcription, offering clues to the pathophysiology of the disorders. The similarity of the expression profile of candidate genes for CDD and SSC+R suggest a pattern relevant to regression.