Phenotypic subgrouping and multi-omics analyses reveal reduced diazepam-binding inhibitor (DBI) protein levels in autism spectrum disorder with severe language impairment.

Chatravee Pichitpunpong,Siriporn Sangsuthum,Wan Mohd Aizat,Surangrat Thongkorn,Valerie Wailin Hu,Waluga Plaingam,Tewin Tencomnao,Tewarit Sarachana,Wasana Yuwattana,Syarul Nataqain Baharum,Songphon Kanlayaprasit,Jon M Jacobs

doi:10.1371/journal.pone.0214198

Abstract

BackgroundThe mechanisms underlying autism spectrum disorder (ASD) remain unclear, and clinical biomarkers are not yet available for ASD. Differences in dysregulated proteins in ASD have shown little reproducibility, which is partly due to ASD heterogeneity. Recent studies have demonstrated that subgrouping ASD cases based on clinical phenotypes is useful for identifying candidate genes that are dysregulated in ASD subgroups. However, this strategy has not been employed in proteome profiling analyses to identify ASD biomarker proteins for specific subgroups.MethodsWe therefore conducted a cluster analysis of the Autism Diagnostic Interview-Revised (ADI-R) scores from 85 individuals with ASD to predict subgroups and subsequently identified dysregulated genes by reanalyzing the transcriptome profiles of individuals with ASD and unaffected individuals. Proteome profiling of lymphoblastoid cell lines from these individuals was performed via 2D-gel electrophoresis, and then mass spectrometry. Disrupted proteins were identified and compared to the dysregulated transcripts and reported dysregulated proteins from previous proteome studies. Biological functions were predicted using the Ingenuity Pathway Analysis (IPA) program. Selected proteins were also analyzed by Western blotting.ResultsThe cluster analysis of ADI-R data revealed four ASD subgroups, including ASD with severe language impairment, and transcriptome profiling identified dysregulated genes in each subgroup. Screening via proteome analysis revealed 82 altered proteins in the ASD subgroup with severe language impairment. Eighteen of these proteins were further identified by nano-LC-MS/MS. Among these proteins, fourteen were predicted by IPA to be associated with neurological functions and inflammation. Among these proteins, diazepam-binding inhibitor (DBI) protein was confirmed by Western blot analysis to be expressed at significantly decreased levels in the ASD subgroup with severe language impairment, and the DBI expression levels were correlated with the scores of several ADI-R items.ConclusionsBy subgrouping individuals with ASD based on clinical phenotypes, and then performing an integrated transcriptome-proteome analysis, we identified DBI as a novel candidate protein for ASD with severe language impairment. The mechanisms of this protein and its potential use as an ASD biomarker warrant further study.

Highlights

Autism spectrum disorder (ASD) is a group of neurodevelopmental disorders thought to result from both genetic and environmental factors [1,2,3,4]
Diazepam-binding inhibitor (DBI) protein was confirmed by Western blot analysis to be expressed at significantly decreased levels in the autism spectrum disorder (ASD) subgroup with severe language impairment, and the diazepam-binding inhibitor (DBI) expression levels were correlated with the scores of several Autism Diagnostic Interview-Revised (ADI-R) items
By subgrouping individuals with ASD based on clinical phenotypes, and performing an integrated transcriptome-proteome analysis, we identified DBI as a novel candidate protein for ASD with severe language impairment

Summary

Introduction

Autism spectrum disorder (ASD) is a group of neurodevelopmental disorders thought to result from both genetic and environmental factors [1,2,3,4]. Some individuals with ASD exhibit severe language deficits, including a lack of verbal communication, whereas many others show little impairment in language and are capable of communication with others [7]. Such heterogeneity in the autism spectrum is thought to result from the combination of multiple molecular mechanisms and environmental factors, leading to the pathobiology specific to the clinical phenotypes of each ASD subpopulation. Recent studies have demonstrated that subgrouping ASD cases based on clinical phenotypes is useful for identifying candidate genes that are dysregulated in ASD subgroups This strategy has not been employed in proteome profiling analyses to identify ASD biomarker proteins for specific subgroups

Methods

Results

Conclusion