Identification of microduplications at Xp21.2 and Xq13.1 in neurodevelopmental disorders

Abstract Background Microduplications are a rare cause of disease in X‐linked neurodevelopmental disorders but likely have been under reported due challenges in detection and interpretation. Methods We performed exome sequencing and subsequent microarray analysis in two families with a neurodevelopmental disorder. Results Here, we report on two families each with unique inherited microduplications at Xp21.2 and Xq13.1, respectively. In the first family, a 562.8‐kb duplication at Xq13.1 covering DLG3, TEX11, SLC7A3, GDPD2, and part KIF4A was identified in a boy whose phenotype was characterized by delayed speech development, mild intellectual disability (ID), mild dysmorphic facial features, a heart defect, and neuropsychiatric symptoms. By interrogating all reported Xq13.1 duplications in individuals affected with a neurodevelopmental disorder, we provide evidence that this genomic region and particularly DLG3 might be sensitive to an increased dosage. In the second family with four affected males, we found a noncontinuous 223‐ and 204‐kb duplication at Xp21.2, of which the first duplication covers exon 6 of IL1RAPL1. The phenotype of the male patients was characterized by delayed speech development, mild to moderate ID, strabismus, and neurobehavioral symptoms. The carrier daughter and her mother had learning difficulties. IL1RAPL1 shows nonrecurrent causal structural variation and is located at a common fragile site (FRAXC), prone to re‐arrangement. Conclusion In conclusion, we show that comprehensive clinical and genetic examination of microduplications on the X‐chromosome can be helpful in undiagnosed cases of neurodevelopmental disease.

Microduplications on the X chromosome are challenging to detect and interpret. However, a careful follow up and investigation of these microduplications may lead to a molecular diagnosis. When screening Finnish families with cases of unexplained NDDs using ES, we identified two families with possible X-chromosomal duplications in areas with known NDD genes. The variants were further characterized and confirmed using chromosome microarray analysis (CMA). Here we report the detailed clinical phenotypes and molecular genetic analyses of the identified families.

| Ethical compliance
Written informed consent was obtained from healthy adult subjects and the parents/legal guardians of minor subjects and ID patients. The study was approved by the ethics committees of the Hospital District of Helsinki and Uusimaa and the Institutional review board of Columbia University, New York (IRB-AAAS3433).

| Exome sequencing
Genomic DNA was extracted from peripheral blood using the NucleoSpin blood XL kit (Macherey Nagel, Germany), according to the manufacturer's instructions. DNA samples from trio FIN15 (FIN15-1; FIN15-2 and FIN15-3) and from family FIN41 (FIN41-5 and FIN41-6) underwent exome sequencing ( Figure 1). Target enrichment was done using the SureSelect Human All Exon V6 kit, and paired-end sequencing was performed on a HiSeq2500/4000 instrument (Illumina Inc, San Diego, CA, USA). Bioinformatic details can be found in the supporting information. In short, data were aligned to the human genome, single nucleotide variant (SNV), Insertion/Deletions (InDels), and Copy number variants (CNV) were called and annotated. Rare SNV, InDel, and CNV variants that fit the appropriate inheritance models based on the pedigree and were predicted to have a functional effect on gene function were retained.

| Microarray analysis
We validated candidate CNVs detected via exome sequencing and tested segregation via a CMA in family members. Microarray analysis was performed using a 50mer-oligochip (HumanCytoSNP-12v2.1, Illumina Inc.) that allowed an effective resolution as small as 30 kb in cytogenetically relevant regions and 200 kb in other areas of the genome. Copy-number changes and regions of SNV-homozygosity were analyzed with GenomeStudio v.2011.1 and KaryoStudio 1.3 programs (Illumina Inc.) using reference genome GRCh37/hg19. Identified CNVs were compared with known CNVs listed in the DGV, dbVar, UCSC genome browser, DECIPHER, and OMIM databases and further interrogated using peerreviewed literature searches in the PubMed database.

