Sporadic male patients with intellectual disability: Contribution of X-chromosome copy number variants

Abstract

Genome-wide array comparative genome hybridization has become the first in line diagnostic tool in the clinical work-up of patients presenting with intellectual disability. As a result, chromosome X-copy number variations are frequently being detected in routine diagnostics. We retrospectively reviewed genome wide array-CGH data in order to determine the frequency and nature of chromosome X-copy number variations (X–CNV) in a cohort of 2222 sporadic male patients with intellectual disability (ID) referred to us for diagnosis. In this cohort, 68 males were found to have at least one X–CNV (3.1%). However, correct interpretation of causality remains a challenging task, and is essential for proper counseling, especially when the CNV is inherited. On the basis of these data, earlier experience and literature data we designed and propose an algorithm that can be used to evaluate the clinical relevance of X-CNVs detected in sporadic male ID patients. Applied to our cohort, 19 male ID patients (0.85%) were found to carry a (likely) pathogenic X–CNV.

Introduction

The human X-chromosome spans 155 Mb and carries about 1250 genes, an excess of which are involved in brain functions [1]. Consequently, since they carry a single X-chromosome, male individuals are more susceptible to X-linked intellectual disability (XLID). Many of these X-linked mutations have been found through studies of XLID families [2]. Screening of XLID families with X-chromosome tiling path arrays has proven to be a successful method to detect X-specific copy number variants (CNVs) [3], [4]. Sporadic male patients with ID are, however, routinely screened by genome-wide CNV detection methods.

Array comparative genome hybridization (array-CGH) has become the first in line routine diagnostic method to screen the genome with a diagnostic yield that is at least twice as high compared to conventional karyotyping [5]. Increased resolution has improved the ability to detect smaller CNVs and targeted array methods enable to screen specific regions of the genome at high resolution, e.g. the X-chromosome [3], [4]. Identification of a novel CNV is followed by an assessment of clinical significance usually commenced with parental testing. A de novo CNV is more likely to be causal, however the opposite is not always true. In the case of CNVs on the X-chromosome (X-CNVs), the mother frequently is a carrier and pathogenicity can be supported by a skewed X-inactivation status [6]. Segregation can further strengthen the causal role of a novel CNV. In addition, several internal and external databases (DECIPHER, ISCA) can be used to investigate whether a similar or overlapping CNV has been reported before, and if so, whether the phenotypes can be compared [7]. Other factors that influence classification are size, kind of arrangement (deletion/duplication) and gene content, as well as data from expression analysis. Different workflows have been described to facilitate interpretation of CNVs [7], [8]. Despite these factors, it often remains difficult to appropriately classify an unreported CNV. This can be attributed to the absence of one or both biological parents, small family size, occurrence of identical CNVs in apparently unaffected controls, or insufficient information available on gene function and/or expression. Finally, a lot of CNVs seem to be clinically heterogeneous, which makes it difficult to extract specific genotype–phenotype correlations.

In this study we present previously unreported as well as recurrent X-CNVs, based on a systematic review of all X-CNVs detected by genome wide array-CGH screening in a large cohort of sporadic male individuals with (non)-syndromic intellectual disability over the last 6 years. Next to clinical and molecular data, we discuss the relevance of these X-CNVs and propose an algorithm to evaluate their causality.