| Review of reported copy number variants identified at Xq13.1 and Xp21.2 in individuals with a NDD
To identify reported duplications on Xq13.1 associated with a NDD, we interrogated Pubmed, the Columbia University catalog and DECIPHER (Firth et al., 2009). The criteria used were the following: (1) duplications with overlap with the duplication from the current study, (2) exact genomic coordinates had been determined, (3) length under 10 Mb, (4) phenotypic overlap between cases, (5) males only were included, and (6) individuals with other genomic variants likely implicated in their disease were excluded. For Xp21.2, a literature search using Pubmed and the Columbia University catalog was performed to identify intragenic variants IL1RAPL1 (# 300143; MRX21) previously associated with neurodevelopmental disease. (1) Both intragenic microduplications and deletions in IL1RAPL1 were identified (<1.3 Mb, the size of IL1RAPL1).
(2) Only variants with detailed coordinates were retained. (3) Individuals with a phenotypic similarity with our patient were retained, including ID and developmental delay, typical for the disorder associated with IL1RAPL1 (Mental retardation, X-linked 21/34). (4) Individuals with other genomic variants likely implicated in their disease were excluded.

|
We next reviewed all cases with a NDD with overlapping features and duplications at Xq13.1, as shown in Table S2 and Figure 2. Via a literature search, we identified three duplications in this area in eight male patients with phenotypic overlap including ID, abnormal behavior and dysmorphic features (Kaya et al., 2012;Bhattacharya et al., 2019;Wentz et al., 2014), and a microduplication in two brothers with a NDD with seizures only (Magini et al., 2019). Seven additional overlapping duplications in patients with phenotypic overlap are currently present in DECIPHER, of which four are shown in Figure 2 for which permission was obtained (Firth et al., 2009). Figure 2 shows that this genomic area might be sensitive to increased dosage. This might particularly be true for DLG3, which is present in a minimal overlapping region. Single nucleotide and InDel variants in DLG3 are known to cause a NDD with variable degrees of ID, dysmorphic features, language delay, and epilepsy in some cases (Philips et al., 2014;Tarpey et al., 2004). Magini et al. also suggested that the microduplication they identified in two affected siblings with epilepsy may be due to increased dosage sensitivity of DLG3 and/or KIF4A (Magini et al., 2019). Our data provide more evidence to show that DLG3 is sensitive to an increased dosage and associated with a similar phenotype as SNV/InDel variants.
The Xp21.2 duplication includes exon 6 (NM_014271.3) of IL1RAPL1 (# 300206), interleukin-1 receptor accessory protein like 1 gene, which could possibly affect the function of its extracellular domain. The in-frame duplication of exon 6 is predicted to lead to an insertion of 25 amino acids in the extracellular domain of IL1RAPL1, between immunoglobulin domain (Ig) 2 and 3 p.(Ala235_Leu259dup). IL1RAPL1 is mostly expressed in the brain where it regulates synapse formation, and it has activity on synaptogenesis and dendrite morphology (Montani et al., 2019;Ramos-Brossier et al., 2015).
The proximal duplication at Xp21.2 (chrX:30643031-30847448, hg19) covers the glycerol kinase (GK) gene (# 300474) and the part of TAB3-gene (# 300480). SNVs, InDels, and large deletions in the GK gene underlie XLR Glycerol kinase deficiency (GKD) (OMIM #307030) known to cause a metabolic disorder. GKD results in hyperglycerolemia, a condition characterized by the accumulation of glycerol in the blood and urine. Isolated glycerol kinase deficiency is believed to be a benign condition (FM Vaz, personal communication).
All four affected males have a similar nonsyndromic phenotype with mild to moderate ID, strabismus and hyperactivity (Table 1). The first symptom was delayed speech development identified at 2-3 years of age. The two carrier females, of them FIN41-3 molecularly tested, only showed learning difficulties.

| DISCUSSION
In this study, we report two families with different inherited X-linked duplications. In the first family, a duplication at chrXq13.1 was first detected using ES, which was confirmed via CMA to be 563 kb. In the duplicated region Xq13.1, two genes, DLG3 and KIF4A (partial), are known to cause ID. We and others have reported SNV and InDel variants in DLG3 (Philips et al., 2014;Tarpey et al., 2004) in cases characterized by mild to severe ID, mild dysmorphic features, language delay, and epilepsy in some patients. KIF4A is involved in cell division and has been reported to underlie mild to moderate nonsyndromic ID, language delay, and epilepsy. TEX11, SLC7A3, and GDPD2 have not been associated with brain function (www.omim.org). Via a detailed analysis of reported cases with duplications at Xq13.1, we provide evidence that DLG3 may have an increased dosage sensitivity in this region (Figure 2). We show that microduplications at Xq13.1 are also associated with a similar phenotype as the phenotype due to SNV/InDel variants in DLG3 (Table S2).
In the second family, a Xp21.2 duplication was identified in four affected males and one female. Single nucleotide, InDel, and copy number variants in IL1RAPL1 have been associated with a variable phenotype ranging from nonsyndromic ID to autism spectrum disorder (ASD) (Ramos-Brossier et al., 2015). The identified Xp21.2 duplication contains exon 6 of the IL1RAPL1 gene, is predicted to lead to p.(Ala235_Leu259dup), and is expected to have a similar effect as other intragenic microCNVs covering exon 6 ( Figure 2; Table S3) (Philips et al., 2014;Tarpey et al., 2004). Affected males with exon 6 duplications and deletions indeed show similar features including language and motor development delay and ID (Table S3).
Female carriers of the exon 6 duplication in our family showed a mild phenotype of learning disability. Similarly, three female carriers of an exon 6 deletion in IL1RAPL1, predicted to lead to p.(Ala235_Leu259del), also had learning difficulties or ID only (Table S3) (Ramos-Brossier et al., 2015). This is not surprising as a female carrier phenotype is present in the majority of X-linked ID disorders, although this is often a milder phenotype (Ziats et al., 2020). One of the females with an exon 6 deletion in IL1RAPL1 showed random X-inactivation based on studies in her fibroblast cells (Ramos-Brossier et al., 2015).
Intragenic deletions of IL1RAPL1 are a common disease mechanism (Whibley et al., 2010); however, intragenic disease-associated duplications in this gene are less common (Laino et al., 2016b). The mechanisms of IL1RAPL1 rearrangement are likely related to its presence in the common fragile site FRAXC, and the implicated mechanisms of SV creation may favor deletions (Whibley et al., 2010). Common fragile sites are common regions of profound genomic instability. FRAXC is a common fragile site containing both DMD and IL1RAPL1; both are genes in which SVs are often involved in Mendelian disease. Instability-induced alterations will primarily occur within intronic regions, and IL1RAPL1 covers a large genomic region (1.37 Mbs) with >99% of its sequence intronic. IL1RAPL1's large genomic size in an area of instability makes it susceptible to DNA breakage and gene rearrangements. Of interest, we also identified a second duplication adjacent to IL1RAPL1 at Xp21.2 in our patient. Similar to previous reports of noncontinuous microSVs in this region (Chatron et al., 2017;Laino et al., 2016b), this is also likely due to this area being prone to breakage and subsequent incorrect rearrangement.
As the duplication we identified in IL1RAPL1 is nonrecurrent ( Figure 2) nor is there significant homology between introns (no low copy repeats), nonhomologous end joining (NHEJ) and microhomology-mediated mechanisms might be the more likely mechanism in our case. Although the exact breakpoints of the duplication in IL1RAPL1 are unknown, the breakpoint sites do primarily contain long interspersed nuclear elements (LINE) elements and some Alu repeats. The disproportionate rate of deletions relative to duplications has also been seen at some nonallelic homologous recombination (NAHR) hotspots (Turner et al., 2008), and Alu-Aluor LINE-LINE-mediated NAHR may also have occurred. Similar to the Xp21.2 duplication, the Xq13.1 duplication is also nonrecurrent (Figure 2), does not contain any low copy repeats, and the breakpoint areas are flanked by a large number of Alu repeats and LINE repeats as well, suggesting NEJH or microhomology-mediated mechanisms.
In conclusion, our study shows that exome sequencing is a useful tool to screen for microCNVs, which can lead to a molecular diagnosis via additional molecular testing and research. By combining careful clinical analysis, literature and further genetic characterization, we report two novel microduplications on chromosome X implicated in X-linked ID and provide evidence that DLG3 is sensitive to increased dosage.

ACKNOWLEDGMENTS
We thank the patients and their families for their participation in this study. We are grateful for Anju K Philips and Shaffaq Raza for excellent technical help and Dr. Bernardini (IRCCS CSS-Mendel Institute, Italy), Dr. Tzschach (University Medical Center Freiburg, Germany), and Dr. McGowan (West of Scotland Centre for Genomic Medicine, UK) for contributing their DECIPHER data. This study makes use of data generated by the DECIPHER community. A full list of contributing centers is available from https://decip her. sanger.ac.uk/about/ stats and via email (decipher@sanger. ac.uk). Funding for the DECIPHER project was provided